Detecting Montane Meadows in the Tahoe National Forest Using LiDAR and ASTER Imagery

In the Sierra Nevada mountains, meadows provide numerous hydraulic and ecosystem functions such as flood attenuation, groundwater storage, and wildlife habitat. However, many meadows have been degraded from historical land use such as water diversion, grazing, and logging. Land managers have altered management strategies for restoration purposes, but there is a lack of comprehensive data on meadow locations. Previous attempts to inventory Sierra Nevada meadows have included several remote sensing techniques including heads up digitizing and pixel based image analysis, but this has been challenging due to geographic variability, seasonal changes, and meadow health. I present a remote sensing method using multiple return LiDAR (Light Detection and Ranging) and ASTER imagery to detect montane meadows in a subset of the Tahoe National Forest. The project used LiDAR data to create a digital terrain model and digital surface model. From these models, I derived canopy height, surface slope, and watercourse for the entire study area. Literature queries returned known values for canopy height and surface slope characteristic of montane meadows. These values were used to select for possible meadows within the study area. To filter out noise, only contiguous areas greater than one acre that satisfied the queries were used. Finally, 15-meter ASTER imagery was used to de-select for areas such as dirt patches or gravel bars that might have satisfied the previous queries and meadow criteria. When using high resolution aerial imagery to assess model accuracy, preliminary results show user accuracy of greater than 80%. Further validation is still needed to improve the accuracy of modeled meadow delineation. This method allows for meadows to be inventoried without discriminating based on geographic variability, seasonal changes, or meadow health.