0

I am trying to map hydrologic capture zones (HCZ). HCZ's are contributing areas upslope of any potential capture feature (i.e. road, pipeline path, railroad line), that would alter/impact the natural drainage pattern in a watershed. A simplified example is show here (yellow area is HCZ, which lies partially in basin- the pink area):

enter image description here

I am having a lot of trouble coming up with a good method to map areas upslope and downslope of roads. My current methodology is based on @nagib's post:

Upslope and downslope distance or area using ArcGIS Desktop

This method may have promise, but details in those steps are sparse, especially for the all important and complex second step- path distance allocation. When I tried to mimic nagib's method, I didn't get a very intuitive result. Not surprising, because I am not very familiar with calculating upslope/downslope using path distance allocation. The result looked like:

enter image description here

In mapping HCZs it seems to me that accurately defining upslope and downslope areas of roads is a critical first step. Any resources or thoughts on a very accurate methodology for getting these areas right? Using ArcMap 10.4 primarily, but have access to QGIS and Whitebox GAT.

EDIT:

Per @nagib's request here is a third pic showing second pic with contours (white lines):

Image with Contours

Also, here is a pic combining upslope/downslope areas, contours, pour point(orange point) at intersection of road(black line) and river(blue line), and the resulting subshed (white area) from that pour point:

enter image description here

  • What is upslope for one feature, might very well be a downslope for another one. – FelixIP Mar 24 '18 at 0:39
  • Can you show us the result of the first image to compare them? How big is the pixel size? – nagib Mar 25 '18 at 7:46
  • @FelixIP. True. Can you expand your thoughts on how this plays out for trying to mapping slopes relative to features that likely have changing slope themselves(in my case road features)? If it's possible to calculate slope along a road feature (which I assume is possible), would it not be possible to compare this feature relative to surrounding slope using elevation rasters? Or are you suggesting this is a futile effort? Thanks! – jbrengel Mar 26 '18 at 13:03
  • @nagib. Thanks! Unfortunately, I don't have spatial data behind the first pic. Just used to show simplest of examples. Think my GIS Admin has previously mapped HCZs with spatial data behind them though. Could ask him. Pixel size is about 27x27. Obviously, not ideal resolution. Before I spent the time creating another hydrological prepared DEM with better resolution figured I'd try this dataset. Can also ask my GIS admin if he has more accurate data. – jbrengel Mar 26 '18 at 13:14
  • Do you have contour map of second picture? We can only compare two images to view what is the problem. Try to interpolate that pixel size maybe 10x10 or 5x5m and maybe you would see differences – nagib Mar 26 '18 at 13:44
1

It's been over a month since any action on this, but I've gotten a fairly acceptable solution to capturing the HCZs using ArcMap 10.4. It requires a hydrologically prepared DEM and accurate stream and road vector data. The solution is:

1) Extract By Mask on the vector data representing your streams and roads.

2) Use Raster Calculator to square the extracted roads raster (ie expand the difference b/w stream and road rasters)

3) Mosaic to New Raster on to combine raster from Step 2 and stream raster from step 1

4) If need be, use Raster Calculator to set all 0 values to NoData on raster from step 3

5) Create a Flow Direction raster from DEM.

6) Run Watershed tool using flow direction raster and raster from Step 3 (or Step 4 if appropriate)

7) Use Reclassify on the watershed raster. The first range is the natural drainage, while all other ranges are the HCZs.

8) Use Raster to Polygon to convert raster to vector polygons.

Some positives: can be automated in arcpy, only requires a Basic license with spatial analyst, relatively quick runtime versus manual delineation.

Some negatives: can create memory issues causing ArcMap to crash/freeze, requires accurate input DEM and stream/road data, differs (10-15% and less) from manually delineated HCZ data we have. This difference seems to be rooted in the fact that the local natural topography (and therefore natural flow pattern) is overcome in areas next to artificially high roads.

Anyways, hope this is helpful to someone. Thanks for the creative suggestions. If anyone has ideas on how to overcome the negatives listed above (or can connect me with another smart person who might have some ideas), that'd be great!

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.