Local Index Of Spatial Association (LISA) Optimization For Secondary Sample Designs In SADA

Sampling programs must take into account balancing the costs of operation with the necessity of accurately characterizing a contaminated site. Sampling designs that call for hundreds of samples will often be pared down drastically due to cost-cutting procedures. The samples that are implemented must then be optimally placed in order to capture the necessary information at a site. This often does not occur and crucial information is lost. The objective of this study was to assess the applicability of the Local Index of Spatial Association (LISA) secondary sampling methods provided by Spatial Analysis and Decision Assistance (SADA) software using actual data from a United States Environmental Protection Agency (USEPA) Superfund site, as well as to identify and optimize the critical parameters of these LISA methods to gain a cost effective, practical, and reliable method to place secondary samples that will ensure the characterization of the spatial distribution of contamination. The limitations of the existing LISA parameters in SADA were observed. The LISA search window greatly affects the outcome of secondary sample designs. Guidelines were developed by mimicking real world conditions and applying them to the SADA parameters and using an iterative function for the LISA search window to develop a potential site sample distribution for each LISA secondary sample design. A methodology is recommended to reduce the redundancies that occur within the site sample distribution and that subsequently occur within the final secondary site sample design. It appears that the guidelines presented in this paper could make SADA a cost effective tool for use in Phase III and Phase IV environmental site assessments, brownfield redevelopment, or other environmental risk management or site remediation situations.