This report was prepared by the U.S. Geological Survey (USGS) in cooperation with the Pennsylvania Department of Agriculture (PDA) as part of the Pennsylvania Pesticides and Ground Water Strategy (PPGWS). Monitoring data and extensive quality-assurance data on the occurrence of pesticides in ground water during 2003–07 are presented and evaluated; decreases in the land area used for agriculture and corresponding changes in the use of pesticides also are documented. In the Pennsylvania ground waters assessed since 2003, concentrations of pesticides did not exceed any maximum contaminant or health advisory levels established by the U.S. Environmental Protection Agency; PPGWS actions are invoked by the PDA at fractions of these levels and were needed only in areas designated by the PDA for special ground-water protection.

Previous investigations through 1998 of pesticides in Pennsylvania ground water identified land use, as a surrogate for pesticide use, and rock type of the aquifer combined with physiography as key hydrogeologic setting variables for understanding aquifer vulnerability to contamination and the common occurrence of atrazine and metolachlor in ground water. Of 20 major hydrogeologic settings in a framework established in 1999 for pesticide monitoring in Pennsylvania, 9 were identified as priorities for data collection in order to change the monitoring status from "inadequate" to "adequate" for the PPGWS.

Agricultural and forested land-use areas are decreasing because of urban and suburban growth. In the nine hydrogeologic settings evaluated using 1992 and 2001 data, decreases of up to 12 percent for agricultural land and 10 percent for forested land corresponded to increases of up to 11 percent for urban land. Changes in agricultural pesticide use were computed from crop data. For example, from 1996 to 2004–05, atrazine use declined by about 15 percent to 1,314,000 lb/yr (pounds per year) and metolachlor use increased by about 20 percent to 895,000 lb/yr; these compounds are the two most-used agricultural pesticides statewide.

In 2003–07, a baseline assessment of pesticides was conducted in five of nine hydrogeologic settings with inadequate monitoring data—the Blue Ridge crystalline and Triassic Lowland siliciclastic, Eastern Lake surficial, Devonian-Silurian carbonate, Great Valley siliciclastic, and Northeastern Glaciated surficial settings. Between 20 and 30 wells in each setting were monitored. Of the 126 wells sampled, 96 well-water samples were analyzed for at least 52 pesticide compounds at the USGS National Water Quality Laboratory (NWQL) using a method with a minimum reporting level (MRL) at or above 0.002 µg/L (micrograms per liter). Of the 96 well waters analyzed by NWQL, 43 had measureable concentrations of one or more pesticides. Atrazine and (or) deethylatrazine (CIAT), a degradation product of atrazine, were reported at or above the MRL in 39 of the 43 well waters. Neither atrazine nor CIAT were reported at concentrations exceeding 0.10 µg/L; all measured concentrations in these five settings were below PPGWS action levels. Metolachlor was present in 7 of the 43 well waters with measureable concentrations of 1 or more pesticides; however, concentrations were below the MRL. The other 30 samples (10 of 20 wells in the Blue Ridge crystalline and Triassic Lowland siliciclastic setting and all 20 wells in the Eastern Lake surficial setting) were analyzed for at least 19 pesticide compounds at the Pennsylvania Department of Environmental Protection Laboratory (PADEPL); the PADEPL reported no concentrations of pesticides at or above an MRL of 0.10 µg/L.

Statistical tests using the NWQL analytical results showed correlations between pesticide occurrence and two indicators of water-quality degradation—the occurrence of total coliform bacteria and nitrate concentration. A 2 × 2 contingency-table test indicated a relation between presence or absence of atrazine or metolachlor and presence or absence of bacteria only for the 10 wells representing the Blue Ridge crystalline and Triassic Lowland siliciclastic setting. Results of Spearman’s rank test showed strong positive correlations in the Devonian-Silurian carbonate setting between 1) the number of pesticides above the MRLs and nitrate concentration, and 2) concentrations of atrazine and nitrate. Atrazine concentration and nitrate concentration also showed a statistically significant positive correlation in the Great Valley siliciclastic setting.

An additional component of baseline monitoring was to evaluate changes in pesticide concentration in water from wells representing hydrogeologic settings most vulnerable to contamination from pesticides. In 2003, 16 wells originally sampled in the 1990s were resampled—4 each in the Appalachian Mountain carbonate, Triassic Lowland siliciclastic, Great Valley carbonate, and Piedmont carbonate settings. Nine of these wells, where pesticide concentrations from 1993 and 2003 were analyzed at the NWQL, were chosen for a paired-sample analysis using concentrations of atrazine and metolachlor. A statistically significant decrease in atrazine concentration was identified using the Wilcoxon signed-rank test (p = 0.004); significant temporal changes in metolachlor concentrations were not observed (p = 0.625).

Monitoring in three areas of special ground-water protection, where selected pesticide concentrations in well water were at or above the PPGWS action levels, was done at wells BE 1370 (Berks County, Oley Township), BA 437 (Blair County, North Woodbury Township), and LN 1842 (Lancaster County, Earl Township). Co-occurrence of pesticide-degradation products with parent compounds was documented for the first time in ground-water samples collected from these three wells. Degradation products of atrazine, cyanazine, acetochlor, alachlor, and metolachlor were commonly at larger concentrations than the parent compound in the same water sample. Pesticide occurrence in water from wells neighboring the hot-spot wells was highly variable; however, the same sets of pesticide compounds that were present in wells BA 437, BE 1370, and LN 1842 were present to some degree in water from neighboring wells. To evaluate temporal changes in concentration, nonparametric statistical tests were used to determine overall and seasonal monotonic trends. Concentrations of alachlor, atrazine, metolachlor, and nitrate were examined using the 5-year (2003–07) and the long-term data from wells BA 437 and LN 1842 (1996–2007 and 1995–2007, respectively), and the long-term data for well BE 1370 (1998–2007); results showed either downward trends or no trends. Trends in acetochlor concentrations were tested only at well LN 1842 using the 5-year data; no trends were observed. Homogeneity of trend tests indicated statistically significant downward concentration trends in the long-term data were due to seasonal trends as follows: BA 437—alachlor and atrazine (summer); BE 1370—atrazine and metolachlor (winter) and alachlor (winter and spring); LN 1842—alachlor (summer and fall) and atrazine (spring and fall).