We have wetlands and riparian areas across our Peace Region that are home to diverse plant and wildlife species. When it comes to conserving and enhancing these areas—which include habitats like riparian forests, bogs, marshes, fens, and swamps—the first step seems pretty straightforward: figuring out where they are. But mapping wetlands and riparian areas isn’t as simple as it sounds. Our region is large, about 7.2 million hectares, and there are remote areas where access is a real challenge.
Here at the Fish & Wildlife Compensation Program (FWCP), it’s our job to pinpoint conservation and enhancement opportunities and get to work on them. So, we invested in a machine learning model to help us predict the abundance, distribution, and connectivity of wetlands and riparian areas throughout our Peace Region. We chose this technological approach because our region is so large that more traditional mapping would not have been cost-effective, or even feasible. This model—which was developed by the B.C. Ministry of Environment and Climate Change Strategy with FWCP funding—displays through an online tool called the Williston Wetland Explorer Tool (WWET).
The model has predicted where wetlands and riparian areas are across our region. It can also predict where uplands and water are, and even further classify wetlands into three riparian classes and five wetland classes: bog, fen, marsh, swamp, and shallow water wetlands.
The model was built using several layers of data, such as elevation data, satellite imagery, and high-resolution spatial climate data. Making the model accurate meant “training” it to interpret the data and accurately predict wetlands and riparian areas. In order to do this, ecosystem-mapping experts interpreted random points using satellite imagery and input their assessments of uplands, water, and wetlands (including wetland and riparian classes) into the model. Over 13,000 training points were expertly interpreted and incorporated into the model and then validated with on-the-ground field data that was collected in 2017, 2018, and 2019.
Comparing the predicted areas with field data showed that the model is 86.7% accurate at predicting where wetlands, water, and uplands occur across our Peace Region, making it a reliable tool for identification, planning, and management. The model is also 53% accurate at predicting the three riparian and five wetland classes—which reflects how challenging it is to make predictions at those levels.
The WWET addresses that first step of conserving and enhancing wetlands and riparian areas: it figures out where they are and what types of disturbances, like roads, are impacting their natural hydrology. It displays disturbance layers—such as roads and linear corridors, cut blocks, wildfire perimeters, mines, and areas of beetle infestation—and shows how those layers overlap with predicted wetlands and riparian areas. This makes it possible to be proactive about conservation and enhancement opportunities. In fact, two actions in our 2020 Peace Region Riparian & Wetlands Action Plan—which lays out priority actions and guides our decision-making when it comes to funding projects—are focused on using the predictive riparian and wetland model and the WWET to prioritize wetlands and riparian areas for conservation and enhancement.
Our Peace Region is home to diverse ecosystems that support countless plant, fish, and wildlife species. Wetlands and riparian areas provide habitat for a disproportionate number of species compared to other ecosystems, and the predictive wetland and riparian model and the WWET will allow us to work more efficiently to conserve and enhance these areas—and provide functioning habitat for the species that depend on them.
The Fish & Wildlife Compensation Program is a partnership between BC Hydro, First Nations, the Province of B.C., Fisheries and Oceans Canada, and public stakeholders to conserve and enhance fish and wildlife in watersheds impacted by BC Hydro dams.
Chelsea Coady is the Peace Region manager for the Fish and Wildlife Compensation Program. Have a question? Email her at firstname.lastname@example.org.