Trends in storm-triggered landslides over southern California

Abstract

Changes in storm-triggered landslide activity for southern California in a future warming climate are estimated using an advanced, fully three-dimensional, process-based landslide model, SEGMENT-Landslide. SEGMENT-Landslide is driven by extreme rainfall projections from the GFDL-HIRAM climate model. Landslide changes are derived from GFDL-HIRAM forcing for two periods: (i) 20th Century (CNTRL); and (ii) 21st Century under the moderate IPCC SRES A1B enhanced greenhouse gas emissions scenario (EGHG). Here, differences are calculated in landslide frequency and magnitude between CNTRL and EGHG projections, and kernel density estimation (KDE) is used to determine differences in projected landslide locations. This study also reveals that extreme precipitation events in southern California are strongly correlated with several climate drivers and that GFDL-HIRAM simulates well the southern (relative to Aleutian synoptic systems) storm tracks in El Nino years, and the rare, ~27 year recurrence period, hurricane-landfalling events. GFDL-HIRAM therefore can provide satisfactory projections of the geographical distribution, seasonal cycle, and interannual variability of future extreme precipitation events (>50 mm) that have possible landslide consequences for southern California. Although relatively infrequent, extreme precipitation events contribute most of the annual total precipitation in southern California. Two findings of this study have major implications for southern California. First is a possible increase in landslide frequency and areal distribution during the 21st Century. Second, the KDE reveals three clusters in both CNTRL and EGHG model mean scarp positions, with a future eastward (inland) shift of ~0.5° and a northward shift of ~1°. These findings suggest that previously stable areas might become susceptible to storm-triggered landslides in the 21st Century.