| Abstract: | “Random” variability in groundwater monitoring data sets reduces the ability to identify long‐term concentration trends. This, in turn, increases the time and cost required to evaluate the effectiveness of natural attenuation and other groundwater remedies. To better understand the factors influencing variability in groundwater monitoring results, we have analyzed three large groundwater monitoring data sets. For the three data sets, the long‐term trend in contaminant concentration in each well accounted for an average of 30% to 40% of the overall variation in contaminant concentration. Understanding the causes of the remaining variability would support the development of improved groundwater monitoring methods. All three data sets show large differences in the temporal monitoring records between individual wells (e.g., coefficient of variation for monitoring results from individual wells ranges from 0.08 to 4.6) indicating that well and aquifer factors are more important contributors to variability than sample collection and analysis factors. However, the depth to groundwater (R2 = 0.020) and distance between water level and screened interval (R2 = 0.049) accounted for only a portion of the differences in variability between wells and other aquifer characteristics evaluated and were not correlated with the observed variability in monitoring results. Unidentified factors were apparently much more important contributors to variability than these factors. The monitoring data sets exhibited two distinct timescales for variability: Time‐independent variability that was apparent even when wells were re‐sampled within a few days and a long‐term variability likely associated with the long‐term concentration trend. The observation of time‐independent variability suggests that frequent monitoring of contaminated monitoring wells serves primarily to characterize sources of variability unrelated to the long‐term trend of primary interest. |
