DRAFT: Modeling Flow-Salinity Relationships in the Sacramento-San Joaquin Delta Using Artificial Neural Networks
 

Nicky Sandhu
Ralph Finch
Francis Chung

SUMMARY AND FINDINGS

A fast, accurate model to estimate salinities in the Sacramento-San Joaquin Delta, given flow inputs, is an important tool but not an easy one to develop. Uses for such a model include estimating Marginal Export Cost (MEC, also known as carriage water); a replacement for the MDO routine in the statewide planning model DWRSIM; realtime flow/salinity estimation; and reservoir release optimization studies. Attempts at developing such a model have been made over many years with less than full success. A fairly recent mathematical/programming technique known as Artificial Neural Networks (ANNs) was applied to the problem with considerable success. ANNs offer several advantages over previous methods: they allow multiple, arbitrary inputs, they easily allow "memory" to be incorporated, they are not confined to pre-determined impulse-response function shapes. ANN models are developed by first calibrating the internal coefficients of an ANN with sequences of flows and salinities at a location of interest. Once calibrated, new flow inputs are provided and estimated salinities produced.

There is a strong need to "model a model" of the Delta; in other words, to have available an ANN calibrated on the salinity output of another numerical model such as the DWR Delta Simulation Model (DWRDSM). Ideally, DWRDSM would be incorporated into DWRSIM directly, but this is impractical because of tremendous differences in running time between the two models. Instead, a faithful, fast imitation of DWRDSM can be developed using ANNs and used in DWRSIM.

ANN models were developed at several western and interior Delta locations, and preliminary studies performed about a variety of topics. Major findings are:

  1. Multiple flow inputs, as opposed to a single, lumped flow parameter such as Net Delta Outflow, provide a significant increase in accuracy of not only salinity estimates given flows (Figure S-1, S-2), but also the inverse problem, that is, to estimate a sequence of flows to meet a required salinity standard (Figures S-3, S-4). It is especially important in the interior Delta to model salinities using separate flows such as the Sacramento and San Joaquin Rivers and Banks and Tracy pumping, instead of lumping them into a single parameter. Furthermore a single-input model must by definition declare the Marginal Export Cost to be zero, at times a significant error in our opinion.
  2. Marginal Export Cost (carriage water) exists, and is not a trivial or negligible quantity. It is highly variable, depending on controlling salinity location, duration and quantity of through-Delta flow, and current and past hydrology. It can range from -100% (Emmaton controlling), to +100% or greater (Rock Slough and Clifton Court Forebay), but typical ranges would seem to be 10% to 30% at Jersey Point, Rock Slough, and Clifton Court Forebay if they are assumed to be the only controlling stations. In other words, the increase in pumping causes salinity to change at that station. In order to bring the salinity back to the historical level, the amount of flows needed in excess of the increase in pumping were estimated with the ANN. The Marginal Export Cost increases moving from the western Delta landward to the interior Delta and the export pumps (Figure S-5, S-6, S-7, S-8).

Further work needs to be done to incorporate barriers and gates in the ANN, and to estimate the amount of carriage water that occured historically.

The full-length version of this report is available by mail; contact Ralph Finch for more information.

Back | Home Available Studies Computer Models Reports

Section Home Pages

Site Map
Hydrology &
Operations
Delta Modeling Computer Assistance

Disclaimer
Comments or Questions
Webmaster: Tawnly Pranger, phone (916) 653-8698

URL http://modeling.water.ca.gov/delta
Last modified: September 30, 2002.

Copyright 2000.
California Department of Water Resources.
All rights reserved.