At noon on April 28, 1997, the USGS released 48 liters of dye
from the Mossdale railroad bridge overlooking the San Joaquin River. Eight
water quality samplers, located at eight different sites in the South Delta,
were used to track the movement of dye. This dye release was simulated
using the Delta Simulation Model 2 (DSM2) and the results, comparing computed to
observed, are shown in this chapter.
The simulation was conducted using the quality portion of
DSM2. The flows and Delta geometry configuration were from the DSM2 1997
hydrodynamics validation, which can be found at http://wwwdelmod/docs/dsm2/calval/valid.html.
During the time frame of the study, the Old River at Head barrier was in
place. It contained two culverts with a capacity of passing approximately
300 cfs. Forty-eight liters of tracer, over a fifteen-minute period, were
released into the San Joaquin River at the DSM2 grid location corresponding to
the Mossdale site. The dye concentration was observed at the eight
locations shown in Figure 3-1 and compared to model results reported at the same
Figure 3-1: USGS Dye Measurement Sites
Discussion of Results
Figures 3-2 through 3-9 show the concentration plots. From these
graphs, travel time, dispersion, and concentrations were analyzed for each
observation site. General comments comparing the model's results to
observed data are shown in Figure 3-1.
There was a strong match between the travel time of the simulated tracer and
the observed data. The timing of the peaks between observed and computed
were within a couple of hours. (This is excluding the Old River at CC
Ferry site, which did not have accurate observed data.)
This study was particularly helpful in showing how well DSM2 models
dispersion in various areas of the Delta. The quality portion of DSM2 is
calibrated using salinity. Since there are several continuous sources of
salinity, it is impossible to determine the local dispersion effects.
Examining the plots shows that in some areas of the Delta, such as the Stockton
site, the model had greater dispersion. At other sites, like Turner Cut or
San Joaquin River at Mandeville, model dispersion was less.
Differences in concentrations between the model and observed data are a
result of differences in channel velocities, flow splits and/or
dispersion. Additionally, the USGS considers concentrations below 0.04 ug/l
to be background concentrations.
This study demonstrated how well DSM2 modeled a conservative
tracer. In general, the model did very well in simulating the travel time
of the dye. Areas for strongest improvement are in the modeling of
Author: Tara Smith
Back to Delta Modeling Section 1999 Annual Report Table of Contents
Last revised: 2001-09-12
Comments or Questions
Webmaster email to