Hello,
I am running the coastal-vulnerability model for the United Arab Emirates (UAE). The UAE only has a small part of the coastline that is exposed to the open seas and therefore the sample wind and wave data (one of the inputs to the model) does not cover the Arabian Gulf.
One option is to create the wind and wave data for the Arabian Gulf near the UAE from other sources, but before embarking on that endeavour, I wanted to test whether the model would work with test data. Find attached the image where I have quickly (and perhaps incorrectly) created a wind and wave point within the Gulf to test the model. You can see this point and its contents in the attached shp. When I run the model I get an error (see attached logfile). Can I get some support to troubleshoot my temporary approach? Then if there is an existing dataset that I can use to create the wind and wave data, that would be great.
Great idea to try this out with a temporary layer. In this case, the error is happening because the model has found a feature in the vector Wind_wave/WaveWatchIII_global_abu_reproj.shp that does not have a defined geometry. Could you verify that you have fully saved the vector, maybe even close down your GIS to make sure everything that should be written to disk has been written? If that doesn’t resolve the issue, could you check that all of your features do have a geometry defined?
Hi James,
Thank you for your response. I think it had something to do with the number of wind_wave points that I had within the project boundary. I only had one originally (I’m 99% sure it had a defined geometry). I created a few more points and it worked.
My next question has to do with the natural habitats that contribute to the exposure index. Are the habitats that contribute to the index located beyond the land mass boundary, or can they also be located within the land mass boundary? I note that we assign a distance of X m for each habitat type, but was wondering if the model was only looking X metres outside the land mass boundary for each habitat?
Cheers,
Reiss
Hi @ReissM , the model naively searches for any habitat present within an X meter buffer of each shore point, regardless of whether the habitat is located on land or offshore. So if you want to exclude inland habitats, for example, you would have to pre-process your habitat layers to clip/remove those features.
Hi Dave,
Thanks for your response. The reason I’m asking is that I would like to attribute the change in the vulnerability index to the ecosystems that are within the buffer zone, such that the contribution of each ecosystem asset can be quantified. For example, if there are two non-contiguous areas of seagrass in the buffer, and two non-contiguous areas of mangrove in the area, I would like to distribute the change in the vulnerability index to the four areas, e.g. if the change in vulnerability index is .4, then seagrass area 1 is attributed .05, seagrass area 2 is attributed .05, mangrove area 1 is attributed .2 and mangrove area 2 is attributed .1. Perhaps it’s as simple as distributing the change based on a weighting which is made up of the rank of each ecosystem type and the proportion of area for each asset within the buffer zone. Thoughts? It’s crude but gives us an idea - of course, noting that the limitations will need to be documented.
