Leveraging the LCMAP Data Products to Distinguish Types of Change in Tree Cover Associated with Drought

Abstract:

California (CA) has in recent years become a hotspot of interannual climatic variability, recording several climate-related disturbances with severe impacts on tree resources and critical habitats of endemic and threatened species. Understanding the patterns of tree cover change is vital for developing strategies to sustain ecosystem services and for projecting future effects of climate and land-use change on similar landscapes. We exploited multiple land cover and land change products from the Land Change Monitoring, Assessment and Projection (LCMAP) dataset along with climate data to detect different types of change in tree cover and examine their relationships with interannual climate variability during 1985–2017. Changes in tree cover identified by the Continuous Change Detection and Classification (CCDC) algorithm were partitioned to track annual estimates of loss, gain and conditional change. Overall, there were increasing trends in tree cover loss and conditional change and decreasing trend in tree cover gain during the study period in CA. The multi-year drought in the state during 2012–2016 was associated with larger changes in tree cover, in which tree cover loss dominated. Annual estimates of conditional change, which indicates change in condition without a thematic change, were generally higher than both gains and losses for the period prior to 2006. There were gradual increases in tree cover gain during 1985–1995 and afterwards gradual decreases until 2017. Spatial variability in climate response was associated with varied patterns in tree cover change across ecoregions in CA. Similar trends in tree cover were observed within the range of Quercus douglasii (Blue Oak), an endemic tree species to CA with a narrow geographic extent and elevation gradient. Within this species range the largest amount of conditional change was recorded during the drought period. Conditional change in tree cover was highly related to tree cover loss, suggesting the need for tracking disturbance trajectories across time. Early detection of subtle changes in tree cover condition may be vital to support proactive resource management decisions aimed at minimizing complete tree cover losses. The LCMAP approach offers a promising capability for early detection of impending losses to forested and tree covered landscapes.