Hi @Nomsy , thanks for sharing your data again. I noticed a few things:
These vectors are huge & complex; they have 2.7 million polygons each. Many of them are overlapping each other, and many have invalid topologies (the warning messages from the invest log).
The Mangrove & Herbaceous layers appear to be identical to each other
The layers appear to have been converted from raster to vector. I can tell because they have a “gridcode” field in the attribute table which probably represents the value of the pixel from the original raster from that location. There are many different values in this gridcode column which probably means the polygons in these layers represent many more landcover types than just “mangrove” or “herbaceous”. To use them in the coastal vulnerability model, you would want to filter & export a layer that contains only the polygons representing a single habitat type, for each habitat you want to include.
I strongly recommend downloading a raster LULC layer from ESA and using that instead. You will need to do some reclassifying to create one individual raster for each habitat. For example, if you want a mangrove layer, reclassify all the LULC codes representing other landcover types to 0 or a Nodata value.
There are a variety of options for downloading ESA data: DATA | WORLDCOVER
I just tried to download the ESA data through the link you provided, and I discovered it’s the same data that I downloaded and used. I queried the data using ArcMap 10.7.1 to get the different habitats I used as my habitat in the study area.
Hmm, I’m not sure which data access method you used. In the “Data Download” section of the page I linked, I used option #1 and downloaded a tile using the webmap interface. I ended up with a geotiff. Their documentation indicates the data is distributed in geotiff format.
Okay, great, when you unzip the file do you find a geotiff with a name like ESA_WorldCover_10m_2021_v200_N03E003_Map.tif?
If so, that’s all you need. Perhaps the steps you did after that in ArcMap generated the shapefile? Instead of creating a shapefile, you will need to reclassify the ESA geotiff (using a GIS “reclassify” tool, perhaps more than once). Create one individual raster for each habitat. For example, if you want a mangrove layer, reclassify all the LULC codes representing other landcover types to 0 or a Nodata value.
I looked at the tile again, and the geotiff has a name like this "ESA_WorldCover_10m_2021_v200_N03E003_Map.tif`.
I attached the geotiff via an email sent to natural capital project notification email, so you could have a look and tell me what I am not doing right, please.
I found this youtube video which shows reclassification in ArcGIS: Raster Reclassify in ArcGIS Pro - YouTube. That might help as a resource for what you’re trying to do. There are MANY video walkthroughs of reclassifying in ArcGIS, some more helpful than others. Hopefully the one I shared helps!
It didn’t work. I got to the point where I reclassified all the LULC layer codes to 0, I made the necessary changes in my habitat folder and the model didn’t run. I tried to reclassify the layer codes to 0 and 1, still reclassifying now. Will come back with update on this later.
Yesterday, I also downloaded a mangrove geotiff online and it still didn’t run. This is log file below.
If all the pixels end up as 0 after reclassifying, then that is telling the model that no habitat exists anywhere.
This is correct.
The error in the log indicates a problem with the information in the habitats CSV file. Your CSV is referencing a file that does not exist: C:/Users/User/Desktop/All documents and folders/SampleData_CVModel/Habitats_RS\E10N10.shp
If you have a mangrove geotiff, then I would expect its filename to end in .tif not .shp
If you have further questions about the results, please consider starting a “New Topic” and please be as specific as possible about your questions, including describing what you have tried so far.
ESA_WorldCover_10m_2020_v100_N03E006_InputQuali…