FSX-MS Coming soon to FSX: Time Capsule - Mount St. Helens Pre-1980

[SteveRosenow 5]

A little update on the current status of this project. This one was a significant alpha testing milestone in the creation of pre-1980 terrain elevations north and west of Mount St. Helens. For the last few days I have been immersing myself into the workings of QGis - an open source geospatial data processing application very similar to the ArcGIS software professional mapping agencies and the USGS use, to create and author DEM and other forms of geospatial data. Yesterday, I began a series of tests on working with ESRI Shapefile data and creating test contour maps, then importing them in another open source geospatial data processing application known as SAGA. It is here, where I applied the elevation data in a TIN (triangulated interface network) format to create and compile a 3D model of those test contour maps.

As a test of those learned skills, earlier this afternoon I decided to trace contours in a small area going up from Spirit Lake's pre-eruption elevation contour of 3,198 feet. For reference, today's lake elevation is 3,406 feet! The small area I chose was the dividing ridge between the east and west lobes of Spirit Lake. Once I selected the area in question, I then proceeded to trace every contour on one gridded section of that ridge in a 1958 topographic map, using a vector layer, with each contour assigned an elevation point corresponding to the 80-foot contour interval of the source reference map. (It was a PAIN IN THE TOOSHIE to do this, since I had to convert the numbers in feet, to meters every time!).

After tracing the final contour in QGis, I then exported it as an ESRI Shapefile, then imported it into SAGA, to which I then did a triangulating extrapolation of that contour data into a working digital elevation model. That elevation model's dataset was then loaded back into QGis as a SAGA DEM file, and subsequently exported as a GeoTIFF DEM.

Once a brief test was initiated in FSX, I was ASTOUNDED at the height difference between present-day Spirit Lake and the former shoreline in the test area. It shows up best at the lake shore (where I had placed an exclusion flatten of the lake to eliminate scenery artifiacting) on the dividing ridge, and where Harry Truman's resort is.
Now the fun part begins... Tracing a word not allowed-ton of contour data, then extracting it into height map data, and then finally, off to the SIM.

For the next month or so, I am going to be hard at work recreating this pre-1980 eruption terrain for both Flight Simulator, and potentially, for the Washington State Department of Natural Resources. It will take approximately one month, as I have to trace each contour by hand using a mouse cursor. It will be extremely laborious work.

Now for a few images showing the sequence of steps I followed, then a series of screenshots in FSX showing the change in topography. Captions are below each image


Selecting the test area in QGis. The base image is a 15-minute USGS Quadrangle of Spirit Lake from 1958. It, and the four additional USGS 15-minute quadrangle maps I am using, all are from 1953 or 1958, and to date are the highest resolution pre-eruption contour maps of Mount St. Helens and the surrounding area.


Contours finalized, and ready for export. One contour was missed in this, but I was told by an expert in geospatial data processing that missing one contour in a quadrangle in which contour line divisions are 80 feet, really doesn't impact the final DEM product much.


Contours imported into SAGA. At this step, the contours have already been assigned a TIN interface.


Converting the TIN interface into a DEM in SAGA. At this stage, the DEM is in a format that cannot be read by Resample.exe. 


Back in QGis, where it is now ready for export as a GeoTIFF file that Resample.exe can read and convert to *.BGL


A raster extraction of hillshade data on the DEM, illustrating the resolution of the test DEM's clarity. At this step, it was exported as a GeoTIFF, then an *.INF file wrote for compilation to a FSX *.BGL 


Loaded into FSX. Note the elevation difference. The former lake shore was set with flatten at the altitude of present-day Spirit Lake. The DEM falls to pre-1980 eruption-level shoreline altitude.


The altitude and elevation difference on Spirit Lake's southeast shore, in the vicinity of Harry Truman's Mount St. Helens Lodge. 

Edited November 9, 2021 by SteveRosenow

[SteveRosenow 5]

PROGRESS REPORT for November 10, 2021.

What was thought to be impossible, has now become a reality.

In the last few days, I have been immersing myself into the workings of Q-GIS, an open-source geospatial data processing suite similar to the ArcGIS software in use by professional mapping agencies, college geology courses, as well as the USGS. Additionally, I have also been exploring SAGA, another open-source geospatial data processing suite. Through a lengthy process of trial-and-error, and much experimentation (which can be seen here: https://www.fsdeveloper.com/forum/threads/sbuilderx-sloping-ground-polygons.454002/#post-893586 ), I have successfully created a test file recreating Spirit Lake's pre-1980 elevation data and former lake shore.

Additionally, this test also successfully eliminated the landslide deposit north of Mount St. Helens in an area immediately west of Spirit Lake itself.

To achieve this I set forth on a test exercise in Q-GIS, by creating an ESRI SHP "shape file" data package. Then, I began an extremely tedious process of tracing every contour below elevation 3,600 feet (I started at the contour line denoting 3,520 feet), in area directly encompassing Spirit Lake  to the north base of Mount St. Helens. This area also went west to just below the western end of present-day Johnston Ridge, at the boundary line for Cowlitz and Skamania Counties.

As to how laborious this process is, this one test sample took seven hours. Six of those were spent doing the "hard" work - tracing those contours - by hand - using a mouse cursor placing a multiple series of lines to which elevation data had to be attached.

After that, it was a trip over to SAGA GIS, where the contour data was then extrapolated into a working digital elevation model. However, since this file was not quite ready for Flight Simulator's Resample SDK application, it had to be imported back into Q-GIS.  It is in Q-GIS, where "the fatty part" (e.g. the unusable data) was then trimmed to meet the test coverage area, and then subsequently exported out to a working GeoTIFF file for Resample to convert into a *.BGL scenery file.

After some initial hiccups, the resulting test of this test file in Flight Simulator X was a resounding success. Initial tests of this file showed large plateaus and large ground spikes where the DEM met the clipped out areas. This was achieved by reworking the*.INF file to the following parameters:

[Source]
Type=GeoTIFF
Layer=Elevation
SourceDir="."
SourceFile="test_BGL_SpiritLake4.tif"
MinValidValue=755
NullCellValue=-32767

[Destination]
DestDir="."
DestBaseFileName="test_BGL_SpiritLake"
DestFileType=BGL
LOD=6,16
FractionBits=6
BaseValue=1000

The MinValidValue line denoted, in meters, the lowest elevation of the DEM. The NullCellValue is derived from a YouTube tutorial here: ([MEDIA=youtube]-PNoaGkbRQg[/MEDIA])

This tutorial was viewed as I was having issues eliminating the initial plateaus in the test DEM. The "MinValidValue" call in the BGL subsequently eliminated that.

As this is a significant update (and a very important learning experience, I am posting in this update, a step-by-step illustration showing what was done. Photo captions below each illustration.

Selecting the coverage area in Q-GIS. The underlying data are georeferenced GeoTIFFs of five USGS 15-minute quadrangles, with a contour interval of 80 feet. These were sourced from TopoView, a USGS public domain download portal.

The traced contours in the test area.

Importing of the traced contour SHP shapefile data into SAGA GIS.

Selecting the interpolation parameters. For this I chose Multi-Level B-Spline, however the final exported DEM will be a TIN interpolation for greater accuracy and 3D display of the DEM.

After a ten-minute wait (it takes this long to compile the DEM, at least for the test area selected), the DEM is now ready for further processing. At this point, it is unusable for Flight Simulator, and must be converted in Q-GIS to a georeferenced GeoTIFF.

Importing the raw data back into Q-GIS. Note that unprocessed DEMS such as this need to be cleaned up for further use and display in GIS data applications, as well as Flight Simulator.

For this, I drew a polygon on the outside edge of the extent of the outermost contour. This was selected because I felt it unnecessary to do the entire area, and I am only after the first five hundred feet above the pre-eruption shoreline.

The final DEM, ready for export to Resample.exe.

Testing in FSX. The shoreline elevation and surrounding hills are restored to their pre-1980 elevation. This view in FSX in fact now matches an aerial of Spirit Lake in my reference archive taken from roughly the same vantage.

View to the NE and NW, overlooking the area of the Dave and Mariam Smith's Spirit Lake Lodge.
Harry Truman's Mount St. Helens lodge is on the SW corner of Spirit Lake at far right in the first image.

These ground spikes and the cliff behind them, illustrate the elevation changes between spring, 1980, and today. The Spirit Lake Lodge, sitting on the spike to the left, and a private cabin (based on photos of the real cabin), sitting on the spike to the right, were placed in the scenery on two AB Flatten polygons to eliminate display issue conflicts caused by the present-day elevation in that area. The cliff behind them, and the height of these spikes, are a direct illustration of the depth of the landslide deposits left behind by the collapsing north flank of Mount St. Helens in the May 18, 1980 eruption. It is in this spot that the landslide rode up and over the ridge, and then into the valley behind.[/URL]
This cliff will be eliminated, and full ground data restored to pre-1980 eruption elevation.[/URL]
For the next few months, I will be hard at work retracing all those contours, then compiling multiple test DEMs, to ensure this displays as accurate as possible.[/URL]

[JayYount731 1]

You’ve done some fantastic work in this for sure. A lot of hours albeit! 
I have found a Government employee that created pre eruption data SFM from aerial photographs taken by the USGS. pretty remarkable DEM but yet again unable to get in touch with her. 
do you no longer have this data

its significantly larger area then I was working with. My main issue now is the odd lateral hatching in the terrain from stitching that was done when they merged maybe? I have some interesting renders of post eruption worth a look. 
 

thanks again for your time Sir and great work! I can’t wait ti see how far you take it. 

[SteveRosenow 5]

6 hours ago, JayYount731 said:

You’ve done some fantastic work in this for sure. A lot of hours albeit! 
I have found a Government employee that created pre eruption data SFM from aerial photographs taken by the USGS. pretty remarkable DEM but yet again unable to get in touch with her. 
do you no longer have this data

its significantly larger area then I was working with. My main issue now is the odd lateral hatching in the terrain from stitching that was done when they merged maybe? I have some interesting renders of post eruption worth a look. 
 

thanks again for your time Sir and great work! I can’t wait ti see how far you take it. 

Unfortunately that was just a *.GIF that I had converted to a *.TIFF. Subsequently I was unable to georeference it to the coordinates I needed, and further to that I was unable to generate a DEM from it.

I'd love to see the data you came across. 

As an update, since my last post in this thread I have made significant progress in recreating an entirely-new pre-eruption DEM of the peak. I will post screenshots and an update tomorrow morning. 

[SteveRosenow 5]

A huge amount of work has been done since the last update. As a word of caution, this post will be somewhat on the image heavy side. 

The first and most important part of the update has been the complete reestablishment of pre-1980 terrain elevation north and west of Mount St. Helens, for a distance of 21 miles downstream from Spirit Lake. This includes every mile of affected terrain which saw elevation changes brought forth due to the landslide deposits off of Mount St. Helens' collapsing north side. This was accomplished by producing a custom digital elevation model in-house using contour extraction from pre-eruption topographic maps. (see illustrations above for the detailed explanation of the processes involved)

Secondly, I have been able to further refine and branch off from the development of this, and create and establish a custom digital elevation model recreating the bulge on Mount St. Helens' north flank as it appeared in late April and early-to-mid-May. This was accomplished by georeferencing a USGS topographic map of Mount St. Helens' north flank with the same 80-foot-interval contour data as the 15-minute topographic maps I am already using as reference material. This map, sourced from Figure 69 of Page 125 in USGS Professional Paper No. 1250, features adjusted contour elevations of the bulge and elevation changes in Mount St. Helens' north side. It also features the 700-foot-deep, 1,600-foot-long crater at the summit. 

To match this custom bulge DEM, I also sourced a high resolution infrared color photo of Mount St. Helens taken on May 2, 1980. The photo shows an ash-stained snowy peak with cracks in the north side. The image, once color corrected, was warped to match the projection of the existing terrain overlay. To add to the package, I recreated a smaller phreatic steam and ash eruption the likes of which occurred before the May 18, 1980 explosion.

And now, for the captures.


Pre-eruption elevation restoration of Spirit Lake. 


Looking SE towards Mount St. Helens. The clearcut in the foreground is a Weyerhaeuser transfer yard, and Highway 504 snakes up the valley. The ridge beyond, then known as South Coldwater Ridge, looms in the background. Today, that ridge is called Johnston Ridge.


Looking north over the summit of Mount St. Helens, at the pre-eruption elevation of Spirit Lake.

The entire coverage area of the DEM in question. 



Mount St. Helens on May 2, 1980. Note the bulge has distended the north flank, and cracks show up. For reference below, I've attached a photo taken by the U.S. Forest Service of the actual peak taken the same day.



And lastly, shots illustrating a minor early May eruption.