How to Statistically Disentangle the Effects of Environmental Factors and Human Disturbances: A Review
Contemporary biological assemblage composition and biodiversity are often shaped by a range of natural environmental factors, human disturbances, and their interactions. It is critical to disentangle the effects of individual natural variables and human stressors in data analysis to support manageme...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-02-01
|
| Series: | Water |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2073-4441/15/4/734 |
| _version_ | 1797617740405538816 |
|---|---|
| author | Yong Cao Lizhu Wang |
| author_facet | Yong Cao Lizhu Wang |
| author_sort | Yong Cao |
| collection | DOAJ |
| description | Contemporary biological assemblage composition and biodiversity are often shaped by a range of natural environmental factors, human disturbances, and their interactions. It is critical to disentangle the effects of individual natural variables and human stressors in data analysis to support management decision-making. Many statistical approaches have been proposed and used to estimate the biological effects of individual predictors, which often correlated and interacted with one another. In this article, we review nine of those approaches in terms of their strengths, limitations, and related r packages. Among those are hierarchical partitioning, propensity score, the sum of AIC weights, structural equation modeling, and tree-based machine learning algorithms. As no approach is perfect, we offer two suggestions: (1) reducing the number of predictors as low as possible by carefully screening all candidate predictors based on biological and statistical considerations; (2) selecting two or more approaches based on the characteristics of the given dataset and specific research goals of a study, and using them in parallel or sequence. Our review could help ecologists to navigate through this challenging process. |
| first_indexed | 2024-03-11T08:00:01Z |
| format | Article |
| id | doaj.art-b65b44a079a14da0828af2d5214f57c3 |
| institution | Directory Open Access Journal |
| issn | 2073-4441 |
| language | English |
| last_indexed | 2024-03-11T08:00:01Z |
| publishDate | 2023-02-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Water |
| spelling | doaj.art-b65b44a079a14da0828af2d5214f57c32023-11-16T23:52:45ZengMDPI AGWater2073-44412023-02-0115473410.3390/w15040734How to Statistically Disentangle the Effects of Environmental Factors and Human Disturbances: A ReviewYong Cao0Lizhu Wang1Illinois Natural History Survey, Prairie Research Institute, University of Illinois, Champaign, IL 61820, USAInternational Joint Commission, P.O. Box 32869, Detroit, MI 48232, USAContemporary biological assemblage composition and biodiversity are often shaped by a range of natural environmental factors, human disturbances, and their interactions. It is critical to disentangle the effects of individual natural variables and human stressors in data analysis to support management decision-making. Many statistical approaches have been proposed and used to estimate the biological effects of individual predictors, which often correlated and interacted with one another. In this article, we review nine of those approaches in terms of their strengths, limitations, and related r packages. Among those are hierarchical partitioning, propensity score, the sum of AIC weights, structural equation modeling, and tree-based machine learning algorithms. As no approach is perfect, we offer two suggestions: (1) reducing the number of predictors as low as possible by carefully screening all candidate predictors based on biological and statistical considerations; (2) selecting two or more approaches based on the characteristics of the given dataset and specific research goals of a study, and using them in parallel or sequence. Our review could help ecologists to navigate through this challenging process.https://www.mdpi.com/2073-4441/15/4/734variance partitioningvariable-importance rankingcollinearityvariable interactionsbioassessmentland use impact |
| spellingShingle | Yong Cao Lizhu Wang How to Statistically Disentangle the Effects of Environmental Factors and Human Disturbances: A Review Water variance partitioning variable-importance ranking collinearity variable interactions bioassessment land use impact |
| title | How to Statistically Disentangle the Effects of Environmental Factors and Human Disturbances: A Review |
| title_full | How to Statistically Disentangle the Effects of Environmental Factors and Human Disturbances: A Review |
| title_fullStr | How to Statistically Disentangle the Effects of Environmental Factors and Human Disturbances: A Review |
| title_full_unstemmed | How to Statistically Disentangle the Effects of Environmental Factors and Human Disturbances: A Review |
| title_short | How to Statistically Disentangle the Effects of Environmental Factors and Human Disturbances: A Review |
| title_sort | how to statistically disentangle the effects of environmental factors and human disturbances a review |
| topic | variance partitioning variable-importance ranking collinearity variable interactions bioassessment land use impact |
| url | https://www.mdpi.com/2073-4441/15/4/734 |
| work_keys_str_mv | AT yongcao howtostatisticallydisentangletheeffectsofenvironmentalfactorsandhumandisturbancesareview AT lizhuwang howtostatisticallydisentangletheeffectsofenvironmentalfactorsandhumandisturbancesareview |