Laserfiche WebLink
<br />• It is anticipated that the draft TM, described above, for this task will be submitted to the State <br />approximately 2 months after State approval of the Phase 3 model and that the final TM will <br />be ready approximately 4 months after State approval of the Phase 3 model. <br />Task 48.3 Model Configuration and Calibration <br />• The Phase 3 model will include definition of the aquifer configuration, land surface, stream <br />locations and initial estimates of hydraulic properties that will be used as a starting point for <br />the Phase 4 calibration activities. <br />• After development of model stress inputs and boundary conditions under Task 48.1 and <br />defining the calibration targets and approach under task 48.2, the model will be calibrated. <br />The calibration process will use a combination of manual calibration and parameter estimation <br />techniques. It is anticipated that appropriate State personnel will be available for bi-weekly to <br />monthly consultation meetings requested by the Groundwater Contractor during the <br />calibration process so the calibration represents a consensus among the project team. <br />However, the Groundwater Contractor will be running the model. <br />• The initial stage of calibration will involve configuring the model to include average stress <br />inputs and boundary conditions to obtain a stable (steady-state) solution prior to initiating <br />transient calibration. The steady-state model inputs will represent average annual conditions <br />with relatively little change in storage extending over 3 to 10 years selected from the 1990- <br />2002 time period. The simulated heads from the pre-development model created in Phase 3 <br />Task 4$ will be used for the starting water levels. <br />This initial stage of calibration will include selection of an appropriate solver package and <br />convergence parameters that result in a solution with mass balance errors of less than l <br />percent, and average convergence errors of less than 0.5 feet. It is anticipated that the PCG2 or <br />GMG solvers will be utilized for most simulations. It is anticipated that both dry and flooded <br />cell issues may arise in areas of minimal saturated thickness in the model. In these areas the <br />data and conceptual model will be reviewed for accuracy. If warranted, the conceptual model <br />will be revised appropriately. If no changes to the conceptual model are warranted then model <br />input parameters will be modified using a variety of techniques to obtain a valid solution as <br />part of the calibration. Stress periods are anticipated to be one month for all transient <br />simulations; however, due to limited information in the early portion of the study period, a <br />longer stress period may be utilized to decrease computation time. Time steps will be selected <br />that result in a stable solution. <br />• Model calibration will proceed until the stated calibration criteria developed under Task 48.2 <br />are met, or no significant improvement in model fit is obtained. It maybe necessary to revise <br />elements of the conceptual flow model to improve the calibration. Modifications of the <br />conceptual model will be implemented, if necessary, after concurrence of the State. <br />• No model calibration is unique, so a verification stage will be used to compare the calibrated <br />model against a portion of the available targets not used in the actual calibration, using the <br />approach that will be defined in Task 48.2. If the validation comparison is poor, then <br />SPDSS Phase4 gw_SOW 09-15-06.doc 13 <br />