Putting the pieces together – and making sure they fit
Canada’s provinces, territories and local governments collect groundwater and surface water data. But there’s little standardization, and hydrologic systems don’t obey political borders anyway. That’s why hydrogeologist Eric Kessel and his colleagues are working hard to bring the disparate datasets together in a harmonized view for Canada1Water.
“I’m extremely proud of the work we’re doing, because our foundational datasets will ultimately help answer important questions about water resources for the entire country.”
Eric Kessel is a hydrogeologist at Aquanty responsible for homogenizing Canada1Water’s provincial/territorial and watershed-scale datasets for use in hydrological models.
Kessel’s work has focused primarily on soils, which in Canada may be mineral, organic or made of peat, each with their own unique hydraulic properties — ways that water migrates through them. Since existing soil maps don’t cover all the various permutations, Kessel and his fellow Canada1Water team members have spent nearly 18 months bridging the gaps.
“Constructing and managing these kinds of datasets requires deep geospatial technical skills,” Kessel says. “I’m extremely proud of the work we’re doing, because our foundational datasets will ultimately help answer important questions about water resources for the entire country.”
Filling in the gaps
Existing machine learning soil maps detail the textures and other properties of mineral soils, and the Canada1Water team has processed these with established tools to derive hydraulic properties for soils across the country. Unfortunately, as Kessel explains, mineral soil maps don’t represent bedrock outcrops or organic and peat soils.
“Wetlands and peatlands account for about 12 percent of the total area of the country, and bedrock outcrops are another 14 percent,” Kessel says. “They all need to be accounted for because they store and pass water differently than mineral soils.”
To incorporate these important hydrological features, the team created new bedrock outcrop and 3D peatland maps and merged them with the existing soil maps. The bedrock outcrops were delineated using a combination of surficial geology and landcover maps, while the peatland was determined using vector peatland maps and machine learning-generated peatland products. Peat depth was estimated from organic soil contents within the peatland coverage.
“We vertically averaged the new soil maps and used them to assign values of porosity and saturated hydraulic conductivity to the modelled soil layers,” Kessel explains.
Going for granularity
DEPTH PERCEPTION - The Canada1Water team integrated custom maps with existing datasets to create accurate models of soil composition at different depths in every corner of the country.
Most global soil datasets have a resolution of about 1 square kilometre. Canada1Water’s soil models are much more refined at 250 square metres. Users will be able to ‘zoom in’ for a precise picture of soil properties at the watershed or municipal level.
“That telescopic functionality will become increasingly important as we begin to answer water resource questions all the way from the regional to the national scale,” says Kessel.
Each pixel of the gridded soil maps has sufficient detail to define water retention and unsaturated hydraulic conductivity curves, which will help reduce pre-processing time for many hydrological modellers.
A resource for the whole country
The soil map data assembly has taken more than a year to complete. When the Canada1Water initiative is completed, all the compiled data will be released to the public.
“I know the hard work we are doing today will continue to be used by future initiatives,” Kessel says. “I’m personally excited for the collaborations we’ll be able to foster with water resource researchers and policymakers across Canada.”
“[The platform’s] telescopic functionality will become increasingly important as we begin to answer water resource questions all the way from the regional to the national scale.”