Cumulative deviation from moving mean precipitation as a proxy for groundwater level variation in Wisconsin

Groundwater management questions often require understanding how water levels respond to precipitation. However, direct observations of precipitation alone rarely explain variation in water levels. This study introduces a modified method for comparing cumulative precipitation anomalies to groundwater level variation. This method transformed gridded monthly precipitation data from 1895 to 2018 into monthly deviations from different moving mean lengths at 90 USGS groundwater monitoring locations in Wisconsin. The precipitation data was then compared to water level variation at each site and correlations were calculated. The average optimal a priori moving mean window length for all sites was identified as 60 months. Fifty-four percent of the monitoring wells were moderately to highly correlated the cumulative deviation from 60-month precipitation and fewer than 30% were uncorrelated. Well depth, aquifer classification and location were tested as potential factors influencing the strength of correlation with groundwater levels and optimal mean length. Aquifer classification had no effect on correlation or optimal moving mean length indicating aquifer response rates are similar in both bedrock and unconsolidated aquifers. Correlation strength was also independent of well location. Well depth and casing in Cambrian/Ordovician formations were the only variables with a weak but statistically significant effect on correlation between the precipitation deviation and water level. This work illustrates how this method can help diagnose factors affecting monitoring well response, identify inconsistencies in a monitoring record and generate hypotheses regarding aquifer response to precipitation. This method leverages easily accessible datasets to serve as a starting point that engineers, groundwater professionals and resources managers can use to generate and test hypotheses about sites without prior knowledge of the geology or aquifer properties, extraction rates, and land cover.