How to create inputs/data for preparing the data catalog?

[Martin20494]

Hi team,

[This question was asked in hydromt and is the source of error of this question in this discussion]

I have been trying to find instructions on how to create data for the data catalog, but I can’t find them. Specifically, I’m unsure how to create the following datasets:

• hydro_lakes.gpkg
• grdc.csv and how many columns are required?
• Files in merit_hydro_ihu/merit_hydro_1k
• merit_hydro_index.gpkg
• rgi.gpkg
• rivers_lin2019_v1.gpkg
• Files in soilgrids
• modis_lai.nc

I’ve looked through the GitHub, documentation, and different versions, which mostly describe how to “prepare a data catalog,” which I understand based on the Piave river example. However, I’m missing guidance for the above datasets and would greatly appreciate any help.

Thank you very much.

Martin Nguyen

[shartgring]

Hi Martin,
Building a Wflow model consists of different layers which are generated via the setup methods. Like setup_basemaps, setup_rivers, etc. Each setup method links which data it needs.

So we probably generated some example data for the piave basin. For your own model, you need to prepare the data yourself. You can take a look at https://github.com/Deltares/hydromt/blob/main/data/catalogs/deltares_data/v1.1.0/data_catalog.yml for some inspiration.

Generally, you will need:

• elevation data and hydrography, using MERIT Hydro https://hydro.iis.u-tokyo.ac.jp/~yamadai/MERIT_Hydro/ see also https://deltares.github.io/hydromt_wflow/stable/_examples/prepare_ldd.html
• river data, see https://zenodo.org/records/3552776#.YVbOrppByUk
• lulc data, CORINE is nice for europe, otherwise something like ESA worldcover. Examples for mapping tables can be found here https://github.com/Deltares/hydromt_wflow/tree/main/hydromt_wflow/data/lulc
• soil data from soilgrids https://isric.org/explore/soilgrids note that you need to define the soilthickness map yourself (easy alternative is to create a layer for that and fix the depth to 2 meters everywhere)
• LAI can be implemented using MODIS https://modis.gsfc.nasa.gov/data/dataprod/mod15.php and should be a map with the LAI as variable, spatial dimensions (x,y or lot,lan) but also a time dimension for the LAI per month
• It is always possible to set fixed values using https://deltares.github.io/hydromt_wflow/stable/_generated/hydromt_wflow.WflowModel.setup_constant_pars.html so to start with only setup_basemaps and setup_rivers, use fixed parameters and see if that works, and then expand the model layer by layer

Do note that we are in the process of switching from version 0.8 (for Wflow verions < 1.) to version 1.0 (for Wflow versions 1.0 and later) which is quite different. So depending on which Wflow version you want to use, you should select the appropiate version of hydromt_wflow

[Martin20494]

Hi shartgring,

Thank you so much for your prompt response. I really appreciate it! I’ve been following your suggestions and have successfully completed many of the steps.

I would like to use my own dataset for setup_basemaps and setup_rivers. I have almost all the necessary data based on your instructions here, but except for the river width and bankfull discharge. I also tried the data from Lin et al. (2019) and MERIT Hydro, but they are either too coarse or miss some main rivers.

Do you know if these data are strictly necessary to generate inputs for Wflow? If so, do you have any suggestions for how I might generate them?

For reference, my versions are: hydromt is 0.10.1, hydromt_wflow is 0.8.0, and Wflow is 0.8.1

Thank you again for your guidance!

Martin Nguyen

[hboisgon]

Hi @Martin20494,

You can use any another DEM than MERIT Hydro for setup_basemaps, depending on what you have available.
To prepare the flow directions, we advise to use then pyflwdir functions as a pre-processing step and you can find some examples on how to use it in the notebook that @shartgring already shared with you: https://deltares.github.io/hydromt_wflow/stable/_examples/prepare_ldd.html
Preparing flow directions can be a little of an iterative depending on the quality of the DEM data and if you need to correct flow directions for example by bruning rivers in.

Same for the river network, you can use any data and not just Lin et al. The river data does need to contain river width information. If you do not have bankfull discharge, then bankfull depth can be used directly instead. Note that if you use the kinematic wave approach for river routing, the model should not be too sensible to bankfull depth so you could just start by using a constant everywhere.

[Martin20494]

Hi @hboisgon,

Thank you very much for your prompt and detailed answer. I have followed the instructions in the notebook that @shartgring and you have just shown me and I successfully have elevtn, basins, flwdir, hnd, lndslp, strord, and uparea just based on my own DEM (without using MERIT Hydro).

I noticed that in your example artifact_data, the merit_hydro contains additional variables - rivwdth and upgrid, which were not mentioned in the notebook. Hence, I am unsure how to generate these data. Similarly, the river_geom_fn also requires rivwth and qbankfull. I can now obtain qbankfull, but I still struggle with generating rivwth and upgrid.

Do you happen to have any suggestions to generate this data? Additionally, should I just need to provide the rivwth for the main river and no need for the tributaries like other branches along the river?

Thank you again for your guidance!

Martin Nguyen

[shartgring]

Hi Martin,

Perhaps also an interesting dataset to consider: https://agupubs.onlinelibrary.wiley.com/doi/full/10.1002/wrcr.20440 http://gaia.geosci.unc.edu/rivers/

[hboisgon]

Not all variables are needed. You can find in the documentation which variables are needed for each setup method:

• setup_basemaps
• setup_rivers

River width will be needed. The quality may influence your results and the accuracy of the modelled river discharge. You can always start with some default values for the tributaries and do some tests to see the influence in your own basin / case (do you want to predict discharge just at the outlet or main river or also in the tributaries)

[Martin20494]

Hi @shartgring and @hboisgon,

Thank you so much for your prompt replies. I really appreciate it. These suggestions are very helpful.

Martin Nguyen