Delineate putting streams on mountaintops

Loving the InVEST modelling suite.

I have a hydro-enforced 1-second DEM for NSW (Australia). I know where my streams are, but somehow, DelieateIt seems to think the render is perhaps inverted, and the streams end up along the ridgetops. It’s quite amusing. But also, wrong.

InVEST-natcap.invest.delineateit.delineateit-log-2022-07-28–22_59_30.txt (3.0 KB)

I have a link to 2 screenshots imgur link taken in QGIS, one which shows the DEM and the scale bar (and the location of the streams overlaid as a blue dashed line, and the other showing the streams.tif file outputted by Delineateit.

Have I had a brainfart? Can a guru out there help me get the streams on the valley floors, like physics suggests they ought to be, in exchange for eternal gratitude?

Log file

InVEST version


Thanks for posting @gismat , can you share the dem2.tif referenced in this logfile so we can take a closer look? Certainly something is a bit off here!

dem2.tif (1.5 MB)

Hi Dave, I surely can, here it is: the exact file.

Thanks for sharing that @gismat . This looks like a 3-band RGB raster. That means that the pixel values in each band represent levels of red, green, and blue color on a scale of 0-255, rather than representing
actual elevation values. That’s useful for visualization, but not so much for hydrological routing. DelineateIt expects the real elevation values to be in the first band of the raster.

Does the original source of your DEM meet this requirement?


My goodness, I didn’t realise. And I don’t know the answer. Looks like I am off to learn more about rasters and DEMs (I assume the source file meets this requirement - it is NASA data).

Time to find out! thank you Dave.

update (later that same morning):
I had imported the original DEM.tif to QGIS, and then gone to the trouble of exporting the DEM layer as a new file dem2.tif to use in DelineateIt. I can’t explain the reason behind my decision to do that. When I reverted to using the original DEM.tif file the DelineateIt model worked as expected. Thank you for troubleshooting this with me. Attached here is the Original DEM.tif file in case someone in the future might learn from my example.

Next for me to learn: how to analyse the bands in a .tif file so I can know when I make that mistake again.

Thanks again,

Hydro_Enforced_1_Second_DEM.tif (1.5 MB)

1 Like

This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.