Lefke Çevre ve Tanıtma Derneği   Environmental Society of Lefka

 Results of Biodegradable Waste Analysis

Programme:

Inventory of Biodegradable Waste Sources in Urban and Rural Areas in Cyprus

Pilot Scale Application in Lefka Municipality and Denizli (Xeros) Community.

Participants:

Environmental Society of Lefka

Akti. Project and Research Centre

Lefka Municipality

XerosCommunities

Aglantzia Municipality

Ergates Community

Lefka, Semptember 2004

1 Introduction

            Biodegradable waste analysis was as a part of the research program Inventory of Biodegradable Waste in Cyprus, funded by UNOPS (Project No PO2-04-003, 2003). 

 The present project included sampling of biodegradable waste from both commercial (supermarkets, fast food restaurants, bakeries, meatmarkets etc) and domestic sources (households) from an urban area in Cyprus (Lefka Municipality). Sampling was also done in a rural area (Xeros community), but only from domestic sources. Sampling from these two locations was carried out to determine the effect of urbanization on food waste production.

2. Materials and Methods

2.1 Sampling

Twenty-four houses from an urban area (Lefka Municipality) and five houses from a rural area (Xeros Municipality) were randomly selected to participate in the food and green waste sampling programme. Two week before commencement of sampling a visit to each house was carried out to inform the houseowners about the objectives of the survey and the importance of their involvement. It was also considered important to explain to the householders what is biodegradable waste to avoid poor segregation. The sampling period was two weeks commencing on 6 June 2004.   Nylon and fabric bins, resistant to tearing, were provided to each house.  Fabric bags were used to contain food waste since in many black large nylon bags were subject to tearing by rodents, dogs and cats. For green waste, large nylon bags were used. In each household stickers were provided with the number of the house and the date for labeling the bags containing the waste. Each householder also completed questionnaires requesting socioeconomic and demographic information, including for instance the number of people permanently living in the house, area of the house, number of children etc. The questionnaire is presented in Appendix 1.

The major sources of organic waste production from commercial establishments were selected after a discussion with the mayor of Lefka municipality. Those included meatmarkets, bakeries, supermarkets, fruitmarkets, restaurants and fast food shops.

All the major commercial sources at Lefka Municipality were included when that was possible. That was done because the markets of the same type (e.g meatmarkets) varied in size and there were difficulties in obtaining representative samples for each category. For example, small restaurants produced less biodegradable waste than large outlets and sampling from a small number of restaurants would not provide representative information for these types of commercial source. The owners of the markets maintained records of their biodegradable waste production over the same sampling period. Questionnaires were also provided to the owners of the markets and these were completed after weighting the organic waste produced each day. In addition, the final destination of the waste was recorded (e.g. disposal to landfill, feeding to animals etc.).

2.2. Collection of food, green waste and measurements

            The food and green waste produced by each house participating in the survey was collected on a daily basis. The bags of waste were weighed on a daily basis.

 The data  collected over a period of 4 days for both urban and the rural areas in Cyprus is shown in Appendix 2.

2.3. Qualitative analysis of food waste

            Food waste is heterogeneous and its composition is difficult to assess. One sample of food waste including the whole amount of waste from 24 houses in the urban area was hand sorted in 9 categories. The same was done separately for the rural area (see Table 1)

                                                            Table 1. Categories of food waste

2.4. Statistical analysis of food waste data

            Appropriate methods of statistical analysis were selected based on the Kolmogorov Smirnov test, used to determine whether the data concerning food waste was normally distributed (Zar, 2000). Normally distributed data can be handled with parametric tests (t-test, ANOVA, etc) and described by the mean as a measure of location and the variance as a measure of scale. In contrast, the above methods are not suitable for data that is not normally distributed. In that case non-parametric tests are used (Man Witney, Kruskal Wallis etc.) and the statistical parameters used to describe this sort of data are the median and the interquartile range (Zar, 2000). The data concerning food waste was normally distributed and consequently, parametric tests were used.

2.4.1 Comparisons between the urban and the rural area

            Previous studies (Cargo, 1978, Kemper and Quigley, 1976) indicate that the degree of urbanization affects waste production from households, and comparisons between the urban and the rural area were carried out, to determine the effect of urbanization in food waste production in Cyprus.

2.4.2 Factors influencing  food waste generation from households

            Food waste production from households can be influenced by socioeconomic factors such as household size, age distribution and household or per capita income (Parfitt, 1997). Behavioural factors, such as the frequency of food preparation, the consumption of food that is not prepared in the house (ready food), and feeding domestic animals with food may also influence waste production (Loizidis, 2004). Due to the lack of data concerning per household income and age distribution, the area of the houses and percentage of children in each household were used as proxy variables of the above factors (Richardson and Havliceck, 1987).

All of the statistical analysis was carried out with the SPSS statistical computer programme using the whole number of houses (both from the rural and the urban area), as there were no statistically significant differences in food waste production between the two sampling locations. Values of food waste production that were out of the range (mean ± 2standard deviation) were excluded as outliers (Statsoft 2003).

Multiple regression models for household waste used two types of dependent variables: (kilograms per capita per week) and (kilograms per household per week). Both units are used widely in the research literature for expressing comparative rates of household waste arisings (Parfitt, 1997). It was therefore important to produce models based on both to compare the results and consider which is likely to provide a representative basis for estimating total arisings of food waste in the two sampling locations. In these models, the major determinants of food waste production were hypothesized to be the area of the house (S), household size (H), number of children (C), frequency of food preparation (F) and consumption of ready food (R). Feeding domestic animals with food waste was not included in the models due to the small numbers of houses were this practice had been observed. The models had the general form:

FW =b₀+b₁S+b₂H+b₃C+b₄F+b₅R

Where:             FW = food  waste production (‘kg per household per week’ or ‘kg per                            

                              capita per week’)

                          S= The area of each house in m² (as a proxy variable of income)

                          H= household size in number of person

                          C= the number of children in each house (below the age of 18)

                          F= the frequency of food preparation in each house (times per week)

                          R= the frequency of consumption of ready food (times per week)        

Regression analysis is useful if the sample is adequately large and reveals only linear relationships between variables (Statsoft 2003). In contrast scatterplots and comparisons of means, such as ANOVA (Analysis of Variance), LSD, and t-tests are suitable for smaller samples and also reveal non-linear relationships between  variables (Statsoft, 2003). This type of statistical analysis was carried out to test the effect of feeding domestic animals on food waste production. In addition, these tests were useful in cases were no significant linear relationships were observed between the variables of the above two models.

The results of regression analysis were not clear in one case since two of the factors that were tested were not independent to each other. As households with higher percentage of children are likely to contain more members than those without, this factor was inter-correlated with household size. To determine the actual effect of each of the above factors on food waste generation statistical analysis was carried out with households of the same size but different percentage of children and vice versa.

2.4.3 Extrapolation of the data from sampling area to the whole Lefka Municipality and Xero Community.

            The whole amount of food waste produced in Lefka municipality and Xero community from domestic sources was estimated using the per-household method. That was considered more reliable than the per-capita method, which assumes a homogeneous population of waste producers. Previous studies (Kahr and Burney, 1983;Richardson and Havliceck, 1978) demonstrated that there is no linear relationship between household size and waste production and the same was shown in the present study (see Section 4.3.1). The dwellings that were included in the sampling were separated into different categories according to the household size and the mean waste production per household was estimated for each category, separately. These values were multiplied by the number of dwellings in each category. According to the official state reports  (1996) in Lefke Municipality and Xeros Community the number of households was 24 and 5, respectively. The number of houses in each category based on the average household size is shown in Table 2.

            Table 2: Household by average size of household at Lefke municipality and Xeros community

Although the initial target was to include all the major commercial sources of food waste 20% of the owners of the markets did not complete the questionnaires. Therefore, mean values for each type of market were estimated and multiplied by the total number of each market type.

2.5. Statistical analysis of green waste data

            A Kolmogorov-Smirnov test (Zaar, 2000) was carried out to determine whether the data concerning the daily green waste production per dwelling were normally distributed. This revealed that the data were not normally distributed and hence non-parametric tests were used in this case.

2.5.1Comparisons between the rural and the urban area

            Man-Whitney test (Zar, 2000) was carried out to determine if there was a statistically significant difference in daily green waste production per dwelling between urban and rural areas in Cyprus. The test compared the equality of means between 2 samples and was equivalent to the independent samples t-test for normally distributed data.

            In accordance with the study that we carried out in Lefke and Xeros area to find out the organic wastes that were collected from gardens and pavements, we   got the following results.

We divided these wastes into two different categories as green wastes and brown wastes. The amount of the wastes collected from pavements were calculated as Kg/m2 and the amount of garden wastes (by pruning) were calculated by weighting and the wastes from house gardens were also calculated by being weighted.

 2.5.2 Relationship between green waste production and garden size

            The relationship between green waste production and garden size was tested by linear regression analysis. Since many different types of vegetation were present at different houses it was difficult to take into account each category separately. Therefore the different types of vegetation were separated to two distinct groups (grass and other types of vegetation). Using SPSS, a linear regression model was derived with the two distinct vegetation types as the independent variables and the total green waste production per household per week in the urban area as the dependent variable. The rural area was excluded from the analysis, since green waste production was significantly different compared to the urban area and the number of houses was too small to be tested separately by linear regression. 

3 Results

3.1 Food Waste

 3.1.1 Comparison between the urban and the rural areas

Table 3: The avarage data about the organic household wastes in Lefka

The avarage daily organic waste production per dwelling in Lefka:  2.510 kg

        The avarage daily organic waste production per capita in Lefka:  0.744 kg

Table 4: The data about the organic household wastes in Xero

The avarage daily organic waste production per dwelling in Xeros: 2.377 kg

         The avarage daily organic waste production per capita in Xeros: 0,699 kg

    The data of food waste production in the two sampling locations is summarized in Table 5.

 Table 5: Summary of food waste production data in the rural and the urban areas

The statistical analysis revealed no significant differences between the two sampling locations for either the per dwelling or the per capita food waste production.  The per-capita and per-dwelling food waste productions in the two sampling locations were relatively constant during the sampling period (see st.dev. in Table 5 and Figures 1 and 2).

Figure 1: Daily variation of per dwelling food waste production in the rural and the urban areas.

Figure 2: Daily variation of per capita food waste production in the rural and the urban areas.

The composition of food waste in the two different sampling locations is shown in Table 6 and in Figures 3 and 4.

Table 6: Comparison of food waste composition beween the urban and the rural areas

Figure 3: Food waste composition in the urban area

                                     Figure 4: Food waste composition in the rural area

The level of contamination of food waste with non-biodegradable material in the urban and the rural area is presented in Table 7.

Table 7: Level of contamination of food waste with non -  biodegradable material in the rural and urban areas

Although the level of contamination in the rural area appeared to be much higher all the contaminated bags were produced from one house that systematically provided poorly segregated food waste.

3.1.2 Commercial sources of food waste

The data concerning production and disposal of food waste from commercial sources in Lefka Municipality is summarized in Table 8.

Table 8: Food waste production from commercial sources in Lefka

Municipality