The Iowa Department of Natural Resources, in conjunction with the University of lowa, Center for Health Effects of Environmental Contamination, conducted the State-Wide Rural Well Water Survey (SWRL) between April 1988 and June 1989. The SWRL survey systematically selected and sampled 686 sites and provided a statistically valid assessment of the proportion of private rural wells and rural Iowa residents affected by various environmental contaminants. The SWRL design framework also systematically selected a subset of 10% (68) of all sites for repeat sampling, to assess temporal changes in water quality during the original survey. The 10% repeat sites yielded a very consistent representation of the statewide data, including proportionately representative detections of pesticides down to about a 1% occurrence interval. These sites provided a representative subset of SWRL for monitoring water quality over time as an indicator of temporal change. The first two samplings of this 10% subset of wells are termed SWRL 10-1 and 10-2, respectively (abbreviated as 10-1 and 10-2). The SWRL 10-1 was part of the full SWRL sampling, and therefore is used as the basis for comparison with subsequent samples.
The SWRL survey was conducted during the driest consecutive two-year period in Iowa's recorded history. The objective of the resampling studies was to resample the subset during more "normal" climatic conditions, and to assess changes in water-quality that may have occurred. The 10% subset was resampled in October 1990 (10-3) and June 1991 (10-4), after weather patterns in Iowa had changed from the drought conditions of 1988-1989, to more normal and wetter-than-normal conditions. Long-term monitoring has shown that a mid-fall period, such as October, often represents conditions near the annual average for many parameters, though typically fewer pesticide detections occur than in late-spring or summer, such as the 10-4 (June 1991) resampling. For cost and technical reasons fewer analytes were included in SWRL 10-3 and 10-4 than the full SWRL survey.
In June 1991, during 10-4, about 19% of the sites showed N03-N >10 mg/L; 57% were positive for total coliform bacteria and 24% positive for fecal coliform bacteria; 20% of wells had a detection of some herbicide compound, about 6% contained detectable atrazine (parent compound only), and about 11% showed detections of atrazine or one of two common metabolites. The pattern of greater nitrate and bacteria occurrences in samples from wells <IOO feet deep and lower contaminant levels in wells > 100 feet deep, continued to be statistically significant in the 10-3 and 10-4 sampling.
Using tritium (3H) analysis as a groundwater dating tool, wells > 100 feet deep were more likely to produce groundwater with <6+/-4 Tritium Units, which averages >20 years old, than wells <IOO feet deep. Groundwater containing detectable tritium showed much higher rates of nitrate, and total- and fecal-coliform contamination, as would be expected. The data indicate a relationship among shallow wells, relatively recently recharged groundwaters, and higher rates of contaminant occurrence. All of the wells > 100 feet deep that exhibited >10 T.U. (i.e., modem recharge water) are from east-central or northeastern Iowa, where the hydrogeologic conditions promote much greater depth of groundwater circulation, and, hence, the depth to which contaminants occur is much greater than elsewhere in the state. This tool provides further insights on the mechanisms of groundwater contamination and may be useful for evaluating pollution potential of well waters.
Relative to prior sampling, the proportion of sites positive for total coliform bacteria and those with any detection of total atrazine or other pesticide increased slightly. The percentage of wells with N03-N > 10 mg/L did not change, but the mean and median concentration did increase somewhat. The only water-quality changes and from the full SWRL and the 10-1 (1988-1989), the 10-3 (October 1990), and 10-4 (June 1991) samplings that were statistically significant as estimates for all rural well waters, statewide were: 1) the decline in the detection of dissolved organic-nitrogen in 10-3 and its increase again in 10-4; 2) the increase in fecal coliform positives in 10-3 and again in 10-4; 3) the decrease in NH4-N in 104; and 4) the increase in detections in 10-3, which subsequently declined in 10-4.

The change from drought to wetter than normal conditions did not affect the 10-3 and 10-4 sample results as noticeably or consistently as it did in long-term monitoring projects in the state. Atrazine (parent) detections increased significantly in 10-3, but declined again in 104. Detections of other herbicides and total atrazine (metabolites included) increased somewhat in the June 10-4 sampling, as might be expected from the seasonal patterns discerned in other Iowa studies. While mean nitrate concentrations increased slightly, the proportion of wells with N03-N > 10mgL, and total coliform detections were largely unchanged. Fecal coliform detections increased, but this increase was unrelated to trends in other contaminants. As a sample of wells state-wide, this less pronounced response might be expected because the 10% subset integrates a variety of well-depths, well types and hydrogeologic settings. The SWRL 10% sample may react more slowly to climatic change; one indication of this is the low tritium content and therefore, relatively old (>20 years) average age of almost half the well water sampled during SWRL 10-3 and 10-4. The change to wetter conditions may explain the increases in total coliform and fecal coliform bacteria positives. Other studies have noted that bacteria, similar to chemicals, may be rapidly transmitted to the water table by preferential flow through the soil. Also, as water tables rise during wet periods, they are closer to the land surface and into contact with soil horizons where coliform bacteria are more abundant and more likely to survive.

One sample was collected from a site using a rural water-supply (RWS) system for its home and farmwater. The RWS system uses surface water. Four herbicides were detected: alachlor (0.7 pg/L), atrazine (1.7 AWL), cyanazine (1.5 pg/L), and metolachlor (1.2 gg/L).

Immunoassay (IMA) methods for scans provided promising results. With further refinements, these methods may provide a tool for inexpensive screening of water supplies for triazine occurrence.