Simulation of a long-term aquifer test conducted near the Rio Grande, Albuquerque, New Mexico

A long-term aquifer test was conducted near the Rio Grande in
Albuquerque during January and February 1995 using 22 wells and
piezometers at nine sites, with the City of Albuquerque Griegos 1
production well as the pumped well. Griegos 1 discharge averaged
about 2,330 gallons per minute for 54.4 days. A three-dimensional
finite-difference ground-water-flow model was used to estimate
aquifer properties in the vicinity of the Griegos well field and the
amount of infiltration induced into the aquifer system from the
Rio Grande and riverside drains as a result of pumping during the
test. The model was initially calibrated by trial-and-error
adjustments of the aquifer properties. The model was
recalibrated using a nonlinear least-squares regression
technique.

The aquifer system in the area includes the middle Tertiary to
Quaternary Santa Fe Group and post-Santa Fe Group valley- and
basin-fill deposits of the Albuquerque Basin. The Rio Grande
and adjacent riverside drains are in hydraulic connection with the
aquifer system.

The hydraulic-conductivity values of the upper part of the
Santa Fe Group resulting from the model calibrated by trial and
error varied by zone in the model and ranged from 12 to 33 feet per
day. The hydraulic conductivity of the inner-valley alluvium was 45
feet per day. The vertical to horizontal anisotropy ratio was
1:140. Specific storage was 4 x 10-6 per foot of aquifer thickness,
and specific yield was 0.15 (dimensionless). The sum of
squared errors between the observed and simulated drawdowns
was 130 feet squared.

Not all aquifer properties could be estimated using nonlinear
regression because of model insensitivity to some aquifer
properties at observation locations. Hydraulic conductivity
of the inner-valley alluvium, middle part of the Santa Fe Group,
and riverbed and riverside-drain bed and specific yield had low
sensitivity values and therefore could not be estimated. Of the
properties estimated, hydraulic conductivity of the upper part of
the Santa Fe Group was estimated to be 12 feet per day, the vertical
to horizontal anisotropy ratio was estimated to be 1:82, and specific
storage was estimated to be 1.2 x 10-6 per foot of aquifer
thickness. The overall sum of squared errors between the
observed and simulated drawdowns was 87 feet squared, a significant
improvement over the model calibrated by trial and error.

At the end of aquifer-test pumping, induced infiltration from
the Rio Grande and riverside drains was simulated to be 13
percent of the total amount of water pumped. The remainder was
water removed from aquifer storage. After pumping stopped,
induced infiltration continued to replenish aquifer storage.
Simulations estimated that 5 years after pumping began (about 4.85
years after pumping stopped), 58 to 72 percent of the total amount
of water pumped was replenished by induced infiltration from the Rio
Grande surface-water system.