Flooded road with yellow lines on a prairie landscape

Prairie Watershed Analytics

Prairie Watershed Analytics (PWA) is focused on providing Canadians in the Prairies with the information they need to understand their local watershed. Watershed models can support local decision-makers and businesses with evaluating ecosystem services and connecting with environmental incentive programs. Prairie Watershed Analytics is making this practice simple and accessible using open data and tools to automate and scale watershed modeling tailored to the Prairies.

Quick Project Facts:

  • Watershed managers use models to understand how water moves across the landscape so they can ensure the healthy provision of ecosystem services.
  • IISD is building a simplified watershed model for practitioners in the Prairies, scheduled for release in 2026.
  • The Prairie Watershed Analytics model will use machine learning to estimate flow and phosphorus concentrations for your selected river.

What is a watershed? 

Watersheds are an area of land where all the water—including rainfall, runoff, and snowmelt—flows and drains toward a common body of water, such as a lake or river. Watersheds are integral to our environmental and social well-being. Every province has a different way of protecting and supporting their regional watersheds. In Manitoba, there are 14 watershed districts (see FAQ’s below for more), and a Manitoba Association of Watersheds that collaborate with communities, farmers, landowners, NGOs, and local, provincial, federal, and Indigenous governments in the protection of watersheds.

Why might I need a watershed model?

Watershed management involves balancing the many uses and ecosystem services from a river’s headwaters to its outlet, across the seasons and over the years. This can mean working with Indigenous communities, land managers, industrial water users, residential and urban water users, fishers, and parks and conservation authorities. When balancing these different water uses and promoting best stewardship practices, it helps to have a clear understanding of how water moves across the landscape and contributes to these different ecosystem services. This can be especially challenging under a changing climate. This is the purpose of a watershed model.

Simplified modeling for watershed managers 

There are many watershed models and datasets available globally and across Canada. However, in the context of the Canadian Prairies, existing models and datasets tend to be inadequate to meeting the needs of local watershed managers. Available models are often:

  • Outdated
  • Inaccurate for the Prairies
  • Too complex to use
  • Paywalled

With the support of the RBC Foundation, the Prairie Watershed Analytics project is providing watershed managers with a tailor-made modeling product that can be easily employed on any watershed in the Prairies. 

Model design and features

The PWA model involves 2 major components: flow and water quality.

PWA rests on prior models to estimate runoff across the watershed, that is, the amount of precipitation water that runs over the surface of the landscape, rather than evaporating, seeping into the soil and plant roots, or remaining on the landscape as frozen snow. These runoff estimates have been previously generated using upstream data such as climate, land cover, and soil datasets. PWA takes these runoff estimates (surface runoff and snow melt) and calculates the amount of water stored on the landscape (in wetlands, ponds, and lakes) as well as the timing of the discharge through the river. This discharge modeling is done using a “routing model”.

Once we know where and when the water is flowing, the model outputs from phase 1 can be used to model water quality. Datasets related to phosphorus exports from the landscape, such as land cover, human population, crop cover, and livestock counts are combined with the data from the flow analysis. This combined dataset can then be used to train a machine learning model against the observed changes in phosphorus concentrations in the river. The results of this second modeling phase have not yet been validated.

Does IISD do this work alone? Of course not...

We are using data provided by the Lake Winnipeg Foundation, DataStream, the University of Waterloo, the University of Saskatchewan - Global Water Futures, and the Province of Manitoba. 

PWA is powered by the Raven modeling framework. We are also working with several governmental and non-governmental organizations at the local, provincial, and federal levels to better meet policy priorities and the needs of water management practitioners.

The image shows channels and ponding of water in the Manning Canal Watershed using PWA’s automated hydro-conditioning tool.
The above image shows channels and ponding of water in the Manning Canal Watershed using PWA’s automated hydro-conditioning tool. The tool relies on elevation data from the Government of Manitoba and stream data from the National Hydrological Network. 

Model development schedule 

The Prairie Watershed Analytics model is currently in development. Following a successful proof-of-concept on the Cypress River in fall of 2024, the first fully automated prototype of the PWA model will be released in the summer of 2025. Model performance will be tested and optimized over the following months. The launch date for the fully operational model is anticipated in the spring of 2026. 

FAQs about the Prairie Watershed Analytics project:

What is a Watershed District and why are they important?
Watershed Districts are independent organizations - typically governed by a board of directors - tasked with the promoting sustainable management practices for land and water resources. Districts can work with communities, landowners, and various levels of government (Municipal, Provincial, Federal, or Indigenous) to promote healthy soils, waters, and habitats.

Why are the adoption of BMPs important for producers, watershed districts, and communities? 
In the context of watershed stewardship, Beneficial Management Practices (BMPs) are agricultural or other land management practices designed to ensure the sustainability of those practices. BMPs like the planting of cover crops can allow landowners and producers to achieve long-term productivity (e.g. by improving soil health) while also providing co-benefits for downstream water users and for local wildlife. BMPs seek to improve ecosystem health (relative to conventional practices) while continuing the productive use of lands and waters.

What does it mean to ‘manipulate a watershed model’?
Watershed models take input data (for example, precipitation) and generate predictions about some new data (for example, phosphorus concentrations). In order to do this, models make many assumptions about the physical behavior of water and chemical species in the watershed. However, there is no “perfect model”, and it is often helpful (or necessary) to make adjustments to the model’s assumptions (model “parameters”) as well as to the datasets being provided, in order to improve the model’s performance. Finding the right model fit depends on the research question, the local watershed context, and other factors.

Is PWA free to use?
Yes! The plan is to design PWA as an open-access model, primarily for use by watershed managers, policy-makers, and the general public. The model will rely entirely on publicly available software and datasets and all code will be shared publicly under a copy-left license.