None Type Error SDR

Hi there,
I am getting the error “‘NoneType’ object is not subscriptable” and cannot decipher what this means. I had successful SDR model runs previously, but now can’t get past this. I did update DEM file, but not sure what could have gone wrong. I also see that the txt file says not enough space on device, I moved all files and workspace to an external hard drive so I am also not sure why it is saying this. The log is attached below. Any insight is appreciated!
Thanks,
Michelle
InVEST-Sediment-Delivery-Ratio-Model-(SDR)-log-2021-05-28–13_06_59.zip (18.2 KB)
InVEST-Sediment-Delivery-Ratio-Model-(SDR)-log-2021-05-28–13_06_59.txt (164.9 KB)

Hi @michelle.vanhilten ,

Thanks for your question. We’re sorry that you are having trouble. If you would please include your input layers in an attached zip, we would be better able to troubleshoot your issue.

I suspect that your Watersheds vector may be the problem. Please confirm that it has features and that its ws_id column includes a unique integer for each one (watershed).

Thanks,
Jesse

Hi Jesse,
Thanks for getting back to me. Some of the Zip files were too large to upload so they are in this drive folder. Let me know if you have any issues accessing. Thank you! SDR - Google Drive

Thanks for sharing your inputs @michelle.vanhilten. Would you please also upload your Biophysical Table to the same GDrive folder? I cannot try to re-create this issue without it. Please let me know when it’s there.

Meanwhile, the “waterproject4.shp” input layer looks a little odd. Is it correct that you only intend to include a single watershed (one feature)? I suggest cleaning up it’s attribute table by deleting these fields: “TNMID”, “METASOURCE”, “SOURCEDATA”, “SOURCEORIG”, “SOURCEFEAT”, and “GLOBALID”. I’m also curious if changing the ws_id value from 0 to 1 would make a difference.

Please keep me updated,
Jesse

Hi Jesse,
I’ve added the biophysical table to the drive. I agree the watershed file is a bit odd, I wasn’t able to edit the attribute table due to data rights but maybe I can export it and edit the tiff then. I’ve used this watershed file in SDR yesterday and it worked, I changed the DEM file today and that’s when I encountered issues, but perhaps it’s this new combination of everything. Thanks for looking into this! Let me know if you need anything else.
-Michelle

And yes I am interested in only the entire basin so I one watershed is correct.

I have been unable to reproduce the error you are receiving. In fact, my model completed successfully, so there is nothing wrong with your input layers.

Now I believe the issue is related to a lack of sufficient space on your device. Is the D: drive the external hard drive you are using? My model results totaled >3 GB when including all of the intermediate outputs, so make sure you have at least 3.5 to 4 GB available before starting the run. Do you have at least that much free space on your external drive? Your Willamette Valley AOI is quite large at ~30,000 sq. km (about the area of Belgium). Everything is resampled to the resolution of the DEM, which is 30 x 30m. This explains the large file sizes. If you do not have adequate storage space, try resampling the DEM to a coarser resolution, like 90 x 90m. What resolution was the DEM version that worked successfully?

Please let me know if this helps or not,
Jesse

1 Like

Hi Jesse,
I’m glad to hear you were able to run it. The D drive is the external hard drive so I will have to figure out why it’s having an issue with storage space on the computer, but I think that’s something I can work out. And I might try changing the resolution as you suggested, thank you! The DEM that worked previously was also 30m but I found out that there was an issue with the file after running it (successfully) and that a lot of data was missing, thus the file sizes were probably smaller. I’ll try it again taking into account storage as the issue and let you know if I have any issues!

Thank you so much for looking into this for me. Have a great weekend!

-Michelle

1 Like

I should also caution you against placing super high confidence in the absolute values of the “watershed_results_sdr_.shp” result without comparing with observations. In the absence of observation data, we recommend that model results should generally be viewed as relative to one another more so than relying on the absolute values. Therefore, you may want to consider splitting the Willamette watershed into smaller sub-watersheds so you could compare those to one another (once you work out the storage limitation issue). Otherwise, refer to the “usle.tif”, “sed_export.tif”, and “sed_retention.tif” raster outputs for relative comparisons throughout the valley.

-Jesse