U.S. flag

An official website of the United States government


Main content area

Automatic calibration and improvements on an instream chlorophyll a simulation in the HSPF model

Lee, Dong Hoon, Kim, Jin Hwi, Park, Mi-Hyun, Stenstrom, Michael K., Kang, Joo-Hyon
Ecological modelling 2020 v.415 pp. 108835
algal blooms, ammonia, chlorophyll, data collection, equations, eutrophication, lakes, models, nitrates, nutrients, phosphates, phytoplankton, prediction, rivers, sediments, streams, summer, surface water, water quality, water temperature
Accurate prediction of chlorophyll a (Chl a) concentration in surface water bodies such as lakes or rivers is crucial for water quality management. This study improved the predictive simulation of instream Chl a with the Hydrological Simulation Program-FORTRAN (HSPF) by adding automatic calibration and modifying the growth-temperature formulation of phytoplankton in the original HSPF model. A total of 62 model parameters, selected from a series of sensitivity analyses, were automatically calibrated in a stepwise manner for different variables in the order of flow, sediment, water temperature, ammonia/nitrate couple, and phosphate/Chl a couple. With finer temporal resolution (5–8 days) data than those of majority of the existing HSPF studies, the automatic calibration procedure provided the model with performance ratings of ‘satisfactory’ or better for all the variables including nutrients and Chl a: The percent bias values ranged from -18% - 54% and -20% – 62% for nutrients and Chl a, respectively. The original linear equation on the growth-temperature relationship of phytoplankton in simulating instream Chl a was modified using a quadratic equation and an exponential equation. The exponential equation outperformed the original linear and quadratic equations, particularly in simulating the excess concentrations of Chl a observed during summer seasons. For the validation data set, the exponential equation predicted 78% of the eutrophic cases while the linear and quadratic equation only predicted 53% and 13% of the eutrophic cases, respectively. The modified HSPF model offers an improved prediction of instream Chl a. This approach will be useful for providing early warning of algal blooms, facilitating the implementation of effective management of stream water quality.