Dynamic calibration of phytoplankton blooms using the modified SWAT model

Abstract

Phytoplankton blooms in natural waters are critical for managing drinking water quality. Thus, estimating and understanding algal systems is important to supply safe drinking water and protect public health. The objectives of this study were to evaluate a modified algal module and calibrate algae-associated parameters annually (i.e., dynamic calibration) with identification of the relationship between parameters and environmental variables. The modified algal module in the Soil and Water Assessment Tool (SWAT) model incorporates temperature rate multipliers to correct the temperature effect on phytoplankton. Three different modules (i.e., original module, modified module, and dynamic calibration) were calibrated using 15-year observations in a watershed that provides drinking water resources. The original module resulted in a Nash-Sutcliffe model efficiency (NSE) of 0.334 and a root mean squared error (RMSE) of 0.578, but the modified module increased up to 0.374 and decreased to 0.561. In addition, the dynamic calibration of the modified module further improved the model performance with an NSE of 0.486 and RMSE of 0.508. The random forest method was used to investigate the relative importance of environmental variables to dynamic parameters. The maximum temperature was a relatively important variable for maximum algal growth, and the nutrient concentration was important for temperature multiplier factors. Dynamic calibration not only improved the simulation accuracy but also provided the relationship between environmental variables and algal parameters. This calibration method will be useful for simulating chlorophyll-a (Chl-a) dynamics and preparing watershed management policies for algal blooms.