Fielding floods for farms and (low) flows : the potential benefit of repurposing small flood reservoirs for drought protection

Droughts have long been recognized as significant natural hazards that impose severe impacts on multiple sectors worldwide. Under climate change, these impacts will likely become more severe and more frequent in the future, prompting a need for increased water resources resilience in even temperate and water rich regions. At the same time, temperate regions have often contended with flooding as the primary hydrological hazard. Over 600 flood reservoirs have been built in the German state of Baden-Württemberg alone, with many of them fitting under the criteria of small reservoirs. Small reservoirs– those with a dam height of ≤ 15 m, a surface area of < 0.1 km2, and / or a storage volume of up to 1-2 million m3– have been proposed as decentralized solutions to water scarcity in arid regions, though their potential in temperate regions is still unexplored. This dissertation aimed to explore the potential water supply benefits of repurposing flood reservoirs as small reservoirs in temperate regions in addition to preserving their flood protection functions. To this end, a variety of models were built to represent the operation of repurposed flood reservoirs for two purposes– streamflow supplementation and irrigation demand– and tested them first in perfect-knowledge conditions and then under uncertain forecasts. In chapter 2, the benefit to streamflow supplementation was measured under perfect knowledge conditions using a function to determine the value of the quantity and timing of water delivered by the modeled reservoirs in relation to low-flow conditions. The reservoirs stored water when water was plenty and released water such that the downstream conditions would meet or exceed the 70th percentile exceedance flow, and emptied in time for a flood event. Under the perfect-knowledge conditions, it was shown that 12 of the 30 reservoirs tested could alleviate at least 70% of drought conditions, and that none of the reservoirs experienced an increase in flooding as a result of the new operating rules. However, there were no strong relationships between how much water was available to the reservoir and performance. In chapter 3, the benefit to irrigation demand was measured under perfect-knowledge conditions by modelling the agricultural irrigation demand around each reservoir individually and determining how much of this demand the reservoirs could fulfill. Similar to chapter 2, the reservoirs stored water when water was plenty, but instead of releasing water, scheduled withdrawals when irrigation demand occurred. This analysis focused on the years 2017-2020, which include the landmark 2018-2020 drought years that caused massive increases in agricultural irrigation demand in the region. Of the 30 tested reservoirs, 20 could provide at least 70% of the needed water during the drought years without increasing flooding conditions. While some of these reservoirs also performed well in chapter 2, the timing of demand is different enough that the two use cases should be considered distinctly. In chapter 4, the main assumption in the previous two studies was challenged: the assumption of perfect-knowledge conditions. Instead, the models in this chapter make a decision based on the current inflow and the forecast available to them at each point in time. If a flood was occurring, that took priority; else, the reservoir could either pre-release in anticipation of a flood in its forecast, store water when water was plenty and no flood was forecasted, release or allow withdrawal upon demand, or simply pass water through. The results demonstrated that most reservoirs were still able to provide tangible benefits (> 40%) when optimized for either use case. The performance decay from perfect-knowledge to forecasted conditions (which in several cases was quite considerable) was explained less by common forecast quality metrics and more by flood incidence, but the timing of flood predictions (i.e. if a forecast caused emptying before demand occurred) was most influential. The most successful operating rules mimicked the rules under perfect-knowledge scenarios, allowing aggressive storage and holding onto the water as long as is safe. Overall, the potential water supply benefits of repurposing flood reservoirs for drought protection can be quite high, particularly for agriculture, and can be operated without increasing flood risk. While calculated metrics were largely unable to identify unsuitable or especially suitable reservoirs, categorization by size and use are possible first filters; however, the sample size remains too small to determine this conclusively. Future investigations of utilization potential in small reservoirs should thus be conducted for each reservoir individually, and with consideration for additional factors such as water quality, river ecology, and any potential long-term effects on water supply and drought, including as a potential mitigation strategy for flash drought. This dissertation has provided the basic tools for such analyses, and in doing so, contributes to informed conversations about ways to increase water supply resilience in a warming world.