Fine-scale remote sensing applications to support vegetation and fire management planning and implementation (At Lava Beds National Monument, LABE)

Strand, Eva; Bunting, Stephen; Farris, Calvin

This Task Agreement details the steps by which Lava Beds National Monument (LABE), and the Rangeland Ecology and Management Program at the University of Idaho will collaborate in development of fine scale remotely sensed digital data layers of vegetation and fuels for LABE. Lava Beds National Monument (LABE) recently initiated a pilot project to develop a comprehensive long‐term vegetation management plan to be integrated with short‐term fire and resource planning. LABE is an ideal setting for such an approach because: (a) it contains a complex mix of intact and degraded sagebrush steppe with differential response to fire, (b) juniper woodlands are expanding into sagebrush and other vegetation types, (c) exotic species, mainly cheatgrass, are present in the park with potential to spread, and (d) the park has experienced a relatively high rate of 20th century burning. Thus, the decision to burn in LABE requires careful consideration of many factors due to these dynamic and altered fuels and expected response to fire. The purpose of this research is to develop state‐of‐the‐art remote sensing techniques and high resolution spatial data layers that will complement and contribute to implementation of an integrated management plan at LABE. We investigated three different remote sensing methods: Classification of 1‐m scale aerial imagery, spectral unmixing of time‐series satellite images, and object extraction using wavelet analysis. Combining these remote sensing methods with field data from the park we produced high thematic resolution layers essential for vegetation and fuels management. Specifically, we developed maps of sagebrush steppe including cover of shrub, annual, and perennial grass; juniper woodland maps differentiating between four phases of woodland development and maps of lava cover. Fuel loading maps were developed via regression of vegetation cover and fuel types with correlation coefficients ranging from 0.4 to 0.8 for different fuel types.