I downloaded the coastal risk reduction thresholded raster (which has either 0 or 1 values), but to visualize it properly in QGIS I need the continuous (non-threshold) file. Do you know how I can obtain it: is it available for download, or should it be generated through the GitHub ? In case of this latter option, could you please indicate me how to do it ?
The nature access data represents the number of people within 1 hour travel from natural areas with spatial resolution around 300 m. However, visualizing the raster file in QGIS as singleband pseudocolor, I see way too high values per cell (until 30 milion) and concentrated only in large cities (where inhabitants are more but probably natural areas are much less). How should I interpret the data ? Am I visualizing it right ?
I see what you mean about the coastal risk layer. It’s been a while since i worked with this, but i think we should be providing continuous values, not just 0/1. Let me look into it.
As for nature access, at first it seems counter-intuitive that high values would be within cities themselves, one would think that the natural pixels around cities would have the high values, and values would be 0 in urban pixels. But here’s how the data authors describe their approach: “…instead of identifying places within access of residential areas, counting people within access of every “natural” pixel” (Supplement to Chaplin-Kramer 2023). So urban areas themselves will show the highest value, not the “natural” pixels that the most people have access to.
@simone.guazzotti I spoke with a colleague who also worked on this analysis, and provided more information about the nature access layer, within the context of the footprinting tool that these data are associated with.
For the footprinting tool, we consider “potential natural vegetation” everywhere, so it’s really just a measure of how many people are around a given pixel, including cities. I would expect the urban areas to have the highest values – losing nature in cities would reduce access for the most people. The interpretation for the footprinting tool is not meant to be who has the highest access, but where does development have the greatest impact on the most people.
Hopefully this helps with interpretation of Nature Access. I’ll also recommend reading the supplement to the published paper on the footprinting tool, which details the “potential natural vegetation” that was used, and how it was mapped to urban areas.
@simone.guazzotti i have reviewed the coastal risk raster, and it looks like i posted the wrong layer, which had been reclassified such that pixels > 0 had been set to 1 - sorry about that, and thanks for catching it. So i’ve updated the downloadable data with the correct layer, which has continuous values. You can download it here.
Of note, it seems to have minimum values that are not exactly 0, but a very small negative value (e-10) that is equivalent to numeric noise. I have tried to set all values < 0 to 0 using both ArcGIS and QGIS, and neither of them worked, they both returned a raster with the e-10 values. This might not affect you, but if it does, maybe you’ll have better luck with the reclassification. We are providing these as a proof of concept, not as final decision-making data, so i’m not concerned enough about it to spend more time on it.
If you have any problems with this layer, let me know.
Thank you vert much Stacie for all the information and data you provided. It was very useful.
For the coastal risk reduction data, I read that is basically the product of an unitless index of the coastal risk reduced by the ecosystem and the number of people within the area protected by the ecosystem… but is there a physical scale reference or interpretation of values ? Is it available a description of how was made the model ?
For example for Europe it goes from 0 untill 600000. Can I assume that if the value is close to 0 then there is no ecosystem and/or there is no coastal flooding risk ? If the values are high, then probably it close to a city, but could I also assume that the risk of flooding is significant ?
Then could you please attach again the reading you mentioned about nature access ?
The output values for the Coastal Vulnerability model are all relative, and do not have a physical scale. A value of zero in the “coastal risk reduction” layer for the footprinting tool means that a point has the lowest combination of coastal risk reduction from nearby ecosystems and number of people benefiting from that risk reduction. It should not be interpreted as there being no flooding risk. It may be that there are no ecosystems nearby. Or it may be that there are ecosystems nearby, but there are no people living near the ecosystems to benefit.
The metadata document provided with the coastal risk data references two studies by Chaplin-Kramer et al that these data were built from. You’ll need to dig around in the supplements to those papers to see their particular methods. The 2023 supplement (attached) says:
The contribution of coastal habitats (such as coral reefs, mangroves, seagrass) to mitigating that risk is considered in terms of the difference in coastal risk with and without that habitat present
This difference in coastal risk due to habitat is calculated by the InVEST Coastal Vulnerability model (the “habitat_role” value), which you can read more about in our User Guide. At least I think that’s the value they are describing, it would be nice if they would say so explicitly.
For the record, here is the DOI for the open access paper that these footprinting tool data were created for. You should be able to access both the paper and the supplement at that link, but i’ve also attached the supplement here.
Hm, it is not successfully uploading the Chaplin-Kramer 2023 supplement. But you should be able to get it at the DOI for the paper, since the paper is open access.