English abstract
Pumping test is commonly used in estimating aquifer properties such as hydraulic conductivity
(KH), anisotropy (KH/KV), specific yield (Sy), and specific storage (Ss). Analytical solutions with
curve fitting software, such as AQTSOLVE, are used for matching the observed drawdown
curves in one or several observation wells including the records in the pumping well. These
analytical solutions often simplified the natural site conditions by making various assumptions
on initial and boundary conditions. The objective of this study is to estimate the effect of these
simplified assumptions on estimated aquifer properties.
The study used the pumping test records collected in Muscat International Airport premises.
They include drawdown recorded in the pumping well (PW-8) and in two monitoring wells
(MW-19 and MW-20) located 10 m and 40 m away from the pumping well. The pumping test
was conducted for 24 hours with a constant pumping rate of 10.04 L/s. The drawdown curves
also included records from the recovery test, which was started immediately after the pumping
test was stopped and lasted for 12 more hours.
Neuman (1975) and Moench (1997) solutions were used with AQTSOLVE software to obtain a
unique set of aquifer properties. These analytical solutions were however unable to include the
effect of the aerial recharge and the initial hydraulic gradient. Therefore, MODFLOW numerical
model was used to constrain the parameters considering all possible initial and boundary
conditions. The three-dimensional groundwater flow model was calibrated and verified by using
observed drawdown records in steady and transient states. Both, auto-calibration and trial and
error techniques were used for constraining the parameters. Uncertainty analysis was conducted
to understand the effects of pumping rate and the constant head boundary on hydraulic
conductivity estimations and hydraulic head variations. Furthermore, a sensitivity analysis was
conducted to understand the significance of aquifer properties in regulating groundwater flow in
the study area. The estimated parameters when averaged in three wells for the horizontal
hydraulic conductivity, vertical anisotropy and recharge rate are 7.8 m / day, 225.7 and 0.005 m /
day, respectively. Moreover, it was found that the uncertainty of the estimated aquifer properties
can be higher due to probable errors in assigning the hydraulic head for constant head boundary.
Such uncertainty due to erroneous measurement of pumping rate remained minimum due to the
reliable procedure practiced during the pumping test. In addition, sensitivity analysis results
show that the hydraulic conductivity is the most significant parameter in the simulation of
groundwater flow. Also, it was found that the sensitivity of the aerial recharge is small in this
site due to low vertical hydraulic conductivity. For the practical purpose of lowering the water
level in the airport area, a relationship was estimated between the pumping rate and the
drawdown in the pumping well. Accordingly, it was estimated that a 1 m drawdown level can be
achieved by changing the pumping rate to 14.3 L/sec.
In conclusion, analytical solutions were found to be less reliable in evaluating aquifer parameters
when used complex site conditions such as aerial recharge and strong hydraulic gradient. In such
conditions numerical modeling is recommended to be used in simulation of groundwater flow
and in estimating aquifer properties.
