Common garden test of range limits as predicted by a species distribution model in the annual plant
            Mimulus bicolor

Jacqueline Grubel* and Christopher G. Eckert (Faculty Supporter) It is widely thought that the size, shape and location of a species’ geographical distribution are a spatial expression of its realized niche, and this assumption is central to evolutionary biology, biogeography and conservation. Yet, the hypothesis that geographical range limits are niche limits is not well supported by experimental translocations of species beyond their range limits. Beyond range populations often exhibit fitness high enough for self-replacement. In contrast, environmental niche models based on bioclimatic data often suggest a decline in habitat suitability beyond range limits, thereby supporting niche limitation. However very few studies have evaluated whether species distribution models (SDMs) accurately predict the viability of populations in nature, and scant results to date are not supportive. Long-term transplant with the short-lived, Pacific costal dune endemic plant Camissoniopsis cheiranthifolia (Onagraceae) suggest that populations are viable beyond the northern range limit over multiple generations. We constructed an SDM based on a large range-wide database of species records plus standard bioclimatic variables and substrate type. We also included sea surface temperature, which greatly modifies the climate of dune habitat. Preliminary results suggest that our SDM reliably predicts the fitness of experimental populations. However, both approaches indicate that something other than niche limitation enforces the northern range limit of this species. Results from this well-studied dune plant suggest that range limitation via constraints on dispersal may play an important role in limiting northern range expansion.

Download Full-text

A Robust Prediction Model for Species Distribution Using Bagging Ensembles with Deep Neural Networks

Remote Sensing ◽

10.3390/rs13081495 ◽

2021 ◽

Vol 13 (8) ◽

pp. 1495

Author(s):

Jehyeok Rew ◽

Yongjang Cho ◽

Eenjun Hwang

Keyword(s):

Neural Networks ◽

Species Distribution ◽

Best Practice ◽

Deep Neural Networks ◽

Species Distribution Models ◽

Species Distribution Model ◽

Environmental Data ◽

Distribution Model ◽

Distribution Models ◽

Robust Prediction

Species distribution models have been used for various purposes, such as conserving species, discovering potential habitats, and obtaining evolutionary insights by predicting species occurrence. Many statistical and machine-learning-based approaches have been proposed to construct effective species distribution models, but with limited success due to spatial biases in presences and imbalanced presence-absences. We propose a novel species distribution model to address these problems based on bootstrap aggregating (bagging) ensembles of deep neural networks (DNNs). We first generate bootstraps considering presence-absence data on spatial balance to alleviate the bias problem. Then we construct DNNs using environmental data from presence and absence locations, and finally combine these into an ensemble model using three voting methods to improve prediction accuracy. Extensive experiments verified the proposed model’s effectiveness for species in South Korea using crowdsourced observations that have spatial biases. The proposed model achieved more accurate and robust prediction results than the current best practice models.

Download Full-text