Search for: All records

The sensitivity of forecast-informed reservoir operating policies to forecast attributes (lead-time and skill) in many-objective water systems has been well-established. However, the viability of forecast-informed operations as a climate change adaptation strategy remains underexplored, especially in many-objective systems with complex trade-offs across interests. Little is known about the relationships between forecast attribute and policy robustness under deep uncertainty in future conditions and the relationships between forecast-informed performance and future hydrologic state. This study explores the sensitivity of forecast-informed policy robustness to forecast lead-time and skill in the outflow management plan of the Lake Ontario basin. We create water supply forecasts at four different subseasonal-to-annual lead-times and two levels of skill and further employ a many-objective evolutionary algorithm to discover policies tailored for each forecast case, historical supply conditions, and six objectives. We also leverage a partnership with decision-makers to identify a subset of candidate policies, which are reevaluated under a large set of plausible hydrologic conditions that reflect stationary and nonstationary climates. Scenario discovery techniques are used to map attributes of future hydrology to forecast-informed policy performance. Results show policy robustness is directly related to forecast lead-time, where policies conditioned on 12-month forecasts were more robust under future hydrology. Policies tailored for noisier long-lead forecasts were more robust under a wide range of plausible futures compared with policies trained to perfect forecasts, which highlights the potential to overfit control policies to historical information, even for a forecast-informed policy with perfect foresight. The relationship between performance and the hydrologic regime is dependent on the complexity of the interactions between control decisions and objectives. A threshold of objective performance as a function of supply conditions can support adaptive management of the system. However, more complex interactions make it difficult to identify simple hydrologic indicators that can serve as triggers for dynamic management. 

Abstract. Deep learning (DL) rainfall–runoff models outperform conceptual, process-based models in a range of applications. However, it remains unclear whether DL models can produce physically plausible projections of streamflow under climate change. We investigate this question through a sensitivity analysis of modeled responses to increases in temperature and potential evapotranspiration (PET), with other meteorological variables left unchanged. Previous research has shown that temperature-based PET methods overestimate evaporative water loss under warming compared with energy budget-based PET methods. We therefore assume that reliable streamflow responses to warming should exhibit less evaporative water loss when forced with smaller, energy-budget-based PET compared with temperature-based PET. We conduct this assessment using three conceptual, process-based rainfall–runoff models and three DL models, trained and tested across 212 watersheds in the Great Lakes basin. The DL models include a Long Short-Term Memory network (LSTM), a mass-conserving LSTM (MC-LSTM), and a novel variant of the MC-LSTM that also respects the relationship between PET and evaporative water loss (MC-LSTM-PET). After validating models against historical streamflow and actual evapotranspiration, we force all models with scenarios of warming, historical precipitation, and both temperature-based (Hamon) and energy-budget-based (Priestley–Taylor) PET, and compare their responses in long-term mean daily flow, low flows, high flows, and seasonal streamflow timing. We also explore similar responses using a national LSTM fit to 531 watersheds across the United States to assess how the inclusion of a larger and more diverse set of basins influences signals of hydrological response under warming. The main results of this study are as follows: The three Great Lakes DL models substantially outperform all process-based models in streamflow estimation. The MC-LSTM-PET also matches the best process-based models and outperforms the MC-LSTM in estimating actual evapotranspiration. All process-based models show a downward shift in long-term mean daily flows under warming, but median shifts are considerably larger under temperature-based PET (−17 % to −25 %) than energy-budget-based PET (−6 % to −9 %). The MC-LSTM-PET model exhibits similar differences in water loss across the different PET forcings. Conversely, the LSTM exhibits unrealistically large water losses under warming using Priestley–Taylor PET (−20 %), while the MC-LSTM is relatively insensitive to the PET method. DL models exhibit smaller changes in high flows and seasonal timing of flows as compared with the process-based models, while DL estimates of low flows are within the range estimated by the process-based models. Like the Great Lakes LSTM, the national LSTM also shows unrealistically large water losses under warming (−25 %), but it is more stable when many inputs are changed under warming and better aligns with process-based model responses for seasonal timing of flows. Ultimately, the results of this sensitivity analysis suggest that physical considerations regarding model architecture and input variables may be necessary to promote the physical realism of deep-learning-based hydrological projections under climate change.