An official website of the United States government
Abstract The Ensemble Streamflow Prediction (ESP) framework combines a probabilistic forecast structure with process‐based models for water supply predictions. However, process‐based models require computationally intensive parameter estimation, increasing uncertainties and limiting usability. Motivated by the strong performance of deep learning models, we seek to assess whether the Long Short‐Term Memory (LSTM) model can provide skillful forecasts and replace process‐based models within the ESP framework. Given challenges inimplicitlycapturing snowpack dynamics within LSTMs for streamflow prediction, we also evaluated the added skill ofexplicitlyincorporating snowpack information to improve hydrologic memory representation. LSTM‐ESPs were evaluated under four different scenarios: one excluding snow and three including snow with varied snowpack representations. The LSTM models were trained using information from 664 GAGES‐II basins during WY1983–2000. During a testing period, WY2001–2010, 80% of basins exhibited Nash‐Sutcliffe Efficiency (NSE) above 0.5 with a median NSE of around 0.70, indicating satisfactory utility in simulating seasonal water supply. LSTM‐ESP forecasts were then tested during WY2011–2020 over 76 western US basins with operational Natural Resources Conservation Services (NRCS) forecasts. A key finding is that in high snow regions, LSTM‐ESP forecasts using simplified ablation assumptions performed worse than those excluding snow, highlighting that snow data do not consistently improve LSTM‐ESP performance. However, LSTM‐ESP forecasts that explicitly incorporated past years' snow accumulation and ablation performed comparably to NRCS forecasts and better than forecasts excluding snow entirely. Overall, integrating deep learning within an ESP framework shows promise and highlights important considerations for including snowpack information in forecasting.
more »
« less
Characterizing the Evolution of Extreme Water Levels with Long Short-Term Memory Station-Based Approximated Models and Transfer Learning Techniques
Extreme water levels (EWLs) resulting from tropical and extratropical cyclones pose significant risks to coastal communities and their interconnected ecosystems. To date, physically-based models have enabled accurate characterization of EWLs despite their inherent high computational cost. However, the applicability of these models is limited to data-rich sites with diverse morphologic and hydrodynamic characteristics. The dependence on high quality spatiotemporal data, which is often computationally expensive, hinders the applicability of these models to regions of either limited or data-scarce conditions. To address this challenge, we present a computationally efficient deep learning framework, employing Long Short-Term Memory (LSTM) networks, to predict the evolution of EWLs beyond site-specific training stations. The framework, named LSTM-Station Approximated Models (LSTM-SAM), consists of a collection of bidirectional LSTM models enhanced with a custom attention layer mechanism embedded in the model architecture. Moreover, the LSTM-SAM framework incorporates a transfer learning approach that is applicable to target (tide-gage) stations along the U.S. Atlantic Coast. The LSTM-SAM framework demonstrates satisfactory performance with “transferable” models achieving average Kling-Gupta Efficiency (KGE), Nash-Sutcliffe Efficiency (NSE), and Root-Mean Square Error (RMSE) ranging from 0.78 to 0.92, 0.90 to 0.97, and 0.09 to 0.18 at the target stations, respectively. Following these results, the LSTM-SAM framework can accurately predict not only EWLs but also their evolution over time, i.e., onset, peak, and dissipation, which could assist in large-scale operational flood forecasting, especially in regions with limited resources to set up high fidelity physically-based models.
more »
« less
- PAR ID:
- 10540740
- Publisher / Repository:
- SSRN
- Date Published:
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
The number of electric vehicles (EV) has increased significantly in the past decades due to its advantages including emission reduction and improved energy efficiency. However, the adoption of EV could lead to overloading the grid and degrading the power quality of the distribution system. It also demands an increase in the number of EV charging stations. To meet the charging needs of 15 million EVs by the year 2030 with limited charging stations, prediction of charging needs, and reallocating charging resources are in emerging needs. In this study, long short-term memory (LSTM) and autoregressive and moving average models (ARMA) models were applied to predict charging loads with temporal profiles from 3 charging stations. Prediction accuracy was applied to evaluate the performance of the models. The LSTM models demonstrated a significant performance improvement compared to ARMA models. The results from this study lay a foundation to efficiently manage charge resources.more » « less
-
Abstract. As a genre of physics-informed machine learning, differentiable process-based hydrologic models (abbreviated as δ or delta models) with regionalized deep-network-based parameterization pipelines were recently shown to provide daily streamflow prediction performance closely approaching that of state-of-the-art long short-term memory (LSTM) deep networks. Meanwhile, δ models provide a full suite of diagnostic physical variables and guaranteed mass conservation. Here, we ran experiments to test (1) their ability to extrapolate to regions far from streamflow gauges and (2) their ability to make credible predictions of long-term (decadal-scale) change trends. We evaluated the models based on daily hydrograph metrics (Nash–Sutcliffe model efficiency coefficient, etc.) and predicted decadal streamflow trends. For prediction in ungauged basins (PUB; randomly sampled ungauged basins representing spatial interpolation), δ models either approached or surpassed the performance of LSTM in daily hydrograph metrics, depending on the meteorological forcing data used. They presented a comparable trend performance to LSTM for annual mean flow and high flow but worse trends for low flow. For prediction in ungauged regions (PUR; regional holdout test representing spatial extrapolation in a highly data-sparse scenario), δ models surpassed LSTM in daily hydrograph metrics, and their advantages in mean and high flow trends became prominent. In addition, an untrained variable, evapotranspiration, retained good seasonality even for extrapolated cases. The δ models' deep-network-based parameterization pipeline produced parameter fields that maintain remarkably stable spatial patterns even in highly data-scarce scenarios, which explains their robustness. Combined with their interpretability and ability to assimilate multi-source observations, the δ models are strong candidates for regional and global-scale hydrologic simulations and climate change impact assessment.more » « less
-
Soil moisture (SM) plays a significant role in determining the probability of flooding in a given area. Currently, SM is most commonly modeled using physically-based numerical hydrologic models. Modeling the natural processes that take place in the soil is difficult and requires assumptions. Besides, hydrologic model runtime is highly impacted by the extent and resolution of the study domain. In this study, we propose a data-driven modeling approach using Deep Learning (DL) models. There are different types of DL algorithms that serve different purposes. For example, the Convolutional Neural Network (CNN) algorithm is well suited for capturing and learning spatial patterns, while the Long Short-Term Memory (LSTM) algorithm is designed to utilize time-series information and to learn from past observations. A DL algorithm that combines the capabilities of CNN and LSTM called ConvLSTM was recently developed. In this study, we investigate the applicability of the ConvLSTM algorithm in predicting SM in a study area located in south Louisiana in the United States. This study reveals that ConvLSTM significantly outperformed CNN in predicting SM. We tested the performance of ConvLSTM based models by using a combination of different sets of predictors and different LSTM sequence lengths. The study results show that ConvLSTM models can predict SM with a mean areal Root Mean Squared Error (RMSE) of 2.5% and mean areal correlation coefficients of 0.9 for our study area. ConvLSTM models can also provide predictions between discrete SM observations, making them potentially useful for applications such as filling observational gaps between satellite overpasses.more » « less
-
Abstract Empirical thresholds for landslide warning systems have benefitted from the incorporation of soil‐hydrologic monitoring data, but the mechanistic basis for their predictive capabilities is limited. Although physically based hydrologic models can accurately simulate changes in soil moisture and pore pressure that promote landslides, their utility is restricted by high computational costs and nonunique parameterization issues. We construct a deep learning model using soil moisture, pore pressure, and rainfall monitoring data acquired from landslide‐prone hillslopes in Oregon, USA, to predict the timing and magnitude of hydrologic response at multiple soil depths for 36‐hr intervals. We find that observation records as short as 6 months are sufficient for accurate predictions, and our model captures hydrologic response for high‐intensity rainfall events even when those storm types are excluded from model training. We conclude that machine learning can provide an accurate and computationally efficient alternative to empirical methods or physical modeling for landslide hazard warning.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Posted Content:
- https://doi.org/10.2139/ssrn.4862037
