3 Results
3.1 Landslide Inventory
We manually identified and mapped 247 active slow-moving landslides in
California (Figure 1c). Many, if not all of these landslides have been
previously identified by other studies (Bennett, Miller, et al., 2016;
Cohen-Waeber et al., 2018; Finnegan et al., 2019; Handwerger, Fielding,
et al., 2019; Jibson, 2006; Kelsey, 1978; Merriam, 1960; Scheingross et
al., 2013; Swirad & Young, 2021; Wills et al., 2001). These landslides
consisted of different types including 71 slumps, 72 earthflows, and 104
landslide complexes (Table S3). As an expected consequence of the
relatively coarse resolution (90 m pixel spacing) of the ARIA
standardized InSAR product, we identified mostly larger landslides with
areas ranging from 0.018 to 11 km2 with a mean area of
0.5 km2 (Table S3). The active landslides are
distributed throughout the mountainous regions in western California,
with the vast majority (230 of 247) in the Coast Ranges, and they
spanned nearly the entire latitudinal extent of the state. Regions with
the highest density of landslides include the well-known landslide
hotspots such as the Eel River catchment, Big Sur coast, and Central San
Andreas Fault (Figure 1c). There are also landslides located in
populated and highly traveled zones such as Los Angeles and Berkeley,
and along California State Highway 1 and State Highway 101 and thus pose
a threat to infrastructure and life. Although we identified a large
quantity of landslides, our inventory is an underestimate of the true
landslide activity in California for three main reasons; 1) we were
unable to identify landslides in regions with high seasonal snow cover
(e.g., Sierra Nevada Mountains). To better explore regions with seasonal
snow requires a different InSAR processing strategy that only utilizes
data from snow-free periods. 2) The coarse resolution prevents us from
imaging many of the smaller landslides that have been identified with
higher resolution InSAR or field data (e.g., Kelsey, 1978; Scheingross
et al., 2013; Nereson and Finnegan, 2018; Schulz et al., 2018;
Handwerger, Fielding, et al., 2019). And 3) InSAR provides a 1D
measurement and cannot detect landslide motion in the direction of the
satellite heading (i.e., observational bias).