# Spatially Varying Drivers of Temporal β Diversity in Forest Avian Communities

## TL;DR

This study shows how elevation and forest changes affect bird diversity in Ohio, highlighting the need for spatially aware conservation strategies.

## Contribution

The study reveals spatially varying drivers of bird community changes and integrates landscape and climate interactions for conservation.

## Key findings

- Elevation variability, mean forest patch size, and total forest area significantly affect temporal β diversity dissimilarity.
- The largest changes in bird assemblages occur in areas with high forest loss and low elevation variability.
- Ecological-environmental relationships vary spatially, emphasizing the need for localized conservation planning.

## Abstract

Preserving biodiversity is crucial for ecosystem function, necessitating its measurement and monitoring to comprehend human impacts on the environment. β‐diversity, a key metric, can be used to measure changes in community diversity and assemblages over time. Multiple lines of evidence show that climate, land use land cover (LULC), and topography independently influence diversity; more recent work also shows that LULC can decouple climate effects, creating refugia. Spatial structures also affect the relationships between ecological processes and environmental factors. This study uses two discrete generations of breeding bird atlas data in Ohio, 24 years apart, to investigate the effects of environmental changes and spatial heterogeneity on statewide forest bird community diversity. Utilizing machine learning, we identified elevation variability, mean forest patch size, and total forest area as the most significant variables affecting temporal β diversity dissimilarity (referred to as temporal dissimilarity). The smallest changes in temporal dissimilarity occurred in landscapes with increased mean forest patch size, forest class area, and varied elevation. Furthermore, the results revealed a weak interaction between total forest area and elevation variability, with the largest temporal dissimilarity occurring in sites with the highest loss of total forest area and invariable elevation. We pinpointed regions in Ohio experiencing the greatest amount of composition changes in intensified agricultural and developed lands. A geographically weighted regression model indicated that the relationships between temporal dissimilarity and elevation variability, mean forest patch size, total forest area, and annual maximum temperature varied spatially within the state. Our study integrates ecological patterns with spatial patterns, showing that landscapes can interact with climate to mitigate regional climatic impacts and suggest effective conservation areas. Our findings also indicate that the relationships between ecological processes and the environment are not always spatially consistent.

Understanding how environmental changes and spatial heterogeneity influence biodiversity is crucial for conservation. Using Ohio Breeding Bird Atlas data and machine learning, we identified elevation variability and forest availability as key drivers of β diversity dissimilarity, with the greatest assemblage changes in areas with high total forest loss and low elevation variability. Our findings highlight spatially non‐stationary relationships between ecological processes and environmental factors, emphasizing the importance of integrating landscape and climatic interactions for effective conservation planning.