Spatial and temporal statistical analysis of a ground-water level network, Broward County, Florida

The U.S. Geological Survey has developed a method to evaluate the spatial and temporal statistics of a continuous ground-water level recorder network in Broward County, Florida. Because the Broward County network is sparse for most spatial statistics, a technique has been developed to define polygons for each well that represent the area monitored by the well within specified criteria. The boundaries of these "confidence polygons" are defined by the endpoints of radial lines oriented toward the other wells. The lengths of these lines are determined as the statistically estimated distances to the points at which ground-water levels can be predicted within specirfied criteria. The confidence polygons indicate: (1) the areal coverage of the network, (2) locations where data are unavailable, and (3) areas of redundant data collection. Comparison with data from a noncontinuous recorder well indicates that the confidence polygons are a good represen- tation of areal coverages. The temporal analysis utilizes statistical techniques similar to those used in the spatial method, defining variations in time rather than in space. Consequently, instead of defining radial distances to points, time intervals are defined over which water-level values can be predicted within a specified confidence. These "temporal confidence intervals" correspond to maximum allowable periods between field measure- ments. To combine all results from the analyses, a single coefficient reflecting the spatial and temporal results has been developed. The coefficient is referred to as the Spatial and Temporal Adequacy and Redundancy Evaluation (STARE) and is determined by three factors: the size of the confidence polygon, the number of times the well is part of a redundant pair, and the temporal confidence interval. This coefficient and the individual results of each analysis are used in evaluating the present network and determining future management decisions.