Regional landslide hazard using Landlab - NOCA Data

This NOCA landslide data repository host the driver code and data files needed to run Landlab's LandslideProbability component, which models annual shallow landslide probability in a steep mountainous region in northern Washington, U.S.A. The model application covers North Cascade National Park Complex (NOCA), using 30-m grid resolution over 2,700 km2. The model use the classic infinite slope, limited equilibrium model driven by contemporary climatology from the Variable Infiltration Capacity (VIC) macroscale hydrology model. Readily available topographic, geophysical, and land cover data are provided to calculate the factor-of-safety stability index in a Monte Carlo simulation, which explicitly accounts for parameter uncertainty.

Data used for this analysis are spatial data on landscape characteristics for NOCA. They include soil, geology, vegetation, topography, and landform data that can be used for quantitative landslides hazard assessment. Elevation was acquired from National Elevation Dataset (NED) at 30 m grid scale; other datasets are matched to scale and location. Slope was derived from the elevation file as "tan theta". Specific contributing area represents the 'upstream' area draining to each cell divided by the cell's width (so minimum value is 30 m). Landform data was developed by Jon Riedel of National Park Service. Landslides were extracted from these data identified as "mass wasting" events. Land use and land cover (LULC) data were acquired from USGS National land Cover Data (NLCD) based on 2011 Landsat satellite data and grouped into eight general categories. Cohesion represent total cohesion, which is equivalent to root cohesion in this application; soils are assumed to be primarily cohesionless, lacking “true cohesion” because of their low clay content in this mountain terrain. Root cohesion is based on the LULC referenced to a look-up table within this resource: (https://www.hydroshare.org/resource/a771ba9bbae24ed8b4673c945fc321a3/). Soil depth comes from Soil Survey Geographic Database (SSURGO) maintained by NRCS processed as soil survey depth-to-restricted layer (weighted-average aggregation) within each soil map unit. An alternative modeled soil depth (SD) described in the accompany paper is also provided, but revisions in the driver notebook would be required to reference this file to see adjusted results. Transmissivity was derived from the soil survey saturated hydraulic conductivity (depth averaged) multiplied by depth-to-restricted layer for each soil map unit; another T file based on the model soil depth is also provided. However, the model can be run using hydraulic conductivity using data file provided to calculate T. All soils within this watershed are sandy loam or loamy sand; therefore, soil surface texture was used as an indicator of internal angle of friction (phi). A header file is provided to understand the spatial details of the ASCII files and to facilitate capability with GIS. Spatial reference for raster mapping is NAD_1983, Albers conical equal area projection.

The model run archived in this resource runs with Landlab version 1.1.0 . The component code (landslide_probability9Jun17.py) is provided as an archive to run a notebook that replicates results in Strauch et al., (in review) . As Landlab is developed with newer versions, the notebook and/or provided component code may need updating to run properly. To run the notebook to replicate results, use the resource "Regional Landslide Hazard Using Landlab - NOCA Observatory", HydroShare resource: https://www.hydroshare.org/resource/07a4ed3b9a984a2fa98901dcb6751954/