Application of the Ecosystem Disturbance and Recovery Tracker in Detection of Forest Health Departure from Desired Conditions in Sierra Nevada National Forests

Three Sierra Nevada national forests have recently released draft revised land management plans to the public for comment. These plans include components guiding land management projects (e.g. timber harvest) over the next decades. Desired conditions for land cover types were identified, and are based on the concept of the natural range of variability (NRV); the agency aims to achieve vegetation structure and composition that is within the range that has occurred in that system during the Holocene, prior to Euro-American settlement. Such systems are identified as having ecological integrity, with the ability to withstand and recover from disturbance. Comparisons between current conditions and the NRV period are often made qualitatively, with quantitative comparisons possible through the reconstruction of forest parameters using paleoecological evidence such as pollen or tree rings. Where such information is lacking, comparisons can be made using more recent data for locations or time periods believed to closely represent NRV. Such evaluations are necessary to develop projects that will be within the agency's desired conditions. As an example, the most recent estimate of the extent of tree mortality in CA continues to rise and expand at an aggressive rate, creating the need to track progress as compared to baseline conditions. The Ecosystem Disturbance and Recovery Tracker (eDaRT) is a system developed by the US Forest Service Region 5 Remote Sensing Lab and UC Davis CSTARS center that uses all available Landsat imagery to detect anomalies in vegetation indices, and has been adapted to detect current tree mortality as compared to modern pre-drought baseline conditions. We present initial results for the southern Sierra Nevada (5,260,913 hectares; 13 million acres) that have been validated using WorldView imagery and field data. The high spatial resolution and large extent of the dataset as compared to other mortality maps enabled identification of factors correlated to disturbance, including topographic position, and position within species' ranges. Further development of eDaRT, including refining the baseline period identification and clustering parameters, may provide a highly informative and efficient method for detecting and monitoring forest health within the NRV framework.