Laserfiche WebLink
<br />South ~,4 54undwater Model Development Report <br /> <br />Oct. 11, 2001 <br />Page 10 <br /> <br />Banta, ] 993). The well log indicated that the aquifer was composed of J S permeable <br />sandstone units with intervening shale stringers. HRS (1997) provided hydraulic <br />conductivity (K) values for each of the permeable units and vertical conductance values <br />(VCONT in MODFLOW parlance) for each of the shaley interbeds. (A sensitivity <br />analysis was undertaken to address uncertainties within the expected magnitudes for <br />VCONT as described in section 4.I.S.) We utilized this sandstone-shale sequence as a <br />representation of typical vertical heterogeneity for the Upper Arapahoe aquifer. <br /> <br />') <br /> <br />To develop vertical heterogeneity for the entire simulation domain of the IS-layer model, <br />we needed to specifY a IS-layer K and VCONT sequence for every corresponding areal <br />"cell" (row and column in MODFLOW jargon) in the single-layer model. Our selected <br />approach involved randomizing the Castle Pines Well A3 layering characteristics (for <br />both K and VCONT; see Fig. 4.1) for each cell. To ensure that the layered model was <br />comparable to the single layer model, we scaled the K values of the permeable layers so <br />that the integrated transmissivity of the layered model equaled the transmissivity for each <br />cell in the single layer model. Thus the resulting IS-layer model exhibited a K <br />distribution with areal heterogeneity for each layer that significantly deviated from the K <br />distribution of the single layer model, but the areal distribution of transmissivity was <br />equivalent between the two models. Furthermore, the thickness of each layer in the 1S- <br />layer model varied from cell to cell such that the total thickness at each areal location was <br />equal to the layer thickness in the single-layer model. <br /> <br />) <br /> <br />In addition to this randomized heterogeneity IS-layer model, we performed additional <br />model runs using a IS-layer model where the K distribution for each layer echoed the K <br />distribution of the single layer model. Vertical conductance values for these models were <br />the same as those used in the randomized-K IS-layer model. <br /> <br />Finally, we developed a three-layer model wherein the total thickness of the three layers <br />was the same as the single-layer model thickness and the K distribution for each layer <br />was the same as that for the single layer model. The VCONT values between model <br />layers were taken as the harmonic average of the vertical conductance values from the <br />Castle Pines Arapahoe Well A3 as reported by HRS (1997). Results for the three-layer <br />model are not included here. <br /> <br />For all layered models (both IS-layer and 3-layer), the boundary conditions and total well <br />pumping stresses were the same as those employed for the single-layer model. <br /> <br />J <br /> <br />Hydrosphere Resource Consultants <br />1002 Walnut Suite 200, Boulder, CO 80302 <br />PO Box 44S, Socorro. NM 87801 <br />