(uhi_results) Table with null values for some classes (?)

Dear all,
I have successfully used the “urban cooling” model, however, when consulting the attribute table (Qgis) linked to “uhi_results” vector, I noticed that class n°1 (trees) and class n°6 (white houses) indicate null values for “avg_cc; avg_tmp_v; and avg_temp_an”. All the other 6 classes on my map (Non-forest vegetation, bare soil, water, dark houses, asphalt, and shade) have values in that table. I don’t understand what could have happened.
I would be very grateful if you could help me.


Hi @jv_martinelli,
You can debug this by checking the intermediate results. Do the intermediate rasters cc.tif and t_air.tif have data? Does this data overlap your building classes 1 and 6?

Your building geometries must be large enough to overlap at least one pixel in order to avoid dividing by zero when calculating averages. How large are your geometries in class 1 and 6 relative to pixel size?

1 Like

Hi @esoth ,
Yes, the intermediate rasters mentioned have data, and they overlap classes 1 and 6. I’m afraid that at the moment I don’t know exactly what the sizes of classes 1 and 6 are in relation to the pixel (which is 4m), so I’ve attached a pdf with images of my results. Perhaps next I’ll be able to be more specific in the information requested. Thanks for your help!

InVEST urban cooling issues.pdf (2.0 MB)

I cannot tell from your PDF what is going on, but if you upload your complete datastack, I will try to reproduce the issue.

The files are larger than the space allowed for uploading. In fact, only the LULC map exceeds 8MB, the other files could be uploaded.

You can upload the datastack to google drive or another file hosting service and share it that way.

1 Like


Great, I hope you find the problem :slight_smile: Thanks again

@jv_martinelli it seems like you’re confusing your LULC raster classes with your AOI vector features. The numbers 1 and 6 in the attribute table refer to features of your AOI vector. These two features (circled below) are outside the area covered by your evapotranspiration raster, so no results can be calculated:

Your evapotranspiration raster should fully cover the AOI. Coverage is also spotty on the features on the right side, but results are still generated because there is at least some data. There is no reason to trim the raster to your AOI, it should extend beyond the AOI for best results.


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