Takeaway: Basins that do not have repeatable seasonal patterns or long-term SWE storage can be more challenging to forecast, as predictions rely more on accurately predicting the weather. Even so, HydroForecast can closely predict the timing and magnitude of spikes in inflow.
Event description: Active periods in the Little Tallapoosa are driven by rain. The most significant rain events include convective thunderstorms in the spring and summer and frontal systems in the winter months. The Little Tallapoosa’s flow is thus more variable throughout the year than snow-driven areas. Climate change is expected to intensify the hydrologic cycle in the Tallapoosa River, increasing the frequency and severity of high and low flows. Both have consequences for water management, but often can be mitigated and better planned for by integrating forecasts.
In rain-driven basins like this, historical averages do not have the benefit of significant and semi-regular seasonal fluctuations such as large annual spring pulses. Spikes in flow are more weather dependent, so long-term averages are less indicative of upcoming flow. For example, during one spring event in 2020 the long-term average was 5,000 CFS below the observed flow.
The 2020 winter/spring period was very active with multiple peaking events, especially as compared with 2021 spring. The National Weather Service archive for Central Alabama noted multiple tornadoes, high-wind, and heavy rain producing events that occurred from January to April, 2020.
The confidence intervals during the winter 2020 events were wider, indicating disagreement in the precipitation sources that the model uses as inputs, namely the ECMWF Hi-Res and NOAA’s Global Forecasting System products. Even with the flashiness of the basin, HydroForecast’s mean prediction and its confidence intervals track closely with the observed flows in the system.
One takeaway for the Upstream Tech team from forecasting in this basin and others like is the importance of incorporating a variety of high-quality meteorological forecasts so that the model can learn and use the most accurate inputs. This informed our SBIR research grant with the Department of Energy to add additional weather forecasting products into HydroForecast.