The characteristics of samples collected from a lot are used to make estimates of the characteristics of that lot. Thus, samples are used to infer properties about the lot in order to make correct decisions concerning that lot. Therefore, for sampling to be meaningful, it is imperative that a sample is as representative as possible of the lot, and more generally, each subsample must be as representative as possible of the parent sample from which it is derived. Subsampling errors propagate down the chain from the largest primary sample to the smallest laboratory analytical subsample. If a collection of samples does not represent the population from which they are drawn, then the statistical analyses of the generated data may lead to misinformed conclusions and perhaps costly decisions.
It is quite a “lot” to ask of the tiny (on the order of a few grams, and often much lower) laboratory analytical subsample to be representative of each of the larger and larger (parent) samples in the chain from which it was derived, up to the entire lot (which could be many tons). Therefore, it is imperative that each subsample is as representative as possible of the parent sample from which it is derived. Any subsampling error is only going to propagate down the chain from the largest sample to the smallest laboratory analytical subsample.
The primary reason that samples are being taken is to make some determination about the lot (e.g., a contaminated site). The study goals and objectives determine the acceptable statistical characteristics for the study. If a decision depends on the analytical results, then the first issue is to determine what type of measurements are needed and how accurate and precise they should be. These goals are referred to as Data Quality Objectives -DQOs (extracts from US EPA 2003).
MSW is a complex material stream. It is particulate, and contains particles which vary substantially in terms of:
· stiffness / flexibility
· surface properties
A lot of particles are almost two-dimensional in nature, for example papers, while others may wrap themselves around other particles, for example textiles. Some particles may be composites of different materials – for example packaging. Often particles of different kinds are contained in several layers of bags.
This complexity is also true for processed fractions of MSW, including mechanically segregated fractions, composts and refined compost product (Barton 1983, Barton and Wheeler 1988). The potential range of variability may, of course, be reduced by processing, but perhaps not to the degree that one might expect. For example, after trommel screening at 50 mm, the undersize should be mostly below 50 mm in size and the oversize mostly greater than 50 mm in size. However, it is quite possible for the undersize fraction to contain materials larger than 50 mm in one dimension and the oversize materials smaller than 50 mm, depending on how the material fell onto the screen. The oversize may still contain smaller particles entrained or contained in or on larger particles. The undersize may contain larger particles which were deformed and forced through the trommel screen. Particles may also break and fall through trommel or flat bed screens, often this is intentional in waste processing, however where screening is used in sample appraisal it is a potential source of error.
There is a link between “information” required by a user and sampling and analysis. The nature of this linkage is often overlooked, but it is critical to determining the approach to sampling and analysis that should be undertaken. The critical factors relate to the type of information needed and the “quality” of information necessary, which in turn are determined by what the information will be used for. These linkages are well explored in other environmental business sectors, for example contaminated land management (Crumbling et al. 2001), but do not appear to be widely considered in the appraisal of waste composting.
Most commonly information from sampling and analysis is used in the composting sector
· as part of a quality monitoring process (including compliance with guidelines and regulations)
· for the evaluation of safety, health and environment (SHE) impacts
· for predictions about likely process performance, especially at the planning and commissioning stages
· for predictions of likely quality and SHE impacts.
Sampling and analysis information may also have a range of uses in research activities beyond the day to day operations of a composting facility.
Most practioners understand that this information is subject to errors, but not all understand the range of potential sources of error and relative importance of these errors to information for decision-making. A very basic distinction is between systematic errors and random errors. A systematic error is one which is a function of the sampling or analysis approach, for example digestion of a compost sample with aqua regia will not liberate all trace elements into solution, hence estimates of “total metals” for example will always be systematically under-estimated. Random errors are unpredictable errors that are a fundamental property of what is being measured – its intrinsic variability. Statistical techniques can be used to compare measurements to determine the probability that they are different given known random error. Often the techniques employed assume that random errors follow the Normal distribution. However, this is not always true, for example distributions may be skewed away from Normal, for example, skewed distributions are often observed for “heavy metals” in organic materials. An EC project (HORIZONTAL) has been investigating the distribution of trace elements and micro-organic pollutants in soils, sewage sludges and composts (including from MSW), and has collated and reviewed the various guidelines available for the sampling and analysis of these materials (Lambkin et al. 2004).
Errors can arise at various stages of the sampling and analysis process:
· during sample collection
· during sample preparation, preservation and storage
· during subsampling
· during analysis.
Understanding the significance of errors can be compounded by a statistically inadequate sampling regime, that prevents an adequate understanding of the variability of the measurement being made.
It is also important to understand the cost of the information being collected versus its utility to the decision maker. A common mistake is to invest a lot of money in few measurements with high analytical precision, when the intrinsic variability of the material being sampled renders a limited set of data points useless or even misleading in cases of compliance with regulatory or guideline standards. In these circumstances it may be better to make many analyses at, say, 20% precision, rather than few at 1% precision. Increasing use is being made of sensors and field based measuring techniques as a means of collecting a large volume of indicative data (e.g. see the EC Project SENSPOL:
Sampling is discussed in a little more detail in a subsection of this Critical Review Section. The other subsections deal with biological, chemical and physical techniques, beginning with physical techniques. For MSW fractions and products it is usually advisable for physical pre-treatment to take place before measurement of chemical parameters, for fraction of size greater than 10 mm, for composts as well as feedstocks (Brunner and Earnst 1986, Wheeler 1993)