Box 2 – Storage Depletion in a Thick Confining Layer: Dakota Aquifer System
Konikow and Neuzil (2007) calculate storage depletion in the low-permeability beds that confine the regionally extensive Dakota Aquifer in South Dakota, USA (Figure Box 2-1), and that example is summarized here. The Dakota Sandstone and related sandstones in western-central North America form what is often considered a classic example of an artesian aquifer system. The Dakota Aquifer system is extensively developed and has played a particularly important role in the settlement and economic development of South Dakota. Study of the aquifer system began with Darton (1896; 1909) and helped shape current ideas about artesian aquifers (Bredehoeft et al., 1983).

The Dakota Aquifer underlies more than 171,000 km2 of South Dakota as described by LeRoux and Hamilton (1985), though it also extends in several adjacent states. Significant recharge to the Dakota and the deeper Madison Aquifers occurs where they crop out on the flanks of the Black Hills (Figure Box 2-1). The Dakota aquifer discharges naturally at low elevations in the eastern part of the state. Discharge from pumped and flowing wells also has become an important source of discharge from the aquifer system (Case, 1984).
Substantial development of the aquifer system began by the early 1880s (Bredehoeft et al., 1983). By 1905, over 1,000 wells were producing water in the portion of South Dakota east of the Missouri River, supplying an estimated 1.2×106 m3/d of water for irrigation and livestock (Bredehoeft et al., 1983). High rates of head decline in the Dakota Aquifer occurred before 1915. For example, eastern South Dakota experienced head declines averaging about 7 m/yr between 1909 and 1915. The rate of decline decreased to less than 0.5 m/yr by 1953 (Schoon, 1971). Estimated withdrawals stabilized at about 150,000 m3/d by 1960 (Helgesen et al., 1984). Pumpage data presented by Bredehoeft et al. (1983), Helgesen et al. (1984), and Case (1984) indicate that the cumulative well discharge from the Dakota Aquifer system in South Dakota from the time before development until 1981 totaled about 19.7 km3 of water. The history of development is incompletely documented, but Bredehoeft et al. (1983) estimate well discharge in 1912 as approximately 1.4 million m3/d and then it declined dramatically to about 300,000 m3/d in 1922; subsequently, it remained at rates less than half of the peak rate into the 1980s.
Groundwater storage depletion from the Dakota Aquifer system in South Dakota was evaluated using potentiometric maps showing predevelopment (Darton, 1909) and 1980 conditions (Case, 1984). The Inyan Kara, Newcastle, and Dakota Sandstones were treated as a single continuous unit forming the Dakota Aquifer. The spatial distribution of differences in head between the predevelopment and 1980 potentiometric surfaces indicated that the maximum decline in head was about 190 m and the average decline was 47 m. Storage coefficient values for the Dakota Aquifer ranged from 1.0×10-5 to 1.0×10-4 (Bredehoeft et al., 1983, table 3); a central value, 5.0×10-5, was used to estimate depletion in this confined aquifer system. These data indicate that a total of about 0.4 km3 of groundwater was derived from storage in the aquifer for the period from predevelopment through 1980, which represents about 2 percent of the estimated cumulative discharge of 19.7 km3. The remaining 98 percent consists mostly of storage depletion in adjacent confining beds as water leaked from the confining beds into the aquifer.
Bredehoeft et al. (1983) used numerical models to analyze flow in the Dakota Aquifer system. They concluded that prior to development, most of the recharge and discharge occurred as steady-state leakage through the thick confining layers. Furthermore, their analyses indicate that since development, most of the water released from storage originated from the confining layers. Using Bredehoeft et al. (1983) model-calibrated estimate of specific storage for confining layers of 1.6×10-4 m-1, Konikow and Neuzil (2007) estimated that the volume of water removed from storage in the confining units in South Dakota by 1980 was 14.9 +/-2.2 km3, which represents approximately 76 percent of the estimated cumulative well discharge. This further implies that about 22 percent of the withdrawals are balanced by capture because only 2 percent was derived from storage in the aquifer. The capture includes both increased inflow (recharge) in or near upgradient outcrop areas and reduced outflow in downgradient discharge areas.