Integrated hydrologic modeling of a transboundary aquifer system — Lower Rio Grande

Abstract

For more than 30 years the agreements developed for the aquifer systems of the lower Rio Grande and related river compacts of the Rio Grande River have evolved into a complex setting of transboundary conjunctive use. The conjunctive use now includes many facets of water rights, water use, and emerging demands between the states of New Mexico and Texas, the United States and Mexico, and various water-supply agencies. The analysis of the complex relations between irrigation and streamflow supply-and-demand components and the effects of surface-water and groundwater use requires an integrated hydrologic model to track all of the use and movement of water. MODFLOW with the Farm Process (MF-FMP) provides the integrated approach needed to assess the stream-aquifer interactions that are dynamically affected by irrigation demands on streamflow allotments that are supplemented with groundwater pumpage. As a first step to the ongoing full implementation of MF-FMP by the USGS, the existing model (LRG_2007) was modified to include some FMP features, demonstrating the ability to simulate the existing streamflow-diversion relations known as the D2 and D3 curves, departure of downstream deliveries from these curves during low allocation years and with increasing efficiency upstream, and the dynamic relation between surface-water conveyance and estimates of pumpage and recharge. This new MF-FMP modeling framework can now internally analyze complex relations within the Lower Rio Grande Hydrologic Model (LRGHM_2011) that previous techniques had limited ability to assess. less

Full text

Integrated Hydrologic Modeling of a Transboundary Aquifer System —
Lower Rio Grande
Randall T. Hanson1, Wolfgang Schmid2,3, Jake Knight1,3, and Thomas Maddock III4
1U.S. Geological Survey, rthanson@usgs.gov, San Diego, CA, USA
2The Commonwealth Scientific and Industrial Research Organisation, Wolfgang.Schmid@csiro.au,
Canberra, AU
3University of Arizona, jakeknight@email.arizona.edu, Tucson, AZ, USA
4University of Arizona, tm3@hwr.arizona.edu, Tucson, AZ, USA
ABSTRACT
For more than 30 years the agreements developed for the aquifer systems of the lower Rio Grande and
related river compacts of the Rio Grande River have evolved into a complex setting of transboundary
conjunctive use. The conjunctive use now includes many facets of water rights, water use, and emerging
demands between the states of New Mexico and Texas, the United States and Mexico, and various
water-supply agencies. The analysis of the complex relations between irrigation and streamflow supply-
and-demand components and the effects of surface-water and groundwater use requires an integrated
hydrologic model to track all of the use and movement of water. MODFLOW with the Farm Process (MF-
FMP) provides the integrated approach needed to assess the stream-aquifer interactions that are
dynamically affected by irrigation demands on streamflow allotments that are supplemented with
groundwater pumpage. As a first step to the ongoing full implementation of MF-FMP by the USGS, the
existing model (LRG_2007) was modified to include some FMP features, demonstrating the ability to
simulate the existing streamflow-diversion relations known as the D2 and D3 curves, departure of
downstream deliveries from these curves during low allocation years and with increasing efficiency
upstream, and the dynamic relation between surface-water conveyance and estimates of pumpage and
recharge. This new MF-FMP modeling framework can now internally analyze complex relations within the
Lower Rio Grande Hydrologic Model (LRGHM_2011) that previous techniques had limited ability to
assess.
INTRODUCTION
Appropriation of surface water to three major water users in the Lower Rio Grande (LRG) Basin in
southern New Mexico (NM) and western Texas (TX) (fig. 1) is regulated by treaty, compact, and Rio
Grande Project operating rules and agreements. The treaty of 1906 distributes the streamflow of the Rio
Grande between the United States (US) and Mexico (MX). And the Rio Grande Project apportions LRG
surface water released from the Caballo Reservoir between two irrigation districts, the upstream Elephant
Butte Irrigation District (EBID) in NM and the downstream El Paso County Water Improvement District
No.1 (EP1) in TX. Recent allocation methods employed by the U.S. Bureau of Reclamation (BOR),
including an operating agreement between the two irrigation districts (EBID and EP1), tie allocations for
EP1 and MX directly to available project water of the LRG. The operating agreement guides the allocation
of surface-water supply to TX and MX regardless of the effect of groundwater pumping in NM on flows to
the LRG and downstream deliveries. Initial disputes in the LRG Basin centered on the effects of
groundwater pumping and spawned the need for model-based analysis that would assess the complex
distributions, volumes and rates of withdrawals, use, and movement. The disputes forced managers to
acknowledge and analyze the relations between river flows, evapotranspiration (ET), drainage of excess
irrigation water, geologic layering, and agriculture induced supply and demand components.
The treaty with MX apportions the Rio Grande below Elephant Butte reservoir with 60,000 acre-ft annually
allocated to MX. After the Rio Grande Project, the proportion of US flows were divided under an operating
agreement based on the proportion of irrigated acreage in NM and TX, the relation between releases of
water from the reservoir and divertible water, where the diversions included the effect of groundwater
pumping for the period 1951–78 (Valdes and Maddock, 2010). A curve called the “D2” curve projects the
amount of water to the farms head gates for various releases of water from the reservoirs after the MX
allocation with 57 and 43% to EBID and EP1, respectively. Intrinsic to the D2 curve is the groundwater
MODFLOW and More 2013: Translating Science into Practice - Conference Proceedings, Maxwell, Hill, Zheng & Tonkin - igwmc.mines.edu
57
MODFLOW and More 2013: Translating Science into Practice - Conference Proceedings, Maxwell, Hill, Zheng & Tonkin - igwmc.mines.edu

pumping for the period 1951–78 (Valdes
and Maddock, 2010). BOR did not
explicitly use the D2 curve for
allocations from the reservoirs, yet the
deliveries to MX are honored except in
severe drought years. Since about
1999, effects of pumping and increased
irrigation efficiencies have resulted in
reductions in deliveries predicted by the
D2 curve. EBID and EP1 have
developed a new operating agreement
and a new relation called the D3 curve
to project EP1 deliveries. The D3 curve
is based on the Texas portion of the D2
curve. Supply-and-demand conditions
have changed with the effects of
climate, increased pumpage and
irrigation efficiency, and reduced river
conveyance efficiency and return flows.
These changing conditions need to be
assessed through simulation of
conjunctive use with a fully coupled
integrated hydrologic model such as
MF-FMP.
This paper highlights how the update of
an existing groundwater flow model
(LRG_2007) through 2009 and
incorporation of some of the features of
the Farm Process (FMP) (Schmid and
Hanson, 2009) for MODFLOW (MF)
(Harbaugh, 2005) starts to facilitate the analysis needed to address conjunctive use with these changing
conditions. The new Lower Rio Grande Hydrologic Model (LRGHM_2011) incorporated the FMP to
eliminate external preprocessing of irrigation well pumping and recharge. The LRGHM_2011 model
calculates irrigation demand, the allocation of irrigation demand to surface-water diversions and
supplemental groundwater pumping, and recharge as interlinked processes that are internal to MF-FMP
simulation.
The model’s implicit estimation of supplemental groundwater pumping is especially valuable for historic
periods in lieu of metered groundwater pumping data. In addition, irrigation demand can now be
influenced by root-zone uptake from groundwater (herein called head-dependency). Because the water
supply and return-flow budget components are implicitly interlinked with irrigation demand, they are also
affected by the head-dependency of irrigation demand. The LRGHM_2011 model provides a tool that can
assess the effects of existing or proposed operating agreements on irrigation demand, streamflow,
surface-water deliveries, supplemental groundwater pumping, and return flows. This assessment is
essential to assess deliveries to TX and MX, and to test the effects of proposed new appropriations within
NM and new groundwater withdrawals in MX.
PROBLEM DESCRIPTION
Because groundwater pumping is an integral part of the water supply, and because groundwater pumping
affects the surface-water supplies, a hydrologic model was needed to simulate all components of the
water budget in LRG. The necessary components in the Rincon Valley and Mesilla Basin include storage,
flow and the interaction between surface water and groundwater in the alluvial aquifers of these basins.
The New Mexico Office of the State Engineer commissioned a groundwater-flow model which included
modeling of relevant groundwater and surface water interactions (SSPA, 2007) to assist with State and
Figure 1. Regional map of Lower Rio Grande and
extent of Hydrologic Model.!
58
MODFLOW and More 2013: Translating Science into Practice - Conference Proceedings, Maxwell, Hill, Zheng & Tonkin - igwmc.mines.edu

local water management and administration. The LRG_2007 model included explicit groundwater
pumping and recharge and simulates related processes of Rio Grande streamflow capture or return flow
to the river. However, these processes were not dynamically linked within the model and were not
implicitly coupled to time varying changes in irrigation demands, supply-constraints of time-varying
allotments, and aquifer head. LRG_2007 used a spreadsheet-based “Farm Budget Model” external to MF
that calculated farm-well pumping and recharge from inefficient irrigation supply, based upon estimated
irrigation demand, precipitation, efficiency, and surface-water delivery data. The spreadsheet and related
software pre-estimated farm-well pumping and recharge into specified MF Well and Recharge package
flux boundaries. This approach did not model return flow to the Rio Grande and was unable to mimic the
features of the D2 and D3 curves. The irrigation demand estimated from the Farm Budget Model of
LRG_2007 is used in the LRGHM_2011 model as the estimated consumptive use; however, the
deliveries, recharge, pumpage, and return flows from irrigation are internally computed by FMP.
LRGHM_2011 replicates the irrigation demand and on-farm efficiencies without modeling surface runoff
and precipitation. Additional upgrades to LRGHM_2011 include time-varying surface-water allotments,
refined agriculture regions, and revised canal geometries, and additional internal simulation of processes
including delivery demands after diversion, estimated head-dependent root-zone uptake of groundwater
controlled by mapped soil properties, estimated supplemental groundwater pumping to meet irrigation
demand, and recharge. Because all components of the water budget are incorporated in LRGHM_2011,
surface-water deliveries conforming with downstream seepage losses are simulated. The LRGHM_2011
simulates all components of the hydrologic water budget such that flow of all the water not subject to
groundwater underflow, ET discharge or river outflow at the American Canal diversion and El Paso
Narrows is simulated.
SUMMARY
Preliminary results from the LRGHM_2011 model indicate a nonlinear relation between groundwater
pumping and surface-water deliveries for irrigation relative to the streamflow in the Rio Grande at the NM-
TX state line during the growing seasons (fig. 2). The relation between reservoir releases and irrigation
deliveries is likely a function of climate variability, groundwater pumping for supplemental irrigation
demand, and variations in groundwater-surface water interaction. It is also likely to be affected by
additional groundwater pumping demand for public-supply well fields from the Canutillo well field in TX,
Santa Teresa well field in NM, and Conejos-Medanos bateria in MX (fig. 1).
Figure 2. Relation between groundwater pumping and surface-water deliveries for irrigation
relative to streamflow in the Rio Grande at the New Mexico-Texas state line during the growing
season.
Results from the two models can be compared to assess the effects of changes in climate, surface-water
allotments, agricultural conditions, and changes in conveyance efficiency. LRGHM_2011 uses the
59
MODFLOW and More 2013: Translating Science into Practice - Conference Proceedings, Maxwell, Hill, Zheng & Tonkin - igwmc.mines.edu

estimated irrigation demand from the LRG_2007 Farm Budget Model as the estimated consumptive use
for the simulated crop-water demand processes in each of the six service areas. The operating
agreements between EBID and EP1 grandfathered in groundwater pumpage within the operating
agreement of the D2 and D3 curves, however recent years have seen increased water demand and
decreased water supply. Although the water demands are comparable in both models the simulated
estimates of the sum of EBID-diverted and divertible water at the Texas state line from LRGHM_2011
more closely match the historical diverted plus divertible water (fig. 3a).
Figure 3. Graph showing (a) measured and simulated streamflows related to the operating
agreement curves from LRG_2007 and LRGHM_2011 models representing variations in water
supply and water demand, and (b) example of measured and simulated estimates of flows related
to the operating agreement with increased irrigation efficiencies using LRGHM_2011 model.
The difference in the water supply and demand conditions between the historical period that forms the
basis for the operating agreement and recent years suggests that increased demands and increased
irrigation efficiencies combined with reduced supply have reduced deliveries, conveyance, and return
flows. Since the operating agreement is based on released water and not releasable water, these
proportions can be simulated with changing levels of surface-water allotments. The most recent years of
drought (2003-07) demonstrate the effects of decreased surface-water supply and increased irrigation
60
MODFLOW and More 2013: Translating Science into Practice - Conference Proceedings, Maxwell, Hill, Zheng & Tonkin - igwmc.mines.edu

efficiencies on deliveries and divertible water and the resulting increased groundwater pumping that also
reduces the efficiency of river conveyance and direct uptake from groundwater to satisfy total agricultural
consumptive use. If irrigation efficiencies are increased from 65 to 90%, the return flows that contribute to
the downstream divertible water at the state line are also diminished for some years (fig. 3b). This
exemplifies how various crops or regions can be evaluated for aspects of best management practices
such as irrigation efficiencies.
As a first step to the ongoing full implementation of MF-FMP by the USGS, the existing model
(LRG_2007) was modified to include some FMP features, demonstrating the ability to simulate the
existing streamflow-diversion relations known as the D2 and D3 curves, departure of downstream
deliveries from these curves during low allocation years and with increasing efficiency upstream, and the
dynamic relation between surface-water conveyance and estimates of pumpage and recharge. This new
MF-FMP modeling framework can now internally analyze complex relations within the Lower Rio Grande
Hydrologic Model (LRGHM_2011) that previous techniques had limited ability to assess. Thus the revised
model with integrated groundwater-surface-water flows in a supply-constrained and demand-driven
framework of MF-FMP allows for the dynamic simulation of changing conditions and simple assessment
of alternative conditions or management practices.
REFERENCES
Harbaugh, A.W., 2005. MODFLOW-2005, the U.S. Geological Survey modular ground-water model —
the Ground-Water Flow Process: U.S. Geological Survey Techniques and Methods 6-A16, v.p.
Schmid, Wolfgang, and Hanson, R.T., 2009. The Farm Process Version 2 (FMP2) for MODFLOW-2005—
Modifications and Upgrades to FMP1: U.S. Geological Survey Techniques and Methods 6-A-32,
102 p.
S.S. Papadopulos & Associates (SSPA), 2007. Groundwater flow model for administration and
management in the Lower Rio Grande Basin: Consultant report prepared for the New Mexico
State Engineer, 63 p.
Valdes, J., and Maddock III, T., 2010. Conjunctive Water Management in the US Southwest, Chapter 15,
Water and Sustainability in Arid Regions, Bridging the gap between physical and social sciences,
Springer, London.
!
61
MODFLOW and More 2013: Translating Science into Practice - Conference Proceedings, Maxwell, Hill, Zheng & Tonkin - igwmc.mines.edu