Development and implementation of subsurface hydrologic thresholds for identifying widespread shallow landslide and debris flow occurrence in the San Francisco Bay area (California, USA)

Precipitation-triggered landslides resulting from prolonged and/or intense storms threaten lives and damage infrastructure throughout the world each year. In the United States, population centers along the West Coast (e.g., Seattle, Washington; Portland, Oregon; San Francisco, California) have particularly high risk from landslides due to the intersection of typical cool season (October through May) Pacific cyclone storm tracks with intense urbanization located on and near steep hillslopes. Past efforts on landslide early warning in California dating to the late 1970s and running through the mid-1990s were initially focused on the development of rainfall intensity-duration thresholds coupled with a recognition that little landsliding tended to occur prior to an antecedent rainfall condition being reached essentially a proxy for soil saturation. However, available technology at the time did not allow for economic and logistically viable subsurface monitoring of in situ hydrologic conditions in steep, landslide prone terrain. In 2009, the U.S. Geological Survey (USGS) began development of a regional subsurface hydrologic monitoring network in the San Francisco Bay area to assist with keeping emergency managers informed about periods of elevated hazard from rainfall-induced shallow landslides. The network currently consists of four stations, each located within hillslopes susceptible to shallow landslide initiation and representative of those that have failed in the past. The stations are spatially arranged to capture the meteorological variability of the approximately 18,000 km2 region. Each station consists of two monitoring nests with soil moisture and piezometric level sensors placed at variable depths within the typically 0.5 to 1-m-deep soil profiles. The goal of the monitoring network is to identify the time frame(s) during which soil saturation may be reached during storms and thus enable generation of positive pore water pressures that can cause shallow landslide initiation. Using the more than 10-year record of soil moisture time series data available from the network stations, combined with piezometric records indicating times of positive pore water pressure and observations of triggered landslides, we developed soil moisture thresholds that are now used for situational awareness to alert for the potential for widespread shallow landsliding and debris flows ahead of incoming storms. Messaging indicating that hillslope soils are nearing or are at saturation is provided to the U.S. National Weather Service (NWS) several days ahead of storms to provide for sufficient time for communication with emergency managers so that they may identify and prepare appropriate resources for response should landslides occur. The monitoring network, combined with established USGS-NWS communication protocols, has been successfully used to alert for expected landslide conditions in numerous storm events over the past 10 years, including the precipitation-record-setting 2022-2023 winter season. Ongoing research is aimed at updating thresholds with recent data and developing semi-automatic routines for monitoring and alerting.