0

I'm trying to calculate the volume of a landform under water in ArcGIS and here's what I decided to do:

  1. Convert the raster of the bathymetry that I had to a gridded point dataset using Raster to Point.
  2. Clip out the points contained within the landform (leaving me with a somewhat large hole). I actually have multiple landforms, some creating smaller holes, but the biggest with a surface area of over 300,000 sq m.
  3. Re-interpolate the surface sans landform.
  4. Use the Cut and Fill tool to compare the original surface to the interpolated surface (without the landform) in order to get the volume.

Makes sense to me --- the trouble is, I've been trying to find the right interpolation method to capture a surface sans landform. I think my hole was way too big for IDW, because as soon as the contours hit the re-interpolated hole they shot off in funky directions. I did a TIN that did fairly well, but I'm not sure how accurate or believable a TIN can be. Lastly I tried kriging, but processing was taking in excess of 24 hours, so I cancelled the process.

Can Anybody give me some tips on the best interpolation method for this task in Arc?

2

The best interpolation method depends on the application and the parameters used. The IDW technique for example requires a minimum and maximum number of surveyed points to predict an interpolated point. The power(exponent) applied determines how quickly the influence of each surveyed point drops off with distance. In regions where data is missing or sparse, IDW generates funky results - you may be seeing this near the holes. And in areas with good data distribution you may see concentric contours or "bull's-eyes" - these get tighter the high the exponent.

Generally, TINs do well for modelling land surfaces(landforms, DTMs etc), however, an understanding of parameters to the algorithm is important. If you have accurate, uniform data, the defaults should give a good representation of the landform without any adjustments. In ArcMap its Delauney triangulation unconstrained. Further reading on this.

Further reading on this

Short answer: it depends on the data you have and application.

1

Agreed, this will be a very dense process for the program to handle. If everything appears to be working (except for the interpolated surface around the edges of your landmass), and you want to use the IDW algorithm, you simply need to set the "Maximum Distance" for the "Search Radius" option in the geoprocessing window in ArcMap, prior to running the tool. The default is no max distance, so any "holes" in the inputted point data will have surfaces created to fill them (elevation values for any cell determined of course by Inverse Distance Weight to surrounding points). This means that, as you're running it now, the tool is filling the landmass area with the weighted elevation value between the closest points, which may be very far away.

If your surface for the bathymetry in the processes you've already run look fine, I wouldn't touch the "Number of Points" in the Search Radius settings, but type a Maximum Distance (in units for coordinate system) that is less than the distance across the landmass, at the shortest point across it. That should exclude the landmass from your interpolated surface.

  • Do you have access to the 3-D Analyst tools? There is Surface Volume tool that will calculate the volume and areas of your surface without all the fussing and interpolating. You may need to use something like Extract by Mask to ensure you are only calculating the landform in question and not adjacent areas. – GBG Nov 1 '17 at 16:41

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.