Considering Climate Change in the Estimation of Long‐Term Flood Risks of Devils Lake in North Dakota

Abstract

Terminal lakes are impacted by regional changes in climate. Devils Lake (DL), North Dakota, United States (U.S.), is a case in which a prolonged shift in the precipitation pattern resulted in a 10-m waterlevel rise over the past two decades, which cost over one billion U.S. dollars in mitigation. Currently, DL is 1.5 m from an uncontrolled overspill to the nearby Sheyenne River, which could lead to unprecedented environmental, social, and economic costs. Water outlets recently implemented in the lake to slow the water-level rise and prevent an uncontrolled overspill are subject to significant concerns over the introduction of invasive species and downstream water quality. We developed a hydrological model of the DL basin using the soil and water assessment tool and analyzed DL's overspill probability using an ensemble of statistically downscaled General Circulation Model (GCM) projections of the future climate. The results indicate a significant likelihood (7.3-20.0%) of overspill in the next few decades in the absence of outlets; some members of the GCM integration ensemble suggest an exceedance probability of over 85.0 and 95.0% for the 2020s and 2050s, respectively. Fullcapacity outlets radically reduce the probability of DL overspill and are able to partially mitigate the problem by decreasing the average lake level by approximately 1.9 and 1.5 m in the 2020s and 2050s, respectively.

Modeling Water Quantity and Sulfate Concentrations in the Devils Lake Watershed Using Coupled SWAT and CE‐QUAL‐W2

“…The results showed that operating two emergency outlets in the past few years has lowered the lake water level by ~0.7 m, but not sufficiently to prevent the lake water level from rising (Figure ). Similar results have been reported in the past (Vecchia, ; Kharel and Kirilenko, ). As there is a ~37% chance for the current wet condition to continue until 2040 (Vecchia, ), these results indicate that Devils Lake will likely reach the natural spill elevation by that time.…”

“…He found that the emergency outlet operation would significantly decrease, but not completely eliminate, the chance of natural spilling. Using the Soil and Water Assessment Tool (SWAT) with an ensemble of projections from 15 General Circulation Models (GCMs), Kharel and Kirilenko (2015) found a significant probability (7.3 to 20.0%) of overspill in the next few decades in the absence of outlets. None of these studies, however, have tried to examine water quality.…”

Mitigating Impact of Devils Lake Flooding on the Sheyenne River Sulfate Concentration

“…Spring floods in the RRNB are typically associated with cold climate conditions in winter and early spring such as harsh blizzards, early snow‐melting events, and/or ice jam processes in the rivers (Todhunter ). SWAT is a watershed‐scale hydrological and water quality model that is not yet fully capable of simulating the complex snowmelt processes under such a harsh continental climate condition in the RRNB (Wang and Melesse ; Wang et al ; Kharel and Kirilenko ). However, it should be noted that the underestimated peak discharges by the SWAT model did not affect the results of this study because: (1) the outlets were not operational during the historical flooding events when the simulated discharge was significantly underestimated (Figures and ), and (2) the SWAT model performed well in capturing the peak discharges within the range of two‐year flood or the maximum impact of the current outlet operation (Figure ).…”

Comparing CMIP-3 and CMIP-5 climate projections on flooding estimation of Devils Lake of North Dakota, USA

“…The amplification effect evident in endorheic watersheds makes them important for detection of regional changes in climate, because even relatively small changes in local temperature and precipitation extended over years leads to highly visible changes in water level and lake area, imposing a direct impact on the livelihood of lake-dependent populations. Kharel & Kirilenko (2015) provided numerous examples of closed lakes across the globe with receding (e.g., Aral Sea, Lake Chad, and Qinghai Lake) or rising (e.g., Caspian Sea and Mar Chiquita) waters related to changing climate and land use (see also Rosenzweig et al, 2007 ). While endorheic lakes most commonly experience receding water level due to human activities, DL, located in a sparsely populated Northern Great Plains area close to the US—Canada border (United States Geological Survey (USGS) Souris-Red-Rainy drainage region, see Fig.…”

“…Those studies however did not consider the impacts of global climate change on regional hydrology. Climate change impacts on DL’s flooding were taken into account by Kharel & Kirilenko (2015) , who estimated the probability of future DL levels and associated uncertainty by employing an ensemble of climate change projections used in the 2007 report of the IPCC ( Solomon et al, 2007 ). Further, Kharel, Zheng & Kirilenko (2016) estimated the role of land use change within the DL watershed in the observed flooding.…”