Will remote sensing shape the next generation of species distribution models?

Abstract

Two prominent limitations of species distribution models (SDMs) are spatial biases in existing occurrence data and a lack of spatially explicit predictor variables to fully capture habitat characteristics of species. Can existing and emerging remote sensing technologies meet these challenges and improve future SDMs? We believe so. Novel products derived from multispectral and hyperspectral sensors, as well as future Light Detection and Ranging (LiDAR) and RADAR missions, may play a key role in improving model performance. In this perspective piece, we demonstrate how modern sensors onboard satellites, planes and unmanned aerial vehicles are revolutionizing the way we can detect and monitor both plant and animal species in terrestrial and aquatic ecosystems as well as allowing the emergence of novel predictor variables appropriate for species distribution modeling. We hope this interdisciplinary perspective will motivate ecologists, remote sensing experts and modelers to work together for developing a more refined SDM framework in the near future.

Spatial and Temporal Dependency of NDVI Satellite Imagery in Predicting Bird Diversity over France

“…As confirmed in this study, continuous (unclassified) predictors derived from satellite imagery contribute to build efficient SDM at large spatial scale [7]. However, because of temporal variability of remotely sensed predictors over the year, acquisition time periods of the data affect the models' performance.…”

“…Remote sensing-based predictors of habitat characteristics may contribute to improve the performances of species distribution models (SDM) [7]. However, the question of the most appropriate data and representation (discrete vs. continuous-based metrics) in SDM still remains [8][9][10][11][12].…”

Priorities to Advance Monitoring of Ecosystem Services Using Earth Observation

“…State-of-the-art and challenges for assessing ecosystem services from space Our current expertise in assessing ecosystem services by means of Earth observation (see Box 1) 130 largely builds on experience gained and methods developed in the context of using satellite data for estimating biodiversity (e.g., [28]) and ecosystem functioning (e.g., [29]). In particular, satellite Earth observation has been used to (1) detect species and assemblages (more recently also functional diversity, [30]), (2) classify the type, extent and variety of habitats [31], and (3) to directly measure ecosystem conditions and functions (e.g., vegetation carbon pools and losses, 135 [32,33].…”

“…For example, multispectral imagery was analyzed to derive forest type and density maps for mapping NTFP provided by trees [49] and for predicting 16 mushroom distributions [50]. The latter study applies a species distribution modeling framework (e.g., [51]) and we expect that this area of research will greatly benefit from new sensors and novel remotely sensed predictor variables (summarized in [28,52]). Direct mapping of NTFP species from satellite data is possible in some cases (e.g., using hyperspectral EO-1 Hyperion data to map specific species of tropical trees, [53]) but needs to be combined with NTFP production and regeneration rates to obtain estimates of NTFP potential.…”

“…It was shown that location characteristics such as LULC diversity and proportions (e.g., % forest cover), special and rare habitat types, terrain, presence and condition 18 of water bodies, as well as abundance of endangered or charismatic species are key for determining recreation potential and supply [65][66][67]. Many of these explanatory variables are routinely mapped from satellite data (e.g., [28,31,63]). In addition, some important recreational infrastructure and facilities (e.g., tourist huts, benches, boat ramps, trail density) can be identified using very-high resolution imagery [68].…”