A hydrologic time machine
MODELLING FRAMEWORK - The Canada1Water project uses three world-class modelling platforms (HGS, WRF and CLMv5) to produce long-term hydrologic predictions at the continental scale.
Once it’s up and running, Canada1Water will allow users to track the evolution of groundwater and surface water flows backward and forward in time — from 1979 through to the year 2100. Dr. Steve Frey provides an overview of the modelling framework driving the Canada1Water project.
The platform will ultimately have three distinct time periods: historical (1979 to 2020), mid-century (2045 to 2060) and end-of-century (2085 to 2100). The past will be simulated with both daily and monthly normal temporal frequency; the future time intervals will be monthly only.
“The advantage of a monthly view for the mid- and end-century simulations is that you can project changes in hydrologic cycles relative to historic conditions without needing to incorporate fine-grained daily weather, which is a highly uncertain endeavour,” explains Steve Frey, Senior Scientist at Aquanty and Canada1Water co-principal investigator.
Many inputs, one ‘clock’
Steve Frey is a Senior Scientist at Aquanty and co-principal investigator on Canada1Water, focused on pushing the limits of what fully integrated state-of-the-art groundwater–surface water models can do.
Because Canada1Water is built on multiple individual models, each has to be aggregated in a different way to arrive at the overall monthly periods. The platform’s regional climate model generates outputs in three-hour increments that will be aggregated into daily values. That daily data will undergo further post-processing so that climate-forcing data (liquid water inputs and potential evapotranspiration) in the end represents monthly average conditions.
The platform’s land surface modelling also generates outputs in three-hour increments. As those are being used to inform models of permafrost extent and active layer freeze/thaw cycles, they will likely be aggregated to five-day or weekly average conditions for the historic, mid-century and end-of-century time periods.
HydroGeoSphere (HGS) — the engine driving Canada1Water groundwater–surface water simulations — will also use daily transient climatological data from 2002 to 2020 as forcing for model calibration simulations. With that daily transient calibration, the HGS models will be evaluated for their ability to reproduce known hydrologic conditions associated with major weather events as well as recorded in-year and year-over-year variability.
Not just time but also space
Canada1Water also has two spatial resolutions, one low and one high. The low-resolution models present surface water features in an unstructured 3D finite element grid with nodal spacing from roughly 1,500 to 5,000 metres; the high-resolution models have nodal spacing from 1,000 to 3,000 metres.
While having two resolutions might seem redundant, Frey says the low-res models provide an experimental testbed for relatively fast simulations while the high-res models will have longer runtimes but provide output data with considerably higher spatial fidelity.
“I see Canada1Water as a fantastic demonstration of how leading-edge scientific tools and methodologies in the hands of a collaborative, world-renowned team can help tackle one of the biggest challenges facing Canada over the next century — identifying impacts and designing mitigation strategies for climate change,” says Frey.
“The advantage of a monthly view is that you can project changes in hydrologic cycles without needing to incorporate fine-grained daily weather, which is a highly uncertain endeavour.”