GIS input parameters for Landlab

[RondaStrauch]

New issue to capture our acquisition of necessary GIS rasters and shape files to run the Landlab Landslide component
Initially storing the in a hydrosphere resource: https://www.hydroshare.org/resource/70d746c7da584ae6bd2f88deb5a4c188/
Files include:

1. DEM
2. Study area boundary (US Skagit River HUC 12 sub basins down to confluence with Sauk
3. HUC 12 and HUC 10 boundaries (might want to aggregate data by these)
4. NOCA park boundary
5. Goodell Creek Fire

[RondaStrauch]

Check Landslide hydrosphere data site and found additional data files needed from Erkan's computer:

• Cohesion (min, mode, max)
• friction angle
• soil depth (?)
• slope (?)
• exclusion mask
• landslide type
• Geology

[ChristinaB]

• Look at Jim's code that clips grids to a Landlab raster model grid

• Add to methods section

• Add clipped files to hydroshare https://www.hydroshare.org/resource/70d746c7da584ae6bd2f88deb5a4c188/

[RondaStrauch]

Added a North Cascades filled DEM at 30 m into Hydroshare.

[RondaStrauch]

Checked on land cover data used to create cohesion. Need to recreate cohesion because the existing files are for NOCA. Check that the LULC is being updated from 2011 to 2016 data, but this is suppose to come out April 2019...not yet though.
So, found the 2011 data for NorthCascades LULC and updating the projection to square up with other data and upload to hydroShare.
Added Washington state geology onto HydroShare.

[RondaStrauch]

Working in Probfail.mxd ArcMap
Found the classification table 'LULC_reclass.xlsx' and reclassed LULC into 8 classes.

Reclasifying the LULC 8 types into cohesion values for mode using the values in the esurf paper, but set ice/snow/water to 1 and wetlands to 2 because we can see where these are. They will be excluded from the analysis with an exclusion mask. Reclass was x100 because has to be integer.

ToDo:

• Need to go back and use raster calculator to change cohesion to real values by /100.
• Create cohesion minimum and max rasters through the same process
• Upload LULC 8 classes and 3 cohesion rasters to hydroshare resource

[RondaStrauch]

Completed todo list in last comment

[RondaStrauch]

5/9/19
Clipping to get rows & columns of study area
This is if you don’t check the geometry box:
Rows = 3220 and columns = 3281
But no data values are -128 even though specified -9999 ???

When we clip to geometry, still get the same number of rows/columns, but this time the no-data values are -9999

Notice that this last clipping looks like the raster is only covering the watershed areas, but in fact it extends to the same extent as the past one, but nodata is showing up as white.
Confirmed rows & columns by converting to ASCII and checking header

[RondaStrauch]

Domain header for 30m grid - but this is in degrees

And Source tab in GIS

[RondaStrauch]

Talked with Nicoleta. We need to decide our spatial domain. We chose the one in 2 comments above. Decided to try reprojecting in UTM 10 coordinates (WGS 1984) in a new dataframe. I also created 3 mask for this domain that has 1 for inside watershed and nodata, -9999, and 0 for outside watershed. Here is the Albers image UTM 10 projection, but the UTM also still shows Albers as spacial reference.
Albers projection

UTM 10 projection

But both of these are still in m units, not lat/long

[ChristinaB]

@RondaStrauch This looks good to me. Can you overlay the full skagit boundary and let's report the fraction of the park and the fraction of the Skagit we are including in the model. It's hard for me to tell what part of the Skagit we are excluding.
C

[ChristinaB]

Methods for clipping DHSVM output to a square domain of the hazard map.

[RondaStrauch]

@ChristinaB Can you load the 'testdomain' GIS layer (or the 01 mask) to Hydroshare GIS folder or is this already in Github?

[ChristinaB]

Click to download grids of climate data elevations

Click to download the SCL mask files

I'm also uploading the clipped DEM of the NorthCascades 30 m data to the Mask- it doesn't totally cover the mask so we will need to merge this with another dataset. How do we want to handle this? I can do a raster calculator and fill in with 150 res data downscaled to 30. Also - use the raster calculator and Environment settings to clip - don't use the clipping tool in ArcGIS!

[ChristinaB]

Click to download sample DEM to see the gap in the North

[NCristea]

I am not sure if the soil depth file you uploaded has the correct data and geographic information. Could you please upload a tif file you used to set up DHSVM? I would need a correct dataset with geographic information of DHSVM grids.

[ChristinaB]

You can download a zipped file of raster and ascii files and headers at this link: SCL Domain

Extent	Resolution/Grid Cell Size	xllcorner	yllcorner	ncols	nrows
Observatory Climate Forcings SCL extent	6 km	591315	5354895	13	13
Skagit Hydrology & Climate Change Study Domain	150 meter	535665	5311395	916	1020
Seattle City Light Skagit Study Domain - Hydrology	150 meter	591315	5354895	545	544
Seattle City Light Skagit Study Domain - Landslide/Fire	30 meter	591315	5354895	2725	2720

[ChristinaB]

The SSURGO shapefile is skagit150DHSVM_soilclasses_20150627_204191130.shp . We attached a reanalysis of soils from LanslidePhi_20190710_subsurface.xls
I converted the shapefile to a grid, and resample to 150m and 30 m grid cell resolution.
I converted the Phi grid into an ascii. The header is consistent with the DEM header.
ncols 2725
nrows 2720
xllcorner 591315
yllcorner 5354895
cellsize 30
NODATA_value -9999

@RondaStrauch Do you want missing Phi values to be -9999, 0 or set to a value like 44?

[ChristinaB]

@RondaStrauch https://www.hydroshare.org/resource/70d746c7da584ae6bd2f88deb5a4c188/ has a new Phi_parameter folder inside the GIS folder. Also adding the ascii to code in this repo.

[RondaStrauch]

@ChristinaB I'm using the new phi you provided and using this to create ascii's for the other parameters in the UTM 10 projection. I'll make the phi have a missing value of -9999