I am having a hard time creating viewsheds on Mars. Originally, I thought it was a coordinate system projection error, so I created a custom projected coordinate system for Mars, went to this site, and made sure all my parameters were correct: http://chrisherwig.org/hirise/Data%20Processing/2012/10/21/mars-map-projections/ (the information that I found most helpful is in the red print and you have to slide the bar to see the parameters they use for the Mars Equirectangular Spatial Reference data). I set my projected coordinate system to Equidistant_Cylindrical, not Plate_Carree, and set the false easting, false westing, central meridian, and the standard parallel1 to 0 (I also made another map with a central meridian set to 180 and all the rest to 0 like the one in this link that uses 180 as its central meridian). However, when I went to check the coordinate systems after I projected them to make sure they were correct, I found that ArcGIS had automatically set their projected coordinate system to Plate_Carree. I didn't understand why it was doing this until I went to re-project it and discovered that the only difference between the Equidistant_Cylindrical projection and the Plate_Carree projection is that the Equidistant_Cylindrical projected coordinate system had an option for using a "Standard_Parallel_1" (it automatically sets this parameter to 60 too). I then realized that since I've been setting the Equidistant_Cylindrical projected coordinate system's "Standard_Parallel_1" to 0, it had just been getting rid of that field all together. That meant that it just projected the data in Plate_Carree since Plate_Carree doesn't have a "Standard_Parallel_1" parameter. I also noticed that this happened only with rasters I used. When I created a custom projected coordinate system for my shapefiles, it acutally made the projected coordinate system I created (the projection was Equidistant_Cylindrical, not Plate_Carree). So, I went ahead and made viewsheds using the shapefiles in this coordinate system and the rasters in the other, and I inevitably got the same questionable viewsheds as I had when I only used the Plate_Carree projection.

Are my projections why my viewsheds are messed up?

The data I am using as my parameters for the viewsheds are hands down correct, and I've read through PDS documents from the NASA website explaining what all the information means.

Explaining what I mean by viewsheds being incorrect:

In the image below, the viewsheds way off to the right should be were the ones off to the left are located (they're kinda small, next to the blue dot, and almost barely visible). Those ones off to the left are also located further west of a bank they are supposed to be on. I know they are incorrect because I am looking at images from the mastcam of the rover, and I am locating the geologic feature I see in the images on this map. The viewsheds are supposed to represent what we see in the images. viewsheds

| improve this question | | | | |
  • What were the precise steps that you performed? – PolyGeo Jul 31 '17 at 21:52
  • Please make sure all the critical information is in the question, as not many people are willing to follow links blindly. Since Earth and Mars have significantly different radiuses (Mars nearly half of Earth), this would certainly impact a viewshed. We need hard numbers. – Vince Jul 31 '17 at 21:55
  • When std parallel = 0, equidistant cylindrlcal == plate carree so that shouldn't be the issue. Bad projection in general for viewshed analysis? How are the viewsheds messed up? – mkennedy Jul 31 '17 at 22:36
  • The viewsheds are sometimes dramatically different like in this example. And other times they are just slightly off, either being too high, too low, or too far left or right from where we expect to see them. I know this would indicate something about my parameters I'm using for producing the viewsheds, but I am very certain I am using the right information. Could I still be wrong? – Mark Jul 31 '17 at 22:59

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.