Role of Groundwater Flow in Tile Drain Discharge

Tile systems drain water applied to agricultural fields as irrigation and precipitation but also may intercept regional groundwater flow. Identification and characterization of the potential sources of tile water is essential for informed management of salinity and contaminants. Factors influencing tile discharge including depth of water applied, evapotranspiration, water storage, drain blockage, and interception of regional groundwater flow were evaluated to determine which may be related to a fivefold variation in cumulative tile discharge among six sumps located 100 km west of Fresno, CA. Cumulative depths drained were calculated for 5 yr of weekly irrigation, precipitation, and discharge data. Evapotranspiration and water storage were estimated using the UnsatchemGeo variably-saturated water flow model. Well water levels measured on 19 dates were spatially-averaged providing spatial variation of depth-towater among the drained areas. Variability in depth of water drained (0.18-0.95 m) was large and was not correlated with either water applied (3.26-4.58 m, r = 0.03) or with computed water flux from the bottom of the soil column (0.050.31 m, r = 0.00). Groundwater interception by tile drains was a factor because depth-to-water was negatively correlated with discharge (r = 0.42) and drawdown of groundwater levels by drains was relatively larger for those drained areas encountered first during regional groundwater flow. For all six sumps, drained water is likely derived from locally applied water and interception of regional groundwater flow implying that standard two-dimensional models of flow to drains, representing only water applied locally, would not be applicable to modeling of drain flows or drain-water solute concentrations. U.S. Salinity Laboratory, 450 W. Big Springs Road, Riverside, CA 92507. Received 18 Nov. 1997. *Corresponding author (pvaughan@ ussl.ars.usda.gov). Published in J. Environ. Qual. 28:403-410 (1999). T ILE DRAINAGE in agricultural areas provides a method of regulating the depth of shallow water tables. The water drained from tile systems is frequently not reusable as irrigation water because of high salinity and contamination. Two contaminants often present in the San Joaquin Valley of California are B and Se. Disposal of tile water in this area is a necessary component of a management system that includes tile drainage. But tile water disposal raises environmental concerns such as the well-known experience of Se contamination at Kesterson reservoir in the San Joaquin Valley. The source of tile water is shallow groundwater which may be derived from water applied during irrigation of a field, from local precipitation, or it may have moved laterally to the tile drain by regional groundwater flow. The determination of the relative importance of these components is significant because drainage of irrigation water represents an economic loss. Drainage of groundwater which has moved laterally from other areas is, however, not locally controllable and cannot be factored into the local economics of production. Regulation of drain water quality is therefore a complex issue because there are, potentially both local and regional components controlling water quality. Tile drain tubes are normally installed in parallel sets with a fixed spacing. This arrangement can be modeled by analytical techniques which lead to expressions for flow rates within tubes based on drain diameter, spacing Abbreviations: CIMIS, California Irrigation Management Information System; CIS, geographic information system. 404 J. ENVIRON. QUAL., VOL 28, MARCH-APRIL 1999 and various boundary conditions for water flux and hydraulic head (Kirkham, 1949, 1958; Jury, 1975a; Fipps and Skaggs, 1991). The lower boundary condition of these models is an impervious layer which permits a solution using the method of images. This assumed boundary implies that all water arriving at a drain originated at the soil surface between the drains. In the analytical solutions, certain flow paths pass through lower elevations than the elevation of the drain resulting in upwards flow to the drains. Such paths may have travel times on the order of 10 yr or more (Jury, 1975b). The utility of the impervious layer assumption is questionable under certain field conditions as, for example, in the western San Joaquin Valley. In this area, a highly permeable sand layer underlies clay-rich soils that have much lower permeability (Deverel and Gallanthine, 1989; Belitz and Phillips, 1995). Groundwater flow is from southwest to northeast and the historic record of artesian wells at the margin of the alluvial fan near the San Joaquin River (Mendenhall et al., 1916) suggested that this groundwater flow was occurring in conditions where the groundwater level was close to the surface. Groundwater levels in this area were lowered by pumping during the first half of the 20th century, but the introduction of irrigation water from northern California, through the Delta-Mendota canal, and the installation of tile drainage have again stabilized the water table at a depth of 1.5 to 3 m (Deverel and Gallanthine, 1989). Water flowing to tile drains in this area is a mixture of local irrigation water and water that has moved upwards from the sand layer (Deverel and Fujii, 1988). Evidence for this partitioning includes isotopic enrichment of g180 with depth, and decreases in tritium concentrations when compared with irrigation water sampled at the same site (Deverel and Fio, 1991). Based on water chemistry data, the proportion of drained water that arrived from greater depths was 30% for a drain at 1.8-m depth and 60% for a 2.7-m drain. Furthermore, Se concentrations indicated that the deep groundwater component had arrived from the west (Deverel and Fio, 1991).