NALMS Conference Paper: Integrating Multi-Source Spatial, and Temporal Data for the Calibration and Sensitivity Analysis of a Water Quality Model in Lake Mendota

Authors: Mehrzad Shahidzadehasadi, Thomas J Mathis, Nguyen T K Linh,
Bui Minh Hoa, Wei-Cheng Wu, Tran Duc Kien, Paul M Craig
Dates: November 5-8, 2024
Location: South Lake Tahoe, CA/NV

This study conducts a sensitivity analysis and calibration of a three-dimensional water quality model for Lake Mendota, a eutrophic lake in Madison, Wisconsin, using EFDC+ software. Lake Mendota exhibits distinct seasonal stratification patterns, with summer stratification and winter inverse stratification under ice. The calibration and sensitivity analysis were conducted by three datasets, including:

  • The first dataset provided spatially distributed surface sonde measurements from 35 randomized grid points across four summer seasons, offering insights into spatial variability and helping evaluate the model's sensitivity in different parts of the lake.
  • The second dataset contained depth profiles from 0 to 20 meters, capturing water temperature, dissolved oxygen and water quality parameters. This dataset helped in assessing the model's ability to accurately represent lake stratification and vertical gradients.
  • The third dataset offered continuous measurements from an instrumented buoy located in a deep part of the lake, supplying minutely data on the lake's temperature and water quality. This dataset was key to understanding the lake's dynamics over time, allowing for a thorough examination of the model's temporal accuracy.

A variety of parameters were examined to understand their impact on the model's accuracy, with particular attention to algal kinetics, nutrient sediment flux, sediment diagenesis, and light attenuation. The sensitivity analysis identified that the maximum growth rate for phytoplankton, and basal metabolism rate were especially sensitive. Furthermore, parameters affecting sediment diagenesis, such as the diffusion coefficient in porewater, sediment thickness, and the sediment burial rate, were found to be critical to the model's performance.

Let us know if you have questions or interest in this topic. We hope to see you at the conference.

← Back to Blog