Print this article    Email this article

Foresight essential
02-JAN-2006

Predicting waste generation to ensure accurate capacity planning

How can waste planners ensure that the new capacities planned will match the amount of waste generated in the future? Socio-economic factors influence waste generation, and a forecast model examines how these factors affect waste generation across some European cities. 

Although capacity planning of waste-processing facilities and infrastructure requires an idea of the future demand, little is known about how to estimate the quantity of future MSW streams. Yet neglecting this first step in the planning process is likely to lead to the wrong answers to questions such as ‘Do we need another new, enlarged waste incineration plant, recycling facility or composting plant?’. A growth forecast that is out by only 1% can lead to a deviation of more than 10% of the total waste generated over a planning period of 10 years. Under- or over-estimate thus has significant consequences in terms of additional investment and operating costs.

This article describes the development of a forecasting method to enable cities to estimate the amounts and composition of MSW they will generate over a period of at least 10 years. The model was developed in an EC project entitled ‘The Use of Life Cycle Assessment Tools for the Development of Integrated Waste Management Strategies for Cities and Regions with Rapidly Growing Economics’ (LCA-IWM).

DATA SOURCES

In order to focus on regional waste management systems in environments with different levels of economic growth, the study considered 91 cities in 44 European countries, including all capitals and cities with more than 500,000 inhabitants. Led by the Institute of Waste Management at BOKU – University of Natural Resources and Applied Life Sciences in Vienna, waste-related data as well as demographic and socio-economic indicators at a city level were collected between October 2002 and February 2003 in co-operation with local representatives. Available data were collected covering the period from 1970 to 2001. The information collected was examined in collaboration with six project partners from different European countries.

TABLE 1. Available socio-economic indicators at city and national level

• Population

• Population density

• Population age structure (0–14; 15–59; > 60 years)

• Sectoral employment (agriculture/industry/services)

• Gross domestic product/regional product

• Infant mortality rate

• Life expectancy at birth

• Average household size

• Unemployment rate

• Overnight stays (tourism)

The availability and comparability of international waste statistics are key issues in studies such as this. With this in mind, the project focused on all waste streams from daily and routine activity of households and businesses, excluding sporadically collected wastes such as construction and demolition waste. Use of this definition (as set by the European Environment Agency) meant that it was possible to evaluate data with an average time-series length of 10 years on MSW quantities from 55 European cities in 10 old EU Member States and five new ones, as well as two applicant countries. Data from other cities were either not available or did not pass the plausibility check with regard to comparability. Only 45 datasets from 31 cities were available for material-related waste generation.

Socio-economic indicators

More than 30 demographic, economic and social indicators were selected for collection at a city and a national level. Of these, the indicators listed in Table 1 were available as time series. These data were obtained from many regional, national and international statistical sources (such as Eurostat, United Nations and OECD) through database investigations, literature and internet reviews, and personal communication.

Increasing amounts of rubbish

Apart from intermittent decreases, generation rates grew steadily throughout Europe between 1995 and 2001 (Figure 1). The data for the eastern European cities of Poznan and Nitra suggest that the gap between old and new Member States is closing.

A comparison of all the 43 cities studied in old EU Member States and 12 cities in central and eastern European (CEE) shows clear differences between these (until only recently) quite distinct economic areas (Table 2). Major EU-15 cities had a far higher MSW generation rate than the CEE cities in 2001, though the average annual growth in CEE cities was more than double that of cities in EU-15 countries between 1995 and 2001. This trend suggests that waste amounts in the two areas will become similar in the future.

All European countries faced increasing quantities of MSW over this period. At the bottom of the scale, we found a growth of 1%–8% over the last six years. However, the highest increases ranging from 20% to more than 50% took place in Slovakia, Poland and Spain – all countries with rapidly growing economies.

TABLE 2. MSW generation in EU cities
	MSW generated	(kg/person/year)	Average annual change
	1995	2001
Old Member States (EU15)	466	519	+1.8%
New Member States and applicant countriesa	287	369	+4.3%
a Bulgaria, Romania

COMPOSITION TRENDS

Poor availability of reliable sorting analyses mean that it was only possible to document changes in MSW composition over a longer period in a few cases; data from three cities are reported in Figure 2. The total percentage of packagingrelated materials (paper/cardboard, plastics and composites, glass and metals) appears to be increasing, while the mass percentage (but not inevitably also the amount generated per person) of the organic fraction is tending to decrease.

Paper and cardboard

A two-fold increase in the amount of paper and cardboard generated per person was observed between CEE and EU-15 cities. The rate in CEE cities ranged from 40 to 80 kg/person/year (mean: 56 kg/person/year) but, in most of the EU-15 cities, it was 90 to 140 kg/person/year (mean: 113 kg/person/year).

Organic waste

No such significant differences were found with regard to organic waste. Despite considerable differences in the amount of MSW generated per person, between 120 and 180 kg/person/year of organic waste were generated in nearly three quarters of the datasets. One discrepancy was observed – considerably higher generation rates were reported for Spanish and Greek cities.

Plastics and composites

In terms of mass percentage, similar amounts of plastics and composites were documented – 10%–15% of total MSW generation for around two thirds of the cities. Much lower values were obtained for cities with a low income (such as Bucharest and Baltic cities) as well as from results of sorting analyses in Polish cities from the early 1990s. These findings are linked to the later introduction of plastics as the main packaging material in these cities.

Other fractions

No significant prosperity-related variations were observed with regard to the generation of glass and metal wastes.

INFLUENCING FACTORS

The trends for total MSW generation and paper/cardboard appear to confirm the assumed relationship to the general level of prosperity. To quantify this impact more accurately, the indicators investigated were subjected to statistical analyses such as correlation and regression analyses or cluster analyses.¹

GDP

This commonly used indicator proved to be a significant factor in cities with a high prosperity but not in cities with a lower economic output. This is because the high regional income inequality in CEE countries causes a big gap between mean values, which are usually available, compared with the clearly lower median values – a more meaningful, but rarely observed indicator for social well-being and living standards.

Health indicators

Although only one previous study² had used infant mortality and life expectancy as parameters to indicate MSW generation, they showed a remarkable ability to serve as an additional or alternative variable for GDP. Their advantages include good availability and high quality of data.

Age structure

This indicator has also far been undervalued as a possible predictor for MSW generation. The greater the percentage of the population aged 15–59 years, the more employed persons; this implies a subsequent rise in overall consumption and waste generation.

It seems more than a coincidence that the three countries with the highest waste increase between 1995 and 2001 also had the highest percentage of people in this age group [Spain and Poland: 64.2%; Slovakia: 65.1% (2000)] as well as the highest increase in this period (Poland and Slovakia: +2.8%; Spain: +1.2%) of all evaluated countries. Comparatively, Germany faced the biggest decrease (–1.7%).

Household size

Household size is a standard indicator for the level of poverty. Our study confirmed that there was a significantly negative relationship between average household size and MSW generation.

HOW TO PREDICT FUTURE WASTE QUANTITIES?

It is always difficult to decide the best way to deal with uncertainties. In planning practice, underlying assumptions or expectations concerning forecasts are, of necessity, not well defined and are, therefore, not transparent and are frequently overlooked. Although exact waste forecasts cannot be achieved, improvements in forecast accuracy are possible. However, they are often limited by:

• a failure to consider social and economic trends related to the region
• a lack of reliable forecasts with regard to these trends.

The approach applied in this study attempts to mitigate these problems. First, the relationship between the set of indicators is well known and assessed statistically on the basis of an investigated period of 22 years. Secondly, the model is based on prioritized parameters represented by factors with a high inertia such as health or demographic indicators. This means that they are easier to predict than highly volatile indicators such as GDP. Forecasts of this economic measure have an early expiry date, whereas forecasts for health or demographic indicators (such as those updated regularly by the United Nations) are available at a national level up to the year 2050.³

Our MSW generation model was used to develop the final forecasting model. An interpolation method was used to consider long-term changes between prosperity levels. Testing the accuracy of forecasts is crucial. Because the real error of a forecast over 10 years cannot be tested for practical reasons, the forecast accuracy was tested as follows. Fictitious forecasts were made on the basis of the underlying independent variables for a historic year. The estimated values were then compared with the real value in this ex-post predicted year. The developed model leads to a median relative error of 5.3% for all time series of 5–10 years and 7.7% for time series of 11–22 years; these errors are related to the error at the end of the time series. The more relevant estimation error of the annual growth rate (derived thereof) amounts to approximately 0.6% per year.

It is always difficult to decide the best way to deal with uncertainties

UNDERLYING TRENDS

All relevant indicators but one foster the assumption of a further growth in MSW generation per person in European cities.

• Economic growth – long-term estimates from the World Bank suggest that moderate GDP growth rates for EU-15 (+2.3%) and higher ones for CEE countries (+3.3%) can be expected.
• Health indicators and household size will follow the previously observed trend of increasing welfare standards and shrinking household sizes.
• Rising urbanization – an increasing share of the population will live in cities, though the population in existing major cities will remain constant.

A contrary, retarding effect on further increases in waste generation could arise from the decreasing share of the population in the mean age group. Following a peak in most European countries between 2000 and 2005, demographic forecasts suggest a steady decline back to the level of the 1980s or even the 1970s.

FORECASTS

Due to regional trends, global forecasts for cities in Europe are neither particularly useful nor helpful. Figure 3 shows forecasts for four European cities – two from the EU-15 and two from the Accession Countries.

Although not guaranteed valid for all cities, the following trends can be deduced:

• further rise in MSW generation
• equalization of MSW generation rates due to stronger growth in CEE countries
• disproportionate increase in the generation of waste paper and cardboard
• rate of increase of organic waste generation less than that for total MSW.

APPLICATION

The use of improved waste generation forecasts enables more adequate dimensioning of the waste collection infrastructure and treatment facilities, thus helping to avoid expensive overor under-capacity. Compliance with the recycling and recovery targets of the EU packaging directive serves as an example. These targets are defined as a percentage of the total futur e generation by material.

There is thus no direct indication of the total volume that will need to be processed. Use of the model to predict the volumes of different materials that could be generated would help to reduce planning uncertainties when designing separate collection systems and recycling facilities to fulfil the prescribed values.

Application of the model presented allows the compilation of regionally adapted forecasts based on the user input of a small set of city-related indicators. Missing data problems are avoided by proposing default values from an extensive background database. To verify its practicability, the software tool has been tested in five European cities from regions with rapidly growing economies.

Peter Beigl is a Research Associate at the Institute of Waste Management, Department of Water, Atmosphere and Environment, BOKU – University of Natural Resources and Applied Life Sciences, Vienna, Austria.
e-mail: peter.beigl@boku.ac.at
web: www.wau.boku.ac.at/abf

The model outlined in this article was implemented in a platform-independent, Java-based decision support tool. The software, related handbook and project report is available for downloading at www.lca-iwm.net.

NOTES

1. Beigl, P., Wassermann, G., Schneider, F. and Salhofer, S. Forecasting municipal solid waste generation in major European cities. In: Pahl-Wostl C., Schmidt S., Jakeman, T. (editors). iEMSs 2004 Int. Congress: Complexity and Integrated Resources Management, Osnabrück, Germany. 2004. www.iemss.org/iemss2004/pdf/regional/beigfore.pdf
2. Bogner, J., Rathje, W., Tani, M. and Minko, O. Discards as measures of urban metabolism: the value of rubbish. Paper presented at an international symposium on urban metabolism, University of Michigan Environment Dynamics Project, held Kobe, Japan. 1993.
3. Beigl, P., Gamarra, P. and Linzner, R. Waste forecasts without ‘rule of thumb’: improving decision support for waste generation estimations. SARDINIA 2005 – Tenth International Waste Management and Landfill Symposium, Paper No. 251 (in press). CISA Environmental Sanitary Engineering Centre, Cagliari. 2005.

 

---

---