Using satellite data to assess spatial drivers of bird diversity

Abstract

Birds are useful indicators of overall biodiversity, which continues to decline globally, despite targets to reduce its loss. The aim of this paper is to understand the importance of different spatial drivers for modelling bird distributions. Specifically, it assesses the importance of satellite‐derived measures of habitat productivity, heterogeneity and landscape structure for modelling bird diversity across Great Britain. Random forest (RF) regression is used to assess the extent to which a combination of satellite‐derived covariates explain woodland and farmland bird diversity and richness. Feature contribution analysis is then applied to assess the relationships between the response variable and the covariates in the final RF models. We show that much of the variation in farmland and woodland bird distributions is explained (R2 0.64–0.77) using monthly habitat‐specific productivity values and landscape structure (FRAGSTATS) metrics. The analysis highlights important spatial drivers of bird species richness and diversity, including high productivity grassland during spring for farmland birds and woodland patch edge length for woodland birds. The feature contribution provides insight into the form of the relationship between the spatial drivers and bird richness and diversity, including when a particular spatial driver affects bird richness positively or negatively. For example, for woodland bird diversity, the May 80th percentile Normalized Difference Vegetation Index (NDVI) for broadleaved woodland has a strong positive effect on bird richness when NDVI is >0.7 and a strong negative effect below. If relationships such as these are stable over time, they offer a useful analytical tool for understanding and comparing the influence of different spatial drivers.

Exploring the Spatiotemporal Influence of Climate on American Avian Migration with Random Forests

“…Future studies could expand the temporal scale and resolution of the dataset or focus on abundance as a target variable to account for population decreases. They could also include other climate features known to affect migrations such as wind speed and direction (Haest 2019), as well as indices derived from satellite data such as Normalized Difference Vegetation and Water Indices, which correspond with food availability and drought conditions (Alessandrini et al, 2022;Hunt et al, 2022). S1 -Observation count per species across eBird checklists Table S2 -Two-sided permutation test (permutations = 100,000) p-values between model ensemble mean absolute errors at various lags of the climate feature sets Table S3 -Two-sided permutation test (permutations = 100,000) p-values and teststatistics between the random forest feature importance of the North American Ecoregion climate features to the South American Ecoregion features for a) precipitation b) maximum temperature, and c) minimum temperature.…”

“…Future studies could expand the temporal scale and resolution of the dataset or focus on abundance as a target variable to account for population decreases. They could also include other climate features known to affect migrations such as wind speed and direction (Haest 2019), as well as indices derived from satellite data such as Normalized Difference Vegetation and Water Indices, which correspond with food availability and drought conditions (Alessandrini et al ., 2022; Hunt et al ., 2022).…”

“…Furthermore, incorporating other climate features that are known to affect migrations such as wind speed and direction (as highlighted by (16)) could contribute to a more comprehensive model. Finally, the inclusion of indices derived from satellite data such as normalized difference vegetation and water indices, which are associated with food availability and drought conditions (as mentioned by (18,43)), could provide a more holistic understanding of the ecological factors impacting bird migrations.…”

“…Still, many papers (7,8,16) show trends from post-hoc analyses and do not directly use climatic features to predict how migratory patterns may change. Machine learning methods such as random forests can leverage non-linear and complex trends in ecological data (17)(18)(19) to model historical trends and make predictions of future behavior. Random forest models are highperforming, efficient to train compared to neural networks under limited data, have built-in feature importance metrics, and have been applied in ecological settings to relate climate and ecological features to biodiversity (18,20) and animal movement and migration (19).…”

“…Machine learning methods such as random forests can leverage non-linear and complex trends in ecological data (17)(18)(19) to model historical trends and make predictions of future behavior. Random forest models are highperforming, efficient to train compared to neural networks under limited data, have built-in feature importance metrics, and have been applied in ecological settings to relate climate and ecological features to biodiversity (18,20) and animal movement and migration (19).…”