Waterfowl Modeling for Prairie Parklands

PHJV’s long term investment in science is the foundation of our success. Through PHJV studies in the Prairie Parklands:

  • 26,979 duck nests found and monitored
  • 3,214 radio-marked female mallards tracked
  • 145 waterfowl nesting study sites – covering 652,000 ha

The Waterfowl Productivity Model (WPM) is the tool that we use for conservation planning across the PHJV.

Decades of research from three, multi-year Ducks Unlimited Canada studies (PHJV Assessment Study, 1993–2000 [Howerter et al. 2014]; Pintail Study, 2005–2007 [Devries et al. 2018]; and the Spatial/Temporal Variability Study 2001–2009 [SPATS, Bortolotti et al. 2022]) focused around hen- and nest-level breeding success related to habitat parameters, has culminated in a dataset that is like no other in the wildlife conservation discipline. The datasets focus on the five main dabbling duck species breeding in prairie Canada; mallard (Anas platyrhynchos), gadwall (Mareca strepera), blue-winged teal (Spatula discors), northern shoveler (Spatula clypeata), northern pintail (Anas acuta). These data have allowed us to build the WPM, a computer simulation that predicts the breeding density and breeding success of these species in response to past and future habitat changes in their key breeding landscapes in prairie Canada.

Figure A5-4 Location of DUC waterfowl nesting study areas (PHJV Assessment, Pintail, SPATS) within Grassland and Aspen Parkland Ecozones of prairie Canada, 1993-2009.

In operation, the WPM inputs breeding pairs by species, habitat availability and other landscape-specific covariates (e.g., breeding location; latitude, longitude) to generate a population of nests which are apportioned among habitats within seasonal periods, and exposed to habitat- and season-specific nest survival rates. Final model output is estimated hatched nests by species, habitat and seasonal period with associated estimates of variation. Thus, in a conservation planning context, the model can estimate the impact of various habitat changes, whether background or conservation-related, affecting the number or species composition of waterfowl pairs present (e.g., wetland loss or restoration), or the distribution and survival of nests (e.g., grassland loss or restoration, increased acreage of winter wheat, etc.).  This conservation planning tool allows us to estimate how much habitat is needed, and where, to achieve waterfowl NAWMP population objectives.

Non-game Birds Modeling for Prairie Parklands

In order to include non-game birds in conservation planning, we developed species distribution models for a variety of landbirds, shorebirds, and waterbirds. Most waterbird models predicted probability of occurrence across the PHJV while shorebird and landbird models predicted species density. Models were developed using data during >70,000 surveys from >28,000 unique locations across the PHJV. Species distribution models were used to identify priority areas and set habitat conservation targets for non-game birds.

Marsh Wren Probability of Occurrence Model
Marsh Wren Probability of Occurrence Model
Sprague's Pipit Species Density Model
Sprague's Pipit Species Density Model

Western Boreal Forest Modeling for Waterfowl

The PHJV partners have directly influenced 50 million acres of some of the best waterfowl habitat in the WBF.

Overall, the WBF contains about 262.5 million acres of wetlands, or 35% of the total area. For the purposes of conservation planning, waterfowl habitat is defined as open water or wetlands along with their adjacent upland habitat located within 400m. Based on this simplistic definition, 49% of the WBF is considered suitable waterfowl habitat with varying distribution among jurisdictions (Table 1).

As waterfowl are distributed at much lower densities in the WBF than in the prairie pothole region, thus,  key waterfowl habitat are identified based on Ducks Unlimited Canada’s Hybrid Wetland Layer (Ducks Unlimited Canada 2010a). This product was derived by merging Canvec and Earth Observation for Sustainable Development earth cover data, and categorizing the landscape into open water, undefined vegetated wetlands and uplands. Waterfowl habitat was then defined as open water and any of the other two habitat classes within 400 metres of the open water boundary.

Distribution of predicted waterfowl breeding population, waterfowl habitat, and proportion of total WBF area among PHJV jurisdictions.
TABLE 1 - Distribution of predicted waterfowl breeding population, waterfowl habitat, and proportion of total WBF area among PHJV jurisdictions.

What’s Next

Current approaches and analyses have focused on guild level grouping of species (i.e., ground, cavity, overwater) to identify broad trends to focus on immediate conservation objectives. There is a need for more species-specific studies given the need to understand changes in community structure (i.e., species composition and abundance) and differing species population trajectories in the WBF, even within guilds.

Several key topics where WBF information gaps will be the priority for future boreal waterfowl modeling studies over the 2021—2025 period include:

  • Gaps in waterfowl population abundance and distribution estimates
  • The cumulative effects of industrial activities on waterfowl distribution and productivity
  • Monitoring and evaluation of current BMPs and mitigation measures to refine policy direction
  • Greater certainty around the suite of climate change implications within the WBF ecoregion