Integrating multi-decadal satellite imagery and species distribution models for conservation planning in oak woodlands
Researchers integrate multi-decadal Landsat and Sentinel-2 imagery with species distribution models to assess blue oak woodland resilience and future habitat suitability at two California preserves. The approach identifies areas that may support restoration, land protection and climate-informed conservation planning under changing conditions.
Subject Tags
- Conservation Planning
- Habitat restoration
- Climate resilience
Abstract
Oak woodlands are among California's least protected yet most ecologically important ecosystems, providing critical habitat for a wide variety of plant and animal species. These foundational plant communities are increasingly threatened by drought, wildfire, and other climate-driven disturbances, which have contributed to significant declines in recruitment and adult tree survival. Understanding the spatial and temporal dynamics of oak woodland resilience and persistence is essential for assessing ecosystem responses under changing conditions. Here, we present a novel approach that integrates vegetation resilience metrics derived from temporal mixture residual analysis of multi-decadal satellite time series with species distribution models (SDMs) forecasting blue oak (Quercus douglasii) habitat suitability under a high-emissions, mid-21st century climate scenario. We apply this approach in two study areas in California managed by The Nature Conservancy, the Randall and Dye Creek Preserves, generating spatially explicit maps that combine historical vegetation patterns with projected climatic suitability. We found that historically resilient areas that are projected to be climatically suitable by the middle of the 21st century were generally located at higher elevations with cooler temperatures and greater winter precipitation, while low-suitability zones were concentrated in warmer, lower elevation areas. To test the applicability of the approach across widely used satellite platforms, we compared results from 30-m Landsat and 10-m Sentinel-2 data. These comparisons showed broadly similar patterns, with Sentinel-2 capturing greater temporal variability. Field validation using 18 blue oak recruitment plots at Randall Preserve confirmed that high-recruitment sites corresponded with areas of high future habitat suitability and generally historically resilient vegetation. This method offers a scalable, transferable method for characterizing historical and projected oak woodland dynamics, supporting land management, restoration, and land protection planning under a changing climate.
Citation
Rose, M. Brooke, Janet Franklin, H. Scott Butterfield, Paige Rohwer, Amelie Beck, Maria C. Viteri, Jeanette K. Howard, Zachary Principe, Andrea Craig, and Daniel Sousa. 2026. “ Integrating Multi-Decadal Satellite Imagery and Species Distribution Models for Conservation Planning in Oak Woodlands.” Ecosphere 17(9): e70752. https://doi.org/10.1002/ecs2.70752
TNC Authors
-
Scott Butterfield
Lead Scientist, Land
The Nature Conservancy
Email: scott_butterfield@tnc.org -
Maria Viteri
Randall Preserve Ecologist
The Nature Conservancy
Email: maria.viteri@tnc.org -
Zachary Principe
Preserve Director, Randall Preserve
The Nature Conservancy
Email: zprincipe@tnc.org -
Andrea Craig
Stewardship Manager
The Nature Conservancy
Email: acraig@tnc.org -
Jeanette Howard
The Nature Conservancy