New Method for Estimating Landslide Losses from Major Winter Storms in California and Application to the ARkStorm Scenario

Abstract

ARkStorm is a statewide emergency planning scenario hypothesizing severe atmospheric river storms that transport huge amounts of moisture from the tropical Pacific to California, where it falls as intense rain. The storm is projected to be similar to historic storms that have struck California in recent decades, but include extended periods of intense rainfall that approach the massive storms of 1861–1862 in total rainfall. A meteorological model for the ARkStorm has been developed from records of the 1969 and 1986 storms, both of which caused widespread landslides and severe damage in California. Estimating the numbers and impact of landslides from the ARkStorm requires models relating rainfall to the number and distribution of landslides and detailed records of the impact of landslides in past storms. Neither of these basic data sets exist. However, models that estimate the areas with the greatest susceptibility to landsliding do exist and can be combined with detailed maps of small areas showing the numbers of landslides triggered in individual storms. This study extrapolated from local landslide damage to private property in past storms by using a new statewide landslide susceptibility map. These methods may provide a framework for other landslide loss estimates, which can be greatly improved with additional data. Records of landslide losses to the state highway system are more complete than records of damage to private property; it was possible to assemble records for only the past 10 years. Even within these short records, the amount of damage to the state highway system appears to be correlated with rainfall. Projecting from these records, the amount of landslide damage to highways and other infrastructure from the ARkStorm can be estimated. Overall, it is estimated that the ARkStorm would trigger landslides that directly cause billions of dollars in damage.