Abstract
AbstractFreshwater ecosystems are experiencing greater variability due to human activities, necessitating new tools to anticipate future water quality. In response, we developed and deployed a real‐time iterative water temperature forecasting system (FLARE—Forecasting Lake And Reservoir Ecosystems). FLARE is composed of water temperature and meteorology sensors that wirelessly stream data, a data assimilation algorithm that uses sensor observations to update predictions from a hydrodynamic model and calibrate model parameters, and an ensemble‐based forecasting algorithm to generate forecasts that include uncertainty. Importantly, FLARE quantifies the contribution of different sources of uncertainty (driver data, initial conditions, model process, and parameters) to each daily forecast of water temperature at multiple depths. We applied FLARE to Falling Creek Reservoir (Vinton, Virginia, USA), a drinking water supply, during a 475‐day period encompassing stratified and mixed thermal conditions. Aggregated across this period, root mean square error (RMSE) of daily forecasted water temperatures was 1.13°C at the reservoir's near‐surface (1.0 m) for 7‐day ahead forecasts and 1.62°C for 16‐day ahead forecasts. The RMSE of forecasted water temperatures at the near‐sediments (8.0 m) was 0.87°C for 7‐day forecasts and 1.20°C for 16‐day forecasts. FLARE successfully predicted the onset of fall turnover 4–14 days in advance in two sequential years. Uncertainty partitioning identified meteorology driver data as the dominant source of uncertainty in forecasts for most depths and thermal conditions, except for the near‐sediments in summer, when model process uncertainty dominated. Overall, FLARE provides an open‐source system for lake and reservoir water quality forecasting to improve real‐time management.
Highlights
FLARE is composed of water temperature and meteorology sensors that wirelessly stream data, a data assimilation algorithm that uses sensor observations to update predictions from a hydrodynamic model and calibrate model parameters, and an ensemble‐based forecasting algorithm to generate forecasts that include uncertainty
The metrics include: root mean square error (RMSE); CRPSforecast (Continuous Ranked Probability Score for the FLARE forecast); CRPSnull (CRPS for the persistence null forecast); CRPS forecast Skill; bias; and confidence interval (CI) reliability
1‐D hydrodynamic models used for hindcasting aim to predict water temperature within an RMSE of 2°C (e.g., Bruce et al, 2018), so the level of accuracy associated with FLARE future forecasts over 475 days at Falling Creek Reservoir exceeds expectations (RMSE = 1.4°C aggregated across all periods and all depths; Table 1)
Summary
As a result of human activities, ecosystems around the globe are increasingly changing (Stocker et al, 2013; Ummenhofer & Meehl, 2017), making it challenging for resource managers to consistently provide vital ecosystem services (West et al, 2009). Managers of freshwater ecosystems, which have been more degraded than any other ecosystem on the planet (Millennium Ecosystem Assessment, 2005), are seeking new tools to anticipate future change and ensure clean water for drinking, fisheries, irrigation, industry, and recreation (Brookes et al, 2014). In response to this need, near‐term, real‐time iterative ecological forecasting has emerged as a solution to provide stakeholders, managers, and policy‐makers crucial information about future ecosystem conditions (Clark et al, 2001; Dietze et al, 2018; Luo et al, 2011). Specifying all of these uncertainty sources provides both an assessment of confidence in a forecast for managers as they interpret the forecasts for decision‐making, and valuable information for researchers about how to improve forecasts (Berthet et al, 2016; Dietze, 2017a; Morss et al, 2008)
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
