Open datasets to simulate pressing global hydrological problems

A new data pub­li­ca­tion on 4TU.ResearchData helps to judge the risk of flood­ing across the globe due to cli­mate change.

Since 2011, researchers at the Fac­ul­ty of Geo­sciences of Utrecht Uni­ver­si­ty have been work­ing on a large-scale hydro­log­i­cal mod­el that pre­dicts the flow of pre­cip­i­ta­tion through the earth. With this glob­al Water Bal­ance Mod­el (PCR-GLOBWB), it becomes pos­si­ble to trace the flow of one droplet of rain through the dif­fer­ent lay­ers of soil and rock, until it reach­es the sea.

The mod­el, pro­duced by the hydrol­o­gy team at the Phys­i­cal Geog­ra­phy depart­ment, is based on data col­lect­ed in the last 120 years. The mod­el itself is a team effort of around 18 mem­bers, who have all con­tin­u­ous­ly devel­oped the mod­el over the last 10 years. This close and effec­tive coop­er­a­tion led to the cur­rent ver­sion of the mod­el (Sutanud­ja­ja et al., 2018). The mod­el is used by gov­ern­ments, NGO’s, researchers and pol­i­cy­mak­ers to do cli­mate sim­u­la­tions and to map the pos­si­ble impact of (the lack of) rain­fall.

In this arti­cle, we zoom in on this ground break­ing ini­tia­tive. We inter­view assis­tant pro­fes­sor, Niko Wan­ders, part of the devel­op­ment team, and Edwin Sutanud­ja­ja, one of the core devel­op­ers of the mod­el and the main sup­port to the model’s user com­mu­ni­ty.

Togeth­er with Pro­fes­sor Marc Bierkens and Assis­tant Pro­fes­sor Rens van Beek, they are coor­di­nat­ing the devel­op­ments on the PCR-GLOBWB.

Niko Wan­ders (right) & Edwin Sutanud­ja­ja (left)
(Pho­to: Anne­miek van der Kuil).

The dataset, in NetCDF for­mat, can be freely con­sult­ed through the 4TU.ResearchData OPeN­DAP serv­er. We asked Niko and Edwin about their efforts to make the data used to build this mod­el open­ly acces­si­ble and FAIR. 

The impact of the research

The Water Bal­ance Mod­el (PCR-GLOBWB) has a clear impact on pol­i­cy-mak­ing, cli­mate pre­dic­tions and oth­er researchers. The Dutch gov­ern­ment, the World Research Insti­tute and the Euro­pean Com­mis­sion are just a few exam­ples of those who prof­it from the pre­dic­tions and sim­u­la­tions made by the project team. 

For instance, the mod­el can help to pre­dict the chances and impact of flood­ing in spe­cif­ic areas under­go­ing tem­per­a­ture increas­es of 2 or 3 degrees. The mod­el can pro­duce fore­casts based on a wide vari­ety of vari­ables, such as tem­per­a­ture, evap­o­ra­tion rates, ele­va­tion, dis­charge and soil mois­ture. Since recent­ly, the mod­el can also account for vari­ables that deal with human inter­ven­tion, such as the impact of reser­voirs, irri­ga­tion and live­stock water use. 

Niko under­lines that the project team works very hard to main­tain the dataset, which is need­ed to run the mod­el.

“The dataset is a real com­mu­ni­ty effort. There are a lot of researchers, some­times also from oth­er uni­ver­si­ties, that con­tribute to it. We orga­nize month­ly meet­ings to stream­line every­thing. In these meet­ings we decide on the way for­ward, for instance on new attrib­ut­es and devel­op­ments, and we make agree­ments on the funds we would like to apply for. Only with this data, the mod­el itself can be adjust­ed to answer local ques­tions.“

The pub­lished dataset con­sists of data that are col­lect­ed by researchers all across the world. Niko explains, “We aggre­gate data that we receive from all over the globe. This gives a unique view on how the rain moves through the sur­face, across the earth and towards the sea. The dataset real­ly keeps track of every loca­tion in the world, and this is the main rea­son that oth­er insti­tu­tions are so keen to use it.“

Edwin adds, “With the data, we can cre­ate month­ly fore­casts that pre­dict the flow of water for the com­ing six months. This helps to pre­dict upcom­ing chal­lenges, such as crop fail­ure and ener­gy prob­lems. To that end, we recent­ly also scaled up the glob­al res­o­lu­tion accu­ra­cy of our mod­el from 50 kilo­me­tres to 10 kilo­me­tres, so we can real­ly go into depth and get answers to very detailed ques­tions.”

NetCDF data and OPeNDAP

The com­plete dataset com­pris­es more than 14 TB of data in dif­fer­ent for­mats, but a sig­nif­i­cant por­tion of it is in the NetCDF for­mat (Net­work Com­mon Data Form), which is wide­ly used in cli­ma­tol­ogy and fore­cast­ing. NetCDF is a self-describ­ing, machine-inde­pen­dent data for­mat that sup­ports the cre­ation, access, and easy shar­ing of array-ori­ent­ed sci­en­tif­ic data. As this is an open for­mat, it can eas­i­ly be used and ana­lyzed with dif­fer­ent soft­ware-tools.

Researchers often con­tact Edwin to ask for the data under­pin­ning the Glob­al Water Bal­ance mod­el. Respond­ing to such demands is dif­fi­cult — it is impos­si­ble to send the data through email. Down­load­ing it from stan­dard repos­i­to­ries is also not an option, as the users would then need to down­load the full dataset on their per­son­al dri­ves.

The OPeN­DAP serv­er, offered and sup­port­ed by 4TU.ResearchData, pro­vides an excel­lent solu­tion. OPeN­DAP makes it pos­si­ble to eas­i­ly access and explore parts of the dataset, before decid­ing which part of the data to down­load.

OPeNDAP creates a lot of opportunities

Niko fur­ther under­lines the val­ue of OPeN­DAP.

“OPeN­DAP is eas­i­ly acces­si­ble, flex­i­ble, and it is a piece of cake to adjust and cus­tomize the area that you want to study. It makes very short and tar­get­ed sim­u­la­tions pos­si­ble.”

For now, Niko and Edwin have pub­lished about 250 GB of the com­plete dataset on OPen­DAP, and they intend to upload more data soon. 

Mak­ing their NetCDF data open­ly avail­able on the OPeN­DAP serv­er of 4TU.ResearchData has oth­er advan­tages. Niko states, “Shar­ing the dataset in a stan­dard­ized for­mat with researchers across the world is essen­tial to fur­ther increase its impact. By shar­ing our mod­el, new knowl­edge can be cre­at­ed upon it. Next to that, the pos­si­bil­i­ty of explor­ing the dataset in OPeN­DAP also opens up a broad­er com­mu­ni­ty of users, who can now use par­tic­u­lar parts of the dataset to answer their own region­al and local ques­tions, for instance focussing on devel­op­ing coun­tries.”

Edwin goes fur­ther to add, “There is a lot of inter­est in access­ing our data, which is why I am devel­op­ing a user man­u­al. It is good to let the researchers know how to use the data and how to nav­i­gate through the over­whelm­ing amount of vari­ables that we offer.”

The use of interactive maps

Despite the model’s suc­cess, Niko also iden­ti­fied an impor­tant chal­lenge when it comes to the research. “Mod­el­ling is not very sexy, so it does not bring in a lot of mon­ey from grants. If we would have the resources, I would def­i­nite­ly want to visu­al­ize the data. That would sim­ply be amaz­ing! Well-visu­al­ized cli­mate data is sim­ply used and analysed the most, as the researcher can then inter­pret the data much faster. The use of such inter­ac­tive maps would mas­sive­ly con­tribute to the usabil­i­ty and the impact of the dataset. And that is what we all do it for: pro­vid­ing answers to press­ing glob­al hydro­log­i­cal prob­lems!”

Link to the dataset on 4TU.ResearchData.
This inter­view was a coop­er­a­tion between 4TU.ResearchData and RDM Sup­port of Utrecht Uni­ver­si­ty.
Author: Daen Smits, Trainee Research Data Offi­cer, 4TU.Centre for Research Data | TU Delft Library.
Cov­er image by Roman Grac from Pix­abay.

References 

Sutanud­ja­ja, E. H., van Beek, R., Wan­ders, N., Wada, Y., Bosmans, J. H. C., Drost, N., van der Ent, R. J., de Graaf, I. E. M., Hoch, J. M., de Jong, K., Karssen­berg, D., López, López, P., Peßen­tein­er, S., Schmitz, O., Straats­ma, M. W., Van­name­tee, E., Wiss­er,   D., and Bierkens, M. F. P. (2018).  PCR-GLOBWB2: A 5 arcmin glob­al hydro­log­i­cal and water resources mod­el. Geosci. Mod­el Dev., 11, 2429–2453. DOI: 10.5194/gmd-11–2429-2018