Landslide Hazard Study of Mercer Island, Washington
Grimm, William Edward
MetadataShow full item record
Landslides are ubiquitous in the Puget Lowland of northwest Washington State for several reasons. These factors include many ìcoarse-over-fineî geologic contacts, steep bluffs and ridges, and a wet climate for much of the year. The rapid expansion of the Seattle area population into areas that are susceptible to landslides puts people who live in such areas at risk for landslide-related damage to life and property. Mercer Island, located just east of downtown Seattle in Lake Washington, is a heavily populated island that has expensive homes and other infrastructure built in numerous landslide-prone areas. In 2009, Troost and Wisher produced a landslide hazard map of the island using field work, an inventory, geomorphic analysis, and geographic information systems (GIS). That map is currently the official landslide hazard map for the City of Mercer Island (City). Landslides have continued to occur on the island since Troost and Wisher (2009) produced their map, so production of updated landslide hazard maps for the island is essential to communicate current landslide hazards to interested stakeholders. Also, the Washington State Growth Management Act requires cities and counties to continually review and update, using the ìbest available science,î laws related to human development in environmentally critical areas, one category of which is landslide hazard areas (Chapter 36.70A RCW). Landslide hazard maps for Mercer Island that communicate present landslide hazards could be used to inform future development decisions and thus help the City comply with the Growth Management Act. Two of the geologic units in the Puget Lowland, deposited by the last major glacial advance (the Vashon Stade of the Fraser Glaciation), are particularly associated with landslides (Galster and Laprade, 1991). These units are the permeable advance outwash deposits and the fairly impermeable Lawton Clay, which lies below the sand (Troost and Booth, 2008; Galster and Laprade, 1991). Water percolates through the till that overlies the sand or directly into the sandy outwash deposits, flows downwards through the sand, and perches on top of the clay layer (Galster and Laprade, 1991). The water then moves laterally along the contact and discharges as springs, but if too much pore pressure builds up in the sand, the saturated hillside fails, and a landslide occurs (Galster and Laprade, 1991). Troost and Booth (2008) found that other coarse-over-fine contacts contribute to increased landslide hazard in the Puget Lowland, and Mercer Island has many such contacts (Troost and Wisher, 2009). My study incorporates GIS, modeling, an inventory, and field work to assess and update landslide hazards on Mercer Island. Protocols from the Oregon Department of Geology and Mineral Industries (DOGAMI) have recently become considered the best available science for mapping landslide susceptibility in GIS; I used these protocols (Burns and Madin, 2009; Burns et al., 2012; Burns and Mickelson, 2016) to present my findings to the City and potentially help the City comply with the Growth Management Act. I first produced an inventory map of landslides on the island using a high-resolution light detection and ranging (LiDAR) digital elevation model (DEM) of the island and field work (Burns and Madin, 2009; Slaughter et al., 2017). I then used the inventory map to produce shallow and deep-seated landslide susceptibility (hazard) maps of the island. I produced the shallow map by layering together areas of existing landslides, a filtered (to remove false positives) and buffered factor of safety (FOS) map, and buffers around shallow landslide head scarps (Burns et al., 2012). To produce the deep-seated landslide hazard map, I combined areas of existing deep landslide deposits and their buffered head scarps, susceptible geologic units, susceptible geologic contacts, and susceptible slope angles for each engineering geologic unit (Burns and Mickelson, 2016). To compare the effectiveness of different mapping methods, I also modeled landslide hazard areas on Mercer Island according to the Mercer Island City Code (MICC; MICC 19.16.010). I did this by combining areas of mapped landslides with hillsides that have steep slopes, coarse-over-fine contacts, and groundwater features. I also produced an additional inventory map using historical landslide data from the GeoMapNW (2005) database, Troost and Wisherís (2009) inventory, and landslides reported to the City of Mercer Island from 2005 ñ 2018. The resulting hazard maps show areas of moderate and high shallow and deep-seated landslide susceptibility across Mercer Island (DOGAMI maps) and areas of landslide hazard on the island (MICC map). Areas of high shallow and deep hazard include areas of existing landslide deposits and scarps as well as areas of steep slope. Areas of moderate shallow and deep hazard include areas of lower slope, and in the case of deep landslides, susceptible geologic units and contacts. Areas of landslide hazard, according to the MICC, include areas of existing landslides as well as many steep slopes and hillsides with susceptible contacts and groundwater features across the island. Comparison of the three maps (combined DOGAMI shallow and deep, MICC, and Troost and Wisher (2009)) shows variable levels of overlapping landslide hazard areas. The DOGAMI and Troost and Wisher (2009) maps show the best correlation, likely due to their similar mapping methods and assumptions. The MICC map overlaps with a small portion of both the DOGAMI and Troost and Wisher (2009) maps, since much of the MICC map consists of areas where groundwater is present. That is, the DOGAMI and Troost and Wisher (2009) maps have areas mapped as susceptible to landslides that do not necessarily have groundwater features. The DOGAMI protocols are now considered the best available science for regional landslide hazard mapping in GIS, and they could be used successfully to assess landslide hazards on Mercer Island henceforth, with two exceptions. The first exception is that the inventory should also include information beyond that derived from LiDAR data (e.g., historical records, field work), and the other is that the shallow hazard map still includes too many false positive hazard areas (e.g., landscaping elements). One main reason that the DOGAMI protocols are more robust than previous mapping methods is that they map shallow and deep-seated landslides separately. Since the frequency and level of damage between shallow and deep landslides is drastically different, the two maps can serve to complement each other. Additionally, the DOGAMI protocols identify areas of moderate and high hazard, whereas the MICC and Troost and Wisher (2009) methods only identify areas of hazard and non-hazard. It is helpful to distinguish between areas of moderate and high landslide hazard, so stakeholders can know where to focus energy and resources. The DOGAMI method captures the most reported and identified landslides of the three methods, which makes it the most successful method of the three. 95% of all 856 landslides reported or identified since the mid-1900s fall within DOGAMI landslide hazard zones, 87% fall within Troost and Wisherís (2009) hazard zones, and only 57% fall within MICC hazard zones. The ability of the DOGAMI method to map moderate and high zones of shallow and deep-seated landslide susceptibility, along with the consideration that the DOGAMI method captures almost all reported and identified landslides since the mid-20th Century, make the DOGAMI methodology superior to other mapping methods. The protocol allows map users to make more informed decisions than they could make with a single map showing only areas of hazard and non-hazard. The landslide inventory, shallow landslide susceptibility, and deep-seated landslide susceptibility maps are intended for regional-scale assessments of landslide hazards on Mercer Island (Burns et al., 2012; Burns and Madin, 2009). These maps could be used by developers, city planners, and other interested parties to assist in making future development decisions and to aid in emergency response planning for the island.