A question of Habitat Quality model

[hongkongdoll]

I have saw some researchs of using InVEST habitat quality model
And I found that some researcher put same land-use type in two table(threat table&sensitive table) such as corpland etc…
Which means any single pixel in raster can regard as threat source and habitat.
Whether the threat source can be used as a habitat.
If yes, Is there anything to pay attention to when setting parameters？

[dcdenu4]

Hi @hongkongdoll,

If I’m understanding your question currently, yes, you could set the model up so that a land cover type of agriculture was both a threat and a habitat. That is, you could have a threat raster which represents agriculture on the landscape and in the sensitivity table, have that land cover class be marked as habitat.

I think it’s important to focus on what species or subject you are trying to model for though and determine if it makes sense that a habitat should also be considered a threat, and alternatively whether a threat should also be considered as habitat for that species.

Also, I believe the model will treat locations of threats on the landscape as completely degraded. So I’m not sure this would even have much an effect in the model. I’d have to check on that. But in most cases I don’t think it makes much sense to consider a threat as habitat.

What might be interesting is trying out a few different scenarios. Running the model where threats do not represent habitat and then running the model when they do.

Cheers,

Doug

[hongkongdoll]

Thanks for reply.
Yeah, that’s why I confuse.
Because if a landuse(for example, cropland) both regard as threat and habitat. That means the “habitat cropland” and “threat cropland” distance are 0. In some condition(like exponential decay) maybe the threat will be overestimated.

To avoid this problem, I saw some researchs let the “Sensitivity of habitat to threats” in 0. That means in same landuse, threat also have “Threat distance” and “Relative impact” but the same habitat have insensitive to this threat. I don’t know if it can solve this problem. Because if we do this, the only factor that can determine the habitat quality of the site is the initial habitat suitability.

I don’t know if my understanding above is wrong.
If there is a problem with this, how can I adjust?

best
Chauncy

[swolny]

Hi @hongkongdoll -

I’m not sure that it makes a lot of sense to set up this model such that a land use type is both threat and habitat. If you’re modeling a particular species, or group of species with the same requirements, then a land use type would be either habitat or a threat, but not really both. It may be habitat that’s not preferred (so gets a lower “habitat” score in the sensitivity table), but that’s different than being a threat.

In reality, I’m sure that there situations where a land use type can be both. For example, a species that will use agricultural fields (so they are habitat), but if the crops are sprayed with pesticides then they are also a threat. But that is a complexity that I don’t think that this simple model can handle clearly.

I’d be curious to know more about the situation that you want to model.

~ Stacie

[hongkongdoll]

Thank you for your reply.

My research did not specifically assess habitat quality for a particular species, but rather evaluated changes in habitat quality within the study area over a temporal scale due to land use changes. The image represents a portion of my study area, where I categorized habitat quality scores into four classes at intervals of 0.25. In this image, the main threats to habitat quality are urban land and agricultural land, both of which are not considered as habitats. If I were to include agricultural land as habitat, the habitat quality scores within the range would be underestimated.

I’m unsure which approach would be more appropriate.