Species Distribution Modeling for Machine Learning Practitioners: A Review

Abstract

Figure 1. Species distribution models describe the relationship between environmental conditions and (actual or potential) species presence. However, the link between the environment and species distribution data can be complex, particularly since distributional data comes in many different forms. Above are four different sources of distribution data for the Von Der Decken's Hornbill [11]: (from left to right) raw point observations, regional checklists, gridded ecological surveys, and data-driven expert range maps. All images are from Map of Life [101].

Mechanistic forecasts of species responses to climate change: The promise of biophysical ecology

“…While these methods are still closer to the statistical end of the spectrum (i.e., the data lead the dance), using mechanistically derived layers that translate time series of environmental conditions into metrics of fitness relevant to the species should help these models predict more reliably to novel conditions. Alternatively, Bayesian statistics and “domain‐aware” or “model‐informed” machine learning models can be used to inform statistical models with biological information and constraints, which can come from biophysical models or experimental results (Beery et al, 2021; Kotta et al, 2019). Additionally, inverse modeling could be used to infer biophysical model parameters from endpoints such as occurrences (Evans et al, 2016; Fordham et al, 2022).…”

Identifying seasonal distribution patterns of fin whales across the Scotia Sea and the Antarctic Peninsula region using a novel approach combining habitat suitability models and ensemble learning methods

“…The combination of species distribution models and machine learning methods is an emerging field in ecology [e.g. (Effrosynidis et al, 2020;Beery et al, 2021)]. In our study, the combination of SDM and a random forest ensemble learning algorithm enabled us to derive information on abundance from a multi-source heterogeneous data set of mainly opportunistic sighting records and to present an additional set of abundance estimates for some of the most frequently visited research areas in the Southern Ocean.…”

Eco-ISEA3H, a machine learning ready spatial database for ecometric and species distribution modeling

“…While species distribution modeling has traditionally focused on local or regional scales, analyses at increasingly large scales are gaining popularity. Machine learning methods are becoming the leading tools for such analyses [25][26][27][28] .…”

“…Human population density is one of several measures of human presence and activity which together define the human "footprint," associated with profound, adverse effects on natural systems 78 . Given this pervasive impact, data characterizing degree of human influence are used as predictors in some ecological models, including SDMs 28 . The Gridded Population of the World (GPW) 68 density dataset represents the global distribution of human population density, developed using census records, population registers, and the administrative boundaries of approximately 13.5 million national and subnational units.…”