Habitat Quality model running slow

When I second time running this model,and it ran very slowly,just like the picture

Hi @Cindy2009, and welcome to the forum!

What is the resolution and extent of your input rasters? If the resolution is very small, and/or it covers a large area, it can take a long time to work. Also, what is the decay distance for your threats? If the distance is very large, that can also take a long time.

~ Stacie

1 Like

I’ll also add that the decay distance in the table should be kilometers (km) not meters. That has sometimes tripped people up.


I made the size was 2*2,and my study area was 700km2.May be I shoule make size bigger.
And,I would like to ask how to set the right decay distance ?
Thank you~

Then how should I make sure the decay distance is KM?

@Cindy2009 the “right” decay distance will depend on the threat you’re modeling, and local conditions related to the threat. So it’s not something that has one correct value, and it’s hard to give specific guidance.

This value is given in the max_dist column of the Threats Table - see the User Guide for more information.

Is 2 meters the original resolution of your spatial data? If you want/need to work with this resolution, the tradeoff might be that it takes a long time to run. If the original resolution is larger than 2 meters, I’d recommend making the pixels larger, since the model will run faster, without losing information. But first, it’s easy to make sure that the max_dist value is set correctly and see if that helps.

~ Stacie


According your suggestion,I choose different size,and modifie some number,it is usefull~I get it,thank you a lot~

Hello, I also have the same problem. My data resolution is 10m*10m and the decay distance as below, so what should I do? change the resolution? or others?
Thank you so much :grin:

Hi @jiaying.li and welcome to the forums!

There are a variety of reasons that the model might be running slowly for you, but at least your MAX_DIST column looks like it’s in km, so it’s not a units issue there! You mention that your data have pixels of 10m by 10m. What are the dimensions of your inputs, in number of pixels in X and Y? If your rasters are large, it’ll take longer to compute.

If you upload your logfile (even if the model run is in progress), we can take a look and confirm why it’s taking a long time.


1 Like

Hi, James, thanks for your reply. My research area is around 11983km2, the X and Y are also 10m. here is my logfile, thanks a lot for your patient help~

And one thing should be mentioned, the file of “filtered_XXX_XXX” is too large, almost 17GB one tiff file, I don’t know maybe that’s the problem.

InVEST-Habitat-Quality-log-2022-06-09–10_00_52.txt (11.3 KB)

Yep, that’s a lot of pixels. So given your research area and assuming your LULC is also using 10m by 10m pixels, then the model is attempting to convolve 119,830,000 pixels once for each threat. That’s just going to take a long time. The easiest way to speed up the habitat quality runtime is to reduce the resolution of your LULC. I would also suggest that you make sure your output workspace is on an internal solid-state hard drive … that makes a noticeable difference.



Thanks a lot! I tried to reduce my resolution to 20m*20m, the processing speed significantly increased. :grin:.


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