Assessing spatio-temporal patterns of forest decline across a diverse landscape in the Klamath Mountains using a 28-year Landsat time-series analysis

Rates of tree mortality in California and the Pacific Northwest have greatly increased in recent years, driven largely by pest and pathogen outbreaks as well as the effects of hotter, warmer droughts. While there have been a multitude of regional-scale assessments of mortality and forest decline, landscape-level studies are necessary to better identify forests that are most vulnerable to decline and to anticipate future changes. This need is particularly notable in the remote and little-studied mountains of northwest California, which are renowned for their diverse, heterogeneous vegetation types. A recent observation of elevated levels of Shasta red fir (Abies magnifica var. shastensis) mortality in a central part of this region – the Russian Wilderness – appears to mirror the timing of these larger forest mortality events and has highlighted the need to investigate if recent levels of mortality are historically unusual. The main objectives of my study were to (1) characterize contemporary tree mortality and determine potential drivers of that mortality using field-measured data, (2) integrate both field-measured data and annual LandTrendr data to assess temporal and spatial patterns of the extent and magnitude of forest decline, (3) assess the relationship between topographic and structural attributes with forest decline, and (4) determine whether climate is a potential driver of forest decline. To characterize contemporary tree mortality and determine potential drivers of that mortality, I established 142 field plots in the summer of 2015 measuring tree health and presence of any pests and pathogens on canopy tree species. Next, I used annualized LandTrendr algorithms across a 28-year time period (1986-2014) coupled with a regional forest type map to determine the timing, extent, and magnitude of canopy decline within each forest type. To assess potential drivers of canopy decline and identify specific vulnerabilities to drought, I used PRISM climate data and random forest classification using topographic and stand structure attributes. Plot data showed the highest proportions of mortality occurred in subalpine fir (Abies lasiocarpa, 35.3%) and Shasta red fir (28.6%), with evidence of fir engraver beetle (Scolytus ventralis) and Wien’s dwarf mistletoe (Arceuthobium abietinum subsp. wiensii) on many Shasta red fir individuals (34.7% and 20.4%, respectively). Forest decline was five times higher in the last two years of the time series (2013-2014) than in the previous twenty-six years. The greatest magnitude of decline was found in the red fir and subalpine conifer forest types, findings supported by my field-measured data. Canopy decline was greater at higher elevations, in denser canopies and in stands with larger trees. I did not detect any relationships between annual climate variables and forest decline, possibly due to a discrepancy between the course spatial scale of the climate data and fine-grained scale of forest disturbance, or because only two years exhibited pronounced canopy decline. My study demonstrates effectiveness in characterizing forest decline in a highly diverse landscape using a remote sensing approach and highlights the complexity of climate, pests and pathogens, stand structure, and topography as they relate to tree mortality and forest decline