Climate change scenarios

CarlaElliffCarlaElliff Member
edited October 2014 in Marine Models

I'm interested in using InVEST in a future project that involves climate change scenarios along the coastline (basically higher and lower sea level changes).
I've had a look at the online user's manual, but it seems only the Coastal Vulnerability model addresses this type of scenario.

Is it possible to use the Scenario Generator tool for this as well? Or is it meant only for land use changes?

I'd like to apply the CV model, as well as the Coastal Protection and Habitat Risk Assessment models.




  • Hi Carla,

    The scenario generator takes a categorical raster and generates a new categorical raster using rules that you specify for how categories should increase in area. Typically this means that you are using a land use land cover raster to create a new land use land cover raster using a set of rules that you specify about how those land use land cover types should increase in size. However, there is no reason that you could not use the scenario generator in a near shore or even ocean environment if that made sense given what you are trying to achieve. For example, you could use the scenario generator to increase the area of shrimp ponds based on their proximity to existing shrimp ponds, locations of suitable mangroves, and distance from transportation routes.

    In order to use the output from the scenario generator with the coastal vulnerability, coastal protection, or habitat risk assessment models you would likely need to convert the scenario raster into vectors. Continuing the example, you could convert the scenario raster into vectors and extract the new diminished locations for mangroves and use them as inputs to the models you listed to see how their results would change based on the scenario.

    Does that answer your questions?

  • Ok, I think I got it.
    So if I have in hand my plausible climate change scenarios I can then model how my coastline would look like with say a 1 m increase in sea level. I can then input this change of land cover (given that there would be less land and habitats) to the Scenario Generator and use its results for the other models (after converting the raster to vectors).

    How would that be different than simply creating a fictitious land polygon under the climate change conditions I'm working with and running this polygon within the CV, CP and HRA models? What does the Scenario Generator provide me that doing this other path does not?
  • Essentially the scenario generator takes a single raster and one or more shapefiles that are used with a set of rules that describe possible changes to the raster, which are then used to create a new raster in accordance with the rules that were defined.

    A land polygon would be a single area with presumably no details of how the land/seascape could vary over space, whereas a raster from the scenario generator would provide you with a possible land/seascape within the entire study area.

    If you created a polygon (likely from a DEM) that covered the area that would be inundated under your sea level rise scenario you could use that as one input to the scenario generator for a factor that influences changes to the land/seascape. It could be used as a factor for the location of sea grasses or other sea plants, or with some other processing as a factor for near shore plants.

    The rules that you would likely want to use would probably be more complex. You could add rules about proximity to existing types. You could add rules about one type overtaking another type. There is quite a bit of flexibility there.

    Not knowing much about what you are trying to do, I guess I would suggest you could use the scenario generator to create a more refined scenario for the CV, CP, and HRA models, where the scenario you run not only has the single variable of sea level, but several other variables like changes to the extent of coral reefs, sea grasses, etc that are determined in a methodical way using rules with the scenario generator that describe how their extents might change over the same time period. For example, you might alter the rules for the scenario generator to take different levels of growth or conservation practices that are in place.

    Does that help?
  • Hi Carla,

    Are you also asking, which InVEST models allow you to look at the influence of sea-level rise on ecosystem services? If so, you are right, coastal vulnerability is the easiest model to input scenarios for sea-level rise. One of the inputs to coastal vulnerability is SLR and you should create a polygon for each scenario that includes spatial variation in sea-level rise across the region for that scenario. Check out the user guide for instructions on creating those polygons. You can also include sea-level rise scenarios in the coastal protection model. But start running coastal vulnerability first. We have also used the habitat risk assessment to ask how sea-level rise increases risk to habitats. For the HRA model, again create a polygon representing sea-level rise as a stressor and include that along with the human activities that pose risk of habitat degradation. Note however, that the HRA model does not include adaptation to climate change by those habitats.

    We are also interested in assessing the influence of climate change on services using other models. For example, the amount of land and water are variables in the recreation model that we've found influence visitation in several of our projects. It might be interesting to think about whether you could convert your sea-level rise scenarios into polygons that depict the relative coverage of land versus water under different scenarios and use the recreation model to assess how climate change might influence tourism values in the future.

    Further if you know or can make predictions about how sea-level rise might influence change in habitats you could then feed those into any of the service models. In that case you're looking at the indirect influence of sea-level rise on services via affects on habitats rather than the direct influence on service as you did by using sea-level rise scenarios in coastal vulnerability.

    We'd love to hear about your creative ideas for using our models to explore impacts of climate change on services. This area needs more work so thanks!

  • Hi, Martin and Katie!

    Thank you, you've been very helpful!

    I'm still at the stage of delineating my project. I hope to use InVEST models along the coast of the state of Bahia, Brazil, (maybe the whole coast, maybe sections - still to be decided) for my PhD.
    During my MSc I used the CV and HRA models for two islands in the same region specifically regarding the ES of coral reefs - I presented this project at the NatCap Meeting back in march of this year. Now I intend to expand and include more habitats and services within the current scenario and a future SLR scenario.

    I didn't have much time to apply the recreation model for my MSc, but I remember that visitation data came from social media. Specifically for the region I was studying I don't think that this type of data would be a reliable proxy (despite the results by Woods et al 2013), considering the local context (I was dealing with natural pools in a place with no wi-fi, so I don't think many pictures could be taken and then posted to flickr...), but I really didn't try running the model, so I'm only assuming.
    With SLR, tourism will definitely be impacted in my study area, which is a very sun-sand-surf oriented industry. I'll read more on this model and see how it can be used in my case.

    Ok, so I think that using the scenario generator is indeed what best suits my objectives.
    Assessing the indirect influence of SLR on services as Katie said by predicting changes to habitats would also suit me - so I'll look in to this.

    You can count on me for coming to the forum every now and then and picking your brains, no problem!!

    Also, I received by e-mail the file I attached which might be of interest to you. It is about a funding opportunity for scientists in developing countries to earn funding from the United States Agency for International Development’s Partnership for Enhanced Engagement in Research (PEER) program. One of the themes they are most interested in is "using geospatial tools to investigate climate change".
    I know you have lots of partners in developing countries (Belize, Colombia...) so this might be useful. Also, if you'd like to expand your efforts to Brazil, you are most welcome of course!

    Thanks again!


Sign In or Register to comment.