Three-dimensional finite element modeling of groundwater flow and solute transport for the lower indus Basin, Pakistan

A relatively thin layer of fresh groundwater exists over a dense saline layer in the study area. A scientific approach is required to harness this fresh water from this layer. Improper well design and indiscriminate pumping might lead to the saline water intrusion from the native dense saline layer. The partially penetrated wells known as skimming wells have been effectively used to restrain the upward movement of the underlying saline water into the overlying fresh zone. The skimming wells not only supplement the scarce irrigation supplies to meet the crop water requirements at critical times but also control the waterlogging. This study was aimed to develop a numerical model that simulates groundwater flow and solute transport. The ultimate goal was to provide feasible solution for skimming the fresh groundwater from the overlying thin layer without saline upconing from the dense layer in the study area. A three dimensional finite element model (FEMGWST) was developed that simulates the groundwater flow and hydrodynamic dispersion of solute in confined and unconfined aquifers under steady and transient flow conditions. The three dimensional finite element method is relatively complex and requires large memory and more computational time but it is flexible to simulate the field system precisely and effectively. The model is capable of handling free surface moving boundaries, aquifer geometry, aquifer heterogeneity and anisotropy, well configurations and constant head and concentration boundaries. The saline and fresh waters are miscible fluids with different densities and a layer of marginal groundwater quality is formed between them, hence the assumption of a sharp interface is avoided in the model. The model is calibrated against field data collected at different agricultural farms located in the Khairpur district in the lower Indus Basin lying between latitudes of 27o20’42” N and 27o19’23” N and longitudes of 68o32’17” E and 68o35’57” E. The main objective of model calibration is to minimize the spatial and temporal difference between observed and model predicted results. The six statistical indices Adjusted R2, mean absolute error (MAE), root mean squared error (RMSE), Nash-Sutcliffe efficiency or model efficiency (ME), BIAS, and index of agreement (d) were employed to evaluate the goodness of the model simulation for groundwater flow and solute transport. The values of these statistical performance indices showed that the overall model performance for groundwater flow and solute transport mirrors closely that of the corresponding observed data. The calibrated model was next used to assess the impacts of different well configurations and boundary conditions such as variation in the pumping rate, tube well operation time, number of well strainers, horizontal distance between well strainers, thickness of fresh saline water interface, well strainer length, water levels in the Rohri and KFE canals and their impact on local groundwater levels and salinity of the surrounding area. It was observed that the well pumping rate is the dominant factor in controlling the waterlogging rather than the fluctuation in canals water level and well screen length. The simulated groundwater salinity was assessed in terms of the temporal variation in the quality of the pumping water and groundwater salinity at the bottom of the well. It is observed that the saline water intrusion into the fresh groundwater layer is directly related to the well discharge and the intermittent pumping. The depth of useable groundwater below the bottom of the well is a key parameter to suppress the salinity mound developed at the bottom of the well. The model results suggest that the multi strainer wells could effectively suppress the development of salinity mound compare to single strainer wells. However, when the thickness of fresh saline water reduces to 4 m then the quadruple strainer wells can induce the salinity and a salinity mound can develop. From the evaluation of different scenarios it is concluded that strainer spaced at the distance of 12 m offer better performances than those spaced at 8, 4 and 2 m distances to reduce pumping water salinity in all configurations, except for the quadruple strainer well with fresh saline water interface of 4 m. The results of this study provide guidelines to the farmers, managers, scientists, and engineers to adopt appropriate sustainable groundwater development policy and skimming well design to harness the fresh water overlying saline water for irrigation without deteriorating the quality of land.

Preview PDF FK 2012 22R.pdf Download (845kB) | Preview

Actions (login required)

View Item