Ecosystem performance monitoring of rangelands by integrating modeling and remote sensing

Monitoring rangeland ecosystem dynamics, production, and performance is valuable for researchers and land managers. However, ecosystem monitoring studies can be difficult to interpret and apply appropriately if management decisions and disturbances are inseparable from the ecosystem's climate signal. This study separates seasonal weather influences from influences caused by disturbances and management decisions, making interannual time-series analysis more consistent and interpretable. We compared the actual ecosystem performance (AEP) of five rangeland vegetation types in the Owyhee Uplands for 9 yr to their expected ecosystem performance (EEP). Integrated growing season Normalized Difference Vegetation Index data for each of the nine growing seasons served as a proxy for annual AEP. Regression-tree models used long-term site potential, seasonal weather, and land cover data sets to generate annual EEP, an estimate of ecosystem performance incorporating annual weather variations. The difference between AEP and EEP provided a performance measure for each pixel in the study area. Ecosystem performance anomalies occurred when the ecosystem performed significantly better or worse than the model predicted. About 14% of the Owyhee Uplands showed a trend of significant underperformance or overperformance (P<0.10). Land managers can use results from weather-based rangeland ecosystem performance models to help support adaptive management strategies.