Research

We are interested in quantifying the changes in floodplain hydrology and vegetation in response to restoration. Because floodplains represent the nexus between terrestrial and aquatic ecosystems, the fluxes of water within and at the boundaries of the floodplain are highly varied in both space and time and play a dominant role in driving ecosystem processes. These hydrologic fluxes include:

Precipitation
Infiltration
Evapotranspiration
Groundwater flow including seepage faces
Hillslope runoff
Streamflow
Hyporheic flow

Several predictions of the effect of restoration include:

decrease in the depth to water table
increase in evapotranspiration
increase in wetland vegetation

Each hydrologic flux shown above will be monitored in the field and modeled in order to test hypotheses and different management scenarios at two different sites owned by TNC. The first site is the floodplain that was restored in 2006 (above figure on right) and the second site is a slightly smaller floodplain located approximately 1.7 km upstream of the 2006 restoration site. This second site was restored in early September 2008 (above figure on left shown before restoration). Because monitoring of the two sites began in the spring of 2007, our strategy is the collection of pre- and post-restoration data at the 2008 site to offer a comparison, while using the 2006 site as a space-for-time substitution in that it may represent conditions at the 2008 site two years in the future.

Riparian vegetation will be monitored at the community-level using ground-based vegetation surveys and remote sensing data collected from a small plane. The remote sensing equipment includes a thermal, 6-band near-infrared, and standard SLR camera.

Vegetation maps will be combined with simulated hydrologic data using a variably-saturated groundwater flow model to generate a dataset of combined hydrologic regime statistics and vegetation observations. Using this spatially-extensive dataset, non-linear vegetation models will be generated based on the hydrologic regime metrics (based on water-table depth and soil moisture) that best explain the occurrence of each vegetation community. For example, the occurrence of a wet-prairie community may be well explained by the median water table depth during the growing season.

These community-specific vegetation models will then be coupled to the variably-saturated groundwater flow model to create a hydroecologic model that will give spatially-extensive predictions of vegetation. Different management or climate scenarios could then be tested to determine impacts on vegetation patterning.

In addition to floodplain hydrology and vegetation, we will also be quantifying the changes in channel geometry and subsequent changes in flood hydraulics. Since the restoration process re-connects the channel with its floodplain during high flows, we predict that stream velocity and depth will decrease and width will increase for a given flood discharge.

Stream Temperature/Groundwater Discharge

Remotely-sensed thermographic profiles and in-stream temperature histories validated a one-dimensional stream temperature model that highlights the threat of global climate change to ecologically valuable and economically significant trout fisheries in southwest Wisconsin (study site – Figure 1). This work demonstrates the value of remotely-sensed thermal data for stream temperature model validation and analyses of thermal heterogeneity. Without thermal infrared data, the significant depression in stream temperature from ~1.5-3 km would not be easily identified (Figure 2). The profile developed with thermal infrared data increases our understanding of spatially-variable groundwater inflow to the stream. This finding will benefit fisheries managers seeking to locate and protect thermal refugia, where maximum daily stream temperatures are depressed due to the inflow of cool groundwater.

Figure 1. Study site location. (A) Location of the East Branch Pecatonica River (B) Sample thermal image collected over the East Branch Pecatonica River (C) Modeled stream reach (10.47 km) and digital elevation model

Figure 2. Comparison of remotely-sensed (RS), in-stream (In) and simulated (Sim) temperatures on July 24, 2008 for the East Branch Pecatonica River. The profile moves downstream left to right (0 km is upstream, 10.47 km is downstream). Remotely-sensed temperatures were sampled using a thermal infrared camera. In-stream records at 2.75 km, 7.3 km and 7.75 km were recorded using HOBO loggers. Simulated temperatures were modeled using Heat Source V.8.0.4 software.

Twelve stream temperature model simulations were used to evaluate the potential threats to brook trout and brown trout at the East Branch Pecatonica. Both increases in air temperature and decreases in baseflow drive maximum daily stream temperature high and in combination threaten the stream’s fishery (Figure 3). In the most extreme climate change scenario, stream temperature may cross thermal tolerance thresholds developed for similar fisheries in Wisconsin and Michigan (Wehrly et al. 2007) (Figure 4). This result has implications beyond this study site. Stream temperature in higher flow streams or those with smaller groundwater input may be more susceptible to change due to changes in stream flow. At this study site, the absolute difference between the base case and the recharge simulations is greater at the downstream location than the upstream location. On a larger stream, cases with stream flow alteration may reach thermal tolerance thresholds. Wider, shallower streams warm faster and may reach critical thermal maxima for trout.

Figure 3. Maximum daily temperature for model simulations of the base case (no change in air/groundwater temperature or recharge), the base case plus 1°C, the base case plus 3°C and the base case plus 5°C at 7.75 km [top]. Maximum daily temperature for model simulations of the base case (no change in air/groundwater temperature or recharge), the base case -30% recharge and the base case +30% recharge at 7.75 km [bottom]. A decrease in recharge decreases the volume of stream flow allowing the stream to warm more readily.

Figure 4. Comparison of simulated cases at an upstream (7.75 km) and downstream (2.75 km) location to the maximum daily maximum curve (solid line) developed by Wehrly et al. (2007). The absolute difference between the base case and the change in recharge cases is greater at the downstream location for all simulations. This is a consequence of greater stream flow downstream; reductions in stream flow are amplified in the simulated stream temperature at locations further downstream. Only the +5°C, -30% recharge case crosses the empirical threshold developed by Wehrly et al. (2007), however this is a relatively upstream segment of the East Branch Pecatonica and ecologically and economically valuable trout fisheries may exist downstream that may be threatened in less-extreme climate change scenarios. Additionally, this has implications for higher order streams and fisheries (wider, higher volume) which may be at risk in less extreme scenarios.

This study demonstrates that climate change poses a risk to trout fisheries in southwest Wisconsin. Site specific conditions, such as stream flow and stream width to depth ratios, are critical controls on stream temperature. Thermal infrared and in-stream data may be used in combination to validate a freely-available one-dimensional stream temperature model to predict the impacts of climate change on stream temperature. Thermal tolerance thresholds, such as the maximum daily maximum temperature developed by Wehrly et al. (2007), may be used to assess species’ risk.

Funding

Wisconsin Groundwater Coordinating Council/Water Resources Institute

References

Wehrly, K.E., Wang, L.Z. and Mitro, M., 2007. Field-based estimates of thermal tolerance limits for trout: Incorporating exposure time and temperature fluctuation. Transactions of the American Fisheries Society, 136(2): 365-374.

Stream & Floodplain Biogeochemistry

Occurrence and generation of nitrite in ground and surface waters in an agricultural watershed

Emily Stanley

Sponsored by the Wisconsin Groundwater Coordination Council and Wisconsin Water Resources Institute

Streams throughout southern and central Wisconsin carry the signal of extensive agricultural activity in the form of high nutrient concentrations. In past studies, we have routinely found high concentrations of nitrogen (N) in stream water, mostly in the form of nitrate (NO₃^–). We also found that these N-rich streams contain low levels of nitrite (NO₂^–), a form of N that is an intermediary in several biological reactions, as well as being a form of N that can be toxic at relatively low concentrations (Stanley and Maxted 2008). Thus, my current work at the East Branch Pecatonica River is intended to identify the locations and the processes that are likely to be responsible for generating NO₂^–in stream water. This has involved measuring concentrations of different forms of nitrogen in groundwater and streamwater stream reaches before and after restoration and conducting a series of experiments to measure the environmental factors that favor NO₂^– production. Preliminary results indicate that prior to discharging into the stream channel, groundwater is NO₂^– poor indicating that NO₂^– is likely generated within the channel. Further, laboratory experiments point to reduction of NO₃^– within sediments as the likely place and process responsible for NO₂^– formation. This summer, we will be investigating how this NO₂^– formation varies among sediment type (e.g., fine silt vs. sand) to test the hypothesis that most NO₂^– is generated in fine, organic-rich silts that are abundant in unrestored reaches. If this hypothesis is supported, it would suggest that the restoration activities at the East Branch should minimize the production of this potent nitrogen byproduct.

Prior to restoration, woody debris from the riparian zone in the East Branch Pecatonica was abundant in the channel (photo, left) and was extremely effective at trapping sediments. As a result, sediments in unrestored (“treatment”) reaches are finer-grained than in restored reaches, and the layer of sediments (measured as the depth a steel rod can be inserted into the sediments, or Depth to Refusal) is significantly greater in unrestored (“treatment” and “reference” reaches) reaches as well. Data from Lottig & Stanley (unpublished). We hypothesize that this thick layer of fine sediments is an ideal environment for nitrite generation in nitrogen-rich streams such as the Pecatonica.

Stanley, E.H. 2008. Potential sources of nitrite in southern Wisconsin agricultural streams. Annual Summer Meeting of the American Society of Limnology and Oceanography, St. John’s, Newfoundland.