This forum is shutting down! Please post new discussions at

Question concerning InVEST sample data for habitat quality model - half saturation constant again

Dear InVEST team,

I studied the manual concerning the habitat quality model and I am about to run the model (Version 3.1.3.) with my own data.

Before doing so, I run the model with the provided sample data (only current land cover) but I need some help with the k-factor. The forum was already a bit of a help (the two posts "Time in Habitat Quality Model" and "Habitat Quality degradation and analyzing output") but still, I don't fully understand how to proceed.

In the manual and in the forums it is always stated that you should run the model once to determine the best k value. Does it mean I run the model first with k=0.5? I was not able to run the degradation part separately (without calculating the final step for habitat quality), so I need a k value for the first run?!

When I did the model run with the sample data (and k=0.5) it resulted in degradation values between 0 and 0.110790881, which seem to be quite low. Should I now rerun the model with k=0.055 (half of 0.11) before I can interpret the habitat quality scores (which are now extremely low, ranging from 0 to 0.111227)? Doing so resulted in degradation values between roughly 0.6 and 1, but the habitat quality scores did not change at all, they still range from 0 to 0.111227!?

The manual says for the half-saturation constant (in the section data needs): "It is important to note that the rank order of grid cells on the habiat quality metric is invariant to your choice of k". But as far as I understand that doesn't mean that the absolute values are the same?

Is there anything that I got fundamentally wrong?

I appreciate your help very much,




  • DougDoug Administrator, NatCap Staff
    Hi Inga,

    I'm one of the software developers behind the model. The approach you took sounds correct to me. The half saturation constant 'K' should have no impact on the degradation score. Therefore running the model through once and taking half the degradation score to use as the half saturation constant going forward is the right approach.

    If your degradation score is changing as half saturation value changes then we might have a bug. We have made some recent changes to this model also. Do you think you could download and run the latest and greatest? 

    Sorry I can't answer your questions completely. Hopefully someone from our science team can pitch in also.
  • Hi Doug,

    thanks for the quick response.

    Before I rerun the sample data with the newest release, just some clarifications, I am afraid I messed things up in my previous question, sorry for that!

    The formula Dxj in the manual calculates the total threat level and is independent from k. The corresponding output from the model is the raster deg_sum_out_c.
    Do you call this the degradation score?! Than it was a misunderstanding , these values didn't change, they range from 0-0,111227.

    The formula Qxj calculates habitat quality, ranging from 0 to 1 (0 being no habitat, 1 being excellent habitat). It depends on the half saturation constant and corresponds to raster quality_out_c.
    And yes, it worked out properly, the results from the first run (with k=0.5) show values which are all close to one, so the map shows excellent habitat except for the roads, which are non-habitat - so only little variation. The results for the second run (with k=0.055) show a more differentiated picture, with more graduation between no and excellent habitat.

    So my problem is solved, sorry again for the confusion, and thanks for your help which served well as eye-opener!
    I will run the model now with my own data.

  • DougDoug Administrator, NatCap Staff
    Glad to hear it Inga! I always enjoy when things work out like this. Let us know if you have any other questions.
This discussion has been closed.