4.3.3. Modeling

Modeling during the early RTDF years focused on developing and enhancing DSM2 forecast simulation capabilities for salinity (specific conductance), bromide, and DOC in the Delta and the downstream aqueduct systems [106,121–123] and conducting tool validation [124]. Most of this work was conducted by CDWR's Delta Modeling Section. As modeling tools matured, CDWR's Division of Operations & Maintenance (O&M) became a program partner and began producing water quality retrospective simulations and 21-day forecasts on a routine basis [125]. The retrospective simulations also characterize Delta water quality through fingerprints that quantify different source contributions (e.g., Sacramento River inflow, San Joaquin River inflow, Delta agricultural drainage) at fixed locations [126,127]. The fingerprint technique has also been used to estimate the fraction of Sacramento and San Joaquin River inflows that are wastewater-derived [22]. The program has also routinely produced "seasonal" water quality forecasts for conditions several months into the future. However, based on several years of experience, the utility of these seasonal forecasts has been questioned (due to large uncertainty) and production of these forecasts was discontinued [128].

Participating water contractors commissioned the development of a DSM2 emulator that employed artificial neural network technology, a common machine learning technique [129]. This model was developed to provide an alternate, easy-to-use tool with greatly reduced run times. However, to date, the tool has not been adopted as part of the MWQI suite of forecasting tools.

Concerns about low alkalinity levels in exported Delta waters have been raised by participating MWQI agencies, prompting the program to develop capabilities to simulate and forecast bicarbonate fate and transport in the Delta. While moderately low alkalinity levels can improve the effectiveness of the coagulation process in water treatment plants, especially when aluminum sulfate is used as a coagulant, extremely low alkalinity levels can depress pH levels and can result in TOC re-stabilization [130]. These new forecasting capabilities were implemented in 2021.

Information from MWQI's routine monitoring and modeling efforts are disseminated in a variety of ways. Discrete and real time data are posted in CDWR's Water Data Library and in CDEC (https://cdec.water.ca.gov/ accessed on 8 October 2022), respectively. A daily water quality report, distributed electronically via email, contains a summary table of real time station data, links to water quality forecasts, and a brief synopsis of current events in and around the Delta. The summary table is updated daily and provides mean daily values, seven-day averages and the percent change over the seven days. Finally, monitoring data and model output, along with the daily water quality report, are posted on the SWC Real Time Data and Forecasting web page at http://rtdf.info/ (accessed on 8 October 2022). This web page provides links to CDEC and the daily MWQI water quality report. Figure 4 provides an example of observed and simulated data at Clifton Court Forebay (located upstream of Banks Pumping Plant) that are routinely updated as part of the MWQI program.
