Hello everyone,
I would like to know if is there any dataset avaiable to run the Nutrient Delivery Ratio model as a test. I have searched for this specific model and did not found, so, I have some questions about the biophysical table content.
Is the source of the eff_x and crit_len_x obtained only by local research or it can be used default values? If it is possible to use default values, is there any example as showed in the user’s guide, but more complete for others types of land user/cover?
Look up Chaplin-Kramer et al. (2019) in Science. There is a reference in the appendix to ipbes-material (ref #40). However, it does only provide eff_p.
Thank you @proma ,
I could take these values in this reference. But the problem is that I can’t find examples of “p” values anywere, even inside the InVEST website plataform… So, usually people use diferent values of “n” to “p”?
Apologies for the missing NDR sample data, since we do provide it for other models. I’m actually in the process of redoing our freshwater sample data, and NDR will be included. In the meantime, here is the sample data that we’ll be providing, from a watershed in Kenya called Gura. NDR_sample_data.zip (797.4 KB)
Note that the values in the biophysical table were selected for this particular watershed, and are only intended to be used for learning the model, they may not apply to your area of interest.
Yes, we usually use different values for N and P. P loading often has lower values than N, but it really depends on what kind of agriculture or other situation is in your watershed. Finding values for phosphorus/nitrogen application in agricultural areas is relatively easy to find. Values for natural vegetation are much less available. As mentioned in another post, the InVEST web page now has a link to a nutrient parameter database, which contains a non-exhaustive list of literature, and the N/P values they provide. It might have studies from your area, or you can use it as a first draft to create more global average values.
No data set available to run the Nutrient Delivery Ratio model but you can get open source raster images such as rainfall and land use images through the websites. Biophysical data are derived from research paper within our context. But now we stopped to run NDR model in coastal area due to tidal influence and other marine activities.
Thank you so much guys… Specially you @swolny and @proma , those data from the world export of N and P data from the study in Africa gave me a direction about the entry values to use. I noticed that in the InVEST user’s guide there is an example csv, table to run NDR. Is it possible to use those values of eff_n and crit_len_n to make tests?
My region of study is in South of Brazil in areas of intense agricultural activity, but unfortunately we don’t have local data of the efficiency or length retention. So, I’d like to know if I can use those data as defalut.
Cheers,
Igor.
The NDR coefficient table that is in the User Guide is intended only to show the format of the table, not to be used for standard or default coefficient values for real projects. Of course, you may use those values to make tests. But then for more official versions, you would want to gather your own data values, and even if they are global average values from the NatCap literature database, keep a record of the source of each value.
It is hard to find efficiency and length values in particular, and I wish I could point you to great sources. Perhaps someone else has done work in Brazil who would be willing to share their information? The User Guide provides a few papers that address riparian buffers, and you can use those values to get started. And it says that if you don’t have data for crit_len_n, you can use the cell size as the default, and we have done that before (assuming that your cell size isn’t very tiny.)
I know that it’s hard to work with this kind of non-optimal data, because we want to get the best local values, and it’s frustrating that even relatively simple models can have inputs that are difficult to find. It’s a lot of time and effort too!
Thank you @swolny !
This answer makes a clearly undertanding about those input data for the biophysical table.
For the crit_len_n, I will use the cell size anyway, which is 30m, and keep searching for the retention efficiency coefficient, even in global data as you suggested before.
