Between 2012 and 2016, California experienced an extreme period of drought and high temperatures. During this period there were two particularly notable disturbances in the southern Sierra Nevada. In 2013, the first major disturbance occurred in the form of the Rim Fire. At 1,041 km2, the Rim Fire is the largest fire ever recorded in the Sierra Nevada. Throughout the drought, but particularly in the latter years, there was also epidemic levels of tree mortality, particularly mortality tied to native bark beetles (Dendroctonus spp.) killing pines (Pinus spp.).

Using the southern Sierra Nevada as a case study, I investigated the novel insights a next-generation imaging spectroscopy satellite would be able to give in understanding fire and tree mortality globally. Specifically, I showed the potential for imaging spectroscopy based spectral mixture analysis (SMA) as an assessment of fire severity. SMA cover fractions allow for a remotely sensed fire severity metric that would be more readily compared at the global level than those currently in use, such as difference normalized burn ratio (dNBR). I also demonstrated that using the random forest machine learning algorithm, a simulated spaceborne imaging spectrometer would be able to more accurately identify the location of red stage tree mortality compared to existing multispectral satellites such as Landsat.

In addition, remote sensing was used to gain an understanding of the impact and drivers of tree mortality. For a 2,240 ha watershed, a model of tree crown locations, with species and height identified, was created based on a combination of high spatial resolution airborne imaging spectroscopy and lidar. Then, high-resolution multispectral imagery was interpreted to determine a tree’s 2016 status. In the area investigated, the net effect of the drought was to reduce the number of live conifer stems taller than 15 m at the crown level by 75%, primarily due to the death of ponderosa pine. Finally, the factors that distinguished conifers that were alive in 2016 from conifers that died between 2015 and 2016, both within the 2,240 ha watershed and across the southern Sierra Nevada were examined. Trees that survived were typically associated with being located in stands with tree species and height class heterogeneity. Stands in wetter, cooler parts of the Sierra Nevada during the drought were also more likely to survive.