Population Vulnerability Index

Hi guys,

Coastal Vulnerability model calculates the exposure index, which doesn't include Population data.
It also gives us a raster of population count along the coast.
I wanted to calculate population vulnerability index. So, I converted this population raster to points and then I joined it with the coastal exposure points.shp. Then I calculated PVI using field calculator. Before that I reclassified population raster into 5 ranks before converting to points. The formula I used was 
(Geomorphology*Relief*Natural_habitats*Wave_exposure*Surge_potential*Population)^(1/6)
Is this the correct method to incorporate population into exposure index.

Sorry if I am asking too many questions.

Thank you in advance.

Comments

  • I also want to calculate the effect of sea level rise for different scenarios.
    In InVEST Coastal Vulnerability model we can give sea level trend in mm/year.
    But if give the trend what does the output mean. Does it calculate the current sea level rise vulnerability.

    And there is another value we can give as input Mean Sea Level Datum.
    With this will I be able to calculate vulnerability at 1ft slr, 2ft slr ... 6ft slr scenarios.

    Again sorry for so many doubts.

    Please do clarify.  
  • Jess_SilverJess_Silver Member
    edited March 2
    Don't be sorry!  
    Happy to answer questions.

    Regarding Sea-level rise, there are a couple main ways to incorporate it into the model.  If you include SLR trends in mm/yr, the model is just looking at the relative difference in rates of SLR across your region, such that regions with relatively higher rates SLR will be more vulnerable than those with relatively lower rates of SLR.  It isn't calculating any information about absolute rise over a given timestep.  

    If there is no difference in RSLR across your region, or you don't have data on it but you still want to look at the effect of rising sea levels in the future, you can create a future scenario where SLR is reflected in a simple additive way (i.e. the relative change will be from current to future SLR using the same blanket value for the entire region).  I can give you more details on how we've done this in the past if that would be helpful.

    Regarding population, people incorporate information about population (and other metrics of community vulnerability) in different ways.  Take a look a the papers I've included below (more details about in the Arkema et al. supplement), and also take a look at a coastal assessment conducted for Maryland that was done using the CV model.  These show some different examples of how people have combined, or in some cases kept separate, the vulnerability and exposure components.  Let me know if you want more details on this as well.

    Thanks,

    Jess
    Post edited by Jess_Silver on
  • Jess
    Thank you very much. I went through all the four papers.I guess it's okay to go with the formula of Population vulnerability I used. I thought it would be nice to use the population data in some meaningful way rather than just displaying population count map. And I am not going to use any parameters other than population count, as I don't have the resources. 

    Regarding SLR vulnerability. The work I need to do is to assess the CVI in future scenarios just like you did in your paper with Arkema. I get how you arrived at SLR values. But what I don't get is how you incorporated those scenarios using InVEST.

    I am planning on using IPCC scenarios to get predicted SLR values for 2100.

    As you suggested, I would like to have more details of how you've done this in the past.

    Thank you again.
  • Hi,

    In order to compare current and future SLR you will doing two scenarios with different SLR ranks. I've found it easiest to do this by doing just a little bit of manual work outside of InVEST.  You will have a coastal_exposure.csv in your model outputs which has all the ranks, the habitat role, the exposure index, etc.  If you import that into excel, you can just create two columns for current/future SLR and assign the ranks you want to.  Then using equation 1 from the UGuide (just the geometric mean), you can compute an exposure index for current and future SLR scenarios.

    Does that make sense? 

    Jess
  • Hello,
    Okay I'll assign rank 1 for current SLR and rank 5 for future SLR. After calculating Geometric mean I will be getting current and future CVI.
    But does that result have meaning? Like if SLR is by 1ft then this is the CVI, if SLR is 2ft then this is the CVI etc.
    Can you understand what I am trying to arrive at.
    Like in your paper you have mentioned that in 2100 for these scenarios (Predicted
    outcomes for the four future scenarios were global rise for 2100 predicted by
    the National Climate Assessment (0.2, 0.5, 1.2, 2 m; ref. 24), multiplied by a
    scaling factor (the ratio of the historical local rate to the historical global rate
    (1.8 mm yr−1
    ); refs 24,28).

    Hope I am not too confusing.

    can you clarify on that for me.

  • Hi,
    I have calculated CVI with current and future SLR values like you suggested.
    I have attached the excel file of the same for your reference.


  • Hello,
    Can you give me any furthur suggestions?
    Thank you.
  • Hi,

    I'm sorry - I was in the midst of writing my response and got distracted and totally forgot. 

    In the US study we had local tide gauge information, so we were able to produce spatially explicit estimated amounts of absolute SLR for different shoreline segments, and the assigned ranks were then taken from that distribution of values.  Where are you running the model?  Do you have any tide gauge information for your area?  Is it a large enough area that you'd expect spatial variability in RSLR?  

    I took a look at your calculations and they look fine.  This is similar to the example that we used in The Bahamas where we did not have spatially explicit tide gauge data so we were looking at the relative change between now and the future.  For that we used the curve of anticipated SLR between now and 2100 and we estimated where along this curve we would fall for a planning horizon of 2040.  In this case, current SLR was assigned a rank of 1, and future a rank of 2.  So, to get at your question of whether this approach has meaning, yes, it definitely has meaning in that you're looking at relative increase in right attributable to increased SLR.  Without having local tide gauge information you're not going to be able to assign ranks to specific absolute amounts of SLR so you won't be able to say 1ft vs. 2ft. But you can estimate relative changes using the method I describe above and reflect those in the ranks of the model.  Is that helpful?
  • Hi,
    No problem for late reply.
    I am doing my study for a state in South India. There are four tide gauges present in the study area and I have downloaded Data from Permanent Service for Mean Sea Level (PSMSL). So I will use it for this study, as the study area is large enough to have RSLR.

    Now I am clear about this.

    Thank you!

  • Great,  sorry it took so long to get this figured out - I should have asked earlier if you had spatial variability in the RSLR in your region.

    The model can accommodate that information directly in terms of local trends. Take a look at the user's guide for some guidance on how to create a shapefile for that input.  You should also look in the sample data for this model (CoastalProtection folder --> SeaLevRise_WCVI.shp) as well.  The tide gauges are points, and you'll have to create polygons from them.

    Let me know if you have additional questions,
    Jess
Sign In or Register to comment.