Enhancing groundwater flow information by understanding of borehole flow

The management of groundwater resources has increasingly become an important task in hydraulic and environmental engineering. For the prediction of any intervention in a groundwater regime the use of numerical models for studying subsurface flows is a common practice in the multidisciplinary frame of hydrology, hydrogeology and water resources engineering. Although modern computers are very powerful and capable of describing the relevant groundwater currents with a precision greater than ever, the main problem consists still in how to calibrate theoretical models. It is obvious that the quality of any numerical model depends on the contribution of the input data. Uncertainties resulting from an incorrect measurement in the aquifer which is done to get fitting parameters for whichever model, carry imperfection to the model itself. Nowadays borehole observations and pumping tests are recommended to be the most carrying procedure to scope the property and capacity of an aquifer, i.e. to provide the most reliable method of obtaining hydraulic characteristics of aquifer systems. In contrast to established technologies like the tracer test, the new (single) borehole observation techniques have not only the advantage to be used easier, faster and cheaper but merely by using them one is able to examine the real flow process in the aquifer locally. The scientific foundation of employing physically reliable data for the close range around a well is not too dependable at the moment. One has to notice that the analysis of experimental borehole data is quite complicated not only for the fact that it is not possible to calculate an equivalent permeability by a single arithmetic mean but additionally by the real state of the flow, e.g. steady, uniform, single-phase, radial or transient. Analytical models are used most frequently for borehole test analysis, although the well-flow equations are always based on a relatively long list of simplifying assumptions. The applied assumptions are less restrictive, when numerical models are used instead, but the construction of such models and the errors associated with numerical techniques are usually involved.