Premium
On the Estimation of Stream Flow Depletion Parameters by Drawdown Analysis
Author(s) -
Christensen Steen
Publication year - 2000
Publication title -
groundwater
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.84
H-Index - 94
eISSN - 1745-6584
pISSN - 0017-467X
DOI - 10.1111/j.1745-6584.2000.tb02708.x
Subject(s) - drawdown (hydrology) , aquifer , hydrogeology , flow (mathematics) , sensitivity (control systems) , environmental science , stream flow , hydrology (agriculture) , mechanics , soil science , groundwater , geology , geotechnical engineering , engineering , physics , drainage basin , cartography , electronic engineering , geography
Hunt (1999) derived a model for computation of stream flow depletion caused by pumping from a well near a stream that does not fully penetrate the aquifer. The model input includes the transmissivity (T) and the storativity (S) of the aquifer and the conductance of the stream bed (λ). We study the experimental and hydrogeological conditions for which drawdown analysis can be expected to produce T, S, and λ estimates that can be used in the model of Hunt (1999) to predict depletion with a specified accuracy. The study is conducted by using a combination of sensitivity analysis and uncertainty analysis. The analysis shows that it is always important to have an accurate estimate of λ in order to accurately predict depletion. Recommendations for the design of a pumping test near a stream are given to achieve this. The pumping well should be located close to the stream. Accurate drawdown measurements should be made both near the pumping well and near the stream, and the measurements should be used simultaneously to estimate T, S, and λ. The duration of the pumping test should be relatively long in order to obtain an accurate estimate of λ. A methodology is proposed in which a desired accuracy of either the estimates of T, S, and λ, or the stream flow predicted from these estimates, is used to estimate the duration of the pumping test. Some examples indicate that in many cases the duration of the test should be from an hour to one or two months. However, if S is large (∼10 −1 ) and λ is small (10‐ 7 m/s) then the duration should be from several months to years. This would be expensive and not practical, so other methods should be used to estimate λ in this situation.