<p>Producing accurate hourly streamflow forecasts in large basins is difficult without a distributed model to represent both streamflow routing through the river network and the spatial heterogeneity of land and weather conditions. HydroForecast is a theory-guided deep learning flow forecasting product that consists of short-term (hourly predictions out to 10 days), seasonal (10 day predictions out to a year), and daily reanalysis models. This work focuses primarily on the short-term model which has award winning accuracy across a wide range of basins.</p> <p>In this work, we discuss the implementation of a novel distributed flow forecasting capability of HydroForecast, which splits basins into smaller sub-basins and routes flows from each subbasin to the downstream forecast points of interest. The entire model is implemented as a deep neural network allowing end-to-end training of both sub-basin runoff prediction and flow routing. The model's routing component predicts a unit hydrograph of flow travel time at each river reach and timestep allowing us to inspect and interpret the learned river routing and to seamlessly incorporate any upstream gauge data. </p> <p>We compare the accuracy of this distributed model to our original flow forecasting model at selected sites and discuss future improvements that will be made to this model.</p>

<p>Providing accurate seasonal (1-6 months) forecasts of streamflow is critical for applications ranging from optimizing water management to hydropower generation. In this study we evaluate the performance of stacked Long Short Term Memory (LSTM) neural networks, which maintain an internal set of states and are therefore well-suited to modeling dynamical processes.</p><p>Existing LSTM models applied to hydrological modeling use all available historical information to forecast contemporaneous output. This modeling approach breaks down for long-term forecasts because some of the observations used as input are not available in the future (e.g., from remote sensing and in situ sensors). To solve this deficiency we train a stacked LSTM model where the first network encodes the historical information in its hidden states and cells. These states and cells are then used to initialize the second LSTM which uses meteorological forecasts to create streamflow forecasts at various horizons. This method allows the model to learn general hydrological relationships in the temporal domain across different catchment types and project them into the future up to 6 months ahead.</p><p>Using meteorological time series from NOAA’s Climate Forecast System (CFS), remote sensing data including snow cover, vegetation and surface temperature from NASA’s MODIS sensors, SNOTEL sensor data, static catchment attributes, and streamflow data from USGS we train a stacked LSTM model on 100 basins, and evaluate predictions on out-of-sample periods from these same basins. We perform sensitivity analysis on the effects of remote sensing data, in-situ sensors, and static catchment attributes to understand the informational content of these various inputs under various model architectures. Finally, we benchmark our model to forecasts derived from simple climatological averages and to forecasts created by a single LSTM that excludes all inputs without forecasts.</p><p> </p>

No abstract

<p>Accurate streamflow forecasts equip water managers to adapt to changing flow regimes and constraints, increase water supply reliability, reduce flood risk, and maximize revenue. Over the 2021 water year, the Upstream Tech team took part in a live, 1-10 day ahead streamflow forecasting competition using our flow forecast system, HydroForecast. The competition was a chance to objectively compare operational forecasts using a range of modeling approaches from national agencies, hydropower utilities’ in-house teams, private forecasters and individual modelers at 19 sites in North America. HydroForecast outperformed both statistical and conceptual models and won the competition. We evaluate HydroForecast’s performance relative to other models to identify its strengths and areas for further research by region, season, and forecast horizon. We also share what our theory-guided machine learning approach to hydrologic modeling means in practice for HydroForecast, focusing on the key facets of our approach which contribute most to our accuracy. Finally, we describe the largest opportunities for further forecast accuracy gains we identified in this competition and some of the research efforts we are working on to meet those opportunities.</p>

No abstract