Predicting species distributions with environmental time series data and deep learning
Smith, A. M.; Capinha, C.; Kramer, A. M.
Show abstract
Species distribution models (SDMs) are widely used to gain ecological understanding and guide conservation decisions. These models are developed with a wide variety of algorithms - from statistic-based approaches to machine learning approaches - but a requirement almost all share is the use of predictor variables that strongly simplify the temporal variability of driving factors. Conversely, novel architectures of deep learning neural networks allow dealing with fully explicit spatiotemporal dynamics and thus fitting SDMs without the need to simplify the temporal and spatial dimension of predictor data. We present and demonstrate a deep learning based SDM approach that uses time series of spatial data as predictors using distribution data for 74 species from a well-established benchmark dataset. The deep learning approach provided consistently accurate models, directly using time series of predictor data and thus avoiding the use of pre-processed predictor sets that can obscure relevant aspects of environmental variation.
Matching journals
The top 4 journals account for 50% of the predicted probability mass.
Similar papers in this journal
Similar papers in this journal
- Multi-scale neural networks enhance species distribution modelling across predictors and taxonomic groups 94%
- Transferability of stream benthic macroinvertebrate distribution models to drought-related conditions 93%
- From local seafloor imagery to global patterns in benthic habitat states: contribution of citizen science to habitat classification across latitudes 93%
Similar papers in this journal
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.