Recalibration of the Lake Okeechobee Water Quality Model (LOWQM) to extreme hydro-meteorological events

• Three hurricanes and two major droughts affected Lake Okeechobee from 2004 to 2012.
• Effects included dramatic changes in water quality and phytoplankton.
• The LOWQM model was recalibrated to this period of time.
• Results were similar to original simulation despite differences in nutrient fluxes.
• These similarities suggest changes in model equations will improve results.

The Lake Okeechobee Water Quality Model (LOWQM) was developed using an enhanced version of WASP5 to improve understanding of nutrient cycling – specifically total phosphorus – and to assess lake-wide averaged nutrient and algal responses to management alternatives. Previously, the LOWQM was calibrated/validated to hydro-meteorological and monitoring data from 1983 to 2000. Subsequently, a series of extreme hydro-meteorological events, including three hurricanes and two major droughts, occurred from 2005 to 2008 that were outside the range of observations in this original LOWQM model calibration (original simulation). These events affected the light environment, nutrients and phytoplankton communities in the lake. The LOWQM was recalibrated and revalidated (new simulation) to the time period 1983–2012. When compared to the original simulation using the same set of forcing functions (e.g. flow, rain, temperature, nutrient loads, and resuspension of sediments), there were some improvements in model predicted nitrogen and phytoplankton, based on goodness-of-fit measures to observed data. Major differences between the two simulations were found in nutrient cycles. The differences in these cycles and the similarity of goodness-of-fit measures between the new and original simulations indicate the importance of sediment-water interactions and the effect that model structure (e.g. equations) have on the ability to improve calibration results. Further evaluation of these two simulations suggest that model accuracy can be improved through changes of model segmentation, additional lake sampling of biological processes simulated by the LOWQM, and revising model equations to allow for variation in algal nutrient to carbon ratios and uptake rates of nitrogen.