Water-level records from an observation-well network are analyzed for utility in estimating probable high ground-water levels in three different geohydrologic environments in Massachusetts. Analyses were made of 83 observation wells with between 8 and 37 years of records. Maximum annual water levels occur most frequently in March and April. The maximum range of water levels equaled or exceeded at 10 percent of randomly selected sites is estimated to be 16 feet in till, 9.2 feet in sand and gravel on terraces, and 4.0 feet in sand and gravel in valleys. An approach to estimating probable high ground-water levels at construction sites is suggested. An estimate of the high water level at a site may be derived by solving the proportion in which the ratio of the potential water-level rise at a test site to the potential water-level rise at an observation well is equal to the ratio of the water-level range at the site to the historic water-level range at an observation well in a similar geohydrologic environment. Precise description of the probabilities of exceedence of estimates made by this approach with the data available has not been made. Assuming that the data are representative of the future, estimates would not be expected to be exceeded at more than 1 in 10 sites over a period of 10 years or longer. (USGS)
