Output Analysis for Habitat Risk Assessment

Hi Everyone.

My name is Aryo. I have run the model successfully for HRA InVEST model. There was no error message. And everything looks fine in QGIS i guess because it can shows different unique values. The visualization of the output from this website (https://joannalcy.github.io/hra-dashboard/) looks fine as well.

However, I am quite surprised of my output from the visualization output because the results shows that my “From all Stressors” of my habitats indicates low exposure and low consequences, for example it indicates (Exposure=0.30 and Consequence= 0.00). I could not believe the value of exposure and consequence are very low.

The output results in QGIS also showed that all habitats are considered very good and good. there is no area with high risk. I expected some areas are in high risk due to agriculture and aquaculture.

So can anybody help me to analyze my results? Whether it is realistic or not. whether it is what is expected from this model or not.

Furthermore, i did run sensitivity test where I put different rating with the same data quality and weight (e.g: i put 2 for all data quality and weight). It shows different classification risk in different area.

For that purpose. Here I attached a zipHRA.zip (668.4 KB) file that consists of:

  1. my output
  2. my visualization files.
  3. consequence exposure criteria.

Thank you.

Have a great day :slight_smile:

Hi @Aryolejandro, I took a quick look at your results and here’s how I see things.

If I look in the Summary Statistics table I see Exposure and Consequence values between 0 and 3 for various habitat-stressor pairs. I’m not sure what the “From all Stressors” values represent. They seem to range from 0 - 2, so maybe they are an average of the individual habitat-stressor pairs rather than a cumulative value? The risk plots on the dashboard website are also a good way to find the instances where habitats have higher exposure & consequence values too.

Here I see risk scores of 0 and 1 (with very few 2s) for most of the habitats. I guess that matches what you describe. I would refer to User’s Guide to help interpret those final scores. http://releases.naturalcapitalproject.org/invest-userguide/latest/habitat_risk_assessment.html#output-folder

I’m not sure what you provided for the “stressor buffer distances”, but if the stressors and habitats don’t directly overlap, you could try adjusting those distances to increase the exposure of habitats to stressors.

Hope that’s helpful,

1 Like

Hi Dave

Thank you so much for your help. It really helps me. I really appreciate it

Stressor buffer distance is used to expand a given stressor’s influence, or footprint. I have increased the “stressors buffer distance” value and the results look the same. So, I concluded that my results are fine.

The next problem that i faced is from the GIS analysis. I could not interpret the results of “Reclass of Habitat” ; “Total risk of habitat” ; “Reclass of Ecosystem” ; “Total risk of ecosystem”. The website has explained its definition, however, i still have difficulties to write the analysis of my results.

I noticed that:

  1. the results of reclass of habitat are some areas are in high risk and some are in medium risk.
  2. the results of total risk of habitat is all area are in high risk
  3. the results of reclass of ecosystem are some areas are in low risk and some areas are in medium risk
  4. the results of total risk of ecosystem is all area is in medium risk.

So my question is why the results of “reclass of habitat” and “total of habitat” have more high risk areas compared to “reclass of ecosystem” and “total of ecosystem”.

For that purpose, to understand my question more clearly. Here I attached my result to have a look of different results of HRA in microsoft word. Discussion.zip (290.7 KB)

Thank you so much.
Have a great Tuesday :slight_smile:

I think this can be explained by the following description of TOTAL_RISK_Ecoystem from the User’s Guide:

“This raster layer depicts the sum of habitat cumulative risk scores divided by the number of habitats occurring in each cell. It is best interpreted as an average risk across all habitats in a grid cell.”

So a specific area can have coral at high risk, but the overall ecosystem at a lower risk because some of the other habitats present may have low risk. Does that make sense?

1 Like

Hi again Dave

Yess, that’s totally make sense. You gave me a new insight. You explained it in a sentences that is easy to understand. Thanks.

Then I realized, for instance, 2 habitats are in high risk in area “A” and another 1 habitat is in medium risk in the same area “A”. However, total ecosystem in that area “A” is in medium Risk. So I am still a bit confuse here.

Or is it me who still do not understand the different between Reclass Habitat ; total_habitat ; reclass ecosystem and total ecosystem. I know it is written in the User’s guide. However, i need more explanation.

So would you please explain it to me in a sentence that is easy to understand please? Maybe by giving an example like you did in the previous comment about the definition of Total Ecosystem.

Sorry if i ask many questions.
Thank you so much.

Hi @Aryolejandro, questions are good. I’ll try to give a useful response.

The numeric risk scores may be more meaningful and useful than the categories of “low”, “medium”, or “high”. Those categories are decided at the very end after all the calculations and they are a little bit arbitrary.

The “medium” label for the total ecosystem risk in your area “A” is not calculated directly from the “low”, “medium”, “high” categories of the individual habitats in area “A”. Instead, the “medium” total ecosystem risk at area “A” is based on the numeric risk score at A relative to the total possible ecosystem risk score.

From the User’s Guide: " Cells are classified as high risk if they have cumulative risk scores of 66%-100% of the total possible cumulative risk score. "

So if we pretend there is somewhere else an area “B” with 4 habitats all with high risk scores, then you can see how area “A” might not be in the highest risk category.