1. Ecological data are collected over vast geographic areas using digital sensors such as camera traps and bioacoustic recorders. Camera traps have become the standard method for surveying many terrestrial mammals and birds, but camera trap arrays often generate millions of images that are time-consuming to label. This causes significant latency between data collection and subsequent inference, which impedes conservation at a time of ecological crisis. Machine learning algorithms have been developed to improve the speed of labelling camera trap data, but it is uncertain how the outputs of these models can be used in ecological analyses without secondary validation by a human.2. Here, we present our approach to developing, testing and applying a machine learning model to camera trap data for the purpose of achieving fully automated ecological analyses. As a case-study, we built a model to classify 26 Central African forest mammal and bird species (or groups). The model generalizes to new spatially and temporally independent data (n = 227 camera stations, n = 23,868 images), and outperforms humans in several respects (e.g. detecting 'invisible' animals). We demonstrate how ecologists can evaluate a machine learning model's precision and accuracy in an ecological context by comparing species richness, activity patternsThis is an open access article under the terms of the Creative Commons Attribution-NonCommercial License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes.

No abstract

Much like humans, chimpanzees occupy diverse habitats and exhibit extensive behavioural variability. However, chimpanzees are recognized as a discontinuous species, with four subspecies separated by historical geographic barriers. Nevertheless, their range-wide degree of genetic connectivity remains poorly resolved, mainly due to sampling limitations. By analyzing a geographically comprehensive sample set amplified at microsatellite markers that inform recent population history, we found that isolation by distance explains most of the range-wide genetic structure of chimpanzees. Furthermore, we did not identify spatial discontinuities corresponding with the recognized subspecies, suggesting that some of the subspecies-delineating geographic barriers were recently permeable to gene flow. Substantial range-wide genetic connectivity is consistent with the hypothesis that behavioural flexibility is a salient driver of chimpanzee responses to changing environmental conditions. Finally, our observation of strong local differentiation associated with recent anthropogenic pressures portends future loss of critical genetic diversity if habitat fragmentation and population isolation continue unabated.

Paleoclimate reconstructions have enhanced our understanding of how past climates have shaped present‐day biodiversity. We hypothesize that the geographic extent of Pleistocene forest refugia and suitable habitat fluctuated significantly in time during the late Quaternary for chimpanzees (Pan troglodytes). Using bioclimatic variables representing monthly temperature and precipitation estimates, past human population density data, and an extensive database of georeferenced presence points, we built a model of changing habitat suitability for chimpanzees at fine spatio‐temporal scales dating back to the Last Interglacial (120,000 BP). Our models cover a spatial resolution of 0.0467° (approximately 5.19 km2 grid cells) and a temporal resolution of between 1000 and 4000 years. Using our model, we mapped habitat stability over time using three approaches, comparing our modeled stability estimates to existing knowledge of Afrotropical refugia, as well as contemporary patterns of major keystone tropical food resources used by chimpanzees, figs (Moraceae), and palms (Arecacae). Results show habitat stability congruent with known glacial refugia across Africa, suggesting their extents may have been underestimated for chimpanzees, with potentially up to approximately 60,000 km2 of previously unrecognized glacial refugia. The refugia we highlight coincide with higher species richness for figs and palms. Our results provide spatio‐temporally explicit insights into the role of refugia across the chimpanzee range, forming the empirical foundation for developing and testing hypotheses about behavioral, ecological, and genetic diversity with additional data. This methodology can be applied to other species and geographic areas when sufficient data are available.