UNCONFINED AQUIFERS A pumped unconfined aquifer:

Differences between confined and unconfined aquifers: 1. A confined aquifer is not dewatered during pumping; it remains fully saturated and the pumping creates a drawdown on the piezometric surface 2. The water produced by a well in a confined aquifer comes from the expansion of the water in the aquifer due to a reduction of the water pressure, and from the compaction of the aquifer due to increased effective stress 3. The flow towards the well in a confined aquifer is and remains horizontal provided, of course, that the well is fully penetrating one; there are no vertical flow components in such an aquifer

Unconfined Time-drawdown curves have an S shape (kind of):

Three distinct segments of the S shape: 1. Steep early time segment: The steep early-time segment covers only a brief period after the start of pumping(often only the first few minutes). At early pumping times, an unconfined aquifer reacts in the same way as a confined aquifer: the water produced by the well is released instantaneously from storage by the expansion of the water and the compaction of the aquifer. The shape of the early-time segment is similar to the Theis curve. 2. Flat intermediate time segment: The flat intermediate-time segment reflects the effect of the dewatering that accompanies the falling watertable. The effect of the dewatering on the drawdown is comparable to that of leakage: the increase of the drawdown slows down with time and thus deviates from the Theis curve. After a few minutes to a few hours of pumping, the time-drawdown curve may approach the horizontal. 3. Relatively steep late time segment: The relatively steep late-time segment reflects the situations where the flow in the aquifer is essentially horizontal again and the timedrawdown curve once again tends to conform to the Theis curve. Reason behind the S shape: In 1954 and 1963, Boulton used the concept of delayed yield to explain the S shape of the curve. He came up with a explicit formula that replicated all three sections of the curve. However, the function required a Boulton delay index that does not represent any physical phenomena.

Conditions under which the methods from chapter 3 may be used to evaluate the data: 1. The early time drawdown data can be fitted to a Theis type curve provided that data from piezometers close to the pumping well are used. However, the storativity SA computed using this method cannot be used to predict long term drawdown. 2. The Theis type curve can be fitted to the late time data to yield the transmissivity and the Specific yield Sy of the aquifer. The Theis method yields realistic values for Sy. 3. If the pumping data does not show phenomena of delayed water table response (another way of saying delayed yield) then the time draw down curve only follows the late time segment of the drawdown curve. Since the flow pattern around the well is the same as in a confined aquifer, the Theis method or Jacobs method can be used. METHODS FOR UNCONFINED AQUIFERS Assumptions: - The aquifer is unconfined; - The aquifer has a seemingly infinite areal extent; - The aquifer is homogeneous and of uniform thickness over the area influenced by - Prior to pumping, the watertable is horizontal over the area that will be influenced - The aquifer is pumped at a constant discharge rate - The well penetrates the entire aquifer and thus receives water from the entire saturated thickness of the aquifer. Remarks: In practice, it is okay to ignore the effect of water flow in the unsaturated zone on the delayed water table response Air entry phenomena may influence the drawdown The condition for uniform thickness will not be met if the drawdown in large compared to the aquifers original saturated thickness. A correction proposed by Jacob must be applied to the drawdown data in this case: S = s - (s2/2D) where: S = corrected drawdown s = observed drawdown D = original saturated aquifer thickness This correction only applies to the late time drawdown data which fall on the Theis curve.

Unsteady-state Nuemans curve fitting method He developed a theory of delayed water on well defined physical parameters. Treats the aquifer as a compressible system and the water table as a moving boundary. General drawdown solution is a function of distance (r) from the well and elevation head. When considering average drawdown, solution is a function of r alone. Elastic storage, SA, and specific yield ,SY, are treated as constants Nuemans drawdown equation:

Early time drawdown is described by this reduced form of the drawdown equation:

Where:

and Sa = volume of water instantaneously released from storage per unit surface area per unit decline in head ( which is elastic early time storativity) Under late time conditions, Neumans equation reduces to:

Where:

and Sy = volume of water released from storage per unit surface area per unit decline of the watertable, i.e. released by dewatering of the aquifer (which is specific yield)

and beta is defined as:

Where: Kv = hydraulic conductivity for vertical flow, in m/d Kh = hydraulic conductivity for horizontal flow, in m/d for isotropic aquifers, Kv=Kh, and beta = r2/D2 Neumans curve-fitting method can be used if the following assumptions and conditions are satisfied: - The assumptions listed at the beginning of this chapter; - The aquifer is isotropic or anisotropic; - The flow to the well is in an unsteady state; - The influence of the unsaturated zone upon the drawdown in the aquifer is negligible; - Sy/Sa > 10; - An observation well screened over its entire length penetrates the full thickness of the aquifer - The diameters of the pumped and observation wells are small, i.e. storage in them can be neglected. Fully penetrating wells short circuit vertical flow so heads observed in observation wells may not always be the same as the average groundwater heads in a vertical slice of the aquifer. The theory should still hold for observation wells of short screen length, provided that the drawdowns are averaged over the full thickness of the aquifer. Procedure: Construct the family of Neuman type curves by plotting W ( Ua ,Ub, beta) versus 1/Ua, l/Ub, for a practical range of values of beta on log-log paper, using Annex 5.1 (pp.303). The left-hand portion of Figure 5.2 shows the type A curves [W(Ua, beta) versus l/Ua] and the right-hand portion the type B curves [W(Ub, beta) versus 1/uB]; - Prepare the observed data curve on another sheet of log-log paper of the same scale by plotting the values of the drawdown s against the corresponding time t for a single observation well at a distance r from the pumped well; - Match the early-time observed data plot with one of the type A curves. Note the beta value of the selected type A curve; - Select an arbitrary point A on the overlapping portion of the two sheets and note the values of s, t, l/Ua, and W(Ua, beta) for this point; - Substitute these values into Equations 5.2 and 5.3 and, knowing Q and r, calculate KhD and Sa; - Move the observed data curve until as many as possible of the late-time observed data fall on the type B curve with the same beta value as the selected type A curve;

- Select an arbitrary point B on the superimposed sheets and note the values of s, t, l/Ub, and W(Ub,beta) for this point; - Substitute these values into Equations 5.4 and 5.5 and, knowing Q and r, calculate KhD and Sy. The two calculations should give approximately the same value for KhD - From the KhD value and the known initial saturated thickness of the aquifer D, calculate the value of Kh; - Substitute the numerical values of Kh, beta, D, and r into Equation 5.6 and calculate Kv; - Repeat the procedure with the observed drawdown data from any other observation well that may be available. The calculated results should be approximately the same. Remarks - To check whether the condition Sy/Sa > 10 is fulfilled, the value of this ratio should be determined; - Gambolati (1976) (see also Neuman 1979) pointed out that, theoretically, the effects of elastic storage and dewatering become additive at large t, the final storativity being equal to Sa + Sy. However, in situations where the effect of delayed watertable response is clearly evident, Sa<< Sy, and the influence of Sa at larger times can safely be neglected. Example:

Values for Khd and Sa are obtained from equations 5.2 and 5.3.

Using Equation 5.6, the vertical hydraulic conductivity can be calculated.

To make sure that Sa/Sy>10 we calculate the ratio:

and the condition for the ratio is nearly satisfied. Sy is unreasonably low though, so determination of Sy from B curves is questionable. STEADY STATE: When the drawdown differences become negligibly small, the Theim-Dupuit method can be used to calculate the transmissivity of an unconfined aquifer. The Thiem-Dupuit method can be used if the following assumptions and conditions are satisfied: - The assumptions listed in the beginning of this chapter; - The aquifer is isotropic; - The flow to the well is in steady state; - The Dupuit (1863) assumptions are satisfied, i.e.: . 1. The velocity of flow is proportional to the tangent of the hydraulic gradient instead of the sine as it is in reality. 2. The flow is horizontal and uniform everywhere in a vertical section through the axis of the well. If these assumptions are met the discharge to a well pumping an unconfined aquifer can be described by

After integration between r1 and r2 with (r2>r1), this yields

(equation 5.7) which is known as the dupuit formula

Since h=D-s equation 5.7 can be transformed into

correcting the drawdown s - s2/2D with s gives:

This method is the same as the Theim method for a confined aquifer so use the methods in 3.1.1.