Waste management is nothing new, nor is the science of working out what’s being thrown away. Julian Parfitt of Resource Futures considers the history and methodology of compositional analysis

Some knowledge of waste composition has always been regarded as essential for effective waste management. For at least 140 years, we have been trying to scientifically analyse what we throw away and it should come as no surprise that this has changed considerably. As, consequentially, has the system for handling and treating it.

Compositional analysis, or ‘bin rummaging’, is a much-maligned science, and collecting such data presents a number of challenges. For planners, there’s a distinct advantage to working with locally gathered evidence, but obtaining it is time-consuming and costly, and waste sorting is often unpleasant to carry out. Yet exactly what’s discarded varies by socio-economic group, season and other temporal factors (such as week-to-week variation), so care is needed to obtain the right sample for an area.

Furthermore, this needs to be repeated in order to assess changes in composition and recycling performance as local policies are implemented or factors influencing arisings change. In recent years, the task has become more difficult with a growing level of confusion and concern among householders about bin rummaging by contractors and general suspicion about what the information is used for.

Bin rummaging is not a new activity, nor is it exclusive to the UK; every industrialised country regularly collects compositional data and has had to tackle the challenges involved. The information gleaned from bin sorts is a fundamental component of designing effective waste management and recycling systems; without such data, planned infrastructure and service improvements cannot be carried out in a cost-effective manner. Imagine trying to design a health care system without analysing the type and frequency of particular illnesses and ailments.

Some of the earliest compositional data for the UK municipal waste stream dates from the 1870s. Although classifications and methodologies have differed since then, a broad picture of changing composition can be reconstructed for kerbside collected household waste. This picture of the national dustbin, and, more recently, kerbside recycling, reflects changing lifestyles and consumption patterns over this period; some of the rapid changes that have occurred also suggest how difficult it might be to predict future changes.

The dominant component of collected household waste in the UK until the 1950s was dust and cinders from coal fires. Following the Clean Air Act of 1956, which sought to curb further London smogs on the scale of the disaster of December 1952 (in which an estimated 4,000 people died), smokeless zones were established and coal fires were largely replaced by gas and electric space heating and more recently by central heating systems.

Looking at the analysis conducted since then, increases in packaging, newspapers and magazines are apparent in the waste stream from the early 1960s. Changes in material use and function in the home appear through the growth of plastics – particularly striking between the late 1960s and late 1970s, where it considerably increased the volume of household waste – and the more recent rise of the ‘putrescible’ fraction. The latter relates largely to food and garden waste. On the face of it, there’s something paradoxical about an increase in food waste alongside packaging that is designed to reduce it, but this illustrates just how much cheaper and expendable food has become in the last 40 years.

Data from the UK-wide ‘National Household Waste Analysis Project’ conducted in 1991-93, estimated food waste to be about 17 per cent of kerbside collected household waste; the latest data compiled for England suggests that the proportion has risen to about 24 per cent (2006/07). Is this suggestive of a trend? Interpretation of such changes in composition can be problematic as the methods used to sort different fractions of waste have evolved in this time.

The 1990s estimate of food waste equates to a mean quantity of about three kilogrammes (kg) per household per week, while the 2006/07 estimate for England suggests a mean of 4.2kg, equivalent to a 39 per cent increase over this period. However, the national household waste analysis project used a ‘trommel’ (a large rotating screened cylinder used to separate materials into different size fractions) as a component of the sorting process, which resulted in more of the material breaking up compared with the hand-sorting techniques that are used currently; as a result ‘fines’ (cinders and dust) were 6.8 per cent in 1991-93, compared with 1.5 per cent in 2006/07. Thus some food waste in the trommel method would have ended up in the ‘fines’ fraction.

Regardless of the actual food waste trend, the rise in the biodegradable fraction of kerbside collected household wastes between these years is beyond doubt, due to the significant increase in kerbside collected garden waste, disposable nappies/sanitary products and textiles. The 1991-93 composition suggests that 63 per cent of kerbside collected household waste was biodegradable, whereas the 2006/07 estimate is 68 per cent.

Profiling areas for sampling

In building up the national compositional profile, or that of a smaller area, it is important to consider how representative the samples are. Techniques developed in the UK, and most other industrialised nations, have addressed the problem of how to classify samples in terms most relevant to waste arisings. The concept behind ‘profiling’ sampled neighbourhoods is no different from a polling company finding out if a representative sample of the population has been taken in predicting an election result. The difference is that most waste samples are ‘bulk’ samples: so each neighbourhood is a single data point, rather than each household. In general, the ways in which this classification is carried out have become more sophisticated over time, but the end-use hasn’t.

The value of classifying neighbourhoods as a starting point for waste analysis was recognised in the first standard procedure that was developed in the UK in 1929 by J C Dawes’s report ‘An investigation into Public Cleansing Service in the Administrative County of London’. At that time London had 20 incinerators and a detailed understanding of waste composition and the fluctuations in calorific value were essential to their proper design and operation. Dawes developed a simple method of arriving at representative samples based on dividing households into three rateable bands related to property value and found significant socio-economic and seasonal variations in waste composition between them.

The next attempt to develop waste sampling strategy was carried out by A E Higginson in the 1960s and ’70s using a classification based on property type (flats, detached, ‘semis’, bungalows, tenement type properties), which remained the standard method well into the 1980s. However, property type is too crude to differentiate family size, disposable income and other demographic factors that significantly influence waste arisings.

Since the 1980s, area stratification for waste analysis has been carried out by use of geo-demographic packages, which use a full range of socio-economic variables, many derived from the Census. ACORN – short for A Classification of Residential Neighbourhoods – is one such package (much used by retailers and financial institutions), which has been applied to the stratification of waste samples. This has improved accuracy and understanding of findings, better enabling planners to identify the right collection and infrastructure and to budget more appropriately.