| Author | Gangopadhyay, Subhrendu |
| Call Number | AIT Diss. no.WM-96-05 |
| Subject(s) | Groundwater--Mathematical models
|
| Note | A dissertation submitted in partial fulfillment of the requirement for the Degree of Doctor of
Engineering |
| Publisher | Asian Institute of Technology |
| Abstract | Detenninistic trial-and-error calibration of groundwater flow models have conventionally been used to simulate potentiometric head variation in regional scale aquifers. Though the problem of groundwater flow simulation has been treated detenninistically, the problem is intrinsically stochastic. This is because of the uncertainties that inherently exist in specifying the input parameters to the flow model. In recent years considerable research has been carried out in the context of stochastic subsurface hydrology. Considerable progress has also been made in the theoretical aspects of this field but practical applications have been to some extent limited. The objective of this research was to apply stochastic concepts to a real world regional flow simulation problem to answer the important question, can we improve regional model predictions by stochastic simulation? The question has been addressed considering the variability of potentiometric head and variability of the transmissivity field; and the study has been conducted on the complex multi-aquifer system underlying the Bangkok metropolis and its adjoining provinces in the Lower Central Plain of Thailand. The conceptual model and mathematical model setup for the aquifer system was based on all available information. First the influence of the variability of potentiometric head was considered. Detenninistic simulations were carried out using the flow simulator MODFLOW, and the model was calibrated to a low fidelity level (measured using a model efficiency index). Potentiometric head residuals were then calculated by comparing these simulated potentiometric heads with field head measurements. Analyzing the head residuals for several time steps showed the existence of spatial correlation. This implies that additional information is embedded in these residuals. Geostatistical analysis was then carried out to identify the underlying spatial correlation structure(s) of the residuals for the individual time steps. Using the spatial correlation model for a particular time step, residual potentiometric head was estimated at model grid locations, and the deterministic head estimates were updated as the simulation progressed in time. It was found that potentiometric head predictions and hence the model efficiencies were significantly improved. This concept has been referred to as the combined deterministic-stochastic approach. On the contrary, experiments with incorporating the stochastic component of transmissivity on the regional trend component of the transmissivity field practically gave little improvement for regional potentiometric head predictions. The variability of transmissivity has been considered because it was found through sensitivity analysis to be the most sensitive parameter in this case. Another important issue for deterministic calibration using a procedure such as trial-and-error is the establishment of appropriate calibration targets. Though to circumvent the problem of parameterization associated with three-dimensional distributed parameter real-world groundwater systems, methods such as geological parameterization have been proposed to identify the parameters, trial-and-error calibration of groundwater models are popular particularly among practitioners. Using geostatistical analysis it has been shown that appropriate calibration targets can be derived from field measurements of potentiometric head. The magnitude of these targets can be used as a guideline for model calibration. Thus the deterministic approach to flow simulation should be supplemented with stochastic analysis. The combined deterministic-stochastic approach also signifies the importance of regional groundwater level monitoring. A new method for groundwater monitoring network evaluation and rationalization has been developed. The method combines the statistical technique of principal component analysis, and a ranking scheme to identify wells that should be continued to monitor in near future. As land subsidence is of major concern in the Bangkok area, some of the flow fields generated in the study were also used to investigate land subsidence using an one-dimensional consolidation model. |
| Year | 1997 |
| Type | Dissertation |
| School | School of Engineering and Technology (SET) |
| Department | Department of Civil and Infrastucture Engineering (DCIE) |
| Academic Program/FoS | Water Engineering and Management (WM) |
| Chairperson(s) | Gupta, Ashim Das; |
| Examination Committee(s) | Huynh, Ngoc Phien;Loof, Rainer ;Nachabe, Mahmood H.;Suphat Vongvisessomjai;Vachi Ramnarong;Yeh, William W.G. |
| Scholarship Donor(s) | GTZ, Federal Republic of German; |
| Degree | Thesis (Ph.D.) - Asian Institute of Technology, 1997 |