ESTIMATING FUTURE ECONOMIC LOSSES USING LANDSLIDE INVENTORY DATA

Quantitative estimates of landslide risk (R) are critical to the efficient management of landslide hazards. However, these estimates are rarely made, especially at regional scales, because of difficulties in determining all of the components in the standard landslide risk equation, R=HVE, where H is probability of hazard occurrence, V is physical vulnerability to elements exposed to H, and E is the cost of elements at risk. Quantitative risk estimation and Economic Loss Estimation (ELE) are closely related, therefore ELE provides an alternative method for determining economic risk from landslides for local or regional areas. We present an ELE method, the Probabilistic Landslide Assessment Cost Estimation System (PLACES), which uses historical inventories of losses from damaging landslides to estimate economic losses from future landslides during a specified time in individual areas. PLACES is based upon conditional probability theory and laws of expectation and variance and is available in the form of a computer spreadsheet program.

We used PLACES with landslide inventory data collected between 1968 and 2008 to estimate the annual, direct economic losses from future, rainfall-triggered landslides in 10 counties of the San Francisco Bay region. The estimated mean annual loss from future landslides in the entire region is about US $14.80 million (year 2000 $). The estimated mean annual loss is highest for San Mateo County ($3.24 million) and lowest for Solano County ($0.18 million). The annual per capita cost for the entire region will be about $2.10. Normalizing costs by dividing by the percentage of land area with slopes equal to or greater than about 10 degrees indicates that San Francisco County will have the highest cost per square km ($7,101), whereas Santa Clara County will have the lowest cost per square km ($229). All of these values are minimum estimates because records of damaging landslides between 1968 and 2008 are incomplete.

PLACES can easily be applied to other geographic locations, but requires an inventory of landslide loss data over an extended (20+ years) period of time. Such inventories require a long-term commitment, and a standardized, sustained effort, which is generally best accomplished by single organizations, rather than by multiple organizations with varying needs and interests.