Preliminary assessment of trends in static water levels in bedrock wells in New Hampshire, 1984 to 2007

Analysis of nearly 60,000 reported values of static water level (SWL, as depth below land surface) in bedrock wells in New Hampshire, aggregated on a yearly basis, showed an apparent deepening of SWL of about 13 ft (4 m) over the period 1984–2007. Water-level data were one-time measurements at each well and were analyzed, in part, to determine if they were suitable for analysis of trends in groundwater levels across the state. Other well characteristics, however, also have been changing over time, such as total well depth, casing length, the length of casing in bedrock, and to some extent, well yield. Analyses indicated that many of the well construction variables are significantly correlated; the apparent declines in water levels may have been caused by some of these factors. Information on changes in water use for the period was not available, although water use may be an important factor affecting water levels.

Multiple regression models were used to determine the simultaneous effects of important variables on SWLs statewide. Models also were generated for each county, and the model-calculated results for counties were generally similar to the results for the state wide models.

The most significant predictors of mean SWL (aggregated by year and quarter) were total depth, the third quarter of the year (July–September), elevation, and height of well above minimum elevation within a 1,640-foot (500-meter) radius (hillslope factor). Casing length was a significant predictor of SWL for igneous-rock models and curvature of the land surface for metamorphic-rock models. Local geologic as well as landscape features appear to provide further explanation of SWL variation. For example, SWLs in wells completed in specific granites appear to be deeper than in other granites; this relation was also observed for different groups of metamorphic rocks.

A more detailed examination of data from six towns that report frequent complaints about low water supply from bedrock aquifer wells showed that hillslope position may play a role in the availability of water in wells. SWLs were commonly deeper (greater depth to water) for wells with more than 100 ft (30 m) of relief between the well and the lowest land-surface elevation within 1,640 ft (500 m) of the well. For these high-relief wells, the depth to SWL was commonly greater for those that were on generally south-facing slopes, compared to wells on north-facing slopes.

Concerns about wells with very little water-producing capacity in localized areas of some towns may be better understood by considering some of the relations identified in this study. For example, the data show that the position of a well on a hillslope affects the SWL depth in that well; however, the data also indicate that the average yield of the well is lower for hillslope wells than for wells in lower relief areas. This relation indicates that as the hillslope factor increases (and SWLs become deeper), the amount of available water in the wells decreases. Knowledge of this relation indicates that deeper wells may be needed in areas of higher relief.

In areas with less extreme relief, the overall depths of wells have increased substantially more than the depths to the SWL. This indicates that the amount of water stored in wells (wellbore storage) has increased, and thus more water is available for use in the average well. As a result, more water may be used because it is available, possibly adding to the problem of local well interference or exacerbating drought-related well problems.

These data provided an opportunity to examine groundwater-level conditions across the state; however, the bedrock wells used in this study would not be suitable for rigorous evaluation of trends in SWL across the state because the locations and characteristics of the wells vary with time. Further, these wells cannot substitute for a carefully designed network of wells selected for the sole purpose of monitoring trends in water levels over time. The SWL data may be useful in the design of a monitoring network, and continued collection of water-level data from the bedrock wells could be used to augment data from monitoring wells.