Why choose Random Forest to predict rare species distribution with few samples in large undersampled areas? Three Asian crane species models provide supporting evidence

Abstract

Species distribution models (SDMs) have become an essential tool in ecology, biogeography, evolution and, more recently, in conservation biology. How to generalize species distributions in large undersampled areas, especially with few samples, is a fundamental issue of SDMs. In order to explore this issue, we used the best available presence records for the Hooded Crane (Grus monacha, n = 33), White-naped Crane (Grus vipio, n = 40), and Black-necked Crane (Grus nigricollis, n = 75) in China as three case studies, employing four powerful and commonly used machine learning algorithms to map the breeding distributions of the three species: TreeNet (Stochastic Gradient Boosting, Boosted Regression Tree Model), Random Forest, CART (Classification and Regression Tree) and Maxent (Maximum Entropy Models). In addition, we developed an ensemble forecast by averaging predicted probability of the above four models results. Commonly used model performance metrics (Area under ROC (AUC) and true skill statistic (TSS)) were employed to evaluate model accuracy. The latest satellite tracking data and compiled literature data were used as two independent testing datasets to confront model predictions. We found Random Forest demonstrated the best performance for the most assessment method, provided a better model fit to the testing data, and achieved better species range maps for each crane species in undersampled areas. Random Forest has been generally available for more than 20 years and has been known to perform extremely well in ecological predictions. However, while increasingly on the rise, its potential is still widely underused in conservation, (spatial) ecological applications and for inference. Our results show that it informs ecological and biogeographical theories as well as being suitable for conservation applications, specifically when the study area is undersampled. This method helps to save model-selection time and effort, and allows robust and rapid assessments and decisions for efficient conservation.

Correcting for bias in distribution modelling for rare species using citizen science data

“…Machine-learning methods, such as random forest (RF), are increasingly being applied to address ecological classification problems (Cutler, Edwards, Beard, Cutler, & Hess, 2007;Mi, Huttmann, Guo, Han, & Wen, 2017). RF is an ensemble method where a large number of individual decision tree models are induced by taking bootstrap samples of the data.…”

Machine Learning Model Analysis of Breeding Habitats for the Black-necked Crane in Central Asian Uplands under Anthropogenic Pressures

“…Further, it features great tolerance to noise 24 , a strong immunity to overfitting 24 and high efficiency in processing a large number of predictors and their interactions 14, 25 . Owing to “recursive partitioning” and bagging, SPM algorithm optimization can properly handle interactions, stopping rules, weighting and complexities in predictor combinations 24–27 . Two modifications were applied in the model settings: we used balanced class weights, a powerful and sophisticated weighting function in SPM, to defend against inequivalent prevalence (183 presence versus 18,300 absence points) 27 ; we set the number of trees to 1,000 to find the best possible model 27, 28 .…”

Parapatric subspecies of Macaca assamensis show a marginal overlap in their predicted potential distribution: Some elaborations for modern conservation management

“…To answer the third research question, species distribution models (SDMs) were developed using MaxEnt (Phillips, Anderson, & Schapire, 2006) and Random Forest (Breiman, 2001a;Liaw & Wiener, 2002) algorithms, which are among the most commonly used machine learning methods (Aguirre-Gutiérrez et al, 2013;Mi, Huettmann, Guo, Han, & Wen, 2017). The models were fitted with binary response data (occurrence data with background data) in the R package sdm (Naimi & Araújo, 2016).…”

“…These problems are often tackled by making an ensemble of multiple models (Araújo & New, 2007;Regmi et al, 2018), but we did not do this here directly. Instead, we used one of the best algorithms in SDM (Aguirre-Gutiérrez et al, 2013;Craig & Huettmann, 2009;Mi et al, 2017), and, due to bagging, Random Forest being an ensemble model (Breiman, 2001a).…”