WSP11706

• File Number: 8271.300
• Description: Colorado River Basin Salinity Control Program - General Information and Publications-Reports
• Basin: Colorado Mainstem
• Water Division: 5
• Date: 6/1/1980
• Title: Development of Procedures to Evaluate Salinity Management Strategies in Irrigation Return Flows
• Water Supply Pro - Doc Type: Report/Study

c,.,)
c,.,)
U1
...;J

out the ~imulation period, was
estimated from the average
salinity level of the base flows
of the streams within the sub-
basin.

Melamed (1975) assumed a lumped "source
s ink" term to represent soil-salt inter-
actions in both the dissolution and precipi-
tation processes. He also assumed that the
rate of the process was directly proportional
to the difference between the concentration
of the surrounding soil water solution, C,
and some equilibrium concentration, R, for
which the rate was zero. In equation form,
the rate of the process is

fn(c,x,t)

(3,4)

K (R - C)

in which

K is a proportionality coefficient re-
lated to soil properties and salt
composition.

Considering TDS as the salinity indicator,
Riley and Judnak (1979) assumed that, in
extensive areas of the Upper Colorado River
basin where the percolating water contacts
saline marine shales underlying the soil, the
volume of salt pickup is proportional to the
volume of percolating water. By inference,
then, this assumption implies that salt
pickup is inversely proportional to irriga-
tion efficiency.

The various assumptions in the detailed
hydrosalinity models discussed in Table 3.4
are made to represent the complex solute
reactions occurring in the soil-water system.
The concentration of. salts in the drain out-
flow is based on assumed chemical equilibrium
conditions in the soil-water system. Shaffer
et 81. (1977) further assumed that no chemi-
cal reactions occurred in the outflow drains.

Limitations of the
EXIsting Models

From the above overview of the basic
relationships used in hydrosalinity models,
the three relationships concluded to be the
most limiting in accurate model representa-
tion are the chemical processes, surface-
groundwater interrelationships, and the salt
pickup phenomena. While better data are
needed to establish just how limited the
models are, some theoretical and empirical
evidence follows.

Chemical processes

The common assumption that total dis-
solved solids is a conservative parameter
may not apply for a wide range of values.
For example, under large fluctuations in
loading conditions (of the order of 10,000
mg/l), several significant mineral constitu-
ents may reach saturation. Precipitation and

dissolution phenomena within the soil profile
and along the stream ma~ sign~fican~~y alter
the proportions of varIOUS IOnS, ~nd as a
result the relationship between constituent
concentrations and the TDS or EC may become
nonlinear beyond a certain range of concen-
trations.

Although TDS concentratons are ade-
quately simulated by the detailed salinity
models (Ayars 1976), the concentrations of
certain individual ionic constituents are
not. The inadequate representation of CaS04
-CaC03 - Ca(CHC03)2 system in the models
appeared to have significantly reduced
prediction accuracy as illustrated in the
following situations.
1) The low correlation between observed
and simulated 802- coocentrat ioos in the
models of Thomas et a1. (1971), Narasimhan
(1975), and Ayars (1976) is attributed to
inadequate representation of the CaS04
- CaC03 s ys t ems.
2) Table 3.6 lists three models pro-
posed by Wi llardson et a1. (1979). Although
the most sophisticated of the three models
considered the precipitation, dissolution,
and cation exchange of chemical constituents,
the concentrat ions of CaZ+ and HC03 were
underestimated by as much as 35 percent.
All three models were observed to have
inherent weakness in representing PCOZ -
HC03 - C03 - pH relationships.

Surface-groundwater interrelationships

Complex interrelationships between the
surface water, soil water, and groundwater
often exist. Representation of these rela-
tionships is quite general in most models and
often based on calibrated percentage param-
eters. Huber et al. (1976) used the propor-
tion of canal seepage that returns to the
stream and the proportion of agricultural
return flow that is available for rediver-
s ion.

Most of the TDS models assumed either
ungaged surface or subsurface flows to
achieve mass balance. The resulting freedom
in model calibration could lead to major
misrepresentation of the relative magnitudes
of the component salt loadings within the
system, particularly between seepage returns
and subsurface return flows. The following
situations illustrated the importance of
identifying and accurately quantifying the
surface-groundwater interrelationships.

1) Salinity modeling studies of the
Duchesne River basin conducted by UWRL (1975)
identified that significant recycling of
stream diversions occurred within the basin.
These findings subsequently were supported by
Mundorff (1977).

2) Weston (1975) identified groundwater
as the primary agent of salt pickup and
transport to the Colorado River from the

29