The Scottish Government

« Previous | Contents | Next »

Listen

11 Detailed Waste Analysis

In addition to the door-to-door surveys carried out by the Scottish Waste Awareness Group ( SWAG), which focussed on perceptions and behaviours as reported by the householders, a detailed pre- and mid-scheme waste composition analysis was carried out on the aggregate waste from a sample of ~150 households within each pilot area and a matched control area.

This analysis determined the detailed waste composition of the residual waste stream in both the pilot and control areas, together with the composition of collected recyclates. Each sampling period was of 2 week's duration. Ideally, a longer period of sampling would have been useful to reduce temporal variability. The actual waste analysis was undertaken using standard procedures.

11.1 General characteristics

The composition and amount of domestic waste appearing in the various components of the waste stream (domestic bins, local recycling schemes etc) will vary from household to household and over time. The aggregate variation across a particular set of domestic properties will be less than at the household level. The literature suggests ( SEPA, 2004) that a sample size of ~150 households is sufficient to reduce the household-to-household variations across an area of a particular housing type. Implicitly, these ~150 samples are taken to be weekly. The literature is less forthcoming with guidance on the duration of the sampling exercise in order to reduce the effects of week-to-week variation in waste generation.

If the waste generation processes were stationary ³ and ergodic ⁴, then the ~150 household figure might also apply - such that ~150 samples in total - over time and households - might be deemed sufficient, notwithstanding seasonal effects. In practice, the occupancy mix (eg single occupants, families, multi-occupancy flats etc) of households within an area of similar household type is likely to vary from property to property, and the ergodic assumption is unlikely to hold. This means that a number of different households need to be included in the sample to ensure that the waste is representative of the area as a whole. And because household habits vary from week to week, waste sampling should ideally be undertaken over an extended period of time.

11.2 Special factors in this study

In this study, the number of weeks of sampling was primarily constrained by the monitoring/sample costs - so that just 2 weeks waste data were sampled during the pre-scheme and mid-scheme periods. The time gap between the pre- and mid-scheme analyses was typically ~3 months, with the mid-scheme being allowed to 'bed in' before starting the waste analysis exercise. It is quite possible that seasonal factors might have affected the pre- and mid-schemes differentially, but this was unavoidable.

The sampled data consisted of a breakdown of the aggregate residual waste stream (other waste + target recyclate) from ~150 households in the pilot and control areas - chosen to be as similar as possible based on proximity, housing type similarity and household ( ACORN) similarity. In the pilot areas, additional data were collected on the recovered target recyclate collected as part of the pilot scheme.

In one of the pilot areas (South Lanarkshire), the recycling scheme was pre-existing, so that whilst a comparison could be made between the pilot and control areas, no comparison could be made within the pilot area between pre- and mid-scheme behaviour.

11.3 Temporal variability

Cursory examination of the data over all 8 schemes revealed substantial variation in all the waste streams in the four sampling periods (2 pre-scheme; 2 mid-scheme). The ~150 household recommendation - as suggested in the literature - therefore needs to be revisited to take account of temporal variations. In this study, the data from each "two week period" (ie pre- and mid- scheme) of the trials were aggregated to provide a slightly more robust set of data. Given just two weeks' data in each phase of the trials, statistically robust estimation of temporal variation/variability was neither possible nor intended. For similar reasons, the focus of the waste data analysis to date has been directed to aggregate measures of target recyclate and the remaining non recyclate fractions, rather than to subcategories.

11.4 Analysis completion efficiency/accuracy

The various waste streams were analysed against a template of 31 waste categories. There is inevitably a rounding error in this process, such that the sum of the component elements does not correspond exactly to the gross weight of the waste recorded on entry to the waste depot. These errors were generally marginal. However, in some of the trials it became apparent from the records that the possibility existed that not all of the waste registered on entry to the depot had been analysed (eg for time limitations). In such cases, even if the waste compositions actually analysed were reflective in percentage terms of the total waste streams, they would not reflect the gross amounts (kg) of the various waste fractions.

To determine the effect of such cases, the data from all the trials has been analysed in both their raw and re-weighted forms. The results cited below refer to the re-weighted data. Attention is drawn to cases where there are different conclusions for the raw data sets.

11.5 Analysis of the scheme data

The collected data consisted of a weight breakdown across 31 categories of waste in the waste streams. Each Local Authority Trial targeted its own mix of target recyclates and their own scheme of recyclate collection.

Notwithstanding these variations, the analysis - and subsequent assessment - of each scheme needs to focus on three key issues:

1. Behaviour Change:
Has the scheme affected a change in behaviour with regard to the disposal of the scheme's target recyclates? Specifically, is less target recyclate being disposed of through the residual waste stream (ie the domestic waste bin)?
2. Scheme Efficiency
How much of the target recyclate is being collected by the scheme?
3. Collateral Effects - Modal Shift in Target Recyclate Disposal
To what extent has the scheme shifted existing target recyclate collection from one scheme (eg recycling points) to another (ie the pilot scheme systems)?

It should be noted that the estimation of the latter effect inevitably has to be by indirect inference, since there was no monitoring - for obvious practical and financial reasons - of the behaviour of the pilot and control areas with respect to such as recycling points.

11.6 Basis of analysis and definitions of terms

As indicated previously, the data relating to the period of the pre-scheme waste analysis and those of the mid-scheme analysis have been separately aggregated. The waste collection schedules between some of the schemes varied (eg weekly, bi weekly, fortnightly), but in each case the data have been aggregated. As also indicated previously, the data have been analysed in both unweighted and reweighted forms. The former better reflects percentage waste compositions; the latter reflects better component waste masses.

Definitions of Terms The data analysis in either case is based on 3 primary quantities (each of which can be broken down into further waste categories as defined by the standardised 31 waste fractions ( SEPA, 2004) Residual Waste (ie waste disposed via the domestic waste stream), comprising: N (kg): Other Waste: All waste not targeted by the recyclate stream T (kg): Target Recyclate Collected Recyclate R (kg): Collected Target Recyclate (Note: The Target Recyclate Collection Scheme will inevitably be contaminated by non-Target material - either material which is perceived by the householder to be target recyclate, or by other incidental contamination by other waste material).

In terms of the three key questions posed earlier, these are addressed as follows:

Behaviour Change:

ie Has the scheme affected a change in behaviour with regard to the disposal of the scheme's target recyclates? Specifically, is less target recyclate being disposed of through the residual waste stream (ie the domestic waste bin)?

This is reflected in the values of T/(T+N) pre- and mid-scheme and compared to behaviour in the control area.

Scheme Efficiency

ie How much of the target recyclate is being collected by the scheme?

This is reflected by the relative amounts of target recyclate collected in the scheme and that still being disposed of in the residual waste stream - ie R/(R+T).

Collateral Effects - Modal Shift in Target Recyclate Disposal

ie To what extent has the scheme shifted existing target recyclate collection from one scheme (eg recycling points) to another (ie pilot scheme systems).

As noted above, the latter can only be assessed indirectly, but it can be estimated by comparing the total target recyclate in the pilot and control areas. In the former this will comprise target recyclate disposed of in either the residual waste stream or the target recyclate collections schemes, and in the latter by the target recyclate disposed of in the residual waste streams. These effects are examined by looking at the relative values of (R+T)/(R+T+N) for the operational pilot scheme versus T/(T+N) in the control and pre-scheme pilot areas.

For all three questions, two broad factors are at work:

1. Random and systemic variations operating either across the pilot and control schemes (eg by virtue of a different household composition), or during the schemes (ie due to some temporal effect such as seasonality, or a common event). The systemic effects might be expected to apply equally to the pre- and mid-scheme behaviours in the pilot and control areas, with any random effects superimposed.

and - in the case of the pilot schemes only,

2. the effect of the schemes themselves.

This second item is the effect of interest, and which, if possible, needs to be filtered out from the first set of factors.

This is addressed as follows for each of the three key questions posed: Behaviour Change, Scheme Efficiency and/or Collateral Effects.

For the present purposes, and to maintain a greater degree of statistical robustness, the waste stream is analysed by splitting it into just two fractions: Target Recyclate (as defined for the scheme in question), and the remaining, Other Waste, rather than by analysing the waste schemes at a finer scale of definition.

11.6.1 Behaviour change

This is estimated as follows:

Other Waste: Irrespective of the particular details of the recyclate scheme adopted, it might be expected (ie 'on average') that the pilot and control areas would behave similarly in terms of other waste (N), such that it might be expected that:

Target Recyclate: Without any recyclate scheme (or equivalently, if the recyclate collection scheme was having no effect whatsoever) it might also be expected (ie 'on average') that the pilot and control areas would also behave similarly over time with respect to the amounts of recyclate generated, such that:

and also that:

On the other hand, if the recyclate collection scheme was working, it might be expected that the proportion of recyclate in the residual waste stream would fall in the pilot areas, such that:

This is examined in the following Figure 11-1 for all 8 schemes, which shows the ratio X:Y.

Figure 11-1 Relative Performance of Pilot and Control Schemes with respect to Target Recyclate Remaining in Residual Waste Stream

(Values << 1 Indicate Behaviour Change in Pilot Area)

Ratios of less than one suggest a change in household behaviour in the areas where a recyclate scheme has been piloted, reflecting a reduction in the target recyclate appearing in the residual waste stream during the period when the target recyclate collection scheme was in operation. This occurred in 4 of the pilot areas:

• Falkirk
• Glasgow City
• Glasgow City SBI
• West Dunbartonshire

Note that in South Lanarkshire there was already a recyclate collection scheme in operation during the so-called 'pre-scheme' period and no change between the pre- and mid-scheme analyses would be expected.

11.6.2 Scheme efficiency

Setting aside for the moment how much of the collected target recyclate is diverted from pre-existing recycling routes (eg recycling points), the efficiency of the pilot schemes can be estimated by measuring the fraction of the total target recyclate that ends up in the collection scheme as opposed to that which remains in the residual waste stream. The ratio of interest is:

and the greater this ratio, then the more effective is the recyclate recovery scheme (notwithstanding recyclate diversion from other collection routes).

As noted above, in South Lanarkshire there was a recyclate collection scheme in operation throughout the monitoring period. This is examined in the following Figure 11-2 for all 8 schemes.

Figure 11-2 Scheme efficiency - percentage of target recyclate collected in scheme = 100R/(R+T) %

As can be seen, 4 schemes out of the 8 achieved a collection efficiency of greater than 50% (ie the target recyclate collection scheme recovered over half of the total recyclate in the pilot area). The areas where this occurred were:

• Falkirk
• Glasgow City
• South Lanarkshire
• West Dunbartonshire

11.6.3 Material capture efficiency

The efficiency of material capture can be estimated by measuring the fraction of the total amount of specific target recyclate that ends up in the collection scheme as opposed to that which remains in the residual waste stream. See Figure 10-3 to 10-4.

Figure 11-3 Aberdeen City: Percentage of target recyclate collected [100R/(R+T)%]

Figure 11-4 Dundee City: Percentage of target recyclate collected [100R/(R+T)%]

Figure 11-5 City of Edinburgh: Percentage of target recyclate collected [100R/(R+T)%]

Figure 11-6 Falkirk: Percentage of target recyclate collected [100R/(R+T)%]

Figure 11-7 Glasgow City: Percentage of target recyclate collected [100R/(R+T)%]

Figure 11-8 Glasgow City SBI: Percentage of target recyclate collected [100R/(R+T)%]

Figure 11-9 South Lanarkshire: Percentage of target recyclate collected [100R/(R+T)%]

Figure 11-10 West Dunbartonshire: Percentage of target recyclate collected [100R/(R+T)%]

11.6.4 Estimated Recycling Rate

The waste analysis data has been used to make an estimate of the percentage recycling rate achieved by each local authority. The recycling rate is defined as the proportion of recyclate collected divided by total waste arisings:

Target recyclate recovered as a % for all waste =100R/(R+T+N)

This has been calculated for each pilot scheme and is shown in Figure 11-11. It should be remembered that this is based on a 'snap-shot' of waste composition, rather than performance over the length of the pilot scheme.

Figure 11-11 Target Recyclate Recovered as a Percentage of all waste = (100*R)/(R+T+N)

The percentage recycling rate could be calculated using the Projected Diversion Rate for each scheme and data of household arisings in each local authority area as submitted to SEPA in the Local Authority Waste Arisings Survey ( LAWAS). However, this would be expected to underestimate the recyclate rate as the amount of waste produced by multi-occupancy households is likely to be lower than the household average submitted by local authorities to the LAWAS ( SEPA, 2005). A comparison of the two results is shown in Table 11-1.

Table 11-1 Percentage rate of recycling

11.6.5 Collateral effects - modal shift in target recyclate disposal

The possibility exists that some of the target recyclate appearing in the recyclate collection scheme has not resulted from a change in householder behaviour with respect to the residual waste mix, but from a diversion of target recyclate previously taken to such other recycling points. This can be estimated by comparing the total amounts of target recyclate collected in the pilot and control areas.

Again, two broad factors are at work - temporal/spatial factors and the effects of the scheme itself: In both the pilot and control areas there may be temporal effects, which, if the recyclate collection scheme has no effect on target recyclate diversion, might be expected (ie 'on the average') to be similar.

Thus, if there is no modal shift effect, it might be expected that

and that

On the other hand, if there is a modal shift effect, it might be expected that:

and that

This is shown in the following Figure 11-12 for all 8 schemes, which shows the ratio V : Z.

Figure 11-12 Extent of target recyclate modal shift in pilot area from other recycling points (Values >> 1 Indicate Possible Target Recyclate Modal Shift)

Ratios greater than one suggest that there appears to be a relative increase in the total target recyclate in the pilot areas during the operational period of the target recyclate collection scheme. This increase may be due to diversion of target recyclate from other recycling points. This could happen for example if the scheme site offered a more convenient recycling option for the householder. There are 3 of the pilot areas in which this might have occurred:

• Aberdeen
• Dundee
• Edinburgh

11.7 Results summary

As can be seen from the charts above and the summary in the Table below, the results are not consistent across all the schemes and in some cases inconclusive. For the reasons alluded to previously, it is not possible to attach particular degrees of statistical significance or statistical confidence to the results. Nevertheless, taken together with co-lateral and related evidence associated with the schemes, it is suggested that some reasonable conclusions can be drawn.

The outcomes of the performance of the schemes according to the criteria outlined above are summarised as follows:

* The results for the unweighted data sets are broadly consistent with those shown above, the only 2 exceptions being some evidence for recyclate modal shift in the Falkirk and Glasgow City SBI schemes.

« Previous | Contents | Next »