Collinearity in ecological niche modeling: Confusions and challenges

Ecology and Evolution, Journal Year: 2019, Volume and Issue: 9(18), P. 10365 - 10376

Published: Aug. 20, 2019

Ecological niche models are widely used in ecology and biogeography. Maxent is one of the most frequently modeling tools, many studies have aimed to optimize its performance. However, scholars conflicting views on treatment predictor collinearity modeling. Despite this lack consensus, quantitative examinations effects modeling, especially model transfer scenarios, lacking. To address knowledge gap, here we quantify under different scenarios training projection. We separately examine collinearity, shifts between testing data, environmental novelty demonstrate that excluding highly correlated variables does not significantly influence find shift significant negative performance transfer. thus conclude (a) robust training; (b) strategy has little impact because accounts for redundant variables; (c) can negatively affect transferability. therefore recommend report better infer accuracy when spatially and/or temporally transferred.

Language: Английский

---

Want to model a species niche? A step-by-step guideline on correlative ecological niche modelling

Ecological Modelling, Journal Year: 2021, Volume and Issue: 456, P. 109671 - 109671

Published: July 19, 2021

Language: Английский

---

Common mistakes in ecological niche models

International Journal of Geographical Information Science, Journal Year: 2020, Volume and Issue: 35(2), P. 213 - 226

Published: July 27, 2020

Ecological niche models (ENMs) are widely used statistical methods to estimate various types of species niches. After lecturing several editions introductory courses on ENMs and reviewing numerous manuscripts this subject, we frequently faced some recurrent mistakes: 1) presence-background modelling methods, such as Maxent or ENFA, if they were pseudo-absence methods; 2) spatial autocorrelation is confused with clustering records; 3) environmental variables a higher resolution than 4) correlations between not taken into account; 5) machine-learning replicated; 6) topographical calculated from unprojected coordinate systems, and; 7) downscaled by resampling. Some these mistakes correspond student misunderstandings corrected before publication. However, other errors can be found in published papers. We explain here why approaches erroneous propose ways improve them.

Language: Английский

---

The area under the precision‐recall curve as a performance metric for rare binary events

Methods in Ecology and Evolution, Journal Year: 2018, Volume and Issue: 10(4), P. 565 - 577

Published: Dec. 17, 2018

Abstract Species distribution models are used to study biogeographic patterns and guide decision‐making. The variable quality of these makes it critical assess whether a model's outputs suitable for the intended use, but commonly evaluation approaches inappropriate many ecological contexts. In particular, unrealistically high performance assessments have been associated with rare species predictions over large geographic extents. We evaluated area under precision‐recall curve ( AUC ‐ PR ) as metric binary events, focusing on assessment models. Precision is probability that present given predicted presence, while recall (more called sensitivity) model predicts presence in locations where has observed. simulated at three levels prevalence, compared receiver operating characteristic ROC when extent was increased assessed how well each reflected utility surveys new populations. robust rarity and, unlike , not affected by an increasing extent. major advantages arise because does incorporate correctly absences therefore less prone exaggerate unbalanced datasets. precision were useful indicators guiding surveys. show important evaluating species, its benefits context responses will make applicable other studies. By considering true negative quadrant confusion matrix, ameliorates issues beyond species’ range or number background points absence information unavailable. However, no single captures all aspects nor provides absolute index can be across Our results indicate provide intuitive metrics sampling, complement help delineate appropriate use.

Language: Английский

---

ENMTML: An R package for a straightforward construction of complex ecological niche models

Environmental Modelling & Software, Journal Year: 2020, Volume and Issue: 125, P. 104615 - 104615

Published: Jan. 6, 2020

Language: Английский

---

Assessing the reliability of species distribution projections in climate change research

Diversity and Distributions, Journal Year: 2021, Volume and Issue: 27(6), P. 1035 - 1050

Published: Feb. 19, 2021

Abstract Aim Forecasting changes in species distribution under future scenarios is one of the most prolific areas application for models (SDMs). However, no consensus yet exists on reliability such drawing conclusions species’ response to changing climate. In this study, we provide an overview common modelling practices field and assess model predictions using a virtual approach. Location Global. Methods We first review papers published between 2015 2019. Then, use approach three commonly applied SDM algorithms (GLM, MaxEnt random forest) estimated actual predictive performance parameterized with different settings violations assumptions. Results Most relied single (65%) small samples ( N < 50, 62%), used presence‐only data (85%), binarized models' output (74%) split‐sample validation (94%). Our simulation reveals that tends be over‐optimistic compared real performance, whereas spatial block provides more honest estimate, except when datasets are environmentally biased. The binarization predicted probabilities presence reduces models’ ability considerably. Sample size main predictors accuracy, but has little influence accuracy. Finally, inclusion ecologically irrelevant violation assumptions increases accuracy decreases projections, leading biased estimates range contraction expansion. Main predict low average, particularly binarized. A robust by spatially independent required, does not rule out inflation assumption violation. findings call caution interpretation projections climates.

Language: Английский

---

Species distribution modelling to support forest management. A literature review

Ecological Modelling, Journal Year: 2019, Volume and Issue: 411, P. 108817 - 108817

Published: Sept. 16, 2019

Language: Английский

---

Spatial prediction of landslide susceptibility using hybrid support vector regression (SVR) and the adaptive neuro-fuzzy inference system (ANFIS) with various metaheuristic algorithms

The Science of The Total Environment, Journal Year: 2020, Volume and Issue: 741, P. 139937 - 139937

Published: June 7, 2020

Language: Английский

---

flexsdm: An r package for supporting a comprehensive and flexible species distribution modelling workflow

Methods in Ecology and Evolution, Journal Year: 2022, Volume and Issue: 13(8), P. 1661 - 1669

Published: April 20, 2022

Abstract Species distribution models (SDM) are widely used in diverse research areas because of their simple data requirements and application versatility. However, SDM outcomes sensitive to input methodological choices. Such sensitivity applications mean that flexibility is necessary create SDMs with tailored protocols for a given set model use. We introduce the r package flexsdm supporting flexible species modelling workflows. functions arguments serve as building blocks construct specific protocol user's needs. The main features flexibility, integration other tools, simplicity objects returned function speed. As an illustration, we define complete workflow California red fir Abies magnifica . This provides by incorporating comprehensive tools structured three steps: (a) Pre‐modelling prepare input, example, sampling bias correction, pseudo‐absences background points, partitioning, reducing collinearity predictors. (b) Modelling allow fitting evaluating different approaches, including individual algorithms, tuned models, ensembles small ensemble models. (c) Post‐modelling include related models' predictions, interpolation overprediction correction. Because comprises large part process, from outlier detection users can delineate partial or workflows based on combination meet

Language: Английский

---

Pervasive gaps in Amazonian ecological research

Current Biology, Journal Year: 2023, Volume and Issue: 33(16), P. 3495 - 3504.e4

Published: July 19, 2023

Biodiversity loss is one of the main challenges our time,1,2 and attempts to address it require a clear understanding how ecological communities respond environmental change across time space.3,4 While increasing availability global databases on has advanced knowledge biodiversity sensitivity changes,5,6,7 vast areas tropics remain understudied.8,9,10,11 In American tropics, Amazonia stands out as world's most diverse rainforest primary source Neotropical biodiversity,12 but remains among least known forests in America often underrepresented databases.13,14,15 To worsen this situation, human-induced modifications16,17 may eliminate pieces Amazon's puzzle before we can use them understand are responding. increase generalization applicability knowledge,18,19 thus crucial reduce biases research, particularly regions projected face pronounced changes. We integrate community metadata 7,694 sampling sites for multiple organism groups machine learning model framework map research probability Brazilian Amazonia, while identifying region's vulnerability change. 15%-18% neglected expected experience severe climate or land changes by 2050. This means that unless take immediate action, will not be able establish their current status, much less monitor changing what being lost.

Language: Английский

---