Groundwater availability of the Mississippi embayment

Abstract

Groundwater is an important resource for agricultural and municipal uses in the Mississippi embayment. Arkansas ranks first in the Nation for rice and third for cotton production, with both crops dependent on groundwater as a major source of irrigation requirements. Multiple municipalities rely on the groundwater resources to provide water for industrial and public use, which includes the city of Memphis, Tennessee. The demand for the groundwater resource has resulted in groundwater availability issues in the Mississippi embayment including: (1) declining groundwater levels of 50 feet or more in the Mississippi River Valley alluvial aquifer in parts of eastern Arkansas from agricultural pumping, (2) declining groundwater levels of over 360 feet over the last 90 years in the confined middle Claiborne aquifer in southern Arkansas and northern Louisiana from municipal pumping, and (3) litigation between the State of Mississippi and a Memphis water utility over water rights in the middle Claiborne aquifer. To provide information to stakeholders addressing the groundwater-availability issues, the U.S. Geological Survey Groundwater Resources Program supported a detailed assessment of groundwater availability through the Mississippi Embayment Regional Aquifer Study (MERAS). This assessment included (1) an evaluation of how these resources have changed over time through the use of groundwater budgets, (2) development of a numerical modeling tool to assess system responses to stresses from future human uses and climate trends, and (3) application of statistical tools to evaluate the importance of individual observations within a groundwater-monitoring network. An estimated 12 million acre-feet per year (11 billion gallons per day) of groundwater was pumped in 2005 from aquifers in the Mississippi embayment. Irrigation constitutes the largest groundwater use, accounting for approximately 10 million acre-feet per year (9 billion gallons per day) in 2000 from the Mississippi River Valley alluvial aquifer in Arkansas, Louisiana, Mississippi, and Missouri, and to a lesser extent in Illinois, Kentucky, and Tennessee. Predevelopment groundwater flow is represented in the MERAS model as a steady-state stress period, assumed to be prior to 1870. The simulated groundwater-flow budget indicates the largest predevelopment inflow to the system is net recharge to the alluvial aquifer. This inflow is balanced by outflow to gaining streams. Overall, water enters as net recharge to the alluvial aquifer or through outcrop areas of the various hydrogeologic units. Away from the outcrop areas, groundwater flow in the deeper formations is primarily upward into overlying units, ultimately discharging to streams through the alluvial aquifer. Total net recharge and discharge (sum of inflows or outflows) for the model ranged from about 0.66 million acre-feet per year during predevelopment to 20.16 million acre-feet per year by the end of the simulation (final simulated irrigation period in summer of 2006). This change in the model budget reflects increases in withdrawals compared to predevelopment conditions. Cumulative storage within aquifers simulated in the MERAS model indicates overall depletion of 140 million acre-feet (equivalent to 2.8 feet of water covering the entire study area). Postdevelopment inflow to the system is still through net recharge to the alluvial aquifer and the outcrop areas of the several hydrogeologic units, however, the flow between each unit is no longer upward to the alluvial aquifer. Groundwater flow during the summer of 2006 was primarily downward to offset demand from pumping. Early in the model simulation (1870-1920s), the primary components of the water budget were simulated as outflow from stream leakage and inflow from net recharge. As pumpage increased through time, water that would otherwise flow to streams reversed, and net stream leakage became an inflow to the system. The largest reversals began in the mid-1980s, but indications of the reversal began in the early 1960s with a trend in loss of streamflow leakage coupled with the first consistent inflow from storage. While groundwater pumped out of the alluvial aquifer was derived primarily from storage, pumpage out of the middle Claiborne aquifer was derived primarily from other aquifers (up to 15 percent from the alluvial aquifer), followed by flow from storage and net recharge. The potential consequences of climate change have been identified as a major concern facing the sustainability of the Nation's groundwater resources. To address this concern, two climate simulations were developed through the use of the MERAS model by extending the simulation period by 30 years to the year 2038 using extrapolated precipitation based on frequency analysis of historic climate cycles. There is little difference between the dry and wet scenarios in terms of percent water-level change. Both scenarios resulted in 14.6 to 13.9 percent of the area containing more than 100 feet of decline, 14.5 to 13.8 percent containing between 75 and 100 feet of decline, and 15.8 to 15.7 percent containing 51 to 75 feet of decline in the alluvial aquifer. The middle Claiborne aquifer water-level changes also were similar between the two scenarios. These scenarios indicate that even with a 25-percent increase in precipitation from that of the dry scenario, there is little difference in the resultant water levels. This is in large part because of the magnitude of differences between changes in net recharge and changes in pumping. When compared to the volume of water pumped out of the system, the effect of this change in net recharge is negligible. The groundwater-level monitoring network used to construct the 2007 middle Claiborne aquifer potentiometric surface was used as an example case to demonstrate statistical technique and to evaluate the importance of individual groundwater-level observations. To calculate the importance of each water-level observation to a prediction, predictions were specified as water-level altitudes near the end of the dry scenario simulation. These predictions were located near the center of cones of depression. Many of the observations that have a high importance are in close proximity to stressed areas of the aquifer.