Interpreting habitat risk assessment outputs and using spatially explicit criteria

astudwelastudwel Member
edited January 2015 in Marine Models
Hi NatCappers,
I have two questions:
1) I'm trying to understand how to interpret the HRA model risk plot outputs. The Sub_Region_Averaged_results output gives tables with headers of habitat name, stressor name, E, C, Risk, and Risk%. The user guide states that 'risk' like the other headers is expressed as a portion of total potential risk. I initially have followed the defaults provided by the HRA model and am ranking stressors on a 1-3 scale. The 'Risk' category in my output table provides values per stressor ranging from 1.66 to 2.06. Do these numbers reflect risk on that same 1 to 3 scale? That being said, if I add all these risk numbers up and relate their sum to the risk sums from each of my future scenarios (all likely to be conservation options), would this be an effective way to represent total decline in risk to the habitat or ecosystem?

2) I am working with a habitat for which I have generated ranked values (of importance to multiplpe species) across a 1x1 km grid in ArcGIS. If I transfer these values from the grid in such a way that each grid cell becomes a polygon associated with a ranked value all within a single shapefile, can I use this as a Spatially Explicit Score that defines the importance of different areas of the habitat?

Thanks for your help!


  • Hi Anna,

    I'm glad you're using the HRA model and able to get outputs for your region.  In terms of your questions,

    1) The Risk column reports results for risk on the regional scale if you input a sub region shapefile or your whole AOI if you did not.  If you chose to use the default method of calculating risk, based on euclidean distance, then Risk is reported on that scale.  The maximum value using a 1-3 scale for E and C is 2.83.  You can calculate that by substituting 3 for E and 3 for C in the eq 3 in the user guide.  Yes, you can sum up the risk for for all the habitat and stressor risks and use that as a total estimate of risk.  You can also consider using the cumulative risk outputs which produce results for risk at the grid cell scale.  Another way we've compared scenarios, is to look at the area of habitat in categories of risk (say high, medium and low).  You can classify your results however you'd like to.  And then you can report how much area of habitat you have in those classifications in the present and future scenarios.

    2) I'm not exactly sure what you are trying to do here.  If you have classified different areas of habitat as more or less important, and if these habitat types might respond different to risk, you could considered dividing your habitat map into these different habitat classifications and running the model for several "different habitats."  This could be useful because it would tell which of those and where is most at risk.  If you don't think these habitat classifications respond different to metrics of risk, I'm wondering if you could just as easily just map your classifications onto your final outputs of risk and see where your risk hotspots overlap with the most important habitat for your bird species.

    Hope this helps!
  • Hi Katie,
    Yes, this helps. I don't think my habitat would be affected differently by stressors based on its importance. It seems like the spatially explicit criteria option is overall more useful for assigning different risk values spatially for stressors than for habitats.

    One other question--Is there a minimum number of exposure/consequence criteria that are required for the model to produce a reliable output? I seem to remember reading that a minimum of four criteria (in total) is required but now I can't find that reference.

     Thanks again!

  • Hi Anna,

    Apologies for the delay in responding to this question.  I just came upon it while searching through the forum.  Yes, we suggest using at least 4 criteria.  Please see Arkema et al. 2014 Environmental Research Letters and references within.

Sign In or Register to comment.