A comparison of geostatistical methods for estimating virus inactivation rates in ground water

Abstract

Enteric viruses are responsible for a large number of ground-waterborne disease outbreaks every year. Septic tanks are the most frequently cited causes of ground-water contamination in disease outbreaks, and are probably the major contributors of viruses to the subsurface environment. Controlling the placement of septic tanks such that all viruses would be non-infective by the time the effluent reached drinking water wells would likely reduce the number of waterborne viral disease outbreaks. This study describes the use of and provides a comparison of different geostatistical techniques (kriging, cokriging, and combined kriging and regression) to estimate virus inactivation rates in ground water. These estimates were used with the regional ground-water flow characteristics to estimate septic tank setback distances over a city-wide area. It was found that combined kriging and regression, which eliminates the need for laboratory values of virus inactivation rates by using the linear regression relationship between temperature and inactivation rates, produced comparable results to the other two methods, yet reduced the cost of the analysis considerably.