WSP05265

• File Number: 8170
• Description: Arkansas Basin Water Quality Issues
• State: CO
• Basin: Arkansas
• Water Division: 2
• Date: 2/1/1998
• Author: Goff Lewis Person Ko
• Title: Simulated Effects of Irrigation on Salinity in the Arkansas River Valley in Colorado
• Water Supply Pro - Doc Type: Report/Study

In these scenarios, flow was maintained in the Fort Lyon Canal to
satisfy downstream inigation demands outside of the study area. In
the four scenarios, average ground water salinity decreases ranged
from 4.610 25% CTable 3). The decrease in ground water salinity that
resulted in scenarios 1 to 3 was related to the decreased irrigated
acreage and the historically applied irrigation water in those subar-
eas. The largest decrease in ground water satinity was observed when
irrigation ceased in all three sub-areas (scenario 4). In scenario 4,
average monthly ground water salinity decreased from 2180 to
1630 mg/L (Table 3). For all scenarios, ground waler salinily
decreased steadily for about 12 years following the beginning of the
simulation period and then approached steady-state conditions
thereafter (Figure 12). This steady-state condition is coincident
with the simulated aquifer residence time of 10 to 15 years as
reponed by Konikow and Person (1985). The decreases in river salin-
ity were related to the magnitude of the decreases in irrigation in sce-
narios 1 to 4 (Table 3), and reached an average maximum decrease
of 4.4%. Salinity decreased in the river because of a decrease in the
quantity and salinity of return flows from the aquifer.
Water levels were relatively insensitive to changes in irri-
gated acreage. Although the river experienced a net gain in flow
between the upslream and downstream ends of the study area, the
decreases In net gams 1Il streamt10w ranged from 7.9 to 64% in the
four scenarios (Table 3). Scenario 4 had the largesl effect on
streamflow gains, decreasing the gains from 0.18 to 0.06 ml/s.
In the preceding four scenarios, the effect of the hydraulic head
in the canal was large, and leakage from the canal was a major com-
ponent of aquifer recharge. Based on model calculations, average
daily recharge to the aquifer from the canal was about 0.17 ml/s.
The canal is 193 kIn long, has a sandy bottom, and delivers water
to about 37,500 hectares of irrigated land (Gronning Engineering
Co. 1993). In the model simulations, 5.5% of the water diverted intc
the canal was applied to 1993 hectares in the study area. Therefore,
most of the water diverted from the river into the canal flows
through the study area for delivery to downstream irrigators.
Consequently, regardless of the modeled scenarios, the large vol-
ume of water remaining in the canal provides a substantial aITIount
(about 22%) of recharge to the aquifer underlying the study area.
A simulation was perfonned to determine the consequence on
the water quality and quantity of removing all the waler from the
canal. Simulation results were quite different when the scenario of
100% decrease in irrigated acreage was run with no water in the
canal, thus eliminating recharge to the aquifer from the canal. This
scenario duplicales conditions that could be associated with the trans-
fer of all w3ter from an entire irrigation system, which could take

.. .:""l

place jf all water rights in a canal system were purchased and that
water was transferred to other uses (e.g., munic~al or industrial).
In thai case, Ute transferred waler typically would be divened from
the river upstream from the study area. Therefore, the transferred
water would not represent a recharge source for the aquifer. In the
scenarios of 100% decrease in irrigated acreage with and \vithou[
flow in the canal, average monthly streamflow gains decreased 64
and 86%, respectively (Table 4). Average monthly water levels in
the aquifer declined aboUI 0.09 and 0.18 m, respectively CTable 4).
This degree of change in the average water level may be the result
of the model assumption that transmissivity doesn'l change over
time. In future modeling effons, it may be advantageous to define
transmissivity as a function of the total saturated thickness. Salinity
in Ihe aquifer and river did nOI vary substantially between the sim-
ulations with or without flow in the canal (Table 4). In the model
simulation with flow in the canal, recharge was supplied from
both the canal and the river; whereas, in the model simulation
with no flow in the canal, most recharge Wa'i supplied from the river.
Both sources of recharge have nearly identical salinity.

-.
.

Discussion
Tl1is study demonstrates the nature and magnitude of changes
in flow and salinity that could occur in response 10 the evaluated
changes in water use practices in the study area. It would be of great
value if model resulls from this sludy can be extended to other por-
tions of me lower Arkansas River valley in orde-r to assess changes
in water quality and quantity at the basin scale. Because the data
requirements (e.g., salinity levels, recharge, ground water pumpage,
aquifer characteristics, etc.) for the model are not available for the
entire valley, it would be difficult to construct a similar model over
the entire valley. '.';bile the states of Colorado and Kansas have each
developed basin-wide models that evaluate the effects of posl-com-
pact wells and the winter water reservoir storage program, many
simplifying assumptions were made t<> develop these models and
both models had their shortcomings (Simpson 1996). Furthennore,
neither of the models addresses the issue of water quality.
It is possible to use a simpler model that would require less data
to estimate the effects that changes in irrigation practices could have
on the quantity and quality of water over the entire valley. For exam-
ple, McLin (1981) used a generalized lumped parameter model on
the same study area as the Konikow and Bredehoelt (I 974b) model.
During the March 197110 February 1972 study period, McLin's
results compared favorably with the results obtained by Konikow
and Bredehoeft (l974b). While limitations exist in representing

Table 4
Model Results for Scenarios of Complete De<:rease in Irrigated A-creage With and Without Flow In the Canal
Alluvial Aquifer Arkansas River
Average Monthly Average Monthly Water A\'erage Monthly Average Monthly
Model Run Salinity (mgIL) Level (m above mean sea level) Salinity (mgIL) Streamflow Gains (m3/s)
Base Condition 2180 1224.42 t810 0.176
Complete Decrease in Irrigated Acreage
With Canal Flow t630 (-25%)] 1224.33 1730 (-4.4%) 0.064 (-64%)
WilhoUl Canal Flow 1650 (-24%\ t224.24 t140(-3.9%). 0.025 (-86%)
INumbers m paTCnl.heses indicate the percent difference belween the modeled base cOlldinon and the indicated scenario.

84