Confined vs unconfined aquifer representation. Technical report 04-020. Scenario document number DDM-016

Design documents are a series of technical papers addressing specific design topics on the eastern Snake River Plain Aquifer Model upgrade. Each design document will contain the following information: topic of the design document, how that topic fits into the whole project, which design alternatives were considered and which design alternative is proposed. In draft form, design documents are used to present proposed designs to reviewers. Reviewers are encouraged to submit suggested alternatives and comments to the design document. Reviewers include all members of the Eastern Snake Hydrologic Modeling (ESHM) Committee as well as selected experts outside of the committee. The design document author will consider all suggestions from reviewers, update the draft design document, and submit the design document to the SRPAM Model Upgrade Program Manager. The Program Manager will make a final decision regarding the technical design of the described component. The author will modify the design document and publish the document in its final form in .pdf format on the SRPAM Model Upgrade web site. Final model documentation will include all of the design documents, edited to ensure that the “as-built” condition is appropriately represented. This document discusses whether the Eastern Snake Plain Aquifer will be represented as a confined or an unconfined aquifer. The decision to use state-of-the-art parameter estimation techniques makes this decision more difficult. The existing IDWR/UI aquifer model (Cosgrove et al, 1999) uses an unconfined representation for the Eastern Snake Plain Aquifer and Garabedian (1992) employed an unconfined representation also. Thus, a choice of an unconfined representation for the Eastern Snake Plain Aquifer would be consistent with tradition. However, a confined aquifer model representation requires fewer computations and behaves more stable numerically than an unconfined aquifer representation. These two factors present significant advantages when using parameter optimization tools that run the model thousands of times to select a parameter suite that minimizes the differences between observations and modeled values