whigg/sliderule_methow

Investigating ICESat-2 SlideRule products for snow depth measurements in the Western U.S.

SlideRule_methow

Investigating ICESat-2 SlideRule products for snow depth measurements in the Western U.S.

I compare ICESat-2 sliderule ATL06 (with atl08 ground classification) elevation point data to an airborne lidar DEM commissioned by the Washington DNR in 2018. This lidar data was accessed through the WA DNR lidar data portal. It was transformed from the NAVD88 geoid to the WGS84 ellipsoid using the script download_dems/methow_32610.sh with help from Eli Schwatt. In the future I would like to coregister them using either command line asp (Ames Stereo Pipeline) tools such as pc_align or the wrapper around this function created by David Shean (dem_coreg https://github.com/dshean/demcoreg).

Notebooks and Descriptions (folder: notebooks):

• Data_Acess_SR-ATL06.ipynb Written by Tyler Sutterly for ICESat-2 Hackweek 2022 hosted by UW's e-science institute. Use widgets to request and download ICESat-2 data using SlideRule. See below for selected parameters in the widget drop down menus. Prior to running this notebook, I created a polygon of the outline of the DEM to pass into SlideRule for my ICESat-2 data download. This notebook then saves the ICESat-2 geodataframe to a geojson for later use.
• HB-Methow_analysis.ipynb The meat of my analysis. Reads in the ICESat-2 geojson created in the Data_Access notebook as a geodataframe. Identifies and gets snow depth data for all SNOTEL sites within a 60km radius from the center of the study site (using the polygon mentioned above - made from the outline of the DEM) and uses these sites to determine snow-on and snow-off dates for the ICESat-2 data. It also pulls in a csv of Citizen Science Observations (https://communitysnowobs.org/). It reads in the 1m resolution DEM, samples the DEM's elevation at each ICESat-2 point, and finds the difference between these elevations. The difference in snow off elevations is consider a bias/offset, and the median of these values is used to correct the DEM sampled elevations to better align with ICESat-2. The winter differences are assumed to be due to snow depth. This notebook then subsets the data to only low slope (<10 degrees slope) areas and plots the ICESat-2 snow depths compared to the Snotel network snow depths. It also plots the heatmaps representing the distribution of differences relative to slope and elevation.
• Snotel_IS2.ipynb An old notebook I pulled code from to make the Methow_analysis notebook.
• Subset_pointcloud.ipynb This is used to create a smaller subset of the ICESat-2 and DEM for use in testing the output of point cloud alignment using the AMES StereoPipeline's pc_align tool. This notebook reads in snow-off ICESat-2 data and clips it to a rectangle within the study area. It then reads in a rasterio window of the DEM and saves that window to a new file. The 1m resolution and 10m resolution files are processed.
• create_lidar_polygon.ipynb Uses code from David Shean's Grand Mesa analysis notebook to create a polygon of the outline of a raster data file, in this case the irregular outline of the Methow Valley DEM. This is the polygon that is passed to SlideRule to download ICESat-2 data within it's outline.

Here I use SlideRule for ICESat-2 data access. SlideRule parameters:

• nsidc-s3
• release = 004
• length = 40
• step = 20
• confidence = 4 (only high confidence photons will be included)
• land class: atl08_ground
• iterations = 1
• spread = 20
• pe count = 10
• window = 3
• sigma = 5
• projection: global