This forum is shutting down! Please post new discussions at

Problems with Explanation of the Biodiversity (Habitat Quality) and Habitat Risk Models

pcutterpcutter Member
edited April 2014 in Terrestrial Models
I want to flag some potentially misleading issues in the way these 2 models are presented.

Usage of and relationship between the terms biodiversity and habitat.
The "Habitat Quality" Model (confusingly called the "Biodiversity" model in certain parts of the NatCap site) states that the model "is most relevant to “coarse filter”, or habitat-based approaches" but then specifies that it is Hall et al.'s (1997) single species-driven definition of habitat that describes the central logic of the model. This seems contradictory. Other language in the explanation seems more consistent with the use of the term "habitat" to indicate a certain recognizable vegetation community or biogeophysical configuration (e.g. seagrasses, kelp forests, mangroves, reefs--(all examples of habitats provided in the explanation of the Habitat Risk model)) (let's call this the "generalized" definition of habitat.

1) Use a single, consistent name for the Habitat Quality / Biodiversity model. Strongly suggest "Habitat Quality" for reasons discussed below.
2) Remove the specification of the Hall et al. species-specific definition of habitat from the explanation of the Habitat Quality model as all other discussion and indeed the model itself seem to indicate the generalized definition of habitat.

The link between habitat, biodiversity, and ecosystem services.
The Habitat Quality (HQ) model description references the interest of investigators in better understanding the congruence of patterns of biodiversity with patterns of ecosystem services. Insomuch as our selected "habitat" actually does represent the needs of a species of interest or of broader patterns of biodiversity, this utility is clear.

However, the HQ model description also references the importance of biodiversity to the production of ecosystem services. This part of the discussion seems to affirm the following :
a) that the habitat we have selected to work with in the model represents biodiversity in a broader sense; and
b) that this biodiversity is meaningfully related to the delivery of ecosystem services (i.e. this model can provide insight into the differential provision of ES's under different habitat conditions).

Both of these affirmations imply a number of assumptions that are not made explicit in the explanation.

Affirmation a) implies that whatever "habitat" (unclear already what is meant by this term (see above)) we have chosen to look at is representative of broader biodiversity. However, without clear guidance as to what "habitats" might in fact be more or less representative of some portion of biodiversity, there is no guarantee that what has been selected has much value at all as a surrogate for broader patterns of biodiversity.

Affirmation b) implies that there is a meaningful relationship between the biodiversity represented by the habitat that we have selected and production of various ES's. There are several issues here: i) this relationship is poorly understood in general; ii) it is certainly not quantified in any way by this model and iii) it is therefore misleading to users to imply that the model tells us something about the delivery of ES's in general or which specific areas relate to this delivery and how.

While the explanation of the Habitat Risk model does not explicitly state a utility related to assessing how habitat relates to ES provision (or, in the case of this model, how ES might be put at risk), references and parallels drawn between the HR and HQ models imply a similar logic.

1) Either provide significantly more explanation of how these model relates to ES production and how they might be used in this context or (better) remove the reference to this relationship.

Overall "Placement" of these models
Given that both models are not meant to model patterns of ES provision, it is misleading to some to embed them, without more explicit explanation of their fundamental distinction, in lists of models that are immediately relevant to ES provision and valuation.

1) Clearly list these models under a unique heading to make more clear that their central utility is to provide practitioners with information on distribution of species habitat or generalized habitat types (e.g. distinct from ES) but of potential interest in assessing congruity of species habitat (or perhaps biodiversity (but see concerns about this above)) with patterns of distribution.

Thanks in advance for any perspectives / ideas.


  • RichRich Administrator, NatCap Staff
    Wow, thank you for this critique. As you might guess, we've been having internal discussions on how to handle this model for some time, plus it looks like there's some historical baggage about the habitat quality vs. biodiversity that you pointed out. We're taking all this into consideration for a possible update in the next InVEST release.
  • Hi Rich,

    Thanks for this response. Glad to hear this will get some attention in the next release.

    My sense is that, if indeed this model is to mostly address questions about habitat quality and risk of decline in that quality, it may be valuable (perhaps in addition to this offering in the InVEST suite) to provide more substantial references and links to some of the many available modeling approaches/tools that have extensive peer-reviewed conceptual and user experience bases.
  • AhmadAhmad Member
    dear NatCappers,

    I HAVE ALREADY POSTED THIS IN "MARINE" DIVISON OF THE FORUM, but since I am dealing with HRA in the terrestrial ecolosystems, I repost it here to be more relevant.

    could you please help me to start on with HRA model.
    I have all the data requried, shapefiles, habitats and stressors. But I can not run the model, particularly, can not understand the process how the modelfulfills its function .

    Imagine, I have a country, through which all the way from east to west there is a pipeline crossing. I would guess this is the stressor. And I have a habitat which it crosses. The pipeline has already been on place for many years, I want to demonstrate risk that it carried/s to the habitat. Since the topic of "Risk" is in general is very popular in Oil-Gas industries, I wanted to apply this model.Questions:

    1) Does it make sense to use HRA for such type of scenario ? Because it is already somehow clear that the habitat which was crossed by the pipeline, has been damaged, and risk has been already in place for many years. Could you please suggest what would be a right question to ask when applying this model ? I would still be pro to utilize the model, but I should be clear in my statements to show a personnel why the model would be useful.

    2) If it is still recommended to run the model, could you please advise me, how to run it.

    I opened a preprocessor, where I crossed "calculate risk to habitat" and "use spatially explicit risk score". Directory for the stressor Layers is automatically directed to the "HabitatLayers" folder. The folder is empty. If it should not be empty, then what kind data should be there, what is the format, size, restrictions? And rest I left as it is, nothing more i redacted. Other rows, "exposure", "sensitivity" and "resilience" are left as they are. I do not clearly understand what are their role, if there is no baseline map was selected, how do they apply if there is nothing yet to be applicable for them? Afterall, I ran the preprocessor. As a result, in the C:\InVEST_3_0_0_x86, I opened HabitatRiskAssess folder, where I have got new folders, such as "habitat_stressor_ratings" and "tmp". In the first folder, there is one file, which is named "stressor_buffers.csv", and which is empty (only in the head its written STRESSOR NAME, and STRESSOR BUFFER (meters). In the guideline, it says that also "habitatname_ratings.csv" file should appear (which did not), as well as that "there will be one file for every habitat, and the one additional file for stressor buffers" (there is no file and habitat; how does buffer function works ? which files represent buffer function?).

    3) Now, I guess I should run HRA model itself, and where all the problems are. First, in the row which is named "criterias scores CSV folder", where I should insert some data, I dont know what and where to insert from. Where is an CSV file where I can fill in the data? it did not appeared after preprocessor was ran.

    I guess there are lots of moments and gaps I missed, otherwise I cant get how the model funtions, how the expected maps could be produced if I am just running empty processors.

    4. The model itself comes together with InVEST when I install it. I click on it, and it opens. So my question, why would I need to use ArcGis for that ? when I usually need to combine InVEST with ArcGIS. Because InVEST interface seem simple enough to be attracted.

    I use GIS 10. When I try to add the InVEST toolbox to ArcCatalog, my GIS does not see InVEST toolbox. Moreover, I myself can not find the toolbox in the InVEST folder, it is not there. Should it be there or it is there only after I apply a scenario for any particular model ?

    I hope I was clear in my questions and issues I raised, and I hope I can get soon some clarifications from your side.

Sign In or Register to comment.