quantitative research paper in environmental science

• Search Menu
• Sign in through your institution
• Advance articles
• Editor's Choice
• Special Collections
• Author Guidelines
• Submission Site
• Open Access
• Reasons to submit
• About BioScience
• Journals Career Network
• Editorial Board
• Advertising and Corporate Services
• Self-Archiving Policy
• Potentially Offensive Content
• Terms and Conditions
• Journals on Oxford Academic
• Books on Oxford Academic

Article Contents

Quantitative environmental science.

• Article contents
• Figures & tables
• Supplementary Data

Scott L Collins, Quantitative Environmental Science, BioScience , Volume 71, Issue 12, December 2021, Page 1199, https://doi.org/10.1093/biosci/biab131

• Permissions Icon Permissions

There is little argument that today's ecologists and conservation biologists are becoming more and more quantitative. A few years ago, I was part of a working group led by Stephanie Hampton that was held at the National Center for Ecological Analysis and Synthesis. The group focused on the training that is needed today for data-intensive environmental research. In a paper from that workshop that was published in BioScience ( https://doi.org/10.1093/biosci/bix025 ), we noted that quantitative skills to perform data-intensive research were generally lacking among most environmental scientists. We argued that, like writing skills, basic math skills should be taught across the curriculum. In addition, to be competitive in an increasingly quantitative world, students and professionals needed to acquire some degree of understanding of data management and processing, analysis, coding, and visualization along with communication skills for presentation and collaboration.

It is somewhat ironic that I was involved in developing these recommendations. I have a confession to make. This will come as no surprise to my colleagues and collaborators (and my graduate students), but my limited quantitative data processing skills have completely eroded over the decades since graduate school, when we ran SAS code on a mainframe computer. And yet, I completely agree that developing quantitative skills needs to be an essential component of undergraduate and graduate training. For example, this year, our National Science Foundation–sponsored Research Experience for Undergraduates (REU) program in dryland ecology held a 2-day Data Carpentries workshop for the REU students (and me). These very well structured and organized workshops provide an excellent entrée into data management and coding in R, the most popular data processing language for ecologists. The students then expanded their coding skills for data analysis and visualization during the rest of the summer session while I proceeded to forget everything I learned.

In this issue of BioScience , Nathan Emery and colleagues ( https://academic.oup.com/bioscience/article-lookup/doi/10.1093/biosci/biab107 ) up the game considerably. Here, the authors primarily focus on data science, per se, including the skills noted by Hampton and colleagues, as well as “being able to scale analyses for high-performance computing, write scripts, and use command line interfaces, version control, and high-performance computing clusters.” That is, environmental scientists could be engaged in training the next generation of data scientists. These authors maintain that teaching such “quantitative literacy” requires competent instructors but that most environmental scientists do not have sufficient data skills to incorporate data science into their courses. Herein lies the problem: A large gap exists between computational needs and the skill set of most environmental scientists. Worse yet, many training opportunities are targeted primarily toward early career scientists reducing the likelihood that more senior scientists will gain little more than a rudimentary understanding of these tool skills.

All in all, the needs are obvious, the intentions are well meaning, but the solutions are complicated and challenging. Time during the semester is limited, and with more and more demand to teach writing, math, and data science across the curriculum, instructors will have to gain new skills and willingly adjust content to meet the needs of students entering an increasingly quantitative world.

Email alerts

Citing articles via.

• Recommend to your Library

Affiliations

• Online ISSN 1525-3244
• Copyright © 2024 American Institute of Biological Sciences
• About Oxford Academic
• Publish journals with us
• University press partners
• What we publish
• New features  
• Open access
• Institutional account management
• Rights and permissions
• Get help with access
• Accessibility
• Advertising
• Media enquiries
• Oxford University Press
• Oxford Languages
• University of Oxford

Oxford University Press is a department of the University of Oxford. It furthers the University's objective of excellence in research, scholarship, and education by publishing worldwide

• Copyright © 2024 Oxford University Press
• Cookie settings
• Cookie policy
• Privacy policy
• Legal notice

This Feature Is Available To Subscribers Only

Sign In or Create an Account

This PDF is available to Subscribers Only

For full access to this pdf, sign in to an existing account, or purchase an annual subscription.

Research Topics & Ideas: Environment

100+ Environmental Science Research Topics & Ideas

Finding and choosing a strong research topic is the critical first step when it comes to crafting a high-quality dissertation, thesis or research project. Here, we’ll explore a variety research ideas and topic thought-starters related to various environmental science disciplines, including ecology, oceanography, hydrology, geology, soil science, environmental chemistry, environmental economics, and environmental ethics.

NB – This is just the start…

The topic ideation and evaluation process has multiple steps . In this post, we’ll kickstart the process by sharing some research topic ideas within the environmental sciences. This is the starting point though. To develop a well-defined research topic, you’ll need to identify a clear and convincing research gap , along with a well-justified plan of action to fill that gap.

If you’re new to the oftentimes perplexing world of research, or if this is your first time undertaking a formal academic research project, be sure to check out our free dissertation mini-course. Also be sure to also sign up for our free webinar that explores how to develop a high-quality research topic from scratch.

Overview: Environmental Topics

• Ecology /ecological science
• Atmospheric science
• Oceanography
• Soil science
• Environmental chemistry
• Environmental economics
• Environmental ethics
• Examples  of dissertations and theses

Topics & Ideas: Ecological Science

• The impact of land-use change on species diversity and ecosystem functioning in agricultural landscapes
• The role of disturbances such as fire and drought in shaping arid ecosystems
• The impact of climate change on the distribution of migratory marine species
• Investigating the role of mutualistic plant-insect relationships in maintaining ecosystem stability
• The effects of invasive plant species on ecosystem structure and function
• The impact of habitat fragmentation caused by road construction on species diversity and population dynamics in the tropics
• The role of ecosystem services in urban areas and their economic value to a developing nation
• The effectiveness of different grassland restoration techniques in degraded ecosystems
• The impact of land-use change through agriculture and urbanisation on soil microbial communities in a temperate environment
• The role of microbial diversity in ecosystem health and nutrient cycling in an African savannah

Topics & Ideas: Atmospheric Science

• The impact of climate change on atmospheric circulation patterns above tropical rainforests
• The role of atmospheric aerosols in cloud formation and precipitation above cities with high pollution levels
• The impact of agricultural land-use change on global atmospheric composition
• Investigating the role of atmospheric convection in severe weather events in the tropics
• The impact of urbanisation on regional and global atmospheric ozone levels
• The impact of sea surface temperature on atmospheric circulation and tropical cyclones
• The impact of solar flares on the Earth’s atmospheric composition
• The impact of climate change on atmospheric turbulence and air transportation safety
• The impact of stratospheric ozone depletion on atmospheric circulation and climate change
• The role of atmospheric rivers in global water supply and sea-ice formation

Topics & Ideas: Oceanography

• The impact of ocean acidification on kelp forests and biogeochemical cycles
• The role of ocean currents in distributing heat and regulating desert rain
• The impact of carbon monoxide pollution on ocean chemistry and biogeochemical cycles
• Investigating the role of ocean mixing in regulating coastal climates
• The impact of sea level rise on the resource availability of low-income coastal communities
• The impact of ocean warming on the distribution and migration patterns of marine mammals
• The impact of ocean deoxygenation on biogeochemical cycles in the arctic
• The role of ocean-atmosphere interactions in regulating rainfall in arid regions
• The impact of ocean eddies on global ocean circulation and plankton distribution
• The role of ocean-ice interactions in regulating the Earth’s climate and sea level

Tops & Ideas: Hydrology

• The impact of agricultural land-use change on water resources and hydrologic cycles in temperate regions
• The impact of agricultural groundwater availability on irrigation practices in the global south
• The impact of rising sea-surface temperatures on global precipitation patterns and water availability
• Investigating the role of wetlands in regulating water resources for riparian forests
• The impact of tropical ranches on river and stream ecosystems and water quality
• The impact of urbanisation on regional and local hydrologic cycles and water resources for agriculture
• The role of snow cover and mountain hydrology in regulating regional agricultural water resources
• The impact of drought on food security in arid and semi-arid regions
• The role of groundwater recharge in sustaining water resources in arid and semi-arid environments
• The impact of sea level rise on coastal hydrology and the quality of water resources

Topics & Ideas: Geology

• The impact of tectonic activity on the East African rift valley
• The role of mineral deposits in shaping ancient human societies
• The impact of sea-level rise on coastal geomorphology and shoreline evolution
• Investigating the role of erosion in shaping the landscape and impacting desertification
• The impact of mining on soil stability and landslide potential
• The impact of volcanic activity on incoming solar radiation and climate
• The role of geothermal energy in decarbonising the energy mix of megacities
• The impact of Earth’s magnetic field on geological processes and solar wind
• The impact of plate tectonics on the evolution of mammals
• The role of the distribution of mineral resources in shaping human societies and economies, with emphasis on sustainability

Topics & Ideas: Soil Science

• The impact of dam building on soil quality and fertility
• The role of soil organic matter in regulating nutrient cycles in agricultural land
• The impact of climate change on soil erosion and soil organic carbon storage in peatlands
• Investigating the role of above-below-ground interactions in nutrient cycling and soil health
• The impact of deforestation on soil degradation and soil fertility
• The role of soil texture and structure in regulating water and nutrient availability in boreal forests
• The impact of sustainable land management practices on soil health and soil organic matter
• The impact of wetland modification on soil structure and function
• The role of soil-atmosphere exchange and carbon sequestration in regulating regional and global climate
• The impact of salinization on soil health and crop productivity in coastal communities

Topics & Ideas: Environmental Chemistry

• The impact of cobalt mining on water quality and the fate of contaminants in the environment
• The role of atmospheric chemistry in shaping air quality and climate change
• The impact of soil chemistry on nutrient availability and plant growth in wheat monoculture
• Investigating the fate and transport of heavy metal contaminants in the environment
• The impact of climate change on biochemical cycling in tropical rainforests
• The impact of various types of land-use change on biochemical cycling
• The role of soil microbes in mediating contaminant degradation in the environment
• The impact of chemical and oil spills on freshwater and soil chemistry
• The role of atmospheric nitrogen deposition in shaping water and soil chemistry
• The impact of over-irrigation on the cycling and fate of persistent organic pollutants in the environment

Topics & Ideas: Environmental Economics

• The impact of climate change on the economies of developing nations
• The role of market-based mechanisms in promoting sustainable use of forest resources
• The impact of environmental regulations on economic growth and competitiveness
• Investigating the economic benefits and costs of ecosystem services for African countries
• The impact of renewable energy policies on regional and global energy markets
• The role of water markets in promoting sustainable water use in southern Africa
• The impact of land-use change in rural areas on regional and global economies
• The impact of environmental disasters on local and national economies
• The role of green technologies and innovation in shaping the zero-carbon transition and the knock-on effects for local economies
• The impact of environmental and natural resource policies on income distribution and poverty of rural communities

Topics & Ideas: Environmental Ethics

• The ethical foundations of environmentalism and the environmental movement regarding renewable energy
• The role of values and ethics in shaping environmental policy and decision-making in the mining industry
• The impact of cultural and religious beliefs on environmental attitudes and behaviours in first world countries
• Investigating the ethics of biodiversity conservation and the protection of endangered species in palm oil plantations
• The ethical implications of sea-level rise for future generations and vulnerable coastal populations
• The role of ethical considerations in shaping sustainable use of natural forest resources
• The impact of environmental justice on marginalized communities and environmental policies in Asia
• The ethical implications of environmental risks and decision-making under uncertainty
• The role of ethics in shaping the transition to a low-carbon, sustainable future for the construction industry
• The impact of environmental values on consumer behaviour and the marketplace: a case study of the ‘bring your own shopping bag’ policy

Examples: Real Dissertation & Thesis Topics

While the ideas we’ve presented above are a decent starting point for finding a research topic, they are fairly generic and non-specific. So, it helps to look at actual dissertations and theses to see how this all comes together.

Below, we’ve included a selection of research projects from various environmental science-related degree programs to help refine your thinking. These are actual dissertations and theses, written as part of Master’s and PhD-level programs, so they can provide some useful insight as to what a research topic looks like in practice.

• The physiology of microorganisms in enhanced biological phosphorous removal (Saunders, 2014)
• The influence of the coastal front on heavy rainfall events along the east coast (Henson, 2019)
• Forage production and diversification for climate-smart tropical and temperate silvopastures (Dibala, 2019)
• Advancing spectral induced polarization for near surface geophysical characterization (Wang, 2021)
• Assessment of Chromophoric Dissolved Organic Matter and Thamnocephalus platyurus as Tools to Monitor Cyanobacterial Bloom Development and Toxicity (Hipsher, 2019)
• Evaluating the Removal of Microcystin Variants with Powdered Activated Carbon (Juang, 2020)
• The effect of hydrological restoration on nutrient concentrations, macroinvertebrate communities, and amphibian populations in Lake Erie coastal wetlands (Berg, 2019)
• Utilizing hydrologic soil grouping to estimate corn nitrogen rate recommendations (Bean, 2019)
• Fungal Function in House Dust and Dust from the International Space Station (Bope, 2021)
• Assessing Vulnerability and the Potential for Ecosystem-based Adaptation (EbA) in Sudan’s Blue Nile Basin (Mohamed, 2022)
• A Microbial Water Quality Analysis of the Recreational Zones in the Los Angeles River of Elysian Valley, CA (Nguyen, 2019)
• Dry Season Water Quality Study on Three Recreational Sites in the San Gabriel Mountains (Vallejo, 2019)
• Wastewater Treatment Plan for Unix Packaging Adjustment of the Potential Hydrogen (PH) Evaluation of Enzymatic Activity After the Addition of Cycle Disgestase Enzyme (Miessi, 2020)
• Laying the Genetic Foundation for the Conservation of Longhorn Fairy Shrimp (Kyle, 2021).

Looking at these titles, you can probably pick up that the research topics here are quite specific and narrowly-focused , compared to the generic ones presented earlier. To create a top-notch research topic, you will need to be precise and target a specific context with specific variables of interest . In other words, you’ll need to identify a clear, well-justified research gap.

Need more help?

If you’re still feeling a bit unsure about how to find a research topic for your environmental science dissertation or research project, be sure to check out our private coaching services below, as well as our Research Topic Kickstarter .

Need a helping hand?

12 Comments

research topics on climate change and environment

Researched PhD topics on environmental chemistry involving dust and water

I wish to learn things in a more advanced but simple way and with the hopes that I am in the right place.

Thank so much for the research topics. It really helped

the guides were really helpful

Research topics on environmental geology

Thanks for the research topics….I need a research topic on Geography

hi I need research questions ideas

Implications of climate variability on wildlife conservation on the west coast of Cameroon

I want the research on environmental planning and management

I want a topic on environmental sustainability

It good coaching

Submit a Comment Cancel reply

Your email address will not be published. Required fields are marked *

Save my name, email, and website in this browser for the next time I comment.

• Print Friendly

• Methodology
• Open access
• Published: 24 April 2023

Quantitative evidence synthesis: a practical guide on meta-analysis, meta-regression, and publication bias tests for environmental sciences

• Shinichi Nakagawa   ORCID: orcid.org/0000-0002-7765-5182 1 , 2 ,
• Yefeng Yang   ORCID: orcid.org/0000-0002-8610-4016 1 ,
• Erin L. Macartney   ORCID: orcid.org/0000-0003-3866-143X 1 ,
• Rebecca Spake   ORCID: orcid.org/0000-0003-4671-2225 3 &
• Malgorzata Lagisz   ORCID: orcid.org/0000-0002-3993-6127 1  

Environmental Evidence volume  12 , Article number:  8 ( 2023 ) Cite this article

8805 Accesses

23 Citations

29 Altmetric

Metrics details

Meta-analysis is a quantitative way of synthesizing results from multiple studies to obtain reliable evidence of an intervention or phenomenon. Indeed, an increasing number of meta-analyses are conducted in environmental sciences, and resulting meta-analytic evidence is often used in environmental policies and decision-making. We conducted a survey of recent meta-analyses in environmental sciences and found poor standards of current meta-analytic practice and reporting. For example, only ~ 40% of the 73 reviewed meta-analyses reported heterogeneity (variation among effect sizes beyond sampling error), and publication bias was assessed in fewer than half. Furthermore, although almost all the meta-analyses had multiple effect sizes originating from the same studies, non-independence among effect sizes was considered in only half of the meta-analyses. To improve the implementation of meta-analysis in environmental sciences, we here outline practical guidance for conducting a meta-analysis in environmental sciences. We describe the key concepts of effect size and meta-analysis and detail procedures for fitting multilevel meta-analysis and meta-regression models and performing associated publication bias tests. We demonstrate a clear need for environmental scientists to embrace multilevel meta-analytic models, which explicitly model dependence among effect sizes, rather than the commonly used random-effects models. Further, we discuss how reporting and visual presentations of meta-analytic results can be much improved by following reporting guidelines such as PRISMA-EcoEvo (Preferred Reporting Items for Systematic Reviews and Meta-Analyses for Ecology and Evolutionary Biology). This paper, along with the accompanying online tutorial, serves as a practical guide on conducting a complete set of meta-analytic procedures (i.e., meta-analysis, heterogeneity quantification, meta-regression, publication bias tests and sensitivity analysis) and also as a gateway to more advanced, yet appropriate, methods.

Evidence synthesis is an essential part of science. The method of systematic review provides the most trusted and unbiased way to achieve the synthesis of evidence [ 1 , 2 , 3 ]. Systematic reviews often include a quantitative summary of studies on the topic of interest, referred to as a meta-analysis (for discussion on the definitions of ‘meta-analysis’, see [ 4 ]). The term meta-analysis can also mean a set of statistical techniques for quantitative data synthesis. The methodologies of the meta-analysis were initially developed and applied in medical and social sciences. However, meta-analytic methods are now used in many other fields, including environmental sciences [ 5 , 6 , 7 ]. In environmental sciences, the outcomes of meta-analyses (within systematic reviews) have been used to inform environmental and related policies (see [ 8 ]). Therefore, the reliability of meta-analytic results in environmental sciences is important beyond mere academic interests; indeed, incorrect results could lead to ineffective or sometimes harmful environmental policies [ 8 ].

As in medical and social sciences, environmental scientists frequently use traditional meta-analytic models, namely fixed-effect and random-effects models [ 9 , 10 ]. However, we contend that such models in their original formulation are no longer useful and are often incorrectly used, leading to unreliable estimates and errors. This is mainly because the traditional models assume independence among effect sizes, but almost all primary research papers include more than one effect size, and this non-independence is often not considered (e.g., [ 11 , 12 , 13 ]). Furthermore, previous reviews of published meta-analyses in environmental sciences (hereafter, ‘environmental meta-analyses’) have demonstrated that less than half report or investigate heterogeneity (inconsistency) among effect sizes [ 14 , 15 , 16 ]. Many environmental meta-analyses also do not present any sensitivity analysis, for example, for publication bias (i.e., statistically significant effects being more likely to be published, making collated data unreliable; [ 17 , 18 ]). These issues might have arisen for several reasons, for example, because of no clear conduct guideline for the statistical part of meta-analyses in environmental sciences and rapid developments in meta-analytic methods. Taken together, the field urgently requires a practical guide to implement correct meta-analyses and associated procedures (e.g., heterogeneity analysis, meta-regression, and publication bias tests; cf. [ 19 ]).

To assist environmental scientists in conducting meta-analyses, the aims of this paper are five-fold. First, we provide an overview of the processes involved in a meta-analysis while introducing some key concepts. Second, after introducing the main types of effect size measures, we mathematically describe the two commonly used traditional meta-analytic models, demonstrate their utility, and introduce a practical, multilevel meta-analytic model for environmental sciences that appropriately handles non-independence among effect sizes. Third, we show how to quantify heterogeneity (i.e., consistencies among effect sizes and/or studies) using this model, and then explain such heterogeneity using meta-regression. Fourth, we show how to test for publication bias in a meta-analysis and describe other common types of sensitivity analysis. Fifth, we cover other technical issues relevant to environmental sciences (e.g., scale and phylogenetic dependence) as well as some advanced meta-analytic techniques. In addition, these five aims (sections) are interspersed with two more sections, named ‘Notes’ on: (1) visualisation and interpretation; and (2) reporting and archiving. Some of these sections are accompanied by results from a survey of 73 environmental meta-analyses published between 2019 and 2021; survey results depict current practices and highlight associated problems (for the method of the survey, see Additional file 1 ). Importantly, we provide easy-to-follow implementations of much of what is described below, using the R package, metafor [ 20 ] and other R packages at the webpage ( https://itchyshin.github.io/Meta-analysis_tutorial/ ), which also connects the reader to the wealth of online information on meta-analysis (note that we also provide this tutorial as Additional file 2 ; see also [ 21 ]).

Overview with key concepts

Statistically speaking, we have three general objectives when conducting a meta-analysis [ 12 ]: (1) estimating an overall mean , (2) quantifying consistency ( heterogeneity ) between studies, and (3) explaining the heterogeneity (see Table 1 for the definitions of the terms in italic ). A notable feature of a meta-analysis is that an overall mean is estimated by taking the sampling variance of each effect size into account: a study (effect size) with a low sampling variance (usually based on a larger sample size) is assigned more weight in estimating an overall mean than one with a high sampling variance (usually based on a smaller sample size). However, an overall mean estimate itself is often not informative because one can get the same overall mean estimates in different ways. For example, we may get an overall estimate of zero if all studies have zero effects with no heterogeneity. In contrast, we might also obtain a zero mean across studies that have highly variable effects (e.g., ranging from strongly positive to strongly negative), signifying high heterogeneity. Therefore, quantifying indicators of heterogeneity is an essential part of a meta-analysis, necessary for interpreting the overall mean appropriately. Once we observe non-zero heterogeneity among effect sizes, then, our job is to explain this variation by running meta-regression models, and, at the same time, quantify how much variation is accounted for (often quantified as R 2 ). In addition, it is important to conduct an extra set of analyses, often referred to as publication bias tests , which are a type of sensitivity analysis [ 11 ], to check the robustness of meta-analytic results.

Choosing an effect size measure

In this section, we introduce different kinds of ‘effect size measures’ or ‘effect measures’. In the literature, the term ‘effect size’ is typically used to refer to the magnitude or strength of an effect of interest or its biological interpretation (e.g., environmental significance). Effect sizes can be quantified using a range of measures (for details, see [ 22 ]). In our survey of environmental meta-analyses (Additional file 1 ), the two most commonly used effect size measures are: the logarithm of response ratio, lnRR ([ 23 ]; also known as the ratio of means; [ 24 ]) and standardized mean difference, SMD (often referred to as Hedges’ g or Cohen’s d [ 25 , 26 ]). These are followed by proportion (%) and Fisher’s z -transformation of correlation, or Zr . These four effect measures nearly fit into the three categories, which are named: (1) single-group measures (a statistical summary from one group; e.g., proportion), (2) comparative measures (comparing between two groups e.g., SMD and lnRR), and (3) association measures (relationships between two variables; e.g., Zr ). Table 2 summarizes effect measures that are common or potentially useful for environmental scientists. It is important to note that any measures with sampling variance can become an ‘effect size’. The main reason why SMD, lnRR, Zr, or proportion are popular effect measures is that they are unitless, while a meta-analysis of mean, or mean difference, can only be conducted when all effect sizes have the same unit (e.g., cm, kg).

Table 2 also includes effect measures that are likely to be unfamiliar to environmental scientists; these are effect sizes that characterise differences in the observed variability between samples, (i.e., lnSD, lnCV, lnVR and lnCVR; [ 27 , 28 ]) rather than central tendencies (averages). These dispersion-based effect measures can provide us with extra insights along with average-based effect measures. Although the literature survey showed none of these were used in our sample, these effect sizes have been used in many fields, including agriculture (e.g., [ 29 ]), ecology (e.g., [ 30 ]), evolutionary biology (e.g., [ 31 ]), psychology (e.g., [ 32 ]), education (e.g., [ 33 ]), psychiatry (e.g., [ 34 ]), and neurosciences (e.g. [ 35 ],),. Perhaps, it is not difficult to think of an environmental intervention that can affect not only the mean but also the variance of measurements taken on a group of individuals or a set of plots. For example, environmental stressors such as pesticides and eutrophication are likely to increase variability in biological systems because stress accentuates individual differences in environmental responses (e.g. [ 36 , 37 ],). Such ideas are yet to be tested meta-analytically (cf. [ 38 , 39 ]).

Choosing a meta-analytic model

Fixed-effect and random-effects models.

Two traditional meta-analytic models are called the ‘fixed-effect’ model and the ‘random-effects’ model. The former assumes that all effect sizes (from different studies) come from one population (i.e., they have one true overall mean), while the latter does not have such an assumption (i.e., each study has different overall means or heterogeneity exists among studies; see below for more). The fixed-effect model, which should probably be more correctly referred to as the ‘common-effect’ model, can be written as [ 9 , 10 , 40 ]:

where the intercept, \({\beta }_{0}\) is the overall mean, z j (the response/dependent variable) is the effect size from the j th study ( j  = 1, 2,…, N study ; in this model, N study  = the number of studies = the number of effect sizes), m j is the sampling error, related to the j th sampling variance ( v j ), which is normally distributed with the mean of 0 and the ‘study-specific’ sampling variance, v j (see also Fig.  1 A).

Visualisation of the three statistical models of meta-analysis: A a fixed-effect model (1-level), B a random-effects model (2-level), and C a multilevel model (3-level; see the text for what symbols mean)

The overall mean needs to be estimated and often done so as the weighted average with the weights, \({w}_{j}=1/{v}_{j}\) (i.e., the inverse-variance approach). An important, but sometimes untenable, assumption of meta-analysis is that sampling variance is known. Indeed, we estimate sampling variance, using formulas, as in Table 2 , meaning that vj is submitted by sampling variance estimates (see also section ‘ Scale dependence ’). Of relevance, the use of the inverse-variance approach has been recently criticized, especially for SMD and lnRR [ 41 , 42 ] and we note that the inverse-variance approach using the formulas in Table 2 is one of several different weighting approaches used in meta-analysis (e.g., for adjusted sampling-variance weighing, see [ 43 , 44 ]; for sample-size-based weighting, see [ 41 , 42 , 45 , 46 ]). Importantly, the fixed-effect model assumes that the only source of variation in effect sizes ( z j ) is the effect due to sampling variance (which is inversely proportional to the sample size, n ; Table 2 ).

Similarly, the random-effects model can be expressed as:

where u j is the j th study effect, which is normally distributed with the mean of 0 and the between-study variance, \({\tau }^{2}\) (for different estimation methods, see [ 47 , 48 , 49 , 50 ]), and other notations are the same as in Eq.  1 (Fig.  1 B). Here, the overall mean can be estimated as the weighted average with weights \({w}_{j}=1/\left({\tau }^{2}+{v}_{j}^{2}\right)\) (note that different weighting approaches, mentioned above, are applicable to the random-effects model and some of them are to the multilevel model, introduced below). The model assumes each study has its specific mean, \({b}_{0}+{u}_{j}\) , and (in)consistencies among studies (effect sizes) are indicated by \({\tau }^{2}\) . When \({\tau }^{2}\) is 0 (or not statistically different from 0), the random-effects model simplifies to the fixed-effect model (cf. Equations  1 and 2 ). Given no studies in environmental sciences are conducted in the same manner or even at exactly the same place and time, we should expect different studies to have different means. Therefore, in almost all cases in the environmental sciences, the random-effects model is a more ‘realistic’ model [ 9 , 10 , 40 ]. Accordingly, most environmental meta-analyses (68.5%; 50 out of 73 studies) in our survey used the random-effects model, while only 2.7% (2 of 73 studies) used the fixed-effect model (Additional file 1 ).

Multilevel meta-analytic models

Although we have introduced the random-effects model as being more realistic than the fixed-effect model (Eq.  2 ), we argue that the random-effects model is rather limited and impractical for the environmental sciences. This is because random-effects models, like fixed-effect models, assume all effect sizes ( z j ) to be independent. However, when multiple effect sizes are obtained from a study, these effect sizes are dependent (for more details, see the next section on non-independence). Indeed, our survey showed that in almost all datasets used in environmental meta-analyses, this type of non-independence among effect sizes occurred (97.3%; 71 out of 73 studies, with two studies being unclear, so effectively 100%; Additional file 1 ). Therefore, we propose the simplest and most practical meta-analytic model for environmental sciences as [ 13 , 40 ] (see also [ 51 , 52 ]):

where we explicitly recognize that N effect ( i  = 1, 2,…, N effect ) >  N study ( j  = 1, 2,…, N study ) and, therefore, we now have the study effect (between-study effect), u j[i] (for the j th study and i th effect size) and effect-size level (within-study) effect, e i (for the i th effect size), with the between-study variance, \({\tau }^{2}\) , and with-study variance, \({\sigma }^{2}\) , respectively, and other notations are the same as above. We note that this model (Eq.  3 ) is an extension of the random-effects model (Eq.  2 ), and we refer to it as the multilevel/hierarchical model (used in 7 out of 73 studies: 9.6% [Additional file 1 ]; note that Eq.  3 is also known as a three-level meta-analytic model; Fig.  1 C). Also, environmental scientists who are familiar with (generalised) linear mixed-models may recognize u j (the study effect) as the effect of a random factor which is associated with a variance component, i.e., \({\tau }^{2}\) [ 53 ]; also, e i and m i can be seen as parts of random factors, associated with \({\sigma }^{2}\) and v i (the former is comparable to the residuals, while the latter is sampling variance, specific to a given effect size).

It seems that many researchers are aware of the issue of non-independence so that they often use average effect sizes per study or choose one effect size (at least 28.8%, 21 out of 73 environmental meta-analyses; Additional file 1 ). However, as we discussed elsewhere [ 13 , 40 ], such averaging or selection of one effect size per study dramatically reduces our ability to investigate environmental drivers of variation among effect sizes [ 13 ]. Therefore, we strongly support the use of the multilevel model. Nevertheless, this proposed multilevel model, formulated as Eq.  3 does not usually deal with the issue of non-independence completely, which we elaborate on in the next section.

Non-independence among effect sizes and among sampling errors

When you have multiple effect sizes from a study, there are two broad types and three cases of non-independence (cf. [ 11 , 12 ]): (1) effect sizes are calculated from different cohorts of individuals (or groups of plots) within a study (Fig.  2 A, referred to as ‘shared study identity’), and (2) effects sizes are calculated from the same cohort of individuals (or group of plots; Fig.  2 B, referred to as ‘shared measurements’) or partially from the same individuals and plots, more concretely, sharing individuals and plots from the control group (Fig.  2 C, referred to as ‘shared control group’). The first type of non-independence induces dependence among effect sizes, but not among sampling variances, and the second type leads to non-independence among sampling variances. Many datasets, if not almost all, will have a combination of these three cases (or even are more complex, see the section " Complex non-independence "). Failing to deal with these non-independences will inflate Type 1 error (note that the overall estimate, b 0 is unlikely to be biased, but standard error of b 0 , se( b 0 ), will be underestimated; note that this is also true for all other regression coefficients, e.g., b 1 ; see Table 1 ). The multilevel model (as in Eq.  3 ) only takes care of cases of non-independence that are due to the shared study identity but neither shared measurements nor shared control group.

Visualisation of the three types of non-independence among effect sizes: A due to shared study identities (effect sizes from the same study), B due to shared measurements (effect sizes come from the same group of individuals/plots but are based on different types of measurements), and C due to shared control (effect sizes are calculated using the same control group and multiple treatment groups; see the text for more details)

There are two practical ways to deal with non-independence among sampling variances. The first method is that we explicitly model such dependence using a variance–covariance (VCV) matrix (used in 6 out of 73 studies: 8.2%; Additional file 1 ). Imagine a simple scenario with a dataset of three effect sizes from two studies where two effects sizes from the first study are calculated (partially) using the same cohort of individuals (Fig.  2 B); in such a case, the sampling variance effect, \({m}_{i}\) , as in Eq.  3 , should be written as:

where M is the VCV matrix showing the sampling variances, \({v}_{1\left[1\right]}\) (study 1 and effect size 1), \({v}_{1\left[2\right]}\) (study 1 and effect size 2), and \({v}_{2\left[3\right]}\) (study 2 and effect size 3) in its diagonal, and sampling covariance, \(\rho \sqrt{{v}_{1\left[1\right]}{v}_{1\left[2\right]}}= \rho \sqrt{{v}_{1\left[2\right]}{v}_{1\left[1\right]}}\) in its off-diagonal elements, where \(\rho \) is a correlation between two sampling variances due to shared samples (individuals/plots). Once this VCV matrix is incorporated into the multilevel model (Eq.  3 ), all the types of non-independence, as in Fig.  2 , are taken care of. Table 3 shows formulas for the sampling variance and covariance of the four common effect sizes (SDM, lnRR, proportion and Zr ). For comparative effect measures (Table 2 ), exact covariances can be calculated under the case of ‘shared control group’ (see [ 54 , 55 ]). But this is not feasible for most circumstances because we usually do not know what \(\rho \) should be. Some have suggested fixing this value at 0.5 (e.g., [ 11 ]) or 0.8 (e.g., [ 56 ]); the latter is a more conservative assumption. Or one can run both and use one for the main analysis and the other for sensitivity analysis (for more, see the ‘ Conducting sensitivity analysis and critical appraisal " section).

The second method overcomes this very issue of unknown \(\rho \) by approximating average dependence among sampling variance (and effect sizes) from the data and incorporating such dependence to estimate standard errors (only used in 1 out of 73 studies; Additional file 1 ). This method is known as ‘robust variance estimation’, RVE, and the original estimator was proposed by Hedges and colleagues in 2010 [ 57 ]. Meta-analysis using RVE is relatively new, and this method has been applied to multilevel meta-analytic models only recently [ 58 ]. Note that the random-effects model (Eq.  2 ) and RVE could correctly model both types of non-independence. However, we do not recommend the use of RVE with Eq.  2 because, as we will later show, estimating \({\sigma }^{2}\) as well as \({\tau }^{2}\) will constitute an important part of understanding and gaining more insights from one’s data. We do not yet have a definite recommendation on which method to use to account for non-independence among sampling errors (using the VCV matrix or RVE). This is because no simulation work in the context of multilevel meta-analysis has been done so far, using multilevel meta-analyses [ 13 , 58 ]. For now, one could use both VCV matrices and RVE in the same model [ 58 ] (see also [ 21 ]).

Quantifying and explaining heterogeneity

Measuring consistencies with heterogeneity.

As mentioned earlier, quantifying heterogeneity among effect sizes is an essential component of any meta-analysis. Yet, our survey showed only 28 out of 73 environmental meta-analyses (38.4%; Additional file 1 ) report at least one index of heterogeneity (e.g., \({\tau }^{2}\) , Q , and I 2 ). Conventionally, the presence of heterogeneity is tested by Cochrane’s Q test. However, Q (often noted as Q T or Q total ), and its associated p value, are not particularly informative: the test does not tell us about the extent of heterogeneity (e.g. [ 10 ],), only whether heterogeneity is zero or not (when p  < 0.05). Therefore, for environmental scientists, we recommend two common ways of quantifying heterogeneity from a meta-analytic model: absolute heterogeneity measure (i.e., variance components, \({\tau }^{2}\) and \({\sigma }^{2}\) ) and relative heterogeneity measure (i.e., I 2 ; see also the " Notes on visualisation and interpretation " section for another way of quantifying and visualising heterogeneity at the same time, using prediction intervals; see also [ 59 ]). We have already covered the absolute measure (Eqs.  2 & 3 ), so here we explain I 2 , which ranges from 0 to 1 (for some caveats for I 2 , see [ 60 , 61 ]). The heterogeneity measure, I 2 , for the random-effect model (Eq.  2 ) can be written as:

Where \(\overline{v}\) is referred to as the typical sampling variance (originally this is called ‘within-study’ variance, as in Eq.  2 , and note that in this formulation, within-study effect and the effect of sampling error is confounded; see [ 62 , 63 ]; see also [ 64 ]) and the other notations are as above. As you can see from Eq.  5 , we can interpret I 2 as relative variation due to differences between studies (between-study variance) or relative variation not due to sampling variance.

By seeing I 2 as a type of interclass correlation (also known as repeatability [ 65 ],), we can generalize I 2 to multilevel models. In the case of Eq.  3 ([ 40 , 66 ]; see also [ 52 ]), we have:

Because we can have two more I 2 , Eq.  7 is written as \({I}_{total}^{2}\) ; these other two are \({I}_{study}^{2}\) and \({I}_{effect}^{2}\) , respectively:

\({I}_{total}^{2}\) represents relative variance due to differences both between and within studies (between- and within-study variance) or relative variation not due to sampling variance, while \({I}_{study}^{2}\) is relative variation due to differences between studies, and \({I}_{effect}^{2}\) is relative variation due to differences within studies (Fig.  3 A). Once heterogeneity is quantified (note almost all data will have non-zero heterogeneity and an earlier meta-meta-analysis suggests in ecology, we have on average, I 2 close to 90% [ 66 ]), it is time to fit a meta-regression model to explain the heterogeneity. Notably, the magnitude of \({I}_{study}^{2}\) (and \({\tau }^{2}\) ) and \({I}_{effect}^{2}\) (and \({\sigma }^{2}\) ) can already inform you which predictor variable (usually referred to as ‘moderator’) is likely to be important, which we explain in the next section.

Visualisation of variation (heterogeneity) partitioned into different variance components: A quantifying different types of I 2 from a multilevel model (3-level; see Fig.  1 C) and B variance explained, R 2 , by moderators. Note that different levels of variances would be explained, depending on which level a moderator belongs to (study level and effect-size level)

Explaining variance with meta-regression

We can extend the multilevel model (Eq.  3 ) to a meta-regression model with one moderator (also known as predictor, independent, explanatory variable, or fixed factor), as below:

where \({\beta }_{1}\) is a slope of the moderator ( x 1 ), \({x}_{1j\left[i\right]}\) denotes the value of x 1 , corresponding to the j th study (and the i th effect sizes). Equation ( 10 ) (meta-regression) is comparable to the simplest regression with the intercept ( \({\beta }_{0}\) ) and slope ( \({\beta }_{1}\) ). Notably, \({x}_{1j\left[i\right]}\) differs between studies and, therefore, it will mainly explain the variance component, \({\tau }^{2}\) (which relates to \({I}_{study}^{2}\) ). On the other hand, if noted like \({x}_{1i}\) , this moderator would vary within studies or at the level of effect sizes, therefore, explaining \({\sigma }^{2}\) (relating to \({I}_{effect}^{2}\) ). Therefore, when \({\tau }^{2}\) ( \({I}_{study}^{2}\) ), or \({\sigma }^{2}\) ( \({I}_{effect}^{2}\) ), is close to zero, there will be little point fitting a moderator(s) at the level of studies, or effect sizes, respectively.

As in multiple regression, we can have multiple (multi-moderator) meta-regression, which can be written as:

where \(\sum_{h=1}^{q}{\beta }_{h}{x}_{h\left[i\right]}\) denotes the sum of all the moderator effects, with q being the number of slopes (staring with h  = 1). We note that q is not necessarily the number of moderators. This is because when we have a categorical moderator, which is common, with more than two levels (e.g., method A, B & C), the fixed effect part of the formula is \({\beta }_{0}+{\beta }_{1}{x}_{1}+{\beta }_{2}{x}_{2}\) , where x 1 and x 2 are ‘dummy’ variables, which code whether the i th effect size belongs to, for example, method B or C, with \({\beta }_{1}\) and \({\beta }_{2}\) being contrasts between A and B and between A and C, respectively (for more explanations of dummy variables, see our tutorial page [ https://itchyshin.github.io/Meta-analysis_tutorial/ ]; also see [ 67 , 68 ]). Traditionally, researchers conduct separate meta-analyses per different groups (known as ‘sub-group analysis’), but we prefer a meta-regression approach with a categorical variable, which is statistically more powerful [ 40 ]. Also, importantly, what can be used as a moderator(s) is very flexible, including, for example, individual/plot characteristics (e.g., age, location), environmental factors (e.g., temperature), methodological differences between studies (e.g., randomization), and bibliometric information (e.g., publication year; see more in the section ‘Checking for publication bias and robustness’). Note that moderators should be decided and listed a priori in the meta-analysis plan (i.e., a review protocol or pre-registration).

As with meta-analysis, the Q -test ( Q m or Q moderator ) is often used to test the significance of the moderator(s). To complement this test, we can also quantify variance explained by the moderator(s) using R 2 . We can define R 2 using Eq. ( 11 ) as:

where R 2 is known as marginal R 2 (sensu [ 69 , 70 ]; cf. [ 71 ]), \({f}^{2}\) is the variance due to the moderator(s), and \({(f}^{2}+{\tau }^{2}+{\sigma }^{2})\) here equals to \(({\tau }^{2}+{\sigma }^{2})\) in Eq.  7 , as \({f}^{2}\) ‘absorbs’ variance from \({\tau }^{2}\) and/or \({\sigma }^{2}\) . We can compare the similarities and differences in Fig.  3 B where we denote a part of \({f}^{2}\) originating from \({\tau }^{2}\) as \({f}_{study}^{2}\) while \({\sigma }^{2}\) as \({f}_{effect}^{2}\) . In a multiple meta-regression model, we often want to find a model with the ‘best’ or an adequate set of predictors (i.e., moderators). R 2 can potentially help such a model selection process. Yet, methods based on information criteria (such as Akaike information criterion, AIC) may be preferable. Although model selection based on the information criteria is beyond the scope of the paper, we refer the reader to relevant articles (e.g., [ 72 , 73 ]), and we show an example of this procedure in our online tutorial ( https://itchyshin.github.io/Meta-analysis_tutorial/ ).

Notes on visualisation and interpretation

Visualization and interpretation of results is an essential part of a meta-analysis [ 74 , 75 ]. Traditionally, a forest plot is used to display the values and 95% of confidence intervals (CIs) for each effect size and the overall effect and its 95% CI (the diamond symbol is often used, as shown in Fig.  4 A). More recently, adding a 95% prediction interval (PI) to the overall estimate has been strongly recommended because 95% PIs show a predicted range of values in which an effect size from a new study would fall, assuming there is no sampling error [ 76 ]. Here, we think that examining the formulas for 95% CIs and PIs for the overall mean (from Eq.  3 ) is illuminating:

where \({t}_{df\left[\alpha =0.05\right]}\) denotes the t value with the degree of freedom, df , at 97.5 percentile (or \(\alpha =0.05\) ) and other notations are as above. In a meta-analysis, it has been conventional to use z value 1.96 instead of \({t}_{df\left[\alpha =0.05\right]}\) , but simulation studies have shown the use of t value over z value reduces Type 1 errors under many scenarios and, therefore, is recommended (e.g., [ 13 , 77 ]). Also, it is interesting to note that by plotting 95% PIs, we can visualize heterogeneity as Eq.  15 includes \({\tau }^{2}\) and \({\sigma }^{2}\) .

Different types of plots useful for a meta-analysis using data from Midolo et al. [ 133 ]: A a typical forest plot with the overall mean shown as a diamond at the bottom (20 effect sizes from 20 studies are used), B a caterpillar plot (100 effect sizes from 24 studies are used), C an orchard plot of categorical moderator with seven levels (all effect sizes are used), and D a bubble plot of a continuous moderator. Note that the first two only show confidence intervals, while the latter two also show prediction intervals (see the text for more details)

A ‘forest’ plot can become quickly illegible as the number of studies (effect sizes) becomes large, so other methods of visualizing the distribution of effect sizes have been suggested. Some suggested to present a ‘caterpillar’ plot, which is a version of the forest plot, instead (Fig.  4 B; e.g., [ 78 ]). We here recommend an ‘orchard’ plot, as it can present results across different groups (or a result of meta-regression with a categorical variable), as shown in Fig.  4 C [ 78 ]. For visualization of a continuous variable, we suggest what is called a ‘bubble’ plot, shown in Fig.  4 D. Visualization not only helps us interpret meta-analytic results, but can also help to identify something we may not see from statistical results, such as influential data points and outliers that could threaten the robustness of our results.

Checking for publication bias and robustness

Detecting and correcting for publication bias.

Checking for and adjusting for any publication bias is necessary to ensure the validity of meta-analytic inferences [ 79 ]. However, our survey showed almost half of the environmental meta-analyses (46.6%; 34 out of 73 studies; Additional file 1 ) neither tested for nor corrected for publication bias (cf. [ 14 , 15 , 16 ]). The most popular methods used were: (1) graphical tests using funnel plots (26 studies; 35.6%), (2) regression-based tests such as Egger regression (18 studies; 24.7%), (3) Fail-safe number tests (12 studies; 16.4%), and (4) trim-and-fill tests (10 studies; 13.7%). We recently showed that these methods are unsuitable for datasets with non-independent effect sizes, with the exception of funnel plots [ 80 ] (for an example of funnel plots, see Fig.  5 A). This is because these methods cannot deal with non-independence in the same way as the fixed-effect and random-effects models. Here, we only introduce a two-step method for multilevel models that can both detect and correct for publication bias [ 80 ] (originally proposed by [ 81 , 82 ]), more specifically, the “small study effect” where an effect size value from a small-sample-sized study can be much larger in magnitude than a ‘true’ effect [ 83 , 84 ]. This method is a simple extension of Egger’s regression [ 85 ], which can be easily implemented by using Eq.  10 :

where \({\widetilde{n}}_{i}\) is known as effective sample size; for Zr and proportion it is just n i , and for SMD and lnRR, it is \({n}_{iC}{n}_{iT}/\left({n}_{iC}+{n}_{iT}\right)\) , as in Table 2 . When \({\beta }_{1}\) is significant, we conclude there exists a small-study effect (in terms of a funnel plot, this is equivalent to significant funnel asymmetry). Then, we fit Eq.  17 and we look at the intercept \({\beta }_{0}\) , which will be a bias-corrected overall estimate [note that \({\beta }_{0}\) in Eq. ( 16 ) provides less accurate estimates when non-zero overall effects exist [ 81 , 82 ]; Fig.  5 B]. An intuitive explanation of why \({\beta }_{0}\) (Eq.  17 ) is the ‘bias-corrected’ estimate is that the intercept represents \(1/\widetilde{{n}_{i}}=0\) (or \(\widetilde{{n}_{i}}=\infty \) ); in other words, \({\beta }_{0}\) is the estimate of the overall effect when we have a very large (infinite) sample size. Of note, appropriate bias correction requires a selection-mode-based approach although such an approach is yet to be available for multilevel meta-analytic models [ 80 ].

Different types of plots for publication bias tests: A a funnel plot using model residuals, showing a funnel (white) that shows the region of statistical non-significance (30 effect sizes from 30 studies are used; note that we used the inverse of standard errors for the y -axis, but for some effect sizes, sample size or ‘effective’ sample size may be more appropriate), B a bubble plot visualising a multilevel meta-regression that tests for the small study effect (note that the slope was non-significant: b  = 0.120, 95% CI = [− 0.095, 0.334]; all effect sizes are used), and C a bubble plot visualising a multilevel meta-regression that tests for the decline effect (the slope was non-significant: b  = 0.003, 95%CI = [− 0.002, 0.008])

Conveniently, this proposed framework can be extended to test for another type of publication bias, known as time-lag bias, or the decline effect, where effect sizes tend to get closer to zero over time, as larger or statistically significant effects are published more quickly than smaller or non-statistically significant effects [ 86 , 87 ]. Again, a decline effect can be statistically tested by adding year to Eq. ( 3 ):

where \(c\left(yea{r}_{j\left[i\right]}\right)\) is the mean-centred publication year of a particular study (study j and effect size i ); this centring makes the intercept \({\beta }_{0}\) meaningful, representing the overall effect estimate at the mean value of publication years (see [ 68 ]). When the slope is significantly different from 0, we deem that we have a decline effect (or time-lag bias; Fig.  5 C).

However, there may be some confounding moderators, which need to be modelled together. Indeed, Egger’s regression (Eqs.  16 and 17 ) is known to detect the funnel asymmetry when there is little heterogeneity; this means that we need to model \(\sqrt{1/{\widetilde{n}}_{i}}\) with other moderators that account for heterogeneity. Given this, we probably should use a multiple meta-regression model, as below:

where \(\sum_{h=3}^{q}{\beta }_{h}{x}_{h\left[i\right]}\) is the sum of the other moderator effects apart from the small-study effect and decline effect, and other notations are as above (for more details see [ 80 ]). We need to carefully consider which moderators should go into Eq.  19 (e.g., fitting all moderators or using an AIC-based model selection method; see [ 72 , 73 ]). Of relevance, when running complex models, some model parameters cannot be estimated well, or they are not ‘identifiable’ [ 88 ]. This is especially so for variance components (random-effect part) rather than regression coeffects (fixed-effect part). Therefore, it is advisable to check whether model parameters are all identifiable, which can be checked using the profile function in metafor (for an example, see our tutorial webpage [ https://itchyshin.github.io/Meta-analysis_tutorial/ ]).

Conducting sensitivity analysis and critical appraisal

Sensitivity analysis explores the robustness of meta-analytic results by running a different set of analyses from the original analysis, and comparing the results (note that some consider publication bias tests a part of sensitivity analysis; [ 11 ]). For example, we might be interested in assessing how robust results are to the presence of influential studies, to the choice of method for addressing non-independence, or weighting effect sizes. Unfortunately, in our survey, only 37% of environmental meta-analyses (27 out of 73) conducted sensitivity analysis (Additional file 1 ). There are two general and interrelated ways to conduct sensitivity analyses [ 73 , 89 , 90 ]. The first one is to take out influential studies (e.g., outliers) and re-run meta-analytic and meta-regression models. We can also systematically take each effect size out and run a series of meta-analytic models to see whether any resulting overall effect estimates are different from others; this method is known as ‘leave-one-out’, which is considered less subjective and thus recommended.

The second way of approaching sensitivity analysis is known as subset analysis, where a certain group of effect sizes (studies) will be excluded to re-run the models without this group of effect sizes. For example, one may want to run an analysis without studies that did not randomize samples. Yet, as mentioned earlier, we recommend using meta-regression (Eq.  13 ) with a categorical variable of randomization status (‘randomized’ or ‘not randomized’), to statistically test for an influence of moderators. It is important to note that such tests for risk of bias (or study quality) can be considered as a way of quantitatively evaluating the importance of study features that were noted at the stage of critical appraisal, which is an essential part of any systematic review (see [ 11 , 91 ]). In other words, we can use meta-regression or subset analysis to quantitatively conduct critical appraisal using (study-level) moderators that code, for example, blinding, randomization, and selective reporting. Despite the importance of critical appraisal ([ 91 ]), only 4 of 73 environmental meta-analyses (5.6%) in our survey assessed the risk of bias in each study included in a meta-analysis (i.e., evaluating a primary study in terms of the internal validity of study design and reporting; Additional file 1 ). We emphasize that critically appraising each paper or checking them for risk of bias is an extremely important topic. Also, critical appraisal is not restricted to quantitative synthesis. Therefore, we do not cover any further in this paper for more, see [ 92 , 93 ]).

Notes on transparent reporting and open archiving

For environmental systematic reviews and maps, there are reporting guidelines called RepOrting standards for Systematic Evidence Syntheses in environmental research, ROSES [ 94 ] and synthesis assessment checklist, the Collaboration for Environmental Evidence Synthesis Appraisal Tool (CEESAT; [ 95 ]). However, these guidelines are somewhat limited in terms of reporting quantitative synthesis because they cover only a few core items. These two guidelines are complemented by the Preferred Reporting Items for Systematic Reviews and Meta-Analyses for Ecology and Evolutionary Biology (PRISMA-EcoEvo; [ 96 ]; cf. [ 97 , 98 ]), which provides an extended set of reporting items covering what we have described above. Items 20–24 from PRISMA-EcoEvo are most relevant: these items outline what should be reported in the Methods section: (i) sample sizes and study characteristics, (ii) meta-analysis, (iii) heterogeneity, (iv) meta-regression and (v) outcomes of publication bias and sensitivity analysis (see Table 4 ). Our survey, as well as earlier surveys, suggest there is a large room for improvement in the current practice ([ 14 , 15 , 16 ]). Incidentally, the orchard plot is well aligned with Item 20, as this plot type shows both the number of effect sizes and studies for different groups (Fig.  4 C). Further, our survey of environmental meta-analyses highlighted the poor standards of data openness (with 24 studies sharing data: 32.9%) and code sharing (7 studies: 29.2%; Additional file 1 ). Environmental scientists must archive their data as well as their analysis code in accordance with the FAIR principles (Findable, Accessible, Interoperable, and Reusable [ 99 ]) using dedicated depositories such as Dryad, FigShare, Open Science Framework (OSF), Zenodo or others (cf. [ 100 , 101 ]), preferably not on publisher’s webpages (as paywall may block access). However, archiving itself is not enough; data requires metadata (detailed descriptions) and the code needs to also be FAIR [ 102 , 103 ].

Other relevant and advanced issues

Scale dependence.

The issue of scale dependence is a unique yet widespread problem in environmental sciences (see [ 7 , 104 ]); our literature survey indicated three quarters of the environmental meta-analyses (56 out of 73 studies) have inferences that are potentially vulnerable to scale-dependence [ 105 ]. For example, studies that set out to compare group means in biodiversity measures, such as species richness, can vary as a function of the scale (size) of the sampling unit. When the unit of replication is a plot (not an individual animal or plant), the aerial size of a plot (e.g., 100 cm 2 or 1 km 2 ) will affect both the precision and accuracy of effect size estimates (e.g., lnRR and SMD). In general, a study with larger plots might have more accurately estimated species richness differences, but less precisely than a study with smaller plots and greater replication. Lower replication means that our sampling variance estimates are likely to be misestimated, and the study with larger plots will generally have less weight than the study with smaller plots, due to higher sampling variance. Inaccurate variance estimates in little-replicated ecological studies are known to cause an accumulating bias in precision-weighted meta-analysis, requiring correction [ 43 ]. To assess the potential for scale-dependence, it is recommended that analysts test for possible covariation among plot size, replication, variances, and effect sizes [ 104 ]. If detected, analysts should use an effect size measure that is less sensitive to scale dependence (lnRR), and could use the size of a plot as a moderator in meta-regression, or alternatively, they consider running an unweighted model ([ 7 ]; note that only 12%, 9 out of 73 studies, accounted for sampling area in some way; Additional file 1 ).

• Missing data

In many fields, meta-analytic data almost always encompass missing values see [ 106 , 107 , 108 ]. Broadly, we have two types of missing data in meta-analyses [ 109 , 110 ]: (1) missing data in standard deviations or sample sizes, associated with means, preventing effect size calculations (Table 2 ), and (2) missing data in moderators. There are several solutions for both types. The best, and first to try, should be contacting the authors. If this fails, we can potentially ‘impute’ missing data. Single imputation methods using the strong correlation between standard deviation and mean values (known as mean–variance relationship) are available, although single imputation can lead to Type I error [ 106 , 107 ] (see also [ 43 ]) because we do not model the uncertainty of imputation itself. Contrastingly, multiple imputation, which creates multiple versions of imputed datasets, incorporates such uncertainty. Indeed, multiple imputation is a preferred and proven solution for missing data in effect sizes and moderators [ 109 , 110 ]. Yet, correct implementation can be challenging (see [ 110 ]). What we require now is an automated pipeline of merging meta-analysis and multiple imputation, which accounts for imputation uncertainty, although it may be challenging for complex meta-analytic models. Fortunately, however, for lnRR, there is a series of new methods that can perform better than the conventional method and which can deal with missing SDs [ 44 ]; note that these methods do not deal with missing moderators. Therefore, where applicable, we recommend these new methods, until an easy-to-implement multiple imputation workflow arrives.

Complex non-independence

Above, we have only dealt with the model that includes study identities as a clustering/grouping (random) factor. However, many datasets are more complex, with potentially more clustering variables in addition to the study identity. It is certainly possible that an environmental meta-analysis contains data from multiple species. Such a situation creates an interesting dependence among effect sizes from different species, known as phylogenetic relatedness, where closely related species are more likely to be similar in effect sizes compared to distantly related ones (e.g., mice vs. rats and mice vs. sparrows). Our multilevel model framework is flexible and can accommodate phylogenetic relatedness. A phylogenetic multilevel meta-analytic model can be written as [ 40 , 111 , 112 ]:

where \({a}_{k\left[i\right]}\) is the phylogenetic (species) effect for the k th species (effect size i ; N effect ( i  = 1, 2,…, N effect ) >  N study ( j  = 1, 2,…, N study ) >  N species ( k  = 1, 2,…, N species )), normally distributed with \({\omega }^{2}{\text{A}}\) where is the phylogenetic variance and A is a correlation matrix coding how close each species are to each other and \({\omega }^{2}\) is the phylogenetic variance, \({s}_{k\left[i\right]}\) is the non-phylogenetic (species) effect for the k th species (effect size i ), normally distributed with the variance of \({\gamma }^{2}\) (the non-phylogenetic variance), and other notations are as above. It is important to realize that A explicitly models relatedness among species, and we do need to provide this correlation matrix, using a distance relationship usually derived from a molecular-based phylogenetic tree (for more details, see [ 40 , 111 , 112 ]). Some may think that the non-phylogenetic term ( \({s}_{k\left[i\right]}\) ) is unnecessary or redundant because \({s}_{k\left[i\right]}\) and the phylogenetic term ( \({a}_{k\left[i\right]}\) ) are both modelling variance at the species level. However, a simulation recently demonstrated that failing to have the non-phylogenetic term ( \({s}_{k\left[i\right]}\) ) will often inflate the phylogenetic variance \({\omega }^{2}\) , leading to an incorrect conclusion that there is a strong phylogenetic signal (as shown in [ 112 ]). The non-phylogenetic variance ( \({\gamma }^{2}\) ) arises from, for example, ecological similarities among species (herbivores vs. carnivores or arboreal vs. ground-living) not phylogeny [ 40 ].

Like phylogenetic relatedness, effect sizes arising from closer geographical locations are likely to be more correlated [ 113 ]. Statistically, spatial correlation can be also modelled in a manner analogous to phylogenetic relatedness (i.e., rather than a phylogenetic correlation matrix, A , we fit a spatial correlation matrix). For example, Maire and colleagues [ 114 ] used a meta-analytic model with spatial autocorrelation to investigate the temporal trends of fish communities in the network of rivers in France. We note that a similar argument can be made for temporal correlation, but in many cases, temporal correlations could be dealt with, albeit less accurately, as a special case of ‘shared measurements’, as in Fig.  2 . An important idea to take away is that one can model different, if not all, types of non-independence as the random factor(s) in a multilevel model.

Advanced techniques

Here we touch upon five advanced meta-analytic techniques with potential utility for environmental sciences, providing relevant references so that interested readers can obtain more information on these advanced topics. The first one is the meta-analysis of magnitudes, or absolute values (effect sizes), where researchers may be interested in deviations from 0, rather than the directionality of the effect [ 115 ]. For example, Cohen and colleagues [ 116 ] investigated absolute values of phenological responses, as they were concerned with the magnitudes of changes in phenology rather than directionality.

The second method is the meta-analysis of interaction where our focus is on synthesizing the interaction effect of, usually, 2 × 2 factorial design (e.g., the effect of two simultaneous environmental stressors [ 54 , 117 , 118 ]; see also [ 119 ]). Recently, Siviter and colleagues [ 120 ] showed that agrochemicals interact synergistically (i.e., non-additively) to increase the mortality of bees; that is, two agrochemicals together caused more mortality than the sum of mortalities of each chemical.

Third, network meta-analysis has been heavily used in medical sciences; network meta-analysis usually compares different treatments in relation to placebo and ranks these treatments in terms of effectiveness [ 121 ]. The very first ‘environmental’ network meta-analysis, as far as we know, investigated the effectives of ecosystem services among different land types [ 122 ].

Fourth, a multivariate meta-analysis is where one can model two or more different types of effect sizes with the estimation of pair-wise correlations between different effect sizes. The benefit of such an approach is known as the ‘borrowing of strength’, where the error of fixed effects (moderators; e.g., b 0 and b 1 ) can be reduced when different types of effect sizes are correlated (i.e., se ( b 0 ) and se ( b 1 ) can be smaller [ 123 ]) For example, it is possible for lnRR (differences in mean) and lnVR (differences in SDs) to be modelled together (cf. [ 124 ]).

Fifth, as with network meta-analysis, there has been a surge in the use of ‘individual participants data’, called ‘IPD meta-analysis’, in medical sciences [ 125 , 126 ]. The idea of IPD meta-analysis is simple—rather than using summary statistics reported in papers (sample means and variances), we directly use raw data from all studies. We can either model raw data using one complex multilevel (hierarchical) model (one-step method) or calculate statistics for each study and use a meta-analysis (two-step method; note that both methods will usually give the same results). Study-level random effects can be incorporated to allow the response variable of interest to vary among studies, and overall effects correspond to fixed, population-level estimates. The use of IPD or ‘full-data analyses’ has also surged in ecology, aided by open-science policies that encourage the archival of raw data alongside articles, and initiatives that synthesise raw data (e.g., PREDICTS [ 127 ], BioTime [ 128 ]). In health disciplines, such meta-analyses are considered the ‘gold standard’ [ 129 ], owing to their potential for resolving issues regarding study-specific designs and confounding variation, and it is unclear whether and how they might resolve issues such as scale dependence in environmental meta-analyses [ 104 , 130 ].

Conclusions

In this article, we have attempted to describe the most practical ways to conduct quantitative synthesis, including meta-analysis, meta-regression, and publication bias tests. In addition, we have shown that there is much to be improved in terms of meta-analytic practice and reporting via a survey of 73 recent environmental meta-analyses. Such improvements are urgently required, especially given the potential influence that environmental meta-analyses can have on policies and decision-making [ 8 ]. So often, meta-analysts have called for better reporting of primary research (e.g. [ 131 , 132 ]), and now this is the time to raise the standards of reporting in meta-analyses. We hope our contribution will help to catalyse a turning point for better practice in quantitative synthesis in environmental sciences. We remind the reader most of what is described is implemented in the R environment on our tutorial webpage and researchers can readily use the proposed models and techniques ( https://itchyshin.github.io/Meta-analysis_tutorial/ ). Finally, meta-analytic techniques are always developing and improving. It is certainly possible that in the future, our proposed models and related methods will become dated, just as the traditional fixed-effect and random-effects models already are. Therefore, we must endeavour to be open-minded to new ways of doing quantitative research synthesis in environmental sciences.

Availability of data and materials

All data and material are provided as additional files.

Higgins JP, Thomas JE, Chandler JE, Cumpston ME, Li TE, Page MJ, Welch VA. Cochrane handbook for systematic reviews of interventions. 2nd ed. Chichester: Wikey; 2019.

Book   Google Scholar  

Cooper HM, Hedges LV, Valentine JC. The handbook of research synthesis and meta-analysis . 3rd ed. New York: Russell Sage Foundation; 2019.

Google Scholar  

Schmid CH, Stijnen TE, White IE. Handbook of meta-analysis. 1st ed. Boca Ranton: CRC; 2021.

Vetter D, Rucker G, Storch I. Meta-analysis: a need for well-defined usage in ecology and conservation biology. Ecosphere. 2013;4(6):1.

Article   Google Scholar  

Koricheva J, Gurevitch J, Mengersen K, editors. Handbook of meta-analysis in ecology and evolution. Princeton: Princeton Univesity Press; 2017.

Gurevitch J, Koricheva J, Nakagawa S, Stewart G. Meta-analysis and the science of research synthesis. Nature. 2018;555(7695):175–82.

Article   CAS   Google Scholar  

Spake R, Doncaster CP. Use of meta-analysis in forest biodiversity research: key challenges and considerations. Forest Ecol Manag. 2017;400:429–37.

Bilotta GS, Milner AM, Boyd I. On the use of systematic reviews to inform environmental policies. Environ Sci Policy. 2014;42:67–77.

Hedges LV, Vevea JL. Fixed- and random-effects models in meta-analysis. Psychol Methods. 1998;3(4):486–504.

Borenstein M, Hedges LV, Higgins JPT, Rothstein H. Introduction to meta-analysis. 2nd ed. Chichester: Wiley; 2021.

Noble DWA, Lagisz M, Odea RE, Nakagawa S. Nonindependence and sensitivity analyses in ecological and evolutionary meta-analyses. Mol Ecol. 2017;26(9):2410–25.

Nakagawa S, Noble DWA, Senior AM, Lagisz M. Meta-evaluation of meta-analysis: ten appraisal questions for biologists. Bmc Biol. 2017;15:1.

Nakagawa S, Senior AM, Viechtbauer W, Noble DWA. An assessment of statistical methods for nonindependent data in ecological meta-analyses: comment. Ecology. 2022;103(1): e03490.

Romanelli JP, Meli P, Naves RP, Alves MC, Rodrigues RR. Reliability of evidence-review methods in restoration ecology. Conserv Biol. 2021;35(1):142–54.

Koricheva J, Gurevitch J. Uses and misuses of meta-analysis in plant ecology. J Ecol. 2014;102(4):828–44.

O’Leary BC, Kvist K, Bayliss HR, Derroire G, Healey JR, Hughes K, Kleinschroth F, Sciberras M, Woodcock P, Pullin AS. The reliability of evidence review methodology in environmental science and conservation. Environ Sci Policy. 2016;64:75–82.

Rosenthal R. The “file drawer problem” and tolerance for null results. Psychol Bull. 1979;86(3):638–41.

Nakagawa S, Lagisz M, Jennions MD, Koricheva J, Noble DWA, Parker TH, Sánchez-Tójar A, Yang Y, O’Dea RE. Methods for testing publication bias in ecological and evolutionary meta-analyses. Methods Ecol Evol. 2022;13(1):4–21.

Cheung MWL. A guide to conducting a meta-analysis with non-independent effect sizes. Neuropsychol Rev. 2019;29(4):387–96.

Viechtbauer W. Conducting meta-analyses in R with the metafor package. J Stat Softw. 2010;36(3):1–48.

Yang Y, Macleod M, Pan J, Lagisz M, Nakagawa S. Advanced methods and implementations for the meta-analyses of animal models: current practices and future recommendations. Neurosci Biobehav Rev. 2022. https://doi.org/10.1016/j.neubiorev.2022.105016:105016 .

Nakagawa S, Cuthill IC. Effect size, confidence interval and statistical significance: a practical guide for biologists. Biol Rev. 2007;82(4):591–605.

Hedges LV, Gurevitch J, Curtis PS. The meta-analysis of response ratios in experimental ecology. Ecology. 1999;80(4):1150–6.

Friedrich JO, Adhikari NKJ, Beyene J. The ratio of means method as an alternative to mean differences for analyzing continuous outcome variables in meta-analysis: A simulation study. BMC Med Res Methodol. 2008;8:5.

Hedges L, Olkin I. Statistical methods for meta-analysis. New York: Academic Press; 1985.

Cohen J. Statistical power analysis for the beahvioral sciences. 2nd ed. Hillsdale: Lawrence Erlbaum; 1988.

Senior AM, Viechtbauer W, Nakagawa S. Revisiting and expanding the meta-analysis of variation: the log coefficient of variation ratio. Res Synth Methods. 2020;11(4):553–67.

Nakagawa S, Poulin R, Mengersen K, Reinhold K, Engqvist L, Lagisz M, Senior AM. Meta-analysis of variation: ecological and evolutionary applications and beyond. Methods Ecol Evol. 2015;6(2):143–52.

Knapp S, van der Heijden MGA. A global meta-analysis of yield stability in organic and conservation agriculture. Nat Commun. 2018;9:3632.

Porturas LD, Anneberg TJ, Cure AE, Wang SP, Althoff DM, Segraves KA. A meta-analysis of whole genome duplication and theeffects on flowering traits in plants. Am J Bot. 2019;106(3):469–76.

Janicke T, Morrow EH. Operational sex ratio predicts the opportunity and direction of sexual selection across animals. Ecol Lett. 2018;21(3):384–91.

Chamberlain R, Brunswick N, Siev J, McManus IC. Meta-analytic findings reveal lower means but higher variances in visuospatial ability in dyslexia. Brit J Psychol. 2018;109(4):897–916.

O’Dea RE, Lagisz M, Jennions MD, Nakagawa S. Gender differences in individual variation in academic grades fail to fit expected patterns for STEM. Nat Commun. 2018;9:3777.

Brugger SP, Angelescu I, Abi-Dargham A, Mizrahi R, Shahrezaei V, Howes OD. Heterogeneity of striatal dopamine function in schizophrenia: meta-analysis of variance. Biol Psychiat. 2020;87(3):215–24.

Usui T, Macleod MR, McCann SK, Senior AM, Nakagawa S. Meta-analysis of variation suggests that embracing variability improves both replicability and generalizability in preclinical research. Plos Biol. 2021;19(5): e3001009.

Hoffmann AA, Merila J. Heritable variation and evolution under favourable and unfavourable conditions. Trends Ecol Evol. 1999;14(3):96–101.

Wood CW, Brodie ED 3rd. Environmental effects on the structure of the G-matrix. Evolution. 2015;69(11):2927–40.

Hillebrand H, Donohue I, Harpole WS, Hodapp D, Kucera M, Lewandowska AM, Merder J, Montoya JM, Freund JA. Thresholds for ecological responses to global change do not emerge from empirical data. Nat Ecol Evol. 2020;4(11):1502.

Yang YF, Hillebrand H, Lagisz M, Cleasby I, Nakagawa S. Low statistical power and overestimated anthropogenic impacts, exacerbated by publication bias, dominate field studies in global change biology. Global Change Biol. 2022;28(3):969–89.

Nakagawa S, Santos ESA. Methodological issues and advances in biological meta-analysis. Evol Ecol. 2012;26(5):1253–74.

Bakbergenuly I, Hoaglin DC, Kulinskaya E. Estimation in meta-analyses of response ratios. BMC Med Res Methodol. 2020;20(1):1.

Bakbergenuly I, Hoaglin DC, Kulinskaya E. Estimation in meta-analyses of mean difference and standardized mean difference. Stat Med. 2020;39(2):171–91.

Doncaster CP, Spake R. Correction for bias in meta-analysis of little-replicated studies. Methods Ecol Evol. 2018;9(3):634–44.

Nakagawa S, Noble DW, Lagisz M, Spake R, Viechtbauer W, Senior AM. A robust and readily implementable method for the meta-analysis of response ratios with and without missing standard deviations. Ecol Lett. 2023;26(2):232–44

Hamman EA, Pappalardo P, Bence JR, Peacor SD, Osenberg CW. Bias in meta-analyses using Hedges’ d. Ecosphere. 2018;9(9): e02419.

Bakbergenuly I, Hoaglin DC, Kulinskaya E. On the Q statistic with constant weights for standardized mean difference. Brit J Math Stat Psy. 2022;75(3):444–65.

DerSimonian R, Kacker R. Random-effects model for meta-analysis of clinical trials: an update. Contemp Clin Trials. 2007;28(2):105–14.

Veroniki AA, Jackson D, Viechtbauer W, Bender R, Bowden J, Knapp G, Kuss O, Higgins JPT, Langan D, Salanti G. Methods to estimate the between-study variance and its uncertainty in meta-analysis. Res Synth Methods. 2016;7(1):55–79.

Langan D, Higgins JPT, Simmonds M. Comparative performance of heterogeneity variance estimators in meta-analysis: a review of simulation studies. Res Synth Methods. 2017;8(2):181–98.

Panityakul T, Bumrungsup C, Knapp G. On estimating residual heterogeneity in random-effects meta-regression: a comparative study. J Stat Theory Appl. 2013;12(3):253–65.

Bishop J, Nakagawa S. Quantifying crop pollinator dependence and its heterogeneity using multi-level meta-analysis. J Appl Ecol. 2021;58(5):1030–42.

Cheung MWL. Modeling dependent effect sizes with three-level meta-analyses: a structural equation modeling approach. Psychol Methods. 2014;19(2):211–29.

Bolker BM, Brooks ME, Clark CJ, Geange SW, Poulsen JR, Stevens MHH, White JSS. Generalized linear mixed models: a practical guide for ecology and evolution. Trends Ecol Evol. 2009;24(3):127–35.

Lajeunesse MJ. On the meta-analysis of response ratios for studies with correlated and multi-group designs. Ecology. 2011;92(11):2049–55.

Gleser LJ, Olkin I. Stochastically dependent effect sizes. In: Cooper H, Hedges LV, Valentine JC, editors. The handbook of research synthesis and meta-analysis. New York: Russell Sage Foundation; 2009.

Tipton E, Pustejovsky JE. Small-sample adjustments for tests of moderators and model fit using robust variance estimation in meta-regression. J Educ Behav Stat. 2015;40(6):604–34.

Hedges LV, Tipton E, Johnson MC. Robust variance estimation in meta-regression with dependent effect size estimates (vol 1, pg 39, 2010). Res Synth Methods. 2010;1(2):164–5.

Pustejovsky JE, Tipton E. Meta-analysis with robust variance estimation: expanding the range of working models. Prev Sci. 2021. https://doi.org/10.1007/s11121-021-01246-3 .

Cairns M, Prendergast LA. On ratio measures of heterogeneity for meta-analyses. Res Synth Methods. 2022;13(1):28–47.

Borenstein M, Higgins JPT, Hedges LV, Rothstein HR. Basics of meta-analysis: I2 is not an absolute measure of heterogeneity. Res Synth Methods. 2017;8(1):5–18.

Hoaglin DC. Practical challenges of I-2 as a measure of heterogeneity. Res Synth Methods. 2017;8(3):254–254.

Higgins JPT, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med. 2002;21(11):1539–58.

Higgins JPT, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. Brit Med J. 2003;327(7414):557–60.

Xiong CJ, Miller JP, Morris JC. Measuring study-specific heterogeneity in meta-analysis: application to an antecedent biomarker study of Alzheimer’s disease. Stat Biopharm Res. 2010;2(3):300–9.

Nakagawa S, Schielzeth H. Repeatability for Gaussian and non-Gaussian data: a practical guide for biologists. Biol Rev. 2010;85(4):935–56.

Senior AM, Grueber CE, Kamiya T, Lagisz M, O’Dwyer K, Santos ESA, Nakagawa S. Heterogeneity in ecological and evolutionary meta-analyses: its magnitude and implications. Ecology. 2016;97(12):3293–9.

Gelman A, Hill J. Data analysis using regression and multilevel/hierarchical models. Cambridge: Cambridge University Press; 2007.

Schielzeth H. Simple means to improve the interpretability of regression coefficients. Methods Ecol Evol. 2010;1(2):103–13.

Nakagawa S, Schielzeth H. A general and simple method for obtaining R2 from generalized linear mixed-effects models. Methods Ecol Evol. 2013;4(2):133–42.

Nakagawa S, Johnson PCD, Schielzeth H. The coefficient of determination R-2 and intra-class correlation coefficient from generalized linear mixed-effects models revisited and expanded. J R Soc Interface. 2017;14(134):20170213.

Aloe AM, Becker BJ, Pigott TD. An alternative to R-2 for assessing linear models of effect size. Res Synth Methods. 2010;1(3–4):272–83.

Cinar O, Umbanhowar J, Hoeksema JD, Viechtbauer W. Using information-theoretic approaches for model selection in meta-analysis. Res Synth Methods. 2021. https://doi.org/10.1002/jrsm.1489 .

Viechtbauer W. Model checking in meta-analysis. In: Schmid CH, Stijnen T, White IR, editors. Handbook of meta-analysis. Boca Raton: CRC; 2021.

Anzures-Cabrera J, Higgins JPT. Graphical displays for meta-analysis: An overview with suggestions for practice. Res Synth Methods. 2010;1(1):66–80.

Kossmeier M, Tran US, Voracek M. Charting the landscape of graphical displays for meta-analysis and systematic reviews: a comprehensive review, taxonomy, and feature analysis. Bmc Med Res Methodol. 2020;20(1):1.

Intout J, Ioannidis JPA, Rovers MM, Goeman JJ. Plea for routinely presenting prediction intervals in meta-analysis. BMJ Open. 2016;6(7): e010247.

Moeyaert M, Ugille M, Beretvas SN, Ferron J, Bunuan R, Van den Noortgate W. Methods for dealing with multiple outcomes in meta-analysis a comparison between averaging effect sizes, robust variance estimation and multilevel meta-analysis. Int J Soc Res Methodol. 2017;20:559.

Nakagawa S, Lagisz M, O’Dea RE, Rutkowska J, Yang YF, Noble DWA, Senior AM. The orchard plot: cultivating a forest plot for use in ecology, evolution, and beyond. Res Synth Methods. 2021;12(1):4–12.

Rothstein H, Sutton AJ, Borenstein M. Publication bias in meta-analysis : prevention, assessment and adjustments. Hoboken: Wiley; 2005.

Nakagawa S, Lagisz M, Jennions MD, Koricheva J, Noble DWA, Parker TH, Sanchez-Tojar A, Yang YF, O’Dea RE. Methods for testing publication bias in ecological and evolutionary meta-analyses. Methods Ecol Evol. 2022;13(1):4–21.

Stanley TD, Doucouliagos H. Meta-regression analysis in economics and business. New York: Routledge; 2012.

Stanley TD, Doucouliagos H. Meta-regression approximations to reduce publication selection bias. Res Synth Methods. 2014;5(1):60–78.

Sterne JAC, Becker BJ, Egger M. The funnel plot. In: Rothstein H, Sutton AJ, Borenstein M, editors. Publication bias in meta-analysis: prevention, assessment and adjustments. Chichester: Wiley; 2005. p. 75–98.

Sterne JAC, Sutton AJ, Ioannidis JPA, Terrin N, Jones DR, Lau J, Carpenter J, Rucker G, Harbord RM, Schmid CH, et al. Recommendations for examining and interpreting funnel plot asymmetry in meta-analyses of randomised controlled trials. Br Med J. 2011;343:4002.

Egger M, Smith GD, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. Brit Med J. 1997;315(7109):629–34.

Jennions MD, Moller AP. Relationships fade with time: a meta-analysis of temporal trends in publication in ecology and evolution. P Roy Soc B-Biol Sci. 2002;269(1486):43–8.

Koricheva J, Kulinskaya E. Temporal instability of evidence base: a threat to policy making? Trends Ecol Evol. 2019;34(10):895–902.

Raue A, Kreutz C, Maiwald T, Bachmann J, Schilling M, Klingmuller U, Timmer J. Structural and practical identifiability analysis of partially observed dynamical models by exploiting the profile likelihood. Bioinformatics. 2009;25(15):1923–9.

Matsushima Y, Noma H, Yamada T, Furukawa TA. Influence diagnostics and outlier detection for meta-analysis of diagnostic test accuracy. Res Synth Methods. 2020;11(2):237–47.

Viechtbauer W, Cheung MWL. Outlier and influence diagnostics for meta-analysis. Res Synth Methods. 2010;1(2):112–25.

Haddaway NR, Macura B. The role of reporting standards in producing robust literature reviews comment. Nat Clim Change. 2018;8(6):444–7.

Frampton G, Whaley P, Bennett M, Bilotta G, Dorne JLCM, Eales J, James K, Kohl C, Land M, Livoreil B, et al. Principles and framework for assessing the risk of bias for studies included in comparative quantitative environmental systematic reviews. Environ Evid. 2022;11(1):12.

Stanhope J, Weinstein P. Critical appraisal in ecology: what tools are available, and what is being used in systematic reviews? Res Synth Methods. 2022. https://doi.org/10.1002/jrsm.1609 .

Haddaway NR, Macura B, Whaley P, Pullin AS. ROSES RepOrting standards for systematic evidence syntheses: pro forma, flow-diagram and descriptive summary of the plan and conduct of environmental systematic reviews and systematic maps. Environ Evid. 2018;7(1):1.

Woodcock P, Pullin AS, Kaiser MJ. Evaluating and improving the reliability of evidence syntheses in conservation and environmental science: a methodology. Biol Conserv. 2014;176:54–62.

O’Dea RE, Lagisz M, Jennions MD, Koricheva J, Noble DWA, Parker TH, Gurevitch J, Page MJ, Stewart G, Moher D, et al. Preferred reporting items for systematic reviews and meta-analyses in ecology and evolutionary biology: a PRISMA extension. Biol Rev. 2021;96(5):1695–722.

Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Plos Med. 2009;6(7):e1000097.

Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, Shamseer L, Tetzlaff JM, Akl EA, Brennan SE, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. Plos Med. 2021;18(3): e1003583.

Wilkinson MD, Dumontier M, Aalbersberg IJ, Appleton G, Axton M, Baak A, Blomberg N, Boiten JW, Santos LBD, Bourne PE, et al. Comment: the FAIR guiding principles for scientific data management and stewardship. Sci Data. 2016;3: 160018.

Culina A, Baglioni M, Crowther TW, Visser ME, Woutersen-Windhouwer S, Manghi P. Navigating the unfolding open data landscape in ecology and evolution. Nat Ecol Evol. 2018;2(3):420–6.

Roche DG, Lanfear R, Binning SA, Haff TM, Schwanz LE, Cain KE, Kokko H, Jennions MD, Kruuk LE. Troubleshooting public data archiving: suggestions to increase participation. Plos Biol. 2014;12(1): e1001779.

Roche DG, Kruuk LEB, Lanfear R, Binning SA. Public data archiving in ecology and evolution: how well are we doing? Plos Biol. 2015;13(11): e1002295.

Culina A, van den Berg I, Evans S, Sanchez-Tojar A. Low availability of code in ecology: a call for urgent action. Plos Biol. 2020;18(7): e3000763.

Spake R, Mori AS, Beckmann M, Martin PA, Christie AP, Duguid MC, Doncaster CP. Implications of scale dependence for cross-study syntheses of biodiversity differences. Ecol Lett. 2021;24(2):374–90.

Osenberg CW, Sarnelle O, Cooper SD. Effect size in ecological experiments: the application of biological models in meta-analysis. Am Nat. 1997;150(6):798–812.

Noble DWA, Nakagawa S. Planned missing data designs and methods: options for strengthening inference, increasing research efficiency and improving animal welfare in ecological and evolutionary research. Evol Appl. 2021;14(8):1958–68.

Nakagawa S, Freckleton RP. Missing inaction: the dangers of ignoring missing data. Trends Ecol Evol. 2008;23(11):592–6.

Mavridis D, Chaimani A, Efthimiou O, Leucht S, Salanti G. Addressing missing outcome data in meta-analysis. Evid-Based Ment Health. 2014;17(3):85.

Ellington EH, Bastille-Rousseau G, Austin C, Landolt KN, Pond BA, Rees EE, Robar N, Murray DL. Using multiple imputation to estimate missing data in meta-regression. Methods Ecol Evol. 2015;6(2):153–63.

Kambach S, Bruelheide H, Gerstner K, Gurevitch J, Beckmann M, Seppelt R. Consequences of multiple imputation of missing standard deviations and sample sizes in meta-analysis. Ecol Evol. 2020;10(20):11699–712.

Hadfield JD, Nakagawa S. General quantitative genetic methods for comparative biology: phylogenies, taxonomies and multi-trait models for continuous and categorical characters. J Evol Biol. 2010;23(3):494–508.

Cinar O, Nakagawa S, Viechtbauer W. Phylogenetic multilevel meta-analysis: a simulation study on the importance of modelling the phylogeny. Methods Ecol Evol. 2021. https://doi.org/10.1111/2041-210X.13760 .

Ives AR, Zhu J. Statistics for correlated data: phylogenies, space, and time. Ecol Appl. 2006;16(1):20–32.

Maire A, Thierry E, Viechtbauer W, Daufresne M. Poleward shift in large-river fish communities detected with a novel meta-analysis framework. Freshwater Biol. 2019;64(6):1143–56.

Morrissey MB. Meta-analysis of magnitudes, differences and variation in evolutionary parameters. J Evol Biol. 2016;29(10):1882–904.

Cohen JM, Lajeunesse MJ, Rohr JR. A global synthesis of animal phenological responses to climate change. Nat Clim Change. 2018;8(3):224.

Gurevitch J, Morrison JA, Hedges LV. The interaction between competition and predation: a meta-analysis of field experiments. Am Nat. 2000;155(4):435–53.

Macartney EL, Lagisz M, Nakagawa S. The relative benefits of environmental enrichment on learning and memory are greater when stressed: a meta-analysis of interactions in rodents. Neurosci Biobehav R. 2022. https://doi.org/10.1016/j.neubiorev.2022.104554 .

Spake R, Bowler DE, Callaghan CT, Blowes SA, Doncaster CP, Antão LH, Nakagawa S, McElreath R, Chase JM. Understanding ‘it depends’ in ecology: a guide to hypothesising, visualising and interpreting statistical interactions. Biol Rev. 2023. https://doi.org/10.1111/brv.12939 .

Siviter H, Bailes EJ, Martin CD, Oliver TR, Koricheva J, Leadbeater E, Brown MJF. Agrochemicals interact synergistically to increase bee mortality. Nature. 2021;596(7872):389.

Salanti G, Schmid CH. Research synthesis methods special issue on network meta-analysis: introduction from the editors. Res Synth Methods. 2012;3(2):69–70.

Gomez-Creutzberg C, Lagisz M, Nakagawa S, Brockerhoff EG, Tylianakis JM. Consistent trade-offs in ecosystem services between land covers with different production intensities. Biol Rev. 2021;96(5):1989–2008.

Jackson D, White IR, Price M, Copas J, Riley RD. Borrowing of strength and study weights in multivariate and network meta-analysis. Stat Methods Med Res. 2017;26(6):2853–68.

Sanchez-Tojar A, Moran NP, O’Dea RE, Reinhold K, Nakagawa S. Illustrating the importance of meta-analysing variances alongside means in ecology and evolution. J Evol Biol. 2020;33(9):1216–23.

Riley RD, Lambert PC, Abo-Zaid G. Meta-analysis of individual participant data: rationale, conduct, and reporting. BMJ. 2010;340:c221.

Riley RD, Tierney JF, Stewart LA. Individual participant data meta-analysis : a handbook for healthcare research. 1st ed. Hoboken: Wiley; 2021.

Hudson LN, Newbold T, Contu S, Hill SLL, Lysenko I, De Palma A, Phillips HRP, Alhusseini TI, Bedford FE, Bennett DJ, et al. The database of the PREDICTS (projecting responses of ecological diversity in changing terrestrial systems) project. Ecol Evol. 2017;7(1):145–88.

Dornelas M, Antao LH, Moyes F, Bates AE, Magurran AE, Adam D, Akhmetzhanova AA, Appeltans W, Arcos JM, Arnold H, et al. BioTIME: a database of biodiversity time series for the anthropocene. Glob Ecol Biogeogr. 2018;27(7):760–86.

Mengersen K, Gurevitch J, Schmid CH. Meta-analysis of primary data. In: Koricheva J, Gurevitch J, Mengersen K, editors. Handbook of meta-analysis in ecology and evolution. Priceton: Princeton university; 2013. p. 300–12.

Spake R, O’Dea RE, Nakagawa S, Doncaster CP, Ryo M, Callaghan CT, Bullock JM. Improving quantitative synthesis to achieve generality in ecology. Nat Ecol Evol. 2022;6(12):1818–28.

Gerstner K, Moreno-Mateos D, Gurevitch J, Beckmann M, Kambach S, Jones HP, Seppelt R. Will your paper be used in a meta-analysis? Make the reach of your research broader and longer lasting. Methods Ecol Evol. 2017;8(6):777–84.

Haddaway NR. A call for better reporting of conservation research data for use in meta-analyses. Conserv Biol. 2015;29(4):1242–5.

Midolo G, De Frenne P, Holzel N, Wellstein C. Global patterns of intraspecific leaf trait responses to elevation. Global Change Biol. 2019;25(7):2485–98.

White IR, Schmid CH, Stijnen T. Choice of effect measure and issues in extracting outcome data. In: Schmid CH, Stijnen T, White IR, editors. Handbook of meta-analysis. Boca Raton: CRC; 2021.

Lajeunesse MJ. Bias and correction for the log response ratio in ecological meta-analysis. Ecology. 2015;96(8):2056–63.

Download references

Acknowledgements

SN, ELM, and ML were supported by the ARC (Australian Research Council) Discovery grant (DP200100367), and SN, YY, and ML by the ARC Discovery grant (DP210100812). YY was also supported by the National Natural Science Foundation of China (32102597). A part of this research was conducted while visiting the Okinawa Institute of Science and Technology (OIST) through the Theoretical Sciences Visiting Program (TSVP) to SN.

Australian Research Council Discovery grant (DP200100367); Australian Research Council Discovery grant (DP210100812); The National Natural Science Foundation of China (32102597).

Author information

Authors and affiliations.

Evolution & Ecology Research Centre and School of Biological, Earth and Environmental Sciences, University of New South Wales, Sydney, NSW, 2052, Australia

Shinichi Nakagawa, Yefeng Yang, Erin L. Macartney & Malgorzata Lagisz

Theoretical Sciences Visiting Program, Okinawa Institute of Science and Technology Graduate University, Onna, 904-0495, Japan

Shinichi Nakagawa

School of Biological Sciences, Whiteknights Campus, University of Reading, Reading, RG6 6AS, UK

Rebecca Spake

You can also search for this author in PubMed   Google Scholar

Contributions

SN was commissioned to write this article so he assembled a team of co-authors. SN discussed the idea with YY, ELM, RS and ML, and all of them contributed to the design of this review. ML led the survey working with YY and ELM, while YY led the creation of the accompanying webpage working with RS. SN supervised all aspects of this work and wrote the first draft, which was commented on, edited, and therefore, significantly improved by the other co-authors. All authors read and approved the final manuscript.

Corresponding authors

Correspondence to Shinichi Nakagawa or Yefeng Yang .

Ethics declarations

Ethics approval and consent to participate.

Not applicable.

Contest for publication

The authors provide consent for publication.

Competing interests

The authors report no competing interests.

Additional information

Publisher's note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Additional file 1:.

The survey of meta-analyses in environmnetal sciences.

Additional file 2:

The hands-on R tutorial.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ . The Creative Commons Public Domain Dedication waiver ( http://creativecommons.org/publicdomain/zero/1.0/ ) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Cite this article.

Nakagawa, S., Yang, Y., Macartney, E.L. et al. Quantitative evidence synthesis: a practical guide on meta-analysis, meta-regression, and publication bias tests for environmental sciences. Environ Evid 12 , 8 (2023). https://doi.org/10.1186/s13750-023-00301-6

Download citation

Received : 13 January 2023

Accepted : 23 March 2023

Published : 24 April 2023

DOI : https://doi.org/10.1186/s13750-023-00301-6

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Hierarchical models
• Robust variance estimation
• Spatial dependency
• Variance–covariance matrix
• Meta-analysis of variance
• Network meta-analysis
• Multivariate meta-analysis

Environmental Evidence

ISSN: 2047-2382

• Submission enquiries: Access here and click Contact Us
• General enquiries: [email protected]

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

• View all journals
• Explore content
• About the journal
• Publish with us
• Sign up for alerts
• Open access
• Published: 15 March 2017

Quantitative analysis of the impacts of terrestrial environmental factors on precipitation variation over the Beibu Gulf Economic Zone in Coastal Southwest China

• Yinjun Zhao 1 , 2 ,
• Qiyu Deng 2 ,
• Qing Lin 1 &
• Chunting Cai 2  

Scientific Reports volume  7 , Article number:  44412 ( 2017 ) Cite this article

9236 Accesses

15 Citations

Metrics details

• Climate and Earth system modelling
• Environmental impact

Taking the Guangxi Beibu Gulf Economic Zone as the study area, this paper utilizes the geographical detector model to quantify the feedback effects from the terrestrial environment on precipitation variation from 1985 to 2010 with a comprehensive consideration of natural factors (forest coverage rate, vegetation type, terrain, terrestrial ecosystem types, land use and land cover change) and social factors (population density, farmland rate, GDP and urbanization rate). First, we found that the precipitation trend rate in the Beibu Gulf Economic Zone is between −47 and 96 mm/10a. Second, forest coverage rate change (FCRC), urbanization rate change (URC), GDP change (GDPC) and population density change (PDC) have a larger contribution to precipitation change through land-surface feedback, which makes them the leading factors. Third, the human element is found to primarily account for the precipitation changes in this region, as humans are the active media linking and enhancing these impact factors. Finally, it can be concluded that the interaction of impact factor pairs has a significant effect compared to the corresponding single factor on precipitation changes. The geographical detector model offers an analytical framework to reveal the terrestrial factors affecting the precipitation change, which gives direction for future work on regional climate modeling and analyses.

Similar content being viewed by others

Vegetation response to precipitation anomalies under different climatic and biogeographical conditions in China

Empirical estimate of forestation-induced precipitation changes in Europe

The human–environment nexus and vegetation–rainfall sensitivity in tropical drylands

Introduction.

Climate change has been a topic of worldwide concern in recent years. Precipitation is the most active parameter of all the meteorological elements. A large number of studies show that precipitation exhibits change in many areas 1 , 2 . Precipitation change caused by the anomalous change of atmospheric circulation is a very complicated phenomenon, which is primarily the result of internal adjustment of the atmosphere itself. However, in terms of the regional scale, the terrestrial environment will respond to precipitation change through land-atmosphere interactions. It should be noted that, to some extent, terrestrial environment impact is comparable to atmospheric circulation and solar radiation.

The terrestrial environment primarily includes natural (vegetation coverage, vegetation type, terrain and terrestrial ecosystem type) and social (human activities) factors. Vegetation coverage impacts the climate through its effect on surface albedo 3 , etc. Compared to surrounding areas, the ground vegetation properties in the region, such as surface albedo, roughness and soil humidity, have a large variation that would influence local thermal and moisture conditions. This, in turn, changes local precipitation through atmospheric circulation on a small to medium scale. Previous research showed that precipitation can increase 4 , 5 , 6 with the growth of forest coverage. Different vegetation types also have distinct impacts on the surrounding climate 7 . For example, a coniferous broad-leaved forest and its analogues have a greater impact than other vegetation types on the change of the average annual precipitation trend 8 . The terrain will affect the entire atmosphere precipitation system 9 . The slope, altitude, latitude and other similar factors will directly affect the precipitation by changing the regional atmospheric circulation 10 , 11 . The terrestrial ecosystem influences the concentration of greenhouse gases and aerosols in the atmosphere, thus affecting climate change through the energy balance between the ground and the atmosphere, the interaction of water vapor exchange and the biogeochemical cycle 12 . At the same time, the ecosystem would respond to the climate change 13 , 14 , namely, both climate interactions have mutual effects. Human activities, such as agricultural irrigation 15 , 16 , 17 , 18 , 19 , 20 , 21 and urbanization 22 , 23 , directly or indirectly exert some impact on the precipitation distribution by changing regional underlying surface hydrothermal conditions to affect atmospheric circulation. For example, urbanization in Guangzhou accounts for 44.7% of the significant precipitation growth since 1991 24 ; irrigation increases precipitation while decreasing the daily average and maximum air temperatures 25 , 26 .

Most of the aforementioned researches have analyzed the change in the characteristics, temporal-spatial trends and impact factors of precipitation (climate) in a certain area 27 , 28 , 29 , 30 , 31 , 32 , 33 ; however, these researches are lacking quantitative analysis on the combined effects of multiple important factors under a unified analysis framework. Based on the spatial variation theory, the geographical detector model 34 was used initially for evaluating the relationship between health and suspicious pathogenic factors. It can measure the spatial consistency and statistical significance between health risk and geographical elements and determine the effectiveness of the spatial correlation without many assumptions. It also effectively overcomes the limitations of processing category variables that exist in the traditional statistical analysis method 35 . Thus, the application of the geographical detector model has been gradually extended to other areas, such as resources and the environment 35 , 36 , 37 , 38 , 39 , 40 , 41 , for quantitative analysis of the mutual relationship between the factor and result variables 42 , 43 . It should be mentioned that the geographical detector model has never been utilized to provide an analysis framework in order to study precipitation change.

Therefore, this paper attempts to answer the following questions: First, what is the major determinant affecting precipitation change? Second, does each factor affect precipitation change independently or interactively? Third, what is the relative importance of these affecting factors? The Guangxi Beibu Gulf Economic Zone was selected as the case study in this research. The Guangxi Beibu Gulf Economic Zone is located in the southwest China coast and consists of the administrative regions of Nanning, Beihai, Qinzhou, Fangchenggang, Yulin, and Chongzuo city ( Fig. 1 ). The land area covers 425,000 km 2 . The Beibu Gulf Economic Zone is located south of the Tropic of Cancer and is a subtropical maritime monsoon climate zone with transitional characteristics from tropical to subtropical. The annual average temperature ranges between 21.5 °C and 23.4 °C, while the average daily temperature stabilizes above 10 °C. The multi-year average precipitation ranges between 1251.27 mm and 2717.87 mm. The Dongxing-Qinzhou region on the southeast of Shiwan Dashan Mountain is one of the three rainy districts in Guangxi. The precipitation amount during flood season, generally from April to September, accounts for 80% of the annual precipitation, with peak precipitation occurring in July and August.

It was generated by ArcGIS10.1( http://www.esrichina.com.cn/softwareproduct/ArcGIS/ ); the locations of the meteorological observatories were obtained from the China meteorological data network ( http://data.cma.cn ).

Results and Interpretation

Precipitation change in the beibu gulf economic zone.

The precipitation trend of each meteorological observation (88 total meteorological observatories in Guangxi Province) was calculated from annual precipitation for the period of 1985~2010 using Equation (1). Linear trends indicate that 76% of the meteorological stations show a positive trend in annual precipitation during 1985–2010, and notably, five of them are statistically significant at the 90% confidence level (see Supplementary Table S1 , Fig. S1 ). The other 24% of the meteorological stations show a negative trend (see Supplementary Table S1 ). A station-by-station analysis was performed and mapped using ArcGIS 10.1 with the Empirical Bayesian Kriging interpolation method in order to explore spatial patterns of precipitation changes in Guangxi Province. Then, the precipitation changes in the Guangxi Beibu Gulf Economic Zone were clipped and are shown in Fig. 2 . This might be better than using direct interpolation of fewer meteorological observatories from the Guangxi Beibu Gulf Economic Zone, especially in the border area, because of the regional characteristics of precipitation.

It was generated using ArcGIS 10.1 ( http://www.esrichina.com.cn/softwareproduct/ArcGIS/ ).

Figure 2 shows that the precipitation trend rate of the Beibu Gulf Economic Zone is between −47 and 96 mm/10a. The character of the spatial precipitation change is primarily in the northwest-southeast direction. The low value zones are located in the southwest Beibu Gulf Economic Zone, while the high values are located in the northeast. The figure also shows that the precipitation changes in the middle zone are relatively lower than those in the neighboring south and north areas. A relative increasing precipitation trend from south to north is observed as a whole. We also found two zero lines of precipitation change in the middle and southwest zone of the Beibu Gulf Economic Zone.

According to Li & Su’s research, the Mann-Kendall mutability test found that precipitation in Guangxi Province had sudden changes in the years of 1984 and 1994. Specifically, from 1984 to 1994, Guangxi Province had less rain, while beginning in 1994, Guangxi Province entered into a relatively pluvial period 44 . Existing research also shows that there was a positive trend center of precipitation in northwest Guangxi. Therefore, in general, the linear trend is increasing.

The feedback of terrestrial environmental factors to precipitation change

The leading factors of precipitation change.

The factor detector ranked the terrestrial environment layers by their influences (P D,H values) on precipitation change in the following order for the study area ( Table 1 ):

FCRC (50.3%) > URC (47.3%) > GDPC (43.5%) > FRC (35.2%) > PDC (27.4%) > VT (10.0%) > DEM (7.3%) > GT (2.4%) > GRD (0.8%) > ASP (0.4%) > LUCC (0.3%) > TET (0.1%).

Among the terrestrial environmental factors, the P D,H value of FCRC is the maximum. Obviously, there is a large break in sorted P D,H values between PDC and VT. The P D,H values of FCRC, URC, GDPC, FRC and PDC are in a group with high values and small differences, while the rest of the factors belong to another group with lower values. In a general sense, if the P D,H value of a factor is larger than 0.2(20%), then the factor can be regarded as a leading factor 39 that strongly explains the spatial pattern. Therefore, FCRC, URC, GDPC, FRC and PDC are potential leading factors that may exert the largest impact on precipitation change (spatial pattern) in this study area. In contrast, the values of VT, DEM, GT, GRD, ASP, LUCC and TET are comparatively small at less than 0.1(10%), which likely reflects their smaller contributions to the precipitation trend (spatial pattern).

The ecological detector ( Table 2 ) shows the differences of the P D,H values. Among the five potential leading factors (FCRC, URC, GDPC, FRC and PDC), approximately 80% of them (FCRC, URC, GDPC and PDC) are not statistically significant with each other, whereas statistically significant differences between FRC and other potential leading factors (URC and GDPC) were found. That is, URC and GDPC have a larger significant effect on the precipitation change than FRC. With the factor detector and the ecological detector, we concluded that FCRC, URC, GDPC and PDC are leading factors, and FRC was eliminated from the potential leading factors. Therefore, FCRC, URC, GDPC and PDC have the largest contribution to the precipitation change, whereas the remaining factors have a relatively weak influence.

In view of the above considerations, we also found that the power of social factors is much larger than that of natural factors in changing precipitation in the study area. This can be seen from the leading factors (FCRC, URC, GDPC and PDC) and whole sorting of the P D,H values, which means that people are likely to be a very powerful factor in changing the terrestrial environment to influence local precipitation on a regional scale.

On the surface, FCRC is the first leading natural factor. However, the increase of forest coverage rate change in the study area is predominantly due to ecological construction for tourism and sustainable development in recent years. The forest coverage rate change of this study area is higher than the rate of other places in China and occurs in a sustainable growth manner. The Shiwan Dashan National Forest Park and the Daming Mountain National Natural Reserve are located in this research area. Some research has already highlighted that the significant role of precipitation may increase or decrease alongside afforestation 4 , 5 , 6 or deforestation.

The Guangxi Beibu Gulf Economic Zone is the first international regional economic cooperation zone in China. According to the statistics, the population of this region was 15.81 million in 1985 and rose by 43 percent to 22.68 million in 2010. The GDP of this region increased from 8.3 billion Yuan (RMB) to 412.1 trillion Yuan (RMB) between 1985 and 2010, which is a huge growth of 49.37 times the initial GDP. With the sustainable growth of population and the increasing development of the economy, the GDP, the population density, urbanization, construction activities, energy consumption and greenhouse gas emissions have experienced a relatively rapid growth in the Beibu Gulf Economic Zone, and among them, the GDP growth rate is the largest.

The urbanization levels of both Nanning City and Beihai City are over 55%. In contrast, Guangxi has a relatively extensive development model for their economy, with an energy consumption per unit of GDP of 1.036 tons of standard coal per ten thousand Yuan (2010), which is 1.28 times the national average of 0.81 tons of standard coal per ten thousand Yuan. A large amount of energy consumption emits a large amount of greenhouse gas, such as carbon dioxide, which is the main source 45 of carbon emissions. This greatly influences the climate of this region and possibly the climate on a larger regional scale. Urbanization is a comprehensive process, which will influence a city’s precipitation, temperature, humidity, visibility and wind, forming a special local meteorological environment and causing material climate changes. The GDP may actually be viewed as a comprehensive result of many human activities. The increase of population density leads to an increase of artificial thermal discharge, directly influencing the change of surficial sensible heat flux, which will influence precipitation significantly 46 . In addition, approximately 45% of China’s farmland is irrigated farmland 47 , whereas Guangxi Province has a higher percentage. The P D,H value of FRC on precipitation change is 35.3% (much higher than 20%). This occurs mainly because the heavy irrigation of farmland affects the distribution of surface net radiation between latent heat flux and sensible heat flux change (latent heat flux increases, but sensible heat flux decreases), and farm irrigation has a cooling effect on the earth’s surface; at the same time, the increase of soil humidity enhances transpiration and further increases the moisture content in the atmosphere and the unstable energy of latent heat, leading to an increase of convective precipitation 48 and producing a marked effect on the region’s precipitation. This finding is supported by other cases. Irrigation over the Ogallala Aquifer in the central United States increased dramatically over the 20th century and has enhanced regional precipitation 49 . The precipitation increase in the Texas Panhandle from 1952 to 1980 was obviously due to the increase in the irrigation area 25 . On the other hand, the amount of precipitation in central and southern India decreased due to a lower surface temperature over the irrigated areas of India in July 26 .

The Beibu Gulf Economic Zone is a relatively small area, which is on a small scale compared to the majority of research on precipitation. In the region, the DEM, geomorphic type, slope aspect, gradient, ecosystem and vegetation form are similar or experience less change, so they probably have a weak effect on precipitation change.

The effect of the interaction of terrestrial environmental factors on precipitation change

The interaction detector was used to check whether or not two factors work independently. The joint impacts of two factors measured by the P D,H values are shown in Table 3 and Table S2 and can be compared with their separate impacts.

It must be noted from Table S2 that the P D,H values of 22 interactive pairs are greater than that of the primary leading factor (FCRC). The max P D,H value comes from interaction of FRC with GDPC (FRC ∩ GDPC = 84.4%). Specifically, all the interactive effects between FCRC and the rest of the factors (FCRC ∩ GDPC = 83.1%, FCRC ∩ FRC = 76.5%, FCRC ∩ URC = 74.1%, FCRC ∩ PDC = 75.0%, FCRC ∩ DEM = 57.3%, FCRC ∩ GRD = 53.6%, FCRC ∩ ASP = 50.9%, FCRC ∩ TET = 51.0%, FCRC ∩ GT = 54.8%, FCRC ∩ VT = 59.8%, FCRC ∩ LUCC = 51.0%) are stronger than the effect of the single FCRC (50.3%, the strongest effect on precipitation changes). We found that FCRC interacting with any other factors is always enhanced. Similarly, all the interaction effects between URC and the rest of the factors are higher than the single URC (47.3%) effect. Even of those factors with the lowest P D,H values, interactions between them enhance their separate effects on precipitation changes. In general, all interactive pairs of impact factors showed enhanced results compared to the corresponding single factor, and among them, 45 interactive pairs have P D,H values larger than 0.2 (20%).

The top P D,H values of interactive pairs are FRC ∩ GDPC = 84.4%, FCRC ∩ GDPC = 83.1%, FRC ∩ URC = 77.0%, FCRC ∩ FRC = 76.5%, FCRC ∩ PDC = 75.0%, FCRC ∩ URC = 74.1%, URC ∩ GDPC = 72.3%, URC ∩ PDC = 71.0%, FRC ∩ PDC = 70.1% and GDPC ∩ PDC = 70.0%, and all of them are larger than 70%. We thought that FCRC is also a social factor because FCRC is mainly due to human ecological construction. Therefore, these factors are all social factors, and it clearly implies that humans are the most important aspect in changing precipitation (similar to the analysis of leading factors) in this region via economic activities such as urban construction, afforestation, changing and developing hillside fields, irrigation and plantation. Under the high pressure of growing population and development, humans are the best medium compared to other natural factors to change and affect the spatial distribution of other factors according to their purposes, and with the development of science and technology, this situation is amplified. For example, large-scale afforestation in the northern mid-latitudes warms the Northern Hemisphere and alters global circulation patterns to redistribute the anomalous energy absorbed in the northern hemisphere, which results in a precipitation decrease over parts of the Amazon basin and an increase over the Sahel and Sahara regions in Africa 50 .

In addition to the above mentioned, we also noted that interactions between social factors and natural factors have two types: nonlinear enhancement and bienhancement ( Table 3 ). Each type indicates that the factors bienhance or nonlinearly enhance each other. As shown in Table 3 , the interactions between social factors and natural factors have predominantly strong, nonlinear synergies. For example, the interactions of PDC and DEM (PD ∩ DEM = 40.3% > 34.7% = PDC (27.4%) + DEM (7.3%)) are larger than the P D,H value sum of PDC and DEM; therefore, the interaction between PDC and DEM has a larger impact on precipitation changes. This is likely due to the city and farmland expansion toward a relatively bad condition of DEM that changes the underlying surface conditions. It also indicates that social factors and natural factors have synergies and can enhance each other’s effect on precipitation change.

In conclusion, social factors have a larger impact on the precipitation change compared to natural factors. Partial natural factors have a relatively small impact on precipitation change but show a strong synergy with the interaction of other factors. The feedback of terrestrial environmental factors on precipitation change mainly arises from interactions of impact factors and interactive pairs of impact factors, which have a larger influence on precipitation change than the single factor does through the feedback. Interactions between factors play an important role in the precipitation change in this region.

Regional analysis of the leading impact range (type) of leading factors on precipitation change

The risk detector shows that the average precipitation change in the different FCRC zones (from I to VI) are −0.66 mm/10a, −9.98 mm/10a, 53.18 mm/10a, 6.88 mm/10a, 23.48 mm/10a and 11.66 mm/10a, respectively, and they are significantly different. It also implies that precipitation will increase or decrease with the increase or decrease of forest coverage. However, higher precipitation change is not consistent with a larger FCRC zone, and precipitation change fluctuates with FCRC values. A similar analysis of other terrestrial environmental factors can be conducted using the risk detector. The small and continued growth of annual urbanization rates will lead to a large increase in annual precipitation. The main impact ranges of FCRC, URC, GDPC and PDC tend to be located at the relatively low-middle value zones. We selected the largest types (ranges) of each leading factor as the main impact types (ranges) by sorting the average precipitation change. The main impact types (range) are tabulated in Table 4 and mapped in Fig. 3 .

The map was generated using ArcGIS 10.1 ( http://www.esrichina.com.cn/softwareproduct/ArcGIS/ ).

From Table 4 , we can see that the main impact types (ranges) of FCRC, URC, GDPC and PDC are 0.7411~4.7979%/10a, −7.7920~2.5006%/10a, 87824~128190ten thousand yuan/10a or 276670~399510ten thousand yuan/10a, and 36.81~52.33person/km 2 /10a, respectively. This means that these ranges probably have more contributions to local zones’ precipitation changes.

As shown in Fig. 3 , the leading impact type or range of each leading factor on precipitation change is predominantly located in the northeast-southeast of the Beibu Gulf Economic Zone. This indicates that the largest precipitation change is in the northeast-southeast of the Beibu Gulf Economic Zone, and the range of the precipitation trend rate is between 39 and 96 mm/ 10a ( Fig. 3 ). Therefore, the main distribution areas of the main impact range (type) of the leading factors on precipitation revealed by the results of the risk detector are consistent with the distribution of the relatively large area of precipitation change trend rate calculated by the linear regression model. This illustrates the flexibility of applying the geographical detector model to obtaining initial detection results of the precipitation change mechanism. Figure 3 shows that the precipitation change for the county of Rongxian is strongly controlled by FCRC, URC and PDC. According to the interaction detector, we also found that the P D,H values of FCRC ∩ URC (74.1%), FCRC ∩ PDC (75.0%), and URC ∩ PDC (71.0%) are very high and enhance each other to increase precipitation change, which emphasizes directions for future work. In conclusion, the largest precipitation change is present in the northeast-southeast region of the Beibu Gulf Economic Zone and is predominantly influenced by the interactions of factors such as FCRC, URC, GDPC and PDC.

Conclusions and Discussion

The causes of precipitation changes are very complicated due to the interaction of the land surface with the atmosphere. In addition, the research resources, such as shared data, are limited in developing countries, creating a high demand for useful detecting and/or analyzing tools. In this study, we used geographical detectors to verify the effects of some of the natural and social factors on precipitation change at a regional scale. We believe that this program is unique because it extracts the interrelationships between precipitation change and terrestrial environmental factors using the correspondence of their spatial distribution and, most importantly, because it is easily implemented.

The feedback of terrestrial environment to precipitation changes can be partially explained by forest cover, urbanization, terrain, irrigation and other single factors. Typically, the comprehensive consequences are the result of interactions of multiple factors. In this study, we found the following:

The precipitation trend rate of the Beibu Gulf Economic Zone is between −47 mm/10a and 96 mm/10a. The minimum and maximum values occur in the southwest and northeast of the Beibu Gulf Economic Zone, respectively.

The results found by the factor detector and the ecological detector show that FCRC, URC, GDPC and PDC, as the leading factors of precipitation change, have a relatively large contribution to the precipitation changes.

The interaction of pairs of impact factors has far larger effects than the corresponding single factor does on precipitation changes.

The precipitation change is predominantly due to human factors, and thus, humans act as an active media linking and enhancing the other impact factors.

The results of the risk detector show that the main impact types (ranges) of the leading factors of FCRC, URC, GDPC and PDC on precipitation change are 0.7411~4.7979%/10a, −7.7920~2.5006%/10a, 87824~128190ten thousand yuan/10a or 276670~399510ten thousand yuan/10a, and 36.81~52.33person/km 2 /10a, respectively.

Our research suggests that the geographical detector offers a quantitative and objective analytical framework that could be used to find the essence of many geosciences phenomena. There are still several aspects for future study. First, spatial scale transformation is an important aspect of geographical detectors. Transforming the administrative regions into the same grid cells might be subjective, as the grid size can have different values. We also found that discretization methods to classify continuous variables into several categories might affect the results because these methods do not currently have standardized rules. Second, due to the limitation of range and data accessibility in this study area, quantitative analysis was not conducted overall based on impact factors in this study. Third, the main impact ranges of leading factors (FCRC, URC, GDPC and PDC) fluctuated with precipitation change, and the largest precipitation change is typically only consistent with the smallest range of URC. In the future, threshold values of the main impact ranges can be overcome by collaboration with climate models. This is likely a better way to integrate geographical detectors with traditional meteorology methods to discover the precipitation change mechanism.

Despite some limitations, we still believe that this study will be meaningful. The geographical detectors are statistical and are not a causality tool; however, they can distinguish high potential impact factors and leading factor ranges to emphasize the next step in research. The results from this study can help researchers to understand the spatial pattern of precipitation change with impact factors and provide clues for further studies by integrating traditional observation, simulation, contrast testing, etc.

Materials and Methods

Research methods.

Tests for trend detection of the climatic element in a time series can be classified as parametric and non-parametric methods (e.g., the Mann-Kendall test). The linear regression method is a very simple and common parametric method 51 , and the trend rate method generally adopts the unitary linear regression model, that is:

where y represents a climatic element or other sequence (e.g., precipitation); x represents a yearly time series (from 1985 to 2010); and b represents a linear trend term, the value of which is a linear trend rate, in mm/10a.

Geographic detector model

Geographical detectors are composed of the factor detector, ecological detector, risk detector and interaction detector 34 , 43 . Factors significantly affecting precipitation change can be selected as the leading factors through analysis using the factor detector and ecological detector models; the risk detector can further analyze leading impact types or scopes (confidence level of 95%) of impact factors that significantly affect precipitation change; and the interaction detector can analyze the interaction among various factors. The core concept of the factor detector is as follows: there is certain differentiation of the factors affecting the development of geographical phenomenon in space. If a certain factor has a remarkable consistency with the change of that geographical phenomenon in space, then the factor will have a definite determinant power on the occurrence and development of a geographical phenomenon 34 , measured by the size of the power determinant value (P D,H ). Details of the geographical detector can be found in the original paper 34 . Here, in our research context, the calculation model for detecting impact factors of precipitation change in the Beibu Gulf Economic Zone is reviewed as follows:

We assume that precipitation change would present a spatial distribution similar to that of an impact factor if the impact factor leads to the change of precipitation (see Supplementary Fig. S2 ). All impact factors are quantified by these power values as follows:

The ecological detector compares which suspected impact factor (e.g., C factor) determinant is more significant than the other (e.g., D factor) in causing precipitation change in the study area. This is measured using the F-test:

Different types or ranges of an impact factor have different influences on precipitation change. The risk detector compares the differences through the t-test. The computational formula is as follows:

The interaction detector shows that when the two different factors of x and y are combined, they either weaken or enhance each another or they are independent in changing precipitation, determined by comparing P D,H (x ∩ y) with the values of P D,H (x) and P D,H (y), where the symbol ‘ ∩ ’ denotes the intersection between the x layer and y layer. If P D,H (x ∩ y) < min (P D,H (x), P D,H (y)), the variables nonlinearly weaken each other; if min (P D,H (x), P D,H (y)) < P D,H (x ∩ y) < max (P D,H (x), P D,H (y)), the variables uniweaken each other; if P D,H (x ∩ y) > max (P D,H (x), P D,H (y)), the variables bienhance each other; and if P D,H (x ∩ y) > P D,H (x) + P D,H (y), the variables nonlinearly enhance each other. If P D,H (x ∩ y) = P D,H (x) + P D,H (y), then the variables are independent of each other.

Based on the precipitation trend rate in the Beibu Gulf Economic Zone from 1985 to 2010, geographical detectors were utilized to explore the impact and indication effect of terrestrial environmental factors on the precipitation change through climate feedbacks.

Technical process

Modeling of the geographical detector mainly involves the following steps: first, determination of the optimal classification method for the factor data; second, determination of the impact of factors on the precipitation change; and third, determination of the leading role of factors in the precipitation change. Regarding the technical process in detail, please see Supplementary Fig. S3 .

Data sources and processing

Precipitation data.

The selected observation data are the mean annual precipitation of 29 meteorological stations from 1985 to 2010 in the Beibu Gulf Economic Zone ( Fig. 1 ). The above data were derived from the China meteorological data network ( http://data.cma.cn ).

Potential natural factors

According to the main impact factors of precipitation change discussed in the introduction, almost all of the environmental factors, except for climate type, were considered as main potential natural factors, such as geomorphic type, the types of terrestrial ecosystem, vegetation type, elevation, gradient, aspect and forest coverage rate, to reveal the feedback. Based on the results of China’s ecological geographic division, the entire study area belongs to the climate type of the south subtropical-humid region; thus, the climate type factor can be ruled out here.

During the study period (1985–2010), the elevation, gradient and aspect of the Guangxi Beibu Gulf Economic Zone remain relatively stable, so SRTM DEM was used and also to produce the gradient and aspect. Similarly, the changes of geomorphic type, the type of terrestrial ecosystems and vegetation type were relatively small and fragmented, so we selected a middle year (around 2000) of these datasets to represent the entire study period. The datasets above were provided by the Data Center for Resources and Environmental Sciences, Chinese Academy of Sciences (RESDC) ( http://www.resdc.cn ). Annual forest coverage rates were collected from the Guangxi Forestry Yearbooks (1958–2003) and the Guangxi Statistical Yearbooks. To ensure the continuity of the dataset in time, regression analysis was used to fix missing data.

Data sorting and pretreatment were conducted in ArcGIS10.1. Based on the input requirements of the geographical detector model ( http://www.sssampling.org/Excel-GeoDetector/ ), projection was unified to the projection coordinate system of Krasovsky-1940-Albers, and raster data were reclassified as 6 to 8 grades 27 , 36 , 37 and then converted to the vector data type. ArcGIS provided some discrete classification methods, such as the Equal Interval Method (EI), Quantile Value Method (QV), Natural Break Method (NB) and Geometrical Interval Method (GI), to reclassify the raster data. Different classification methods result in different P D,H values for the classified factor. The highest P D,H value result will indicate that this impact factor classification, using the discrete method, can be more representative as the classification of a geographical phenomenon, thus better revealing spatial distribution laws of the geographical phenomenon 38 . Natural factors were processed and classification methods were selected after many experiments ( Fig. 4 ).

This map was generated by ArcGIS 10.1 ( http://www.esrichina.com.cn/softwareproduct/ArcGIS/ ). ( a ) DEM with the sea-level elevation data at the resolution of 90 m of SRTM, which was divided into 8 types through QV; ( b ) Gradient data results obtained from the gradient analysis on DEM in ArcGIS, which was divided into 7 types through QV; ( c ) Aspect map results from the analysis of aspect on DEM in ArcGIS, which was divided into 8 types through NB; ( d ) Geomorphic type derived from a 1: 1,000,000 geomorphic map at the spatial resolution of 1000 * 1000 m; ( e ) Data of the terrestrial ecosystem type were derived from spatial distribution data of the Chinese terrestrial ecosystem types at the spatial resolution of 1000 * 1000 m; ( f ) Forest coverage rate change was calculated through Equation (1), and then, the trend of the forest coverage rate of each county (b value) was mapped and classified into 6 types through NB; ( g ) Data of the vegetation types were derived from a 1: 1,000,000 vegetation map at the spatial resolution of 1000 * 1000 m.

Potential social factors

Population density, GDP, farmland rate, urbanization rate and land use were selected as potential social factors that likely caused regional precipitation change because of changes in them, as described in the introduction. The population density, GDP and urbanization rate were derived from the Guangxi Statistical Yearbooks (1986~1991, 1993~1999 and 2001~2010), while the farmland rate was derived from the Guangxi Rural Statistical Yearbooks (1985~2010), and regression analysis methods were used to fill the entire 26-year period (1985–2010). Land use data (1980s and 2010) were collected from the Data Center for Resources and Environmental Sciences, Chinese Academy of Sciences (RESDC) ( http://www.resdc.cn ). We used the trend rates of these factors, derived from Equation (1), to express change because these factors have changed greatly over the 26-yearperiod. Similarly, the social factors adopted the same processing method as the potential natural factors ( Fig. 5 ).

This map was generated by ArcGIS 10.1. ( http://www.esrichina.com.cn/softwareproduct/ArcGIS/ ). ( a ) Population density change was calculated using Equation (1), and then, the trend of the annual population density of each county from 1985 to 2010 (b value) was mapped and classified into 8 types through QV; ( b ) GDP change was calculated using Equation (1), and then, the trend of the annual GDP of each county from 1985 to 2010 (b value) was mapped and divided into 8 types through QV; ( c ) Farmland rate change was calculated using Equation (1), and then, the trend of the annual farmland rate of each county from 1985 to 2010 (b value) was mapped and divided into 8 types through QV; ( d ) Urbanization rate change was calculated using Equation (1), and then, the trend of the annual urbanization rate of each county from 1985 to 2010 (b value) was mapped and divided into 6 types through GI; ( e ) Landuse and landcover change was derived from the subtraction of land use maps between the 1980 s and 2010.

Additional Information

How to cite this article : Zhao, Y. et al . Quantitative analysis of the impacts of terrestrial environmental factors on precipitation variation over the Beibu Gulf Economic Zone in Coastal Southwest China. Sci. Rep. 7 , 44412; doi: 10.1038/srep44412 (2017).

Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Wang, S., Gong, D. & Ye, J. Seasonal precipitation series of eastern China since 1880 and the variability. ActaGeographicaSinica-Chinese Edition 55 , 292–300 (2000).

Google Scholar  

Goswami, B. N., Venugopal, V., Sengupta, D., Madhusoodanan, M. S. & Xavier, P. K. Increasing trend of extreme rain events over India in a warming environment. Science 314 , 1442–1445 (2006).

ADS   CAS   PubMed   Google Scholar  

Charney, J. G. Dynamics of deserts and drought in the Sahel. Quarterly Journal of the Royal Meteorological Society 101 , 193–202 (2006).

ADS   Google Scholar  

Mcalpine, C. et al. Modeling the impact of vegetation cover change on Australia’s regional climate. Land & Water Australia 4 , 206–207 (2006).

Chen, G., Notaro, M., Liu, Z. & Liu, Y. Simulated local and remote biophysical effects of afforestation over Southeast United States in boreal summer. Journal of Climate 25 , 4511–4522 (2012).

Ma, D. et al. Short-term climatic impacts of afforestation in the East Asian monsoon region. Chinese Science Bulletin 58 , 2073–2081 (2013).

Olah, A. B. The Effects of Different Vegetation Types on Local Climate. Journal of Landscape Ecology 4 , 67–84 (2011).

Shi, Z. J. et al. Effect of vegetation on changes of temperature and precipitation Inner Mongolia, China. Ecology and Environmental Sciences 20 , 1594–1601 (2011).

Houze, R. A. Orographic effects on precipitating clouds. Reviews of Geophysics 50 , 294–295 (2012).

Chen, H., Li, Y. Y., Yang, Z. F. & Shen, Z. Y. Research on the Relationship between Terrain Factors and Precipitation. Research of Soil and Water Conservation 14 , 119–122 (2007).

Shu, S. J., Wang, Y. & Xiong, A. Y. Estimation and analysis for geographic and or graphic influences on precipitation distribution in China. Chinese Journal of Geophysics 50 , 1482–1493 (2007).

Pielke, R. A. et al. Interactions between the atmosphere and terrestrial ecosystems: influence on weather and climate. Global change biology 4 , 461–475 (1998).

Morales, P., Hickler, T., Rowell, D. P., Smith, B. & Sykes, M. T. Changes in European ecosystem productivity and carbon balance driven by regional climate model output. Global Change Biology 13 , 108–122 (2007).

Doherty, R. M., Sitch, S., Smith, B., Lewis, S. L. & Thornton, P. K. Implications of future climate and atmospheric CO2 content for regional biogeochemistry, biogeography and ecosystem services across East Africa. Global Change Biology 16 , 617–640 (2010).

Boucher, O., Myhre, G. & Myhre, A. Direct human influence of irrigation on atmospheric water vapour and climate. Climate Dynamics 22 , 597–603 (2004).

Sacks, W. J., Cook, B. I., Buenning, N., Levis, S. & Helkowski, J. H. Effects of global irrigation on the near-surface climate. Climate Dynamics 33 , 159–175 (2009).

Lobell, D. B., Bala, G. & Duffy, P. B. Biogeophysical impacts of cropland management changes on climate. Geophysical Research Letters 33 , 272–288 (2006).

Puma, M. J. & Cook, B. I. Effects of irrigation on global climate during the 20th century. Journal of Geophysical Research: Atmospheres 115 , 751–763 (2010).

Wen, L. J. & Jin, J. M. Modelling and analysis of the impact of irrigation on local arid climate over northwest China. Hydrological Processes 26 , 445–453 (2012).

Luo, W. et al. Spatial association between dissection density and environmental factors over the entire conterminous United States.Geophys. Res. Lett . 43 , 1–9 (2015).

Zhu, X. F., Zhao, A. Z., Li, Y. Z., Cao, S. & Li, M. Y. Impact of Irrigation on Climate. ActaEcologicaSinica 34 , 4816–4828 (2014).

Rosenfeld, D. Suppression of rain and snow by urban and industrial air pollution. Science 287 , 1793–1796 (2000).

Fang, F., Sun, L. D., Guo, J. Q. & Feng, J. Y. The Impact of Urban Economic Development on Precipitation Changing Trends over Northwestern China. Journal of natural Resources 29 , 1878–1887 (2014).

Liao, J. B., Wang, X. M., Li, Y. X. & Xia, B. C. An analysis study of the impacts of urbanization on precipitation in Guangzhou. Journal of the Meteorological Sciences 31 , 384–390 (2011).

Moore, N. & Rojstaczer, S. Irrigation-indued rainfall and the Great Plains. Journal of Applied Meteorology 40 , 1297–1309 (2001).

Lee, E. et al. Effects of irrigation and vegetation activity on early Indian summer monsoonvariability. International Journal of Climatology 29 , 573–581 (2009).

Wang, F. Y. & Liu, P. X. Characteristic Analysis on Sunshine Duration Change in Jiuquan Oasis in recent 45 Years. Plateau Meteorology 29 , 999–1004 (2010).

Su, B. D., Jiang, T. & Jin, W. B. Recent trends in observed temperature and precipitation extremes in the Yangtze River basin, China. Theoretical and Applied Climatology 83 , 139–151 (2006).

Ren, G. W., Hong, W. & Chen, W. Spatial patterns of change trend in rainfall of china. Quarterly Journal of Applied Meteorology 11 , 322–330 (2000).

Xu, Z. X. & Zhang, N. Long-term trend of precipitation in the Yellow River basin during the past 50 years. Geographical Research 25 , 27–34 (2006).

King, D. Climate change: the science and policy. Journal of Applied Ecology 42 , 779–783 (2005).

Zhang, Q., Han, Y. X. & Song, L. C. The Summarize of Development of Global Climate Change and Its Effect Factors. Advances In Earth Science 20 , 990–998 (2005).

Yang, D. H. & Yang, X. X. Study on cause of formation in Earth’s climatic changes. Progress In Geophysics 28 , 1666–1677 (2013).

Wang, J. F. et al. Geographical detectors-based health risk assessment and its application in the neural tube defects study of the Heshun region, China. International Journal of Geographical Information Science 24 , 107–127 (2010).

CAS   Google Scholar  

Zhan, D. S., Zhang, W. Z., Yu, J. H., Meng, B. & Dang, Y. X. Analysis of influencing mechanism of residents’ livability satisfactionin Beijing using geographical detector. Progress in Geography 34 , 966–975 (2015).

Hu, Y., Wang, J. F., Li, X. H., Ren, D. & Zhu, J. Geographical detector-based risk assessment of the under-five mortality in the 2008 Wenchuan earthquake, China. PLoS ONE 6 , 2592–2599 (2011).

Tong, L. G., Xu, X. L., Fu, Y. & Wei, F. H. Impact of environmental factors on snail distribution usinggeographical detector model. Progress in Geography 33 , 625–635 (2014).

Bi, S. B., Ji, H., Chen, C. C., Yang, H. R. & Shen, X. Application of geographical detector in human-environment relationshipstudy of prehistoric settlements. Progress in Geography 34 , 118–127 (2015).

Li, X. W., Xie, Y. F., Wang, J. F., Christakos, G. & Si, J. L. et al. Influence of planting patterns on Fluoroquinolone residues in the soil of an intensive vegetable cultivation area in north China. Science of the Total Environment 458 , 63–69 (2013).

Ding, Y., Cai, J. M., Ren, Z. P. & Yang, Z. S. Spatial disparities of economic growth rate of China’s National-level ETDZsand their determinants based on geographical detector analysis. Progress in Geography 33 , 657–666 (2014).

Wei, L. et al. Spatial association between dissection density and environmental factors over the entire conterminous United States. Geophysical Research Letter 43 , 692–700 (2016).

Wu, R., Zhang, J. Q., Bao, Y. H. & Zhang, F. Geographical Detector Model for Influencing Factors of Industrial Sector Carbon Dioxide Emissions in Inner Mongolia, China. Sustainability 8 , 149 (2016).

Wang, J. F., Zhang, T. L. & Fu, B. J. A measure of spatial stratified heterogeneity. Ecological Indicators 67 , 250–256 (2016).

Li, Z. & Su, Y. X. The analysis on precipitation variation characteristic in Guangxi from 1961 to 2004. Chinese Agricultural Science Bulletin 25 , 268–272 (2009).

Lin, B. & Liu, X. China’s Carbon Dioxide Emissions under the Urbanization Process :Influence Factors and Abatement Policies. Economic Research Journa l8 , 66–78 (2010).

Thielen, J., Wobrock, W., Gadian, A. M., Mestayer, P. G. & Creutin, J. D. The possible influence of urban surfaces on rainfall development: A sensitivity study in 2D in the meso-gamma-scale. Atmos Res 54 , l5–39 (2000).

Wu, P. T. & Zhao, X. N. Impact of climate change on agricultural water use and grain production in China. Transaction of the Chinese Society of Agricultural Engineering 26 , 1–6 (2010).

Mao, H. Q., Yan, X. D., Xiong, Z. & Tian, H. Q. Modeled impact of irrigation on regional climate in India. ActaEcologicaSinica 31 , 1038–1045 (2011).

DeAngelis, A., Dominguez, F., Fan, Y. & Robock, A. Evidence of enhanced precipitation due to irrigation over the Great Plains of the United States. Journal of Geophysical research 115 , 1–14 (2010).

Swann, A. L. S., Fung, I. Y. & Chiang, J. C. H. Mid-latitude afforestation shifts general circulation and tropical precipitation. PNAS 109 , 712–716 (2012).

Asddieh, B. & krakauer, N. Y. Global trends in extreme precipitation:climate models versus observations. Hydrology and Earth System Sciences 19 , 877–891 (2015).

Download references

Acknowledgements

This study is supported by the project of Natural Science Foundation of China (41461021,41661085), Guangxi Natural Science Foundation (2016GXNSFAA380094) and Opening fund of Key Laboratory of Environment Change and Resource Use in Beibu Gulf, Ministry of Education (Guangxi Teachers Education University) (2014BGERLXT15).

Author information

Authors and affiliations.

Key Laboratory of Environment Change and Resources Use in Beibu Gulf, Ministry of Education, Guangxi Teachers Education University, 175 Mingxiu east st, Nanning, 530001, China

Yinjun Zhao & Qing Lin

School of Geography and Planning, Guangxi Teachers Education University, 175 Mingxiu east st, Nanning, 530001, China

Yinjun Zhao, Qiyu Deng & Chunting Cai

You can also search for this author in PubMed   Google Scholar

Contributions

All authors contributed to the development of this manuscript led by Yinjun Zhao who designed the conception and carried out the analysis of results. The manuscript was primarily written by Yinjun Zhao and Qiyu Deng. Qing Lin contributed analysis ideas. Qiyu Deng and Chunting Cai collected and processed the data.

Corresponding author

Correspondence to Yinjun Zhao .

Ethics declarations

Competing interests.

The authors declare no competing financial interests.

Supplementary information

Supplementary information (pdf 338 kb), rights and permissions.

This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/

Reprints and permissions

About this article

Cite this article.

Zhao, Y., Deng, Q., Lin, Q. et al. Quantitative analysis of the impacts of terrestrial environmental factors on precipitation variation over the Beibu Gulf Economic Zone in Coastal Southwest China. Sci Rep 7 , 44412 (2017). https://doi.org/10.1038/srep44412

Download citation

Received : 08 July 2016

Accepted : 10 February 2017

Published : 15 March 2017

DOI : https://doi.org/10.1038/srep44412

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

This article is cited by

Embankment deformation characteristics analysis of an expressway widening project near a pond.

• Dufeng Zhang

Scientific Reports (2023)

Spatiotemporal distribution of flood disasters in Asia and influencing factors in 1980–2019

• Xiaojie Wang
• Junqiang Xia
• Shanshan Deng

Natural Hazards (2021)

Ecological analysis of Kashin-Beck osteoarthropathy risk factors in Tibet’s Qamdo City, China

Scientific Reports (2019)

Quantitatively evaluate the environmental impact factors of the life expectancy in Tibet, China

• Chengqun Yu

Environmental Geochemistry and Health (2019)

By submitting a comment you agree to abide by our Terms and Community Guidelines . If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Quick links

• Explore articles by subject
• Guide to authors
• Editorial policies

Sign up for the Nature Briefing: Anthropocene newsletter — what matters in anthropocene research, free to your inbox weekly.

Quantitative Methods for Current Environmental Issues

• Conference proceedings
• © 2002
• Clive W. Anderson 0 ,
• Vic Barnett 1 ,
• Philip C. Chatwin 2 ,
• Abdel H. El-Shaarawi 3

School of Mathematics and Statistics, University of Sheffield, Sheffield, UK

You can also search for this editor in PubMed   Google Scholar

School of Mathematical Sciences, University of Nottingham, Nottingham, UK

National water research institute, burlington, canada.

Features the most recent quantitative methods for studying environmental issues

Includes contributions from the foremost researchers in the field

8457 Accesses

236 Citations

This is a preview of subscription content, log in via an institution to check access.

Access this book

Subscribe and save.

• Get 10 units per month
• Download Article/Chapter or eBook
• 1 Unit = 1 Article or 1 Chapter
• Cancel anytime
• Available as PDF
• Read on any device
• Instant download
• Own it forever
• Compact, lightweight edition
• Dispatched in 3 to 5 business days
• Free shipping worldwide - see info

Tax calculation will be finalised at checkout

Other ways to access

Licence this eBook for your library

Institutional subscriptions

About this book

Similar content being viewed by others, the insignificance of thresholds in environmental impact assessment: an illustrative case study in canada, basic mathematical errors may make ecological assessments unreliable.

What Is Sociological About Environmental Sociology?: Qualitative Methods in an Era of Rapid Environmental Change

• environment
• environmental science
• environmental sciences

Table of contents (12 papers)

Front matter, spatial and temporal models and methods, modeling spatio-temporally misaligned areal and point process environmental data.

• Bradley P. Carlin, Andrew S. Mugglin, Li Zhu, Alan E. Gelfand

Space and Space-Time Modeling using Process Convolutions

• Dave Higdon

Multivariate Kriging for Interpolating with Data from Different Sources

• H. Wackernagel, L. Bertino, J. P. Sierra, J. González del Río

Environmental Sampling and Standards

Distance sampling: recent advances and future directions.

• S. T. Buckland, L. Thomas, F. F. C. Marques, S. Strindberg, S. L. Hedley, J. H. Pollard et al.

Setting Environmental Standards: A Statistical Approach

• Vic Barnett, Marion Bown

Atmosphere and Ocean

The interpretation and validation of measurements of the ocean wave directional spectrum.

• Lucy R. Wyatt

Thermal Energy Emission and Propagation from Accidents

• A. Pelliccioni, F. Altavilla, S. Berardi

Development and Application of an Extended Methodology to Validate Short-Range Atmospheric Dispersion Models

• Harry Eleveld, Harry Slaper

Uncertainty and Sensitivity of Dispersion Model Results to Meteorological Inputs: Two Case Studies

• Joseph C. Chang

Risk and Uncertainty

Statistics and the environmental sciences: approaches to model combination.

• Gudmund Høst

Bayesian Analysis of Computer Code Outputs

• Marc C. Kennedy, Anthony O’Hagan, Neil Higgins

The Realities of Decision Making on Risks

• Jim McQuaid

Back Matter

Editors and affiliations.

Clive W. Anderson, Philip C. Chatwin

Vic Barnett

Abdel H. El-Shaarawi

Bibliographic Information

Book Title : Quantitative Methods for Current Environmental Issues

Editors : Clive W. Anderson, Vic Barnett, Philip C. Chatwin, Abdel H. El-Shaarawi

DOI : https://doi.org/10.1007/978-1-4471-0657-9

Publisher : Springer London

eBook Packages : Springer Book Archive

Copyright Information : Springer-Verlag London Limited 2002

Hardcover ISBN : 978-1-85233-294-5 Due: 15 January 2002

Softcover ISBN : 978-1-4471-1171-9 Published: 15 September 2011

eBook ISBN : 978-1-4471-0657-9 Published: 06 December 2012

Edition Number : 1

Number of Pages : X, 270

Topics : Statistical Theory and Methods , Math. Appl. in Environmental Science , Statistics for Engineering, Physics, Computer Science, Chemistry and Earth Sciences , Applications of Mathematics

• Publish with us

Policies and ethics

• Find a journal
• Track your research

Study at Cambridge

About the university, research at cambridge.

• Undergraduate courses
• Events and open days
• Fees and finance
• Postgraduate courses
• How to apply
• Postgraduate events
• Fees and funding
• International students
• Continuing education
• Executive and professional education
• Courses in education
• How the University and Colleges work
• Term dates and calendars
• Visiting the University
• Annual reports
• Equality and diversity
• A global university
• Public engagement
• Give to Cambridge
• For Cambridge students
• For our researchers
• Business and enterprise
• Colleges & departments
• Email & phone search
• Museums & collections
• Subject information
• 2nd Year (Part IB) Subject information
• Natural Sciences Tripos
• Prospective students overview
• Course structure
• Maths in the first year
• Admission requirements and advice
• Reading List
• Course costs
• Subject information overview
• 1st Year (Part IA) Subject information overview
• Biology of Cells
• Earth Sciences
• Evolution and Behaviour
• Introduction to Experimental Psychology: From Brain to Cognition
• Materials Science
• Mathematical Biology
• Mathematics
• Physiology of Organisms
• 2nd Year (Part IB) Subject information overview
• Biochemistry & Molecular Biology
• Biology of Disease
• Cell & Developmental Biology
• Chemistry A
• Chemistry B
• Earth Sciences A
• Earth Sciences B
• Ecology & Conservation
• Evolution & Animal Diversity
• Experimental Psychology
• History and Philosophy of Science
• Mathematical and Computational Biology
• Neurobiology
• Pharmacology
• Plant & Microbial Sciences
• Quantitative Environmental Science
• 3rd Year (Part II) Subject information overview
• Astrophysics*
• Biochemistry*
• Materials Science*
• Physiology, Development and Neuroscience
• Plant Sciences
• Biological & Biomedical Sciences
• Physical Sciences
• 4th Year (Part III) Subject information overview
• Astrophysics
• Biochemistry
• Systems Biology
• Quantitative Climate and Environmental Science
• Programme specification for the Natural Sciences Tripos overview
• Aims and Outcomes

Teaching and Learning Methods

• Support for Students and their Learning
• Criteria for Admission
• Mechanisms for evaluating and improving
• Employment and Careers
• Management of Education Quality and Standards
• Current students overview
• New students overview
• Learning and teaching - methods and mechanisms
• How hard should I work?
• General tips and further help
• A year-by-year guide
• 1st Year (Part IA)
• 2nd Year (Part IB)
• 3rd Year (Part II) overview
• Part II Subject Allocation overview
• Allocation Procedure
• Further Advice to Tutors and Directors of Studies
• Selection Criteria
• Part II Physical Sciences
• 4th year (Part III)
• Academic Misconduct
• Student Support
• Examinations overview
• Parts II and III
• Form and Conduct
• Calculator Notice
• Results and Marks
• Extensions and Late Submissions
• Data Retention Policy
• Contacts overview
• Departmental websites
• Directors of Studies and Staff overview
• Directors of Studies email list
• IA Practical Allocation - Timeline, Induction and System Access overview
• IA Practical Allocation - Information for Colleges
• 1st Year (Part IA) Subject information
• 3rd Year (Part II) Subject information
• 4th Year (Part III) Subject information
• Programme specification for the Natural Sciences Tripos

Subject Summary

Quantitative Environmental Sciences (QES) is designed to integrate the physics, maths, biology, and chemistry behind environmental issues and models of environmental systems. Students explore how Natural Sciences are instrumental in solving the real-world problems that we are currently facing and will continue to face in the coming decades. The course is broadly structured around the presentation of key environmental systems, and the fundamental mathematical principles underlying designing models to study environmental challenges. Students will also gain experience in breaking down scientific challenges to policy makers through writing a policy paper. The practical component will teach students how to build simple numerical climate models in Python and to visualise environmental datasets.

The course begins with an overview of how the surface of the planet functions, with introductory lectures on the global carbon cycle, hydrological cycle, the source of energy on the planet and how oceans move. Michaelmas term focuses on groundwater flow, surface water, ice dynamics, and atmospheric chemistry, with consideration of challenges in groundwater contamination, flooding, ice melt and sea level rise, and air pollution. The focus is on the applied mathematics and introductory fluid dynamics behind these systems. There are also several lectures in Michaelmas dedicated to presentation of environmental datasets and the role of science in government.

Student choose a policy paper topic and write a modified policy paper that is due early in Lent term. In the Lent term lecture course we consider the global environment, how the oceans, and atmosphere serves to redistribute heat, energy and carbon around the planet. This includes modules on the physics of atmosphere and ocean transport as well as the chemistry of carbon in the ocean, case studies on the study of the ocean system in polar regions, and an understanding of the land carbon system. The goal of the Lent term is to explore the role of the oceans, land, and atmosphere in the global climate system. In the Lent term practicals students will be build a simple climate model, which contributes to a lab report due in the Easter vacation. In the final term, students will learn about how the energy transition will impact environmental systems, and where solutions to the future climate crisis may lie.

QES is a multidisciplinary course, taught by lecturers from Maths (DAMTP), Chemistry, Earth Sciences, and the British Antarctic Survey.

Programme Specification

This course is taught jointly by DAMPT, the Departments of Earth Sciences and Chemistry, with guest lectures from the British Antarctic Survey. It is administered through DAMTP.

This course aims to:

• teach a cross disciplinary course on the use of Mathematics in environmental studies and in solving environmental challenges.
• engender an understanding of the role that natural scientists (in particular) will play in designing solutions to environmental challenges (air pollution, groundwater pollution, climate change) that students will face over the course of their lifetimes;
• learn to apply the maths that students have been taught to environmental problems, and how to write simple code to understand what data is telling us, and to build a simple climate model;
• build an understanding of how Earth’s surface environment functions, where energy comes from, where there are environmental challenges and what the nature of the solutions to these environmental challenges might be;
• understand how policy makers can benefit from the outcomes of environmental models and where and how science can inform policy.

Learning Outcomes

At the end of the course students should:

• understand how knowledge of physics, chemistry and biology informs models of environmental systems;
• be able to write code in python to build simple environmentally relevant models (e.g. a box model for carbon, or a flow model for groundwater);
• be able to work with large environmental datasets (e.g. data incorporation, visualisation, regression);
• understand how to disseminate scientific concepts to a general audience (e.g. policy makers or the public).

These include lectures, supervisions and online practicals. The practicals are done in the students own time, but with drop-in sessions with demonstrators available during the week.

Assessment for this course is through:

• one unseen written examination (for aims 1-5 and learning outcomes 1 and 4);
• one unseen computational exam (for aims 1, 3 and learning outcomes 1-3);
• one lab report base on the practical component (for aims 1-4 and learning outcomes 1-3);
• one policy paper (for aims 2,3 and 5, and learning outcome 4).

Courses of Preparation

Essential: A level Mathematics.

Recommended: A Level Further Mathematics.

Additional Information

Further information is available on the Course Websites pages.

• Information provided by:     [email protected]
• Site Privacy & Cookie Policies

© 2024 University of Cambridge

• Contact the University
• Accessibility
• Freedom of information
• Privacy policy and cookies
• Statement on Modern Slavery
• Terms and conditions
• University A-Z
• Undergraduate
• Postgraduate
• Research news
• About research at Cambridge
• Spotlight on...

• General & Introductory Civil Engineering & Construction
• Environmental Engineering

Ecological Numeracy: Quantitative Analysis of Environmental Issues

ISBN: 978-0-471-18309-9

Robert A. Herendeen

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings

Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .

• Advanced Search
• Journal List

Using qualitative approaches to improve quantitative inferences in environmental psychology

Neil a. lewis, jr..

a Cornell University

Mario Bravo

b Environmental Defense Fund

Sarah Naiman

Adam r. pearson.

c Pomona College

Rainer Romero-Canyas

Jonathon p. schuldt, hwanseok song.

d Purdue University

This article describes the qualitative approach used to generate and interpret the quantitative study reported by Song and colleagues’ (2020) in their article, “What counts as an ‘environmental’ issue? Differences in environmental issue conceptualization across race, ethnicity, and socioeconomic status.” Song and colleagues (2020) describe the results of a survey documenting that, in the United States, White and high-SES respondents perceive environmental issues differently than their non-White and lower-SES counterparts, reflecting structural differences in environmental risks. While Song and colleagues (2020) discuss the survey results in detail, the discussion of the qualitative research that led to the creation of that survey was limited due to space constraints. The current article provides a more holistic account of the methods behind the Song and colleagues (2020) study by discussing the qualitative component of the research in detail. In addition to discussing how the qualitative research complements and critically informs the findings reported by Song et al., we also consider the broader implications and value of integrating qualitative and quantitative methods in environmental psychology.

• • Conduct qualitative study to inform quantitative design.
• • Use qualitative patterns to make inferences about quantitative indicators.

Graphical abstract

Specifications table

Background and study rationale

Research in social psychology has consistently documented that people's embodied traits and characteristics, social relationships, roles, and group memberships shape not only how we come to define ourselves [28] , but also how we perceive the world around us, and act on those perceptions. That is, our social contexts and identities provide a lens through which to interpret and make meaning of the world around us [20 , 29] . They also allow us to determine which groups we belong to [42] . And, because social institutions make group differences salient through differential treatment [34] , group memberships end up shaping how we process and make sense of our experiences [19] , and understand the differences between our experiences and the experiences of others with different attributes [36] . This finding—that our social contexts and identities shape perception, cognition, and action has been well-document across a variety of domains (c.f., [30] ), including environmental psychology.

Research over the past few decades increasingly documents differences in the ways that individuals think about environmental issues as a function of their race, ethnicity, and socioeconomic status (SES) group membership, variables that serve as proxies for the differential experiences that people have in segregated and otherwise stratified social contexts [33 , 35 , 37 , 45] . Take, for instance, the sizable literature on racial/ethnic differences in environmental attitudes. Numerous studies reveal that—contrary to prevailing stereotypes—racial/ethnic minorities express comparable and oftentimes greater levels of environmental concern and elevated environmental risk perceptions, relative to Whites (e.g., [13 , 14 , 17 , 25 , 41] ). At the same time, work in public opinion finds that Americans substantially underestimate the environmental concerns of non-Whites and lower-income respondents —an “environmental belief paradox” observed even among non-White and lower-income respondents [31] . That is, in the United States for example, nationally representative probability sample studies have documented that the groups that are most concerned about the environment are Latina/os, Low-income Americans, Asian Americans, and African Americans (in that order), but those are the very groups that Americans misperceive to be the least concerned about the environment [31] . Although different explanations for these misperceptions have been offered, a common account involves the assumption that these groups have more immediate concerns, such as concerns about employment, that preclude them from prioritizing the environment (for a review, see [12] ).

Research on environmental justice (see [26] ) and environmental racism [33 , 45] offers an alternative perspective on group differences in environmental concern. In particular, the environmental deprivation hypothesis posits that heightened levels of environmental concern expressed by minority and lower-SES communities reflect awareness of the disproportionate environmental risks (e.g., pollution exposure; [45] ) these communities face, as well as recognition of social conditions that can exacerbate these risks [33] . Supporting this reasoning, a substantial literature documents that minorities and lower-SES groups are exposed to relatively greater levels of environmental hazards (e.g., air and water pollution) [24 , 48] .

Moreover, minority and lower-income communities face a “double jeopardy” when it comes to environmental threats – not only are they often more exposed to environmental hazards, but, due to systemic social inequities, they are also more sensitive to that exposure, particularly in urban areas [38 , 47] . For instance, low-income housing sites are often more exposed to environmental hazards, and poor infrastructure, and limited access to transportation and other public services can further amplify social vulnerability, and reduce the capacity of communities to mobilize to address environmental problems (for reviews, see [10] ; and [38] ).

Whereas early risk research on environmental hazards emphasized vulnerability due to the hazard itself (e.g., flooding due to coastal erosion, the nature and composition of pollutants), more recent work on “social vulnerability” has focused on the underlying social conditions, including systemic political and economic inequality, segregation, structural racism, unequal access to transportation, limited opportunities for education, weak infrastructure, and existing health disparities, that make humans vulnerable – the focus of environmental justice [6] and environmental racism research [45] . Neighborhood-level factors, including access to transportation, healthy food, and critical services, including healthcare, parks, open spaces, social environments, crime rates, and physical features of urban environments (e.g., traffic density, housing quality) can further shape vulnerability to environmental hazards [23] .

Additional work has revealed the ways in which environmental hazards, in turn, can expose and exacerbate existing social inequities, as illustrated by the disproportionate effects of Hurricane Katrina on minority and low-income communities in the U.S. Gulf Coast [8] . A sizable literature on vulnerability to environmental impacts over the past decade now documents dual influences of both biophysical factors and social conditions that make communities vulnerable. However, few studies have examined how these broader understandings of environmental threats may be perceived among different segments of the public – the focus of the present research.

Group differences in the experience of environmental risks, or awareness that these disparities exist, may hold implications for how different groups conceptualize environmental issues. Although theories from the environmental justice literature have discussed this possibility [1 , 22 , 44] , limited empirical research in the environmental psychology literature speaks to this possibility directly. Whereas environmental science scholarship and advocacy has traditionally focused on ecocentric environmental issues (e.g., industrial pollution, flooding, drought), reflecting biophysical hazards, environmental justice broadens this focus to consider social conditions that magnify human harm (e.g., racism, poverty), as well as the disparate impact of environmental harms across groups [25 , 26] . Indeed, communities of color have long protested issues such as inadequate sanitation, lead poisoning in urban areas, and asbestos in schools and work at the local level, and placed these on the agenda of the Civil Rights Movement [32] . According to Taylor [44] , due to their cultural roots, mainstream environmentalists appealing to the White middle-class tend to associate images of wilderness and wildlife protection with romanticized 19th century experiences; in contrast, environmental justice advocates may evoke cultural images of racism, land appropriation, and community destruction associated with the same era.

Because previous research documented that White and wealthier Americans had different exposure to environmental hazards and political experiences with environmental issues than their racial and ethnic minority and lower socioeconomic status peers, Song and colleagues [40] examined whether those differences have implications for how different groups of Americans construe environmental issues—that is, whether they have different definitions of what “counts” as an environmental issue. They conducted a national survey to explore that research question in a quantitative manner. That survey was developed not only in light of these previous findings in the literature, but also due to a qualitative study conducted with residents in a low-income racial and ethnic minority community in a major US city. The goal of this article is to describe that qualitative study, and its implications for understanding the quantitative findings reported in Song and colleagues [40] , as well as the broader implications of incorporating qualitative approaches for improving quantitative inferences in environmental psychology research.

Our investigation of racial and economic group differences in environmental issue conceptualization began, admittedly, by accident. After discovering that racial and ethnic minority and low-income Americans express the highest levels of environmental concern in national opinion polling but are stereotyped (even among themselves) as having the lowest levels of concern [31] , we wondered whether this misperception could be addressed with a social norm intervention [43] , and whether correcting the misperception would have downstream consequences for environmental engagement. To develop our normative intervention, we decided to conduct focus group research with Latina/o community group members to gather information that could be used to create motivational intervention messages [7] . We chose to focus on Latina/o participants because in our previous research on environmental concern, that was the group whose environmental concerns were misperceived the most. That is, although other groups also face environmental hazards and are misperceived as not caring about the environment, our previous nationally representative studies documented that Latina/o Americans are the group of Americans who are most concerned about the environment, but are perceived as being the least concerned [31] . This misperception occurs even among Latina/o participants—Latina/o participants report high levels of concern about the environment, but when asked how concerned other Latina/o people are about the environment, they perceive fellow Latina/os as significantly less concerned than the self-report data suggest. Correcting such misperceptions is important given that misperceiving ingroup norms may deter groups from partaking in collective action (see [31] ).

Given these previous findings, we partnered with the Environmental Defense Fund to leverage their network of Latina/o community organizations to recruit participants for our research study. Through this partnership we decided to recruit participants in San Antonio, Texas. San Antonio is a majority Latina/o city that is home to one of the largest Hispanic populations in the U.S., which made it an ideal location to recruit a sufficiently sized sample to gather information to create the messages we hoped to generate and test. Thus, the primary goal of the focus group study was to elicit leading environmental and sustainability issues that could be used in the creation of normative messages (see [7] ). As described in the remainder of this article however, once we conducted the focus groups, the goal of the research changed based on what we learned, as sometimes happens when researchers step outside of the laboratory and interact with people rather than just the datapoints in our spreadsheets [5 , 11 , 21] . These focus group findings introduced a different set of questions, which led us to first investigate potential group differences in environmental issue conceptualization and their antecedents and consequences, using a semi-structured interview protocol adapted from Krueger and Casey [16] – a comprehensive and practical guidebook for conducting focus groups for applied research purposes.

Interview procedure

Our team worked with community partners to recruit 24 representatives from 16 Hispanic and Latino community organizations working in and around San Antonio. The focus groups were hosted in a community center in a low-income neighborhood on the city's Westside (zip code 78207). The organizations themselves were diverse in mission and scope, and included those working on: voter registration, business development, culture and arts, and promoting access to healthy food. Invitees were originally contacted by a member of our research team and informed about our plan to hold small, informal focus groups focused on the following questions: “What are the top issues facing your community?” “What environmental changes would you like to see?” “Do your friends and family care about the environment?” These questions were designed to inductively illuminate specific topics that participants associated with environmental degradation in their communities. Invitees were further informed that the output from these focus groups would shape the next steps of our research, including our plan to investigate how different ways of communicating about the environmental beliefs and concerns of different communities would affect people and motivate them to take environmental action.

Participants attended one of three semi-structured focus groups discussions, each of which involved between six and nine participants, in addition to the moderator. For observation and data-recording purposes, three other members of the research team were present in the room but did not sit at the main table or participate in the discussion – they took notes as the conversations progressed. One of the three 1 focus groups was audio-taped for transcription, to compare to written notes of the research team. Upon arrival, participants provided informed consent and were reminded that the purpose of the discussions was for our research team to learn about “issues facing your community in San Antonio.” Moreover, participants were informed that the purpose was not to establish consensus and that there were no right or wrong answers, but rather, our goal was to have them generate as broad a list of issue priorities as possible.

The moderator then proceeded to walk the group through a semi-structured interview protocol (adapted from [16] ), which began with asking participants to generate “three issues that seem most important” to the Latina/o community in San Antonio. After about a 1-minute pause to allow participants to privately reflect, the moderator invited each participant to share the three issues they identified, which were recorded as a list on a large notepad by another member of our team (the notetaker). As the moderator went from participant to participant and novel issues were shared, they were added to the growing list which was visible to all participants. This was intended to encourage discussion about the ways in which different issues were the same, related but not the same, or entirely different. Throughout this reporting process, the moderator and notetaker were mainly passive observers; they did, however, prompt participants to clarify their comments if they were unclear in and of themselves, or in their relationship to other issues that had already been recorded.

After hearing from everyone, the moderator proceeded to the second prompt, which asked participants to generate, specifically, the most important environmental issues facing their communities. Once again, the protocol asked the moderator to pause to allow time for each participant to identify issues and reflect, although in some cases, the discussion was proceeding organically by this point and participants simply volunteered issues as they occurred to them. As before, the notetaker listed novel issues on the large, visible sheet of paper.

Following the discussion of important issues in general and of important environmental issues specifically, the protocol called for the moderator to prompt participants to name any issues that had not yet been named, and as a final question, to share “one thing” they would tell the mayor of San Antonio, if they were given the opportunity. This closing question was intended to encourage participants to think concretely about community problems and their potential solutions—to mitigate the possibility that framing questions in terms of “issues” might have invited responses that were too abstract for the purposes of our research questions; it also allowed participants to highlight which issues they thought were the most important. Finally, participants were thanked, provided with a paper survey to report any additional comments, and given a debriefing form with information to contact us if there was anything else they wished to discuss. Altogether, each focus group session lasted about 1.5 hours.

Coding and analysis procedure

Following the focus group discussions, the research team that was on site for data collection (all senior members of the research team) returned to our hotels to transcribe our handwritten notes and upload our typed notes to our secure Server. After we returned to campus, the graduate student members of our research team took the large notepad posters which contained participant responses from each focus group and transcribed the issues that were discussed in each focus group. They also transcribed the audio file from the one group for which we had an audio recording, but we decided not to include that in our coding as that would have introduced an imbalance in the amount of information available for each group (i.e., one focus group would have had more information, and of a different source, than the other focus groups).

After all files were transcribed, one of the graduate student members of our research team went through all transcription files to extract key themes from the focus groups. She began by conducting a frequency count of issue mentions in each group, and then identified the convergence of issues across groups. Specifically, groups varied in the number and type of issues mentioned: the first group generated 34 different issues, the second group generated 22 different issues, and the third group generated 16 different issues. We discussed these issues as a larger research team to determine whether there was convergence across the groups that could be generative. Although issues varied across groups as a function of the particular people who were present, there were some consistent issues that were raised in every group and emerged as common themes. These themes informed the development of the research questions and survey described by Song and colleagues [40] .

Of the 34 issues mentioned in the first focus group the issues that came up most frequently were: poverty, low wages, and employment (6 unique mentions), educational opportunities (5 unique mentions), disparities and economic inequality (3 unique mentions), and health (3 unique mentions); other issues only received one unique mention from respondents. Of the 22 issues brought up in the second focus group, the issues that came up most frequently were: education (5 unique mentions), economic inequality (3 unique mentions), access to nutritious food (3 unique mentions), walkability in the city (3 unique mentions), issues with mass transit (3 unique mentions), and people not voting (3 unique mentions); other issues received only one unique mention from respondents. Of the 16 issues mentioned in the third focus group, the issues that came up most frequently were: air quality (4 unique mentions), flood and drainage issues in the city (3 unique mentions), educational opportunities (3 unique mentions), and poverty (3 unique mentions). After seeing both what participants brought up consistently across groups as well as how they discussed these issues, it became clear to us that our participants’ environmental concerns could be synthesized into two emergent themes.

One emergent theme can be characterized as environmental issues are social issues. When participants were asked to identify the most important environmental issues facing Latina/os in San Antonio, they were quick to highlight the links between what are typically viewed as environmental issues and the social issues that had been previously discussed. For example, issues like air pollution and lack of green spaces were explicitly discussed as intertwined with a host of social issues, including racism, economic inequality, and health problems (notably, obesity). One notable example concerns discarded drug needles on lawns and walkways in poorer neighborhoods, which multiple participants cited as a leading environmental issue (as a form of litter, and because it discourages residents from safely accessing and enjoying outdoor spaces) that is inextricable from inequities in education, employment, and health care access in one of the poorest zip codes in the U.S. A related example concerns the lack and poor maintenance of sidewalks in the city's lower-income neighborhoods, which participants linked to multiple effects—including air pollution due to an increased reliance on cars, restricted access to (already limited) green space, and limited opportunity for physical exercise.

Another emergent theme can be characterized as environmental issues are local issues. When asked to name leading environmental issues facing their communities, focus group participants tended not to mention issues playing out at the national or global scale (e.g., climate change, fracking), but rather, they tended to name issues that they perceived as affecting certain neighborhoods in San Antonio more than others, as demonstrated by the drug litter example above. Other examples endorsed by multiple participants included localized flooding resulting from inadequate storm-water infrastructure in low-income neighborhoods of mostly Latina/o residents and differential exposure to harmful soil and air pollution originating from a nearby military base. Indeed, participants explicitly acknowledged that members of their communities might be primarily concerned about localized issues that they encounter on a daily basis (“[It's] not like global warming,” one participant noted when describing the drug litter issue), while members of more-affluent communities elsewhere in San Antonio might be more concerned about broader-scale environmental issues (“someone from the North side isn't going to talk about drug needles,” one participant explained, to which another replied, “somebody from the North side would probably say ‘fracking’”).

Significance and applications

Prior to conducting this focus group study with members of Latina/o community organizations in San Antonio, we expected that discussions would revolve around the impacts of environmental issues that are discussed more frequently in the environmental psychology literature—issues such as climate change, air quality, and water quality. Although those issues did surface, the discussions were dominated by issues that have received greater attention in the environmental justice and environmental racism literatures—including issues that represent social determinants and consequences of environmental risks, with lack of education, poverty, built environments (i.e., lack of sidewalks) and obesity being notable examples. Moreover, much of the discussion focused on drawing out the interconnections that participants saw between what we ended up terming human-oriented issues (e.g., poverty, lack of access to grocery stores) rather than more eco-oriented issues (e.g., climate change, industrial pollution) that are more frequently discussed in the environmental psychology literature [40] . For example, in our participants’ minds, inadequate sidewalks were inextricably linked to poor air quality because the lack of sidewalks (and other reliable public transit) forced residents to drive more frequently, and the cars that people in their neighborhoods could typically afford tended to be older and more polluting than those in wealthier San Antonio neighborhoods.

Conducting this focus group study made salient and explicit a phenomenon that environmental justice and environmental racism scholars have written about, but has received limited empirical attention, to-date, in environmental psychology: that group-based differences in vulnerabilities to environmental hazards that have emerged due to structural inequities [2 , 25 , 26 , 45] may lead members of different groups to different conceptions of what “counts” as an environmental issue [40] . The principal finding from Song and colleagues [40] —that issues including poverty, drug abuse, and racism are more likely to be counted as environmental issues by non-White and lower-income survey respondents, compared to their White and higher-income counterparts—is consistent not only with environmental justice perspectives, but also research in psychology on how differences in social structure shape the lenses through which people interpret the world around them [30] . Had we not held those initial focus groups, we likely would have surveyed our respondents about a narrower set of issues, or have been less equipped to make sense of our quantitative results. Understanding how the qualitative patterns in empirical data emerge from models of psychological processes is essential for making inferences about quantitative indicators [27] . In this way, our team came to more fully appreciate the value of combining qualitative examinations with quantitative approaches, and in particular, the importance of linking environmental psychology, environmental justice, and environmental racism perspectives when seeking to deepen our understanding of whether and how different groups may think differently about environmental issues [3] . We limit our ability to generate useful explanations and predictions when we employ one approach without the other.

Limitations and Constraints on Generalizability

Although informative for improving the quantitative study reported by Song and colleagues [40] , and more generally for thinking about quantitative inferences in environmental psychology, the current study has some important limitations that must be considered when interpreting the findings and their generalizability [39] . We recruited a sample of Latina/o individuals from San Antonio, Texas to participate in our study. As noted earlier, this decision was driven by findings from previous research documenting misperceptions of the environmental concerns of this group among the US public [31] . San Antonio is a city wherein the vast majority of residents are Latina/o; in other words, it is a “majority-minority” city. The bulk of environmental justice and environmental racism literatures sample participants from low SES minority populations in settings in which they are minorities within a majority race population. While we do not yet know how these contextual differences might affect the results found in the current paper, a growing body of meta-scientific research suggests that there is often substantial variability in social scientific research by context (e.g., [15] ), and thus future research should explore this more systematically due to its implications for both theory advancement and practical application [9 , 39] .

Despite the limitations of the current findings, we believe there are important theoretical, practical, and ethical gains associated with combining qualitative and quantitative methods in (environmental) psychology. We discussed the theoretical and practical elements in the previous sections, and want to end by discussing the ethical argument—the argument about the implications of our methodological practices for our research's influence in society [4 , 18] . Over the past few years, as occurs every few decades [4] , professional psychological science organizations have been encouraging psychologists to “give psychology away” in hopes that psychologists can have similar influences on public policy as our colleagues in economics [46] . If research in environmental psychology is to inform social policies about environmental issues, then it ought to not only include broad representation of groups that are disproportionately affected by environmental hazards [18] , it also needs to reflect deeper perspectives of people from those groups—insights gained by combining quantitative techniques with qualitative techniques.

Acknowledgements

We thank Matthew Ballew, Julie Davydova, H. Oliver Gao, Robert Garcia, and Sofia Hiltner for their assistance with this research project. This research was funded by an Academic Venture Fund grant from the David R. Atkinson Center for a Sustainable Future (PI: J. Schuldt; Co-PIs: N. Lewis, Jr. & H. O. Gao) and by a David L. Hirsch III and Susan H. Hirsch Research Initiation Grant (A.R.P.).

Declaration of interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

1 One group did not consent to being recorded; in another group the moderator forgot to press record at the beginning of the session.

• Privacy Policy

Home » Quantitative Research – Methods, Types and Analysis

Quantitative Research – Methods, Types and Analysis

Table of Contents

Quantitative Research

Quantitative research is a type of research that collects and analyzes numerical data to test hypotheses and answer research questions . This research typically involves a large sample size and uses statistical analysis to make inferences about a population based on the data collected. It often involves the use of surveys, experiments, or other structured data collection methods to gather quantitative data.

Quantitative Research Methods

Quantitative Research Methods are as follows:

Descriptive Research Design

Descriptive research design is used to describe the characteristics of a population or phenomenon being studied. This research method is used to answer the questions of what, where, when, and how. Descriptive research designs use a variety of methods such as observation, case studies, and surveys to collect data. The data is then analyzed using statistical tools to identify patterns and relationships.

Correlational Research Design

Correlational research design is used to investigate the relationship between two or more variables. Researchers use correlational research to determine whether a relationship exists between variables and to what extent they are related. This research method involves collecting data from a sample and analyzing it using statistical tools such as correlation coefficients.

Quasi-experimental Research Design

Quasi-experimental research design is used to investigate cause-and-effect relationships between variables. This research method is similar to experimental research design, but it lacks full control over the independent variable. Researchers use quasi-experimental research designs when it is not feasible or ethical to manipulate the independent variable.

Experimental Research Design

Experimental research design is used to investigate cause-and-effect relationships between variables. This research method involves manipulating the independent variable and observing the effects on the dependent variable. Researchers use experimental research designs to test hypotheses and establish cause-and-effect relationships.

Survey Research

Survey research involves collecting data from a sample of individuals using a standardized questionnaire. This research method is used to gather information on attitudes, beliefs, and behaviors of individuals. Researchers use survey research to collect data quickly and efficiently from a large sample size. Survey research can be conducted through various methods such as online, phone, mail, or in-person interviews.

Quantitative Research Analysis Methods

Here are some commonly used quantitative research analysis methods:

Statistical Analysis

Statistical analysis is the most common quantitative research analysis method. It involves using statistical tools and techniques to analyze the numerical data collected during the research process. Statistical analysis can be used to identify patterns, trends, and relationships between variables, and to test hypotheses and theories.

Regression Analysis

Regression analysis is a statistical technique used to analyze the relationship between one dependent variable and one or more independent variables. Researchers use regression analysis to identify and quantify the impact of independent variables on the dependent variable.

Factor Analysis

Factor analysis is a statistical technique used to identify underlying factors that explain the correlations among a set of variables. Researchers use factor analysis to reduce a large number of variables to a smaller set of factors that capture the most important information.

Structural Equation Modeling

Structural equation modeling is a statistical technique used to test complex relationships between variables. It involves specifying a model that includes both observed and unobserved variables, and then using statistical methods to test the fit of the model to the data.

Time Series Analysis

Time series analysis is a statistical technique used to analyze data that is collected over time. It involves identifying patterns and trends in the data, as well as any seasonal or cyclical variations.

Multilevel Modeling

Multilevel modeling is a statistical technique used to analyze data that is nested within multiple levels. For example, researchers might use multilevel modeling to analyze data that is collected from individuals who are nested within groups, such as students nested within schools.

Applications of Quantitative Research

Quantitative research has many applications across a wide range of fields. Here are some common examples:

• Market Research : Quantitative research is used extensively in market research to understand consumer behavior, preferences, and trends. Researchers use surveys, experiments, and other quantitative methods to collect data that can inform marketing strategies, product development, and pricing decisions.
• Health Research: Quantitative research is used in health research to study the effectiveness of medical treatments, identify risk factors for diseases, and track health outcomes over time. Researchers use statistical methods to analyze data from clinical trials, surveys, and other sources to inform medical practice and policy.
• Social Science Research: Quantitative research is used in social science research to study human behavior, attitudes, and social structures. Researchers use surveys, experiments, and other quantitative methods to collect data that can inform social policies, educational programs, and community interventions.
• Education Research: Quantitative research is used in education research to study the effectiveness of teaching methods, assess student learning outcomes, and identify factors that influence student success. Researchers use experimental and quasi-experimental designs, as well as surveys and other quantitative methods, to collect and analyze data.
• Environmental Research: Quantitative research is used in environmental research to study the impact of human activities on the environment, assess the effectiveness of conservation strategies, and identify ways to reduce environmental risks. Researchers use statistical methods to analyze data from field studies, experiments, and other sources.

Characteristics of Quantitative Research

Here are some key characteristics of quantitative research:

• Numerical data : Quantitative research involves collecting numerical data through standardized methods such as surveys, experiments, and observational studies. This data is analyzed using statistical methods to identify patterns and relationships.
• Large sample size: Quantitative research often involves collecting data from a large sample of individuals or groups in order to increase the reliability and generalizability of the findings.
• Objective approach: Quantitative research aims to be objective and impartial in its approach, focusing on the collection and analysis of data rather than personal beliefs, opinions, or experiences.
• Control over variables: Quantitative research often involves manipulating variables to test hypotheses and establish cause-and-effect relationships. Researchers aim to control for extraneous variables that may impact the results.
• Replicable : Quantitative research aims to be replicable, meaning that other researchers should be able to conduct similar studies and obtain similar results using the same methods.
• Statistical analysis: Quantitative research involves using statistical tools and techniques to analyze the numerical data collected during the research process. Statistical analysis allows researchers to identify patterns, trends, and relationships between variables, and to test hypotheses and theories.
• Generalizability: Quantitative research aims to produce findings that can be generalized to larger populations beyond the specific sample studied. This is achieved through the use of random sampling methods and statistical inference.

Examples of Quantitative Research

Here are some examples of quantitative research in different fields:

• Market Research: A company conducts a survey of 1000 consumers to determine their brand awareness and preferences. The data is analyzed using statistical methods to identify trends and patterns that can inform marketing strategies.
• Health Research : A researcher conducts a randomized controlled trial to test the effectiveness of a new drug for treating a particular medical condition. The study involves collecting data from a large sample of patients and analyzing the results using statistical methods.
• Social Science Research : A sociologist conducts a survey of 500 people to study attitudes toward immigration in a particular country. The data is analyzed using statistical methods to identify factors that influence these attitudes.
• Education Research: A researcher conducts an experiment to compare the effectiveness of two different teaching methods for improving student learning outcomes. The study involves randomly assigning students to different groups and collecting data on their performance on standardized tests.
• Environmental Research : A team of researchers conduct a study to investigate the impact of climate change on the distribution and abundance of a particular species of plant or animal. The study involves collecting data on environmental factors and population sizes over time and analyzing the results using statistical methods.
• Psychology : A researcher conducts a survey of 500 college students to investigate the relationship between social media use and mental health. The data is analyzed using statistical methods to identify correlations and potential causal relationships.
• Political Science: A team of researchers conducts a study to investigate voter behavior during an election. They use survey methods to collect data on voting patterns, demographics, and political attitudes, and analyze the results using statistical methods.

How to Conduct Quantitative Research

Here is a general overview of how to conduct quantitative research:

• Develop a research question: The first step in conducting quantitative research is to develop a clear and specific research question. This question should be based on a gap in existing knowledge, and should be answerable using quantitative methods.
• Develop a research design: Once you have a research question, you will need to develop a research design. This involves deciding on the appropriate methods to collect data, such as surveys, experiments, or observational studies. You will also need to determine the appropriate sample size, data collection instruments, and data analysis techniques.
• Collect data: The next step is to collect data. This may involve administering surveys or questionnaires, conducting experiments, or gathering data from existing sources. It is important to use standardized methods to ensure that the data is reliable and valid.
• Analyze data : Once the data has been collected, it is time to analyze it. This involves using statistical methods to identify patterns, trends, and relationships between variables. Common statistical techniques include correlation analysis, regression analysis, and hypothesis testing.
• Interpret results: After analyzing the data, you will need to interpret the results. This involves identifying the key findings, determining their significance, and drawing conclusions based on the data.
• Communicate findings: Finally, you will need to communicate your findings. This may involve writing a research report, presenting at a conference, or publishing in a peer-reviewed journal. It is important to clearly communicate the research question, methods, results, and conclusions to ensure that others can understand and replicate your research.

When to use Quantitative Research

Here are some situations when quantitative research can be appropriate:

• To test a hypothesis: Quantitative research is often used to test a hypothesis or a theory. It involves collecting numerical data and using statistical analysis to determine if the data supports or refutes the hypothesis.
• To generalize findings: If you want to generalize the findings of your study to a larger population, quantitative research can be useful. This is because it allows you to collect numerical data from a representative sample of the population and use statistical analysis to make inferences about the population as a whole.
• To measure relationships between variables: If you want to measure the relationship between two or more variables, such as the relationship between age and income, or between education level and job satisfaction, quantitative research can be useful. It allows you to collect numerical data on both variables and use statistical analysis to determine the strength and direction of the relationship.
• To identify patterns or trends: Quantitative research can be useful for identifying patterns or trends in data. For example, you can use quantitative research to identify trends in consumer behavior or to identify patterns in stock market data.
• To quantify attitudes or opinions : If you want to measure attitudes or opinions on a particular topic, quantitative research can be useful. It allows you to collect numerical data using surveys or questionnaires and analyze the data using statistical methods to determine the prevalence of certain attitudes or opinions.

Purpose of Quantitative Research

The purpose of quantitative research is to systematically investigate and measure the relationships between variables or phenomena using numerical data and statistical analysis. The main objectives of quantitative research include:

• Description : To provide a detailed and accurate description of a particular phenomenon or population.
• Explanation : To explain the reasons for the occurrence of a particular phenomenon, such as identifying the factors that influence a behavior or attitude.
• Prediction : To predict future trends or behaviors based on past patterns and relationships between variables.
• Control : To identify the best strategies for controlling or influencing a particular outcome or behavior.

Quantitative research is used in many different fields, including social sciences, business, engineering, and health sciences. It can be used to investigate a wide range of phenomena, from human behavior and attitudes to physical and biological processes. The purpose of quantitative research is to provide reliable and valid data that can be used to inform decision-making and improve understanding of the world around us.

Advantages of Quantitative Research

There are several advantages of quantitative research, including:

• Objectivity : Quantitative research is based on objective data and statistical analysis, which reduces the potential for bias or subjectivity in the research process.
• Reproducibility : Because quantitative research involves standardized methods and measurements, it is more likely to be reproducible and reliable.
• Generalizability : Quantitative research allows for generalizations to be made about a population based on a representative sample, which can inform decision-making and policy development.
• Precision : Quantitative research allows for precise measurement and analysis of data, which can provide a more accurate understanding of phenomena and relationships between variables.
• Efficiency : Quantitative research can be conducted relatively quickly and efficiently, especially when compared to qualitative research, which may involve lengthy data collection and analysis.
• Large sample sizes : Quantitative research can accommodate large sample sizes, which can increase the representativeness and generalizability of the results.

Limitations of Quantitative Research

There are several limitations of quantitative research, including:

• Limited understanding of context: Quantitative research typically focuses on numerical data and statistical analysis, which may not provide a comprehensive understanding of the context or underlying factors that influence a phenomenon.
• Simplification of complex phenomena: Quantitative research often involves simplifying complex phenomena into measurable variables, which may not capture the full complexity of the phenomenon being studied.
• Potential for researcher bias: Although quantitative research aims to be objective, there is still the potential for researcher bias in areas such as sampling, data collection, and data analysis.
• Limited ability to explore new ideas: Quantitative research is often based on pre-determined research questions and hypotheses, which may limit the ability to explore new ideas or unexpected findings.
• Limited ability to capture subjective experiences : Quantitative research is typically focused on objective data and may not capture the subjective experiences of individuals or groups being studied.
• Ethical concerns : Quantitative research may raise ethical concerns, such as invasion of privacy or the potential for harm to participants.

About the author

Muhammad Hassan

Researcher, Academic Writer, Web developer

You may also like

Textual Analysis – Types, Examples and Guide

Ethnographic Research -Types, Methods and Guide

Triangulation in Research – Types, Methods and...

Experimental Design – Types, Methods, Guide

Exploratory Research – Types, Methods and...

Observational Research – Methods and Guide

Information

• Author Services

Initiatives

You are accessing a machine-readable page. In order to be human-readable, please install an RSS reader.

All articles published by MDPI are made immediately available worldwide under an open access license. No special permission is required to reuse all or part of the article published by MDPI, including figures and tables. For articles published under an open access Creative Common CC BY license, any part of the article may be reused without permission provided that the original article is clearly cited. For more information, please refer to https://www.mdpi.com/openaccess .

Feature papers represent the most advanced research with significant potential for high impact in the field. A Feature Paper should be a substantial original Article that involves several techniques or approaches, provides an outlook for future research directions and describes possible research applications.

Feature papers are submitted upon individual invitation or recommendation by the scientific editors and must receive positive feedback from the reviewers.

Editor’s Choice articles are based on recommendations by the scientific editors of MDPI journals from around the world. Editors select a small number of articles recently published in the journal that they believe will be particularly interesting to readers, or important in the respective research area. The aim is to provide a snapshot of some of the most exciting work published in the various research areas of the journal.

Original Submission Date Received: .

• Active Journals
• Find a Journal
• Proceedings Series
• For Authors
• For Reviewers
• For Editors
• For Librarians
• For Publishers
• For Societies
• For Conference Organizers
• Open Access Policy
• Institutional Open Access Program
• Special Issues Guidelines
• Editorial Process
• Research and Publication Ethics
• Article Processing Charges
• Testimonials
• Preprints.org
• SciProfiles
• Encyclopedia

Article Menu

• Subscribe SciFeed
• Recommended Articles
• Google Scholar
• on Google Scholar
• Table of Contents

Find support for a specific problem in the support section of our website.

Please let us know what you think of our products and services.

Visit our dedicated information section to learn more about MDPI.

JSmol Viewer

Refined land use classification for urban core area from remote sensing imagery by the efficientnetv2 model.

1. Introduction

2.1. efficientnetv2 model, 2.2. transfer learning, 2.3. model validation, 3.1. data set construction, 3.2. experimental environment and parameter configuration, 4. results and discussion, 4.1. convolutional network basic model comparison, 4.2. transfer learning process, 4.3. accuracy and loss analysis, 4.4. model confusion matrix analysis, 4.5. inference results, 5. conclusions, author contributions, institutional review board statement, informed consent statement, data availability statement, conflicts of interest.

• Alcock, I.; White, M.; Cherrie, M.; Wheeler, B.; Taylor, J.; McInnes, R.; Im Kampe, E.O.; Vardoulakis, S.; Sarran, C.; Soyiri, I. Land cover and air pollution are associated with asthma hospitalisations: A cross-sectional study. Environ. Int. 2017 , 109 , 29–41. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Chan, I.Y.; Liu, A.M. Effects of neighborhood building density, height, greenspace, and cleanliness on indoor environment and health of building occupants. Build. Environ. 2018 , 145 , 213–222. [ Google Scholar ] [ CrossRef ]
• Hassan, A.M.; Lee, H. Toward the sustainable development of urban areas: An overview of global trends in trials and policies. Land Use Policy 2015 , 48 , 199–212. [ Google Scholar ] [ CrossRef ]
• Xia, C.; Yeh, A.G.-O.; Zhang, A. Analyzing spatial relationships between urban land use intensity and urban vitality at street block level: A case study of five Chinese megacities. Landsc. Urban Plan. 2020 , 193 , 103669. [ Google Scholar ] [ CrossRef ]
• Gong, P.; Li, X.; Zhang, W. 40-Year (1978–2017) human settlement changes in China reflected by impervious surfaces from satellite remote sensing. Sci. Bull. 2019 , 64 , 756–763. [ Google Scholar ] [ CrossRef ]
• Li, W.; Dong, R.; Fu, H.; Wang, J.; Yu, L.; Gong, P. Integrating Google Earth imagery with Landsat data to improve 30-m resolution land cover mapping. Remote Sens. Environ. 2020 , 237 , 111563. [ Google Scholar ] [ CrossRef ]
• Gao, J.; O’Neill, B.C. Mapping global urban land for the 21st century with data-driven simulations and Shared Socioeconomic Pathways. Nat. Commun. 2020 , 11 , 2302. [ Google Scholar ] [ CrossRef ]
• He, J.; Li, X.; Liu, P.; Wu, X.; Zhang, J.; Zhang, D.; Liu, X.; Yao, Y. Accurate estimation of the proportion of mixed land use at the street-block level by integrating high spatial resolution images and geospatial big data. IEEE Trans. Geosci. Remote Sens. 2020 , 59 , 6357–6370. [ Google Scholar ] [ CrossRef ]
• Li, W.; Wu, C. Incorporating land use land cover probability information into endmember class selections for temporal mixture analysis. ISPRS J. Photogramm. Remote Sens. 2015 , 101 , 163–173. [ Google Scholar ] [ CrossRef ]
• Li, W.; Wu, C.; Zang, S. Modeling urban land use conversion of Daqing City, China: A comparative analysis of “top-down” and “bottom-up” approaches. Stoch. Environ. Res. Risk Assess. 2014 , 28 , 817–828. [ Google Scholar ] [ CrossRef ]
• Zhang, W.; Tang, P.; Zhao, L. Remote sensing image scene classification using CNN-CapsNet. Remote Sens. 2019 , 11 , 494. [ Google Scholar ] [ CrossRef ]
• Yao, Y.; Yan, X.; Luo, P.; Liang, Y.; Ren, S.; Hu, Y.; Han, J.; Guan, Q. Classifying land-use patterns by integrating time-series electricity data and high-spatial resolution remote sensing imagery. Int. J. Appl. Earth Obs. Geoinf. 2022 , 106 , 102664. [ Google Scholar ] [ CrossRef ]
• Zhong, Y.; Zhu, Q.; Zhang, L. Scene classification based on the multifeature fusion probabilistic topic model for high spatial resolution remote sensing imagery. IEEE Trans. Geosci. Remote Sens. 2015 , 53 , 6207–6222. [ Google Scholar ] [ CrossRef ]
• Gomroki, M.; Hasanlou, M.; Reinartz, P. STCD-EffV2T unet: Semi transfer learning EfficientNetV2 T-unet network for urban/land cover change detection using sentinel-2 satellite images. Remote Sens. 2023 , 15 , 1232. [ Google Scholar ] [ CrossRef ]
• Afrin, S.; Gupta, A.; Farjad, B.; Ahmed, M.R.; Achari, G.; Hassan, Q.K. Development of land-use/land-cover maps using landsat-8 and MODIS data, and their integration for Hydro-Ecological applications. Sensors 2019 , 19 , 4891. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Dastour, H.; Hassan, Q.K. A comparison of deep transfer learning methods for land use and land cover classification. Sustainability 2023 , 15 , 7854. [ Google Scholar ] [ CrossRef ]
• Kavran, D.; Mongus, D.; Žalik, B.; Lukač, N. Graph neural network-based method of spatiotemporal land cover mapping using satellite imagery. Sensors 2023 , 23 , 6648. [ Google Scholar ] [ CrossRef ]
• Li, W. Mapping urban land use by combining multi-source social sensing data and remote sensing images. Earth Sci. Inform. 2021 , 14 , 1537–1545. [ Google Scholar ] [ CrossRef ]
• Yin, J.; Dong, J.; Hamm, N.A.; Li, Z.; Wang, J.; Xing, H.; Fu, P. Integrating remote sensing and geospatial big data for urban land use mapping: A review. Int. J. Appl. Earth Obs. Geoinf. 2021 , 103 , 102514. [ Google Scholar ] [ CrossRef ]
• Gong, X.; Su, H.; Xu, D.; Zhang, Z.; Shen, F.; Yang, H. An overview of contour detection approaches. Int. J. Autom. Comput. 2018 , 15 , 656–672. [ Google Scholar ] [ CrossRef ]
• He, K.; Zhang, X.; Ren, S.; Sun, J. Deep residual learning for image recognition. In Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, NV, USA, 27–30 June 2016; pp. 770–778. [ Google Scholar ]
• Qi, K.; Wu, H.; Shen, C.; Gong, J. Land-use scene classification in high-resolution remote sensing images using improved correlatons. IEEE Geosci. Remote Sens. Lett. 2015 , 12 , 2403–2407. [ Google Scholar ] [ CrossRef ]
• Zhang, Q.; Zhu, S. Visual interpretability for deep learning: A survey. Front. Inf. Technol. Electron. Eng. 2018 , 19 , 27–39. [ Google Scholar ] [ CrossRef ]
• He, F.; Tao, D. Recent advances in deep learning theory. arXiv 2020 , arXiv:2012.10931. [ Google Scholar ] [ CrossRef ]
• Sutskever, I.; Vinyals, O.; Le, Q.V. Sequence to Sequence Learning with Neural Networks. arXiv 2014 , arXiv:1409.3215. [ Google Scholar ]
• Zhang, R.; Li, W.; Mo, T. Review of deep learning. arXiv 2018 , arXiv:1804.01653. [ Google Scholar ] [ CrossRef ]
• Wang, M.; Tan, K.; Jia, X.; Wang, X.; Chen, Y. A deep siamese network with hybrid convolutional feature extraction module for change detection based on multi-sensor remote sensing images. Remote Sens. 2020 , 12 , 205. [ Google Scholar ] [ CrossRef ]
• Wen, D.; Huang, X.; Zhang, L.; Benediktsson, J.A. A novel automatic change detection method for urban high-resolution remotely sensed imagery based on multiindex scene representation. IEEE Trans. Geosci. Remote Sens. 2015 , 54 , 609–625. [ Google Scholar ] [ CrossRef ]
• Zhang, Y.; Govindaraj, V.V.; Tang, C.; Zhu, W.; Sun, J. High performance multiple sclerosis classification by data augmentation and AlexNet transfer learning model. J. Med. Imaging Health Inform. 2019 , 9 , 2012–2021. [ Google Scholar ] [ CrossRef ]
• Simonyan, K.; Zisserman, A. Very deep convolutional networks for large-scale image recognition. arXiv 2014 , arXiv:1409.1556. [ Google Scholar ] [ CrossRef ]
• Tan, M.; Le, Q.V. EfficientNet: Rethinking model scaling for convolutional neural networks. In Proceedings of the 36th International Conference on Machine Learning, Long Beach, CA, USA, 9–15 June 2019. [ Google Scholar ]
• Tan, M.; Le, Q.V. EfficientNetV2: Smaller models and faster training. In Proceedings of the 38th International Conference on Machine Learning, Online, 18–24 July 2021. [ Google Scholar ]
• Li, A.; Qi, J.; Lu, H. Multi-attention guided feature fusion network for salient object detection. Neurocomputing 2020 , 411 , 416–427. [ Google Scholar ] [ CrossRef ]
• Raza, A.; Huo, H.; Fang, T. EUNet-CD: Efficient UNet++ for change detection of very high-resolution remote sensing images. IEEE Geosci. Remote Sens. Lett. 2022 , 19 , 1–5. [ Google Scholar ] [ CrossRef ]
• Yang, L.; Chen, Y.; Song, S.; Li, F.; Huang, G. Deep siamese networks based change detection with remote sensing images. Remote Sens. 2021 , 13 , 3394. [ Google Scholar ] [ CrossRef ]
• Zhang, M.; Shi, W. A feature difference convolutional neural network-based change detection method. IEEE Trans. Geosci. Remote Sens. 2020 , 58 , 7232–7246. [ Google Scholar ] [ CrossRef ]
• Krizhevsky, A.; Sutskever, I.; Hinton, G.E. ImageNet classification with deep convolutional neural networks. Commun. ACM 2017 , 60 , 84–90. [ Google Scholar ] [ CrossRef ]
• Bello, I.; Fedus, W.; Du, X.; Cubuk, E.D.; Srinivas, A.; Lin, T.-Y.; Shlens, J.; Zoph, B. Revisiting resnets: Improved training and scaling strategies. Adv. Neural Inf. Process. Syst. 2021 , 34 , 22614–22627. [ Google Scholar ]
• Cao, G.; Wang, B.; Xavier, H.-C.; Yang, D.; Southworth, J. A new difference image creation method based on deep neural networks for change detection in remote-sensing images. Int. J. Remote Sens. 2017 , 38 , 7161–7175. [ Google Scholar ] [ CrossRef ]
• Liu, S.; Shi, Q. Local climate zone mapping as remote sensing scene classification using deep learning: A case study of metropolitan China. ISPRS J. Photogramm. Remote Sens. 2020 , 164 , 229–242. [ Google Scholar ] [ CrossRef ]
• Yuan, Q.; Shen, H.; Li, T.; Li, Z.; Li, S.; Jiang, Y.; Xu, H.; Tan, W.; Yang, Q.; Wang, J. Deep learning in environmental remote sensing: Achievements and challenges. Remote Sens. Environ. 2020 , 241 , 111716. [ Google Scholar ] [ CrossRef ]
• Zhang, X.; Du, S. A Linear Dirichlet Mixture Model for decomposing scenes: Application to analyzing urban functional zonings. Remote Sens. Environ. 2015 , 169 , 37–49. [ Google Scholar ] [ CrossRef ]
• Yu, J.; Zeng, P.; Yu, Y.; Yu, H.; Huang, L.; Zhou, D. A combined convolutional neural network for urban land-use classification with GIS data. Remote Sens. 2022 , 14 , 1128. [ Google Scholar ] [ CrossRef ]
• Seydi, S.T.; Hasanlou, M.; Amani, M. A new end-to-end multi-dimensional CNN framework for land cover/land use change detection in multi-source remote sensing datasets. Remote Sens. 2020 , 12 , 2010. [ Google Scholar ] [ CrossRef ]
• Gong, P.; Chen, B.; Li, X.; Liu, H.; Wang, J.; Bai, Y.; Chen, J.; Chen, X.; Fang, L.; Feng, S.; et al. Mapping essential urban land use categories in China (EULUC-China): Preliminary results for 2018. Sci. Bull. 2020 , 65 , 182–187. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Yao, Y.; Liu, X.; Li, X.; Zhang, J.; Liang, Z.; Mai, K.; Zhang, Y. Mapping fine-scale population distributions at the building level by integrating multisource geospatial big data. Int. J. Geogr. Inf. Sci. 2017 , 31 , 1220–1244. [ Google Scholar ] [ CrossRef ]
• Zhu, Q.; Zhong, Y.; Zhao, B.; Xia, G.-S.; Zhang, L. Bag-of-visual-words scene classifier with local and global features for high spatial resolution remote sensing imagery. IEEE Geosci. Remote Sens. Lett. 2016 , 13 , 747–751. [ Google Scholar ] [ CrossRef ]
• Dai, D.; Yang, W. Satellite image classification via two-layer sparse coding with biased image representation. IEEE Geosci. Remote Sens. Lett. 2010 , 8 , 173–176. [ Google Scholar ] [ CrossRef ]
• Yang, Y.; Newsam, S. Bag-of-visual-words and spatial extensions for land-use classification. In Proceedings of the 18th SIGSPATIAL International Conference on Advances in Geographic Information Systems, San Jose, CA, USA, 2–5 November 2010; pp. 270–279. [ Google Scholar ]
• Helber, P.; Bischke, B.; Dengel, A.; Borth, D. Eurosat: A novel dataset and deep learning benchmark for land use and land cover classification. IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens. 2019 , 12 , 2217–2226. [ Google Scholar ] [ CrossRef ]
• Li, J.; Lin, D.; Wang, Y.; Xu, G.; Zhang, Y.; Ding, C.; Zhou, Y. Deep discriminative representation learning with attention map for scene classification. Remote Sens. 2020 , 12 , 1366. [ Google Scholar ] [ CrossRef ]

Click here to enlarge figure

Share and Cite

Wang, Z.; Liang, Y.; He, Y.; Cui, Y.; Zhang, X. Refined Land Use Classification for Urban Core Area from Remote Sensing Imagery by the EfficientNetV2 Model. Appl. Sci. 2024 , 14 , 7235. https://doi.org/10.3390/app14167235

Wang Z, Liang Y, He Y, Cui Y, Zhang X. Refined Land Use Classification for Urban Core Area from Remote Sensing Imagery by the EfficientNetV2 Model. Applied Sciences . 2024; 14(16):7235. https://doi.org/10.3390/app14167235

Wang, Zhenbao, Yuqi Liang, Yanfang He, Yidan Cui, and Xiaoxian Zhang. 2024. "Refined Land Use Classification for Urban Core Area from Remote Sensing Imagery by the EfficientNetV2 Model" Applied Sciences 14, no. 16: 7235. https://doi.org/10.3390/app14167235

Article Metrics

Article access statistics, further information, mdpi initiatives, follow mdpi.

Subscribe to receive issue release notifications and newsletters from MDPI journals

ORIGINAL RESEARCH article

The characteristics and mechanisms of carbon finance development on green economic efficiency: an empirical analysis based on endogenous economic growth model.

• University of Southampton, Southampton, United Kingdom

The final, formatted version of the article will be published soon.

Select one of your emails

You have multiple emails registered with Frontiers:

Notify me on publication

Please enter your email address:

If you already have an account, please login

You don't have a Frontiers account ? You can register here

In this paper, I explore the impact mechanisms and characteristics of carbon finance on economic sustainability. The results indicate that green total factor productivity has a spatial spillover effect and there are two mediators between carbon finance and green total factor productivity, including technological progress and the increase in technology market turnover. My findings can provide policy implications for sustainable economic development, environmental protection, and resource conservation. Different from previous academic research to study the relationship between carbon finance and green total factor productivity from the theoretical aspect, this paper studies the relationship between carbon finance and green total factor productivity from the empirical aspect and applies spatial econometric models to analyze spatial heterogeneity by using the Romer Endogenous Economic Growth Model. In addition, early scholars were more interested in green finance rather than carbon finance, and just viewed carbon finance as an indicator to construct a green finance index This paper provides suggestions for environmental protections, carbon finance development, and the development of green total factor productivity through data analysis.

Keywords: Green total factor productivity, spatial Durbin model, Sustainable economic growth, mediating effects, spatial spillover effects. (Min

Received: 21 Apr 2024; Accepted: 13 Aug 2024.

Copyright: © 2024 Chen. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY) . The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

* Correspondence: Yiru Chen, University of Southampton, Southampton, United Kingdom

Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.

Please enable Javascript in your web browser in order to use the features on this website.

QSS - Environmental Studies Track

Total credits towards qss: 50+ required .

• Current Checklist for Environmental Studies Track  (Read on before downloading)
• Checklist if major was declared prior to fall 2022
• MATH 111, MATH 221: Prerequisites to the QSS major; Do not count towards major GPA but counts toward 50 credit requirements

QTM Electives

• QTM electives include 300- & 400-level lectures and seminar style classes of 3+ credit hours
• QTM 398R, 496R, 497R, and 499R do not satisfy the QTM elective requirements

Track Electives

• Ecology & Conversion
• Earth and Atmospheric Sciences
• Social Science & Policy; additional 300+ level courses required
• Consult track department’s website for courses designation/acceptability

• DOI: 10.69997/sct.176005
• Corpus ID: 271836661

An Update on Project PARETO - New Capabilities in DOE's Produced Water Optimization Framework

• Miguel A. Zamarripa , Elmira Shamlou , +4 authors Markus Drouven
• Published in Systems and Control… 9 July 2024
• Environmental Science, Engineering
• Systems and Control Transactions

4 References

Pareto: an open-source produced water optimization framework, multiobjective optimization model for minimizing cost and environmental impact in shale gas water and wastewater management, the idaes process modeling framework and model library—flexibility for process simulation and optimization, quantitative assessment of induced seismicity from hydrocarbon production and produced water disposal in azle area, north texas, related papers.

Showing 1 through 3 of 0 Related Papers

Academia.edu no longer supports Internet Explorer.

To browse Academia.edu and the wider internet faster and more securely, please take a few seconds to  upgrade your browser .

•  We're Hiring!
•  Help Center

Environmental Science

• Most Cited Papers
• Most Downloaded Papers
• Newest Papers
• Last »
• Environmental Sustainability Follow Following
• Environmental Studies Follow Following
• Climate Change Follow Following
• Earth Sciences Follow Following
• Ecology Follow Following
• Environmental Sciences Follow Following
• Environmental Management Follow Following
• Environmental Education Follow Following
• Water quality Follow Following
• Water Pollution Follow Following

Enter the email address you signed up with and we'll email you a reset link.

• Academia.edu Journals
•   We're Hiring!
•   Help Center
• Find new research papers in:
• Health Sciences
• Earth Sciences
• Cognitive Science
• Mathematics
• Computer Science
• Academia ©2024