|Abstract||One of the significant benefits of the tiered Ontario Source Water Protection water budget approach was the opportunity for significant improvement in numerical model analysis at each progressive level.
The concurrent improvements in water use data, advances in computing and storage, and the release of a practical, open-source integrated surface water/groundwater model in 2008 (USGS GSFLOW) further supported the technical advancements. Most
important, however, was the recognition that a holistic "one water" approach, addressing the entire hydrologic cycle, was necessary to address the cumulative effects of increased water use, drought, storage, and land use change on groundwater levels,
streamflow, and wetland viability.|
Recognizing this challenge and opportunity, Earthfx strongly advocated conducting fully-integrated surface water and groundwater modelling studies for all the Tier 3 studies. A number of common response patterns
and insights emerged from the six fully integrated Tier 3 and Lake Simcoe Protection models that we created. First, we found that groundwater feedback (Dunnian rejected recharge) was the dominant form of interaction, occurring in as much as 30
percent of the watershed areas. Hortonian runoff was found to be relatively rare, due to the infrequency of intense storms, summer ET deficits, and actively-vegetated loose soil conditions.
Fully represented headwater streams and springs, high
resolution surface topography, and detailed land cover were needed to represent spatially variable and often highly-focussed recharge. The need for detail extended into the conceptualization of the shallow subsurface, where detailed representations
of the soil zone and shallow geology were needed to properly simulate subsurface stormflow and seasonal flow through highly permeable shallow aquifer units (weathered tills, epi-karst, etc.). Detailed representations of reservoir operations, quarry
dewatering, irrigation water takings, and return flow were also found to be important to simulate overall watershed functions and, ultimately, producing a defensible risk assessment. Based on this experience and insight gained, we are convinced that
the key to successful integrated modelling is in the details.
Perhaps the most significant conclusion is that practical, engineering-scale integrated analysis can be accomplished within a watershed context. At too large a scale, many of the key
process details and complex shallow system interactions would be oversimplified and generalized. Similarly, at too small a scale, such as limiting the model to the extended area of influence of a wellfield, would require oversimplification of model
boundaries and neglect of the transient nature of surface and groundwater flow in the surrounding area.
In 2010, Refsgaard et al. predicted that by 2020 all modelling in Denmark would consist of fully integrated analysis. Perhaps, due to the
challenges and opportunities of the Tier 3 process, that future has arrived early in Ontario.