Loading...

Development of an Improved Wet/Drying Algorithm for Hydrodynamic Models to Ppredict Water Salinity in a Hypersaline Lake: the case of Lake Urmia

Zarei Beydokhty, Alireza | 2023

52 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 56419 (09)
  4. University: Sharif University of Technology
  5. Department: Civil Engineering
  6. Advisor(s): Safaie, Ammar
  7. Abstract:
  8. Considering the significant impact of salinity and temperature on the life of aquatic ecosystems, the growth of plants and animals, and the health of people; it is necessary to predict the spatiotemporal distribution of these qualitative characteristics in aquatic environments by using precise algorithms through hydrodynamic models. This research presents an improved algorithm for calculating qualitative characteristics of aquatic environments, such as salinity, during the wet/drying processes. The developed algorithm in this study has been incorporated into FVCOM which is a three-dimensional hydrodynamic model. The current algorithm governing the FVCOM model is not capable of accurately predicting the salinity values in cells and nodes that undergo changes between dry and wet states during the wet/drying processes. In fact, when there is a significant difference between the initial salinity of a lake and the salinity of inflowing rivers, the FVCOM model encounters difficulties in accurately estimating the salinity values of coastal nodes and cells. Therefore, in this study, a new algorithm compatible with saline and hypersaline aquatic environments has been developed. In the first step, an incomplete quadrilateral pyramid reservoir has been designed to evaluate the accuracy of the developed algorithm. The reservoir had brackish water and its only input was a river with fresh water which entered the reservoir and caused the wet/drying processes over time. The output salinity results from the new algorithm have been compared with the former algorithm and manual calculations, and various statistical indices indicate that the new algorithm significantly outperforms the previous one in simulating salinity. For example, using the new algorithm has led to an increase in the R² value from 0.30 to 0.97 and a decrease in the RMSE value from 132 PSU to 19 PSU in the model. Furthermore, through sensitivity analysis on various factors such as mesh resolution, internal and external time steps, depth of viscous sublayer, and inflowing river discharge, the accuracy and stability of the model have been assessed. The results of the sensitivity analysis show that increasing the mesh resolution, reducing the internal and external time step, decreasing the viscous sublayer depth, and reducing the inflowing river discharge all contribute to enhancing the modeling accuracy and bringing the output results closer to manual calculations. After the successful implementation of the new algorithm on the designated reservoir, a three-dimensional model has been developed for Lake Urmia to assess the performance of this algorithm. Bathymetric data, precipitation, evaporation, long and short wavelengths, relative humidity, temperature, air pressure, and the characteristics of inflowing rivers were considered as the model inputs for three-month simulations from 2/18/2016 to 5/18/2016. The salinity output results were compared with the measured values at six different points within the lake which indicated an enhancement in the simulation of salinity variations during the occurrence of wet/drying processes in Lake Urmia. For instance, using the new algorithm has led to an increase in the R² value for the water salinity near the causeway from 0.25 to 0.99 and a decrease in the RMSE value from 86 PSU to 16 PSU. The developed algorithm in this study can be applied to other qualitative characteristics such as temperature, chlorophyll, bacteria, etc., and can also be incorporated into other hydrodynamic models
  9. Keywords:
  10. Hydrodynamic Modeling ; Flamingo ; Urumieh Lake ; Finite Volume Ocean Comuntty Model (FVCOM) ; Wet/Drying Process

 Digital Object List

 Bookmark

No TOC