The MODSIM river basin network flow model is extended to directly incorpora
te constraints on concentrations of conservative water quality constituents
. The extended model MODSIMQ is linked with the U.S. Environmental Protecti
on Agency's QUAL2E streamflow water quality model and a model for estimatin
g quality of irrigation return flows. An iterative procedure based on the F
rank-Wolfe nonlinear programming algorithm links MODSIMQ and the water qual
ity models to assure convergence to solutions satisfying water right priori
ties, while attempting to maintain minimum water quality requirements. Irri
gation return flows, canal seepage, reservoir seepage, deep percolation, an
d river depletion due to groundwater pumping are modeled using stream deple
tion factors from the U.S. Geological Survey. Application of MODSIMQ to the
lower Arkansas River basin in Colorado successfully models the complex leg
al and administrative issues under Colorado water law and the Arkansas Rive
r Compact, including the many water exchange mechanisms governing use of of
f-stream reservoirs in the basin. Model calibration exercises conducted for
the case study area confirm that MODSIMQ reasonably reproduces both histor
ical flows and salinity levels for the calibration period. Results from var
ious management scenarios indicate that appropriate conjunctive use of surf
ace and groundwater can simultaneously satisfy water demands for users whil
e enhancing control of salinization.