Remote Sensing in Ecology and Conservation (Mar 2020)

Parched pines: a quantitative comparison of two multi‐year droughts and associated mass mortalities of bishop pine (Pinus muricata) on Santa Cruz Island, California

  • Annalise Taylor,
  • Tanushree Biswas,
  • John M. Randall,
  • Kirk Klausmeyer,
  • Brian Cohen

DOI
https://doi.org/10.1002/rse2.123
Journal volume & issue
Vol. 6, no. 1
pp. 20 – 34

Abstract

Read online

Abstract Extreme weather events such as droughts are expected to increase in severity and frequency as the climate changes; it is imperative that land managers be able to monitor associated changes in vegetation health efficiently and across large scales in order to mitigate or prepare for these events. This need motivated deeper study of the die‐off of bishop pine (Pinus muricata) on California's Santa Cruz Island during the 2012–2016 drought. These pines play a keystone role within the island's ecosystem and have experienced two severe droughts and associated mass die‐offs in the past 40 years. In an effort to compare these events, we used meteorological data to track changes in drought severity from 1985 to 2018 coupled with novel methods for forest monitoring to reveal dynamics not detectable by shorter‐duration studies. Leveraging 34 years of 30 m resolution Landsat imagery, we compared vegetation mortality between the two most severe droughts of that time period: 1987–1991 and 2012–2016. We used the slope of decline in the annual median value of three different Vegetation Indices (VIs; NBR, NDMI and NDVI) to compare mortality between the two drought events and to reveal spatial patterns of mortality within the bishop pine forests. Our results indicated that the 2012–2016 drought was the island's harshest in over a century and that it resulted in greater and more widespread mortality of vegetation within bishop pine stands than the 1987–1991 drought. The average VI decline was significantly greater during the 2012–2016 drought than the 1987–1991 drought by a factor of 1.89, 2.09 and 1.84 for NBR, NDMI and NDVI, respectively. Our results aligned with projections of increasing drought severity and associated tree mortality across the region. The temporal monitoring methods developed here can be adapted to study similar landscape scale changes over multiple decades in other forest ecosystems facing similar threats.

Keywords