Statistical Sampling Strategies for Survey of Soil Contamination - Dealing with Contaminated Sites: From Theory Towards Practical Application - page 166

4.2.2 Estimation .........................................

175

4.2.3 Using Ancillary Information in Estimation ..................... 181

4.2.4 Composite Sampling .................................. 182

4.2.5 Required Number of Sampling Locations ...................... 183

4.3 Estimating Mean Concentrations for Delineated Blocks .................

185

4.3.1 Design-Based Approach ................................

186

4.3.2 Model-Based Approach ................................. 188

4.3.3 Required Number of Sampling Locations ...................... 188

4.4 Mapping Concentrations at Point Locations ........................ 190

4.4.1 Sampling Patterns ....................................

191

4.4.2 Spatial Interpolation ................................... 197

4.4.3 Required Number of Sampling Locations ......................

199

4.5 Detecting and Delineating Hot Spots ............................. 199

4.5.1 Detecting Hot Spots ................................... 200

4.5.2 Delineating Hot Spots .................................. 202

References ................................................ 205

4.1 Introduction

This chapter is about statistical methods for sampling contaminated soils. I advo-
cate statistical methods for survey of soil contamination as they provide estimated
concentrations of contaminants with quantified uncertainty . This uncertainty can
be used to optimize the number of sampling locations and/or the geographical
co-ordinates of the sampling locations, i.e., the sampling pattern. From the many
survey methods described in the literature, I selected some methods whose sim-
plicity makes them attractive for application in common, real-world situations.
More advanced sampling designs and estimation methods, requiring surveyors more
skilled in statistics, will only be touched on.
This chapter is structured as follows. In Section 4.2 sampling methods are
described for estimating the entire histogram or spatial cumulative distribution func-
tion (SCDF) of contaminant concentrations, or parameters of this SCDF such as the
mean or median. Section 4.3 describes methods for estimating the means of several
blocks that have been delineated before sampling. Section 4.4 describes sampling
methods for mapping the concentrations of contaminants in the soil. In a final sec-
tion I will go into the sampling aspects of detecting and delineating hot spots, i.e.,
relatively small areas where the concentration is relatively high.
The spatial resolution of the survey result increases from Section 4.2 , via Sections
4.3 to 4.4 . Section 4.2 considers reconnaissance surveys with the aim of estimating
a single number (e.g. the mean) or several numbers (e.g. several percentiles) or
the entire SCDF of the concentration for one or several contaminants. Information
on where specific concentrations occur is not (yet) asked for. The other extreme
is when a map is required, depicting the contaminant concentration at any point
location in the study area (contaminated site) (Section 4.4 ). Section 4.3 is about an
intermediate situation, requiring spatial information on the concentrations, but with