Habitat Quality in the Floodplain Following a Channel Reconfiguration Project: Stewart River, MN

Abstract

Low flows and elevated temperatures threaten many cold-water habitats. Restoration work is commonly implemented to improve a variety of habitat degradations like elevated temperatures and lack of deep pool habitat. However, the influence of restoration work on groundwater storage and connectivity in systems with low human impacts is not well understood, especially restoration projects that involve the installation of off-channel ponds. This study looks at the connectivity and habitat quality of seven floodplain ponds with very different morphology and connectivity with the channel. The ponds were constructed as part of a large channel reconfiguration project to decrease width-to depth ratios, and improve/create coldwater fish habitat in the Stewart River, a trout stream near Two Harbors, Minnesota. During May-September of 2016 and 2017 (one year and two years following the realignment work), I studied the constructed ponds influence on water storage and water temperatures by installing water level and temperature sensors in the ponds, in shallow groundwater wells, and in the main channel. I also measured dissolved oxygen in the channel and ponds and surveyed the locations of thermal refuges between 2016 and 2018. Only one of seven floodplain ponds had suitable low-flow habitat for coldwater fish. This pond (P761) was directly connected to the channel at all flows. Three of six ponds in which temperature was measured throughout the summer were warmer than the channel, and six of seven ponds exhibited low dissolved oxygen levels (≤ 7 mg/L) in summer. Habitat quality in ponds, as measured by temperature and dissolved oxygen, was related to shading, pond morphology, and connectivity with the channel. Sunny locations displayed warmer water temperatures than shaded locations (p = 0.0001). Dissolved oxygen levels increased with pond surface area and connectivity with the channel. This study illustrates that pond geometry, connectivity with the channel, and canopy cover primarily drive temperature and dissolved oxygen levels in constructed ponds. Water elevations in five ponds and several channel locations showed that water levels in the channel peaked earlier and receded slightly quicker than ponds following precipitation events. It took about a day for the channel to reach its peak water elevation following summer precipitation events and two days for constructed ponds to peak. Pond water elevations receded to half their elevation ~3-6 days slower than the channel, but not gradually enough to augment low flows long term. During storm events, channel water elevations peaked earlier and were higher than nearby pond water elevations making the upper part of the realignment behave like a losing reach. Overall, the Stewart River reconfiguration project did not generate substantial long-term water storage or lasting high-quality floodplain fish habitats. This study illustrates the challenges and risks of channel re-meandering in flashy, relatively healthy river systems.