DEM Reconditioning: Burning culverts into a DEM -- Z units?

[AstoriaTershy]

Hello, 

I'm a graduate student using the InVEST NDR model to estimate nutrient retention in riparian buffers. One of the model inputs is a DEM. The DEM doesn't capture agricultural canals and ditches that are a significant part of nutrient loading in the watershed I'm studying. I used the DEM Reconditioning tool in Arc Hydro to burn in canals. Disclaimer: I apologize in advance if I am blatantly ignorant about this subject; please correct me.

I wanted to get opinions on my methodology to see if I've made errors in my geoprocessing. My main concern is Z units. My understanding is that they are vertical units that are usually equal to the horizontal dimensions of whatever raster one is working with.

The DEM reconditioning tool prompts the user to choose the number of Z units that the elevation will be dropped to burn in the stream raster. Z units need to be integers greater than or equal to 1. My DEM was a 10x10 m raster, so if I chose a Z unit value of 1, that would equal a 10 m drop-- which is way too deep for an agricultural canal! Since I'm working with agricultural canals that aren't too deep (usually about 1 m), I needed my Z units to be able to reflect an accurate depth of these canals. As a work-around, I resampled both my DEM and my stream raster layer to 1 x1 m resolution. Then I choose a Z unit of 1 in my DEM Reconditioning tool (which hopefully is equal to 1 m).

I've attached photos of my DEM Reconditioning tool and an overall model of my work.

1. Is the above a viable approach for DEM reconditioning?
2. How can I check if the Z units of my 1x1 m raster are actually equal to 1 m?

[KyleGonterwitz · Accepted solution]

You can look at the environments tab there and check your horizontal and vertical coordinate system reference, I think you can even change your desired output reference there - but there are separate vertical and horizontal coordinate system references.  Also consider you are reconditioning for archydro and accurate channel depth doesn't really matter in the hydrologic analysis aspect, other than it needs to be deep enough to cut through say a road embankment.

[MarkBoucher · Accepted solution]

As for units of the burn, I have always experienced that the units are feet or meters, whatever your project is in. It is not a function of the raster grid size. You can always do a trial and error on that one.

I agree with KyleGonterwitz that the depth of the burn doesn't mean anything... that is, if what you are trying to do is come up with a flow direction grid. The bottoms of the burned channels will have elevations relative to the original DEM. If you run the fill function, the sinks w/in the stream bottoms will be filled to make the flow direction correct.

If you are trying to lower the known channels that don't show well in the DEM by just a small amount to get a good elevation grid, you can burn the streams to some deep depth and process the DEM one more time using a CON() function where if the elevation is below below 0.00 (negative), then make the elevation 0 (or whatever elevation you need). The burned channel bottoms will all be at the same elevation. You may need to check the flow direction to make sure after you create the flow direction grid that the flows are going in the right direction. There is a tool to correct the flow direction using polylines drawn in the direction of the flow you need.

[AstoriaTershy]

Hi Mark, 

Thanks very much. This is quite helpful! Using my original 10 x 10 m DEM instead of resampling to 1x1 m saved a ton of processing time and hard drive space.

To clarify my understanding-- Since my project is in meters, then 2 Z units should equal 2 meters?  My results support this notion! 🙂 Below is my original DEM "DEM_Nooksack" and my DEM Reconditioning tool output "DEM_AGREE". 

 
Follow-up question: I already filled sinks in the original DEM before using it as input for the DEM Reconditioning tool. Do I need to fill sinks again? My best guess is no. I'm studying a coastal watershed with lots of low-lying land, so it seems accurate to me that some of the agricultural canals that I burned would be changed from 0 to -2 m elevation.

Thanks again for your help!