Warmer air temperatures predicted to result in wetland drying in the Upper Columbia River Valley, British Columbia, Canada

Climatic warming is likely to affect the Canadian Rockies, leading to changes in the land cover (LC) and hydrological cycles. This study estimates climate-induced changes in LC (open water, marsh, wet meadow, and woody/shrub) in the Upper Columbia River Wetlands (UCRW), British Columbia, Canada, from 1984 to 2040. An artificial Neural Network (ANN) approach was used with Landsat series archive data from 1984 to 2022 to project seasonal LC change from 2020s to 2040s. Concurrently, hydroclimatic-based models (using air temperature and precipitation to predict river discharge at the UCRW, 1984–2022) were developed (average Nash Sutcliffe: training 0.75 and validation of 0.70) to predict (1984–2040) river discharge forced by Representative Concentration Pathway (RCP) 4.5 and 8.5. The 1984–2022 regression between river discharge and UCRW open water area was forced by RCP scenario river discharge results, calculating open water area for both scenarios. ANN-predicted LC with a Kappa of 0.85 (average of all seasons) for 2020s reference and projected LC, and 0.82 for reference and projected LC change maps (2000s–2020s). From 2020s to 2040s, the ANN projected a reduction (−5 %) of open water areas during late summer (August to mid-September) in the UCRW, consistent with RCP 4.5 forecasts. The peak of the open water area in the UCRW is projected to shift from summer (late-May to July) to spring (April to mid-May) in both RCP scenarios. The projected changing hydrological conditions reduced the marsh area (−1 % to −12 %) and increased the wet meadow (+1 % to +4 %) mostly in the summer and late summer. Meanwhile, woody and shrubby vegetation on the floodplain increased (3 % to 5 %), indicating that the floodplain is projected to dry out. [Display omitted] •ANN-predicted seasonal land cover from historical Landsat time series.•Hydroclimatic models (air temperature and precipitation) to predict river discharge.•The open water area decreased while woody vegetation expanded.•Peak river discharge and open water area projected in the spring.•ANN projected late summer open water decline, similar to hydroclimatic projections.