Pesticide hotspot probability model

Published: 22-02-2021| Version 3 | DOI: 10.17632/5wst5krs5f.3
Leena Jaakola,
Poul Larsen,
Per Loll


A high density of shallow pore water screening samples taken for a pesticide contamination study were applied to the development of a probability model spreadsheet. The JAGG model 1.5, developed by the Danish Environmental Protection Agency, was modified to pore water sampling instead of soil gas, to output the number of pore water samples necessary to detect a pesticide hotspot with a defined level of certainty. The spreadsheets calculate the grid size (representing one pore water sample per grid cell) required to detect a pesticide of a defined concentration at a defined certainty. In this study, the probability model was run for several concentration levels of two dominant pesticides, desphenyl choridazone (DPC) and mechloroprop (MCPP). The results of the spreadsheet provide a starting point for similar studies to locate and delineate pesticide contamination areas, and investigate the conceptual understanding of the contamination area.


Steps to reproduce

The JAGG model can calculate the probability of locating a cylindrical volume of contaminated soil with a specified size, based on the grid spacing of boreholes placed in a quadratic grid. The model was modified to apply to pore water samples with a (small) radius of influence and output the probability level associated with locating pesticide hotspots of a certain size, given the grid spacing. User input, indicated by gray cells with blue text, includes the assumed aerial rectangular investigation area (length and width in meters), radius of contamination (meters) and desired probability of detecting the contamination hotspot. Iso-contours produced from the sample data provided an initial estimate of the contamination area, which was transformed into an equivalent cylindrical volume of soil, assuming contamination spread concentrically from the source. A pore water volume is removed corresponding to the volume of pore water extracted from a spherical volume of soil. The model calculates the grid size (i.e. sampling density) required to locate a cylindrical volume containing contaminated soil to the user-defined certainty. Contamination is detected by the model if a grid cell overlaps the contaminated volume of soil. The number of pore water sampling points required to detect an unknown hotspot for a certain contaminant concentration is found by adjusting grid spacing and probability that the hotspot falls within that area. The count based on the actual area is then normalized to sample count per 100 m2 of investigation area. Several pesticide concentration levels are included as screening criteria for discovery to find the number of samples needed to theoretically detect a hotspot with an unknown concentration. This model can be a starting point for the number of samples needed to locate and delineate a contamination area at future sites.