1. Estimating canopy cover loss with Landsat dense time series: a Mortality Magnitude Index for the Ecosystem Disturbance and Recovery Tracker (eDaRT)
- Author
-
Slaton, Michèle R., Koltunov, Alexander, Evans, Kirk, Kohler, Tanya, and Young-Hart, Laura
- Abstract
Recent drought and insect-induced tree mortality across the western United States has increased the need among land managers and scientists for improved estimation of disturbance timing, type and magnitude, accelerating improvements in high density time series algorithms to map disturbance. We present a novel severity index for tree mortality, called Mortality Magnitude Index (MMI), which expands functionality of a Landsat-processing system, the Ecosystem Disturbance and Recovery Tracker (eDaRT) to estimate tree canopy cover loss as a pixel area proportion. The MMI model utilizes time series of anomaly metrics generated by the eDaRT disturbance detection process, which represent robust normalized statistics of spectral change related to vegetation cover and health. We describe the MMI methodology that integrates the eDaRT metrics over pre- and post-disturbance windows and fits a generalized linear model by beta-regression using training data from the Sierra Nevada, California, including tree mortality that occurred during 2010–2017. The selected model included seven predictor variables (RMSE = 13.0%) and was further validated in southern California for tree mortality events spanning the last two decades. The MMI model exhibited robust temporal and spatial transferability, which have led to subsequent applications of MMI datasets in forest health assessment, wildlife habitat analyses and wildland fire incident support. [ABSTRACT FROM AUTHOR]
- Published
- 2024
- Full Text
- View/download PDF