GIS and Evidence-Based Policy Making - Chapter 13 pdf

13 A GIS-Based Methodology to Support the Development of Local Recycling Strategies Andrew Lovett, Julian Parfitt, and Gilla Su ¨ nnenberg CONTENTS 13.1 Introduction 299 13.2 Research Methodology 300 13.3 Results 306 13.4 Conclusions 309 Acknowledgments 310 References 310 13.1 Introduction Following the publication of the Waste Strategy 2000 and its associated performance targets (DETR and National Assembly for Wales, 2000), there is presently particular interest among local authorities in improving their household waste recycling rates. The governmen t has set a national house- hold waste and composting target of 25% by 2005–2006, to be met through the cumulative effect of individual statutory targets for each local authority (Cabinet Office Strategy Unit, 2002). Many authorities have introduced kerbside collections of recyclables or established facilities at Civic Amenity (CA) sites, smaller neighborhood recycling centers, or supermarkets where households can deposit materials such as glass, paper, or cans. The form and extent of such recycling infrastructure provision varies greatly between local authorities (Parfitt et al., 2001; Community Recycling Network, 2002), and it is not straightforward to assess the effectiveness of different initia- tives. Partly this is because recycling rates also tend to differ according to the socioeconomic characteristics of the population involved (e.g., Tucker et al., 1998; Perrin and Barton, 2001; Barr, 2002; EnCams, 2002), but there are additional difficulties arising from the need to examine the interaction between the recycling scheme and use of other household waste stream ß 2007 by Taylor & Francis Group, LLC. outle ts such as conven tional house-to -house refus e vehicl e collectio ns. A further comp lication is that opera tional weight data often relates to a diver sity of zon es (includ ing collectio n rou nds or catchmen t areas of differ- ent sizes around ‘‘b ring’’ sites) , so making the task of reco nstructi ng the hou sehold waste stream and linkin g suc h informati on to define d popu la- tion s a challe nging one. Ge ographical informati on system s (GIS) can mak e a sign ificant contri bu- tion to the problem s of evaluat ing recycl ing strat egies because of their abili ty to integrat e data from opera tional entitie s such as CA sites or refuse colle ction rounds with details of the popu lation residen t in par ticular loca- tion s (Lovett, 2000). This capaci ty to link da ta sour ces ba sed on geogr aphi cal posi tion, co upled with the me ans to calcul ate ad ditional inform ation such as acces sibility me asures or catch ment boun darie s, in turn permits the recon- struct ion of waste st ream charact erist ics at the local scal e. To illustrat e the poten tial for enhanci ng waste man agem ent and plan ning through the use of GIS , a res earch project was carr ied out in one local aut hority (South Norf olk Cou ncil) to link subdistr ict opera tiona l da ta (from both collecti on round and site- based sour ces ) with pop ulation det ails. This database was then used to pre dict the impa ct of sever al poss ible new schemes for impro ving recycl ing per formance and assess the extent to whic h they wou ld help the aut hority meet thei r recycli ng targe ts for 2005–200 6. 1 3.2 Rese arc h Meth od olog y Sou th Norfo lk Cou ncil covers a predom inantly rural a rea of 908 km 2 extend- ing from the south ern subur bs of Norwi ch to the county borde r with Suff olk (Fig ure 13.1). In the 2001 census , the res ident popul ation was recorded at 110,710, an increas e of som e 8% since 19 91 (Offic e for Nationa l Stat istics, 2003). When the resea rch began in 1999, the distri ct used black plastic sack s for refus e colle ctions and had no ker bside recycl ing scheme s but a high densit y of bring sites through the pro vision of a netw ork of minirecy- clin g center s and facil ities at CA sites or local supermarke ts. South Norf olk was chosen as a focus for the stu dy on several grounds that incl uded the pre sence of a consi stent and wel l-maintai ned waste statisti cs database (Parfitt, 1997), previous involvement in other GIS projects (e.g., Lovett et al., 2003a), and proximity to the research team based in Norwich. The three main pha ses of the resea rch are sum marized in Figure 13.2. In the first phase, a GIS database was compiled that included details of recyc- ling facility locations (Figure 13.1) and a road network with attributes such as speed estimates to allow the calculation of car travel times (e.g., Lovett et al., 2004). Boundaries of refuse collections were defined by generating Thiessen polygons around addresses in the street listings used by contract- ors, and then dissolving the dividing lines between areas served on the same round day. Overall, there were 40 round days in the district ß 2007 by Taylor & Francis Group, LLC. Norwich Wymondham Bungay Diss South Norfolk minirecycling centers Urban areas 0 5 10 15 20 25 km A Roads B Roads Minor roads South Norfolk district Supermarket recycling centers Civic amenity sites Merchant recycling centers FIGURE 13.1 Recycling facilities in South Norfolk, spring 2000. (ß Crown copyright=database right 2007. An Ordnance Survey=EDINA supplied service.) Phase 1: Desktop Phase 2: Fieldwork Phase 3: Refinement and scenarios Data preparation • • • • • • • Area classification of waste variation (Cluster analysis) Evaluation of recycling scenarios Household data: model refined Subdistrict waste variations defined and mapped Compost bin assessments 600 house-to-house interviews across different round types focusing on recycling activity and use of facilities Final GIS Inititial GIS: combining waste statistics and population characteristics FIGURE 13.2 The overall research methodology. ß 2007 by Taylor & Francis Group, LLC. (eight rounds each weekday), but only 13 of these (32.5%) represented a single contiguous polygon and there were many irregularly shape d bound- aries. Figure 13.3 illustrates a typical situation using the example of Round 5 in the southeast of the district. Operational factors (particularly vehicle capacities) were the main reason for the complex configuration of round days. For instance, some villages were served when a vehicle was traveling to=from a landfill site, while in other cases full loads were generated by collecting from part of a town followed by a section of the surrounding rural hinterland. Given the nature of the round-day boundari es it was difficult to reliably estimate the populations of these areas from even the finest resolution census statistics (i.e., enumeration districts) and there was also the problem that the most recent details (from 1991) were somewhat outdated. As an Loddon Beccles Bungay Monday Tuesday South Norfolk district Other rounds Round 5 0 5 10 15 20 25 km Norwich Beccles Bungay Thursday Friday A roadsWednesday FIGURE 13.3 Boundaries of daily refuse collection areas in Round 5. (ß Crown Copyright=database right 2007. An Ordnance Survey=EDINA supplied service.) ß 2007 by Taylor & Francis Group, LLC. alternative source, South Norfolk Council provided data from their Council Tax register as of August 1999. The Council Tax is a compulsory local taxation system based on residential property values and the register is updated annually for each local authority (Eastham, 1993). Several studies have identified associations between the propensity of households to participate in recycling activities and the Council Tax valuations of their properties (Mansell, 2001; Lovett et al., 2003b). There were 47,474 properties on the South Norfolk register, of which 47,291 (99.6%) had useable postcodes. By applying geocoding and point- in-polygon techniques to the data (household totals by unit postcode and Council Tax band) it was straightforward to estimate the numbers of house- holds in each round day, while the proportion of properties in different valuation bands also provided a simple measure of socioeconomic status. Figure 13.4 maps the percentage of households in bands A and B (the lowest value properties) for each round day and indicates a general pattern of more affluent areas around the southern fringe of Norwich. Details of the refuse wei ghts (in metric tons) collected from each round- day during the period March 1999–February 2000 were supplied by South Norfolk Council. Other data on the weights of materials collected at Norwich Bungay Harleston Diss 10.0–24.99 Incomplete data 2.50 5 7.5 10 12.5 km 50.0–75.00 40.0–49.99 25.0–39.99 Wymondham Beccles FIGURE 13.4 Percentage of households in Council Tax bands A and B. (ß Crown Copyright=database right 2007. An Ordnance Survey=EDINA supplied service.) ß 2007 by Taylor & Francis Group, LLC. diffe rent types of bri ng faciliti es (e.g., mi nirecyc ling centers or CA sites) were obtain ed from the re levant authorities . Thi s incl uded in formatio n from sites outsi de the district that were kno wn to be used by South Norf olk residen ts. All these sets of opera tional re cords encompas sed at least a 12-m onth per iod which helped to take account of kno wn season al variati ons in co llection weig hts and so pro vide a more rep resentat ive pict ure of overall trends. An areal interpol ation method was then used to appor tion the site wei ght da ta among refuse co llection round da ys. This in volved first defin- ing sets of catch ment area s (bas ed on shortest travel time) for eac h typ e of bring site. Figure 13.5 shows an isochrone (travel time) map for CA sites and Figure 13.6 the resu lting catchmen t boun darie s. The travel -time map high- lights the influence of main roads (such as the A11, A140, and A146) radiating out of Norwich, while the catchment areas indicate that in several parts of South Norfolk (actually encompassing over 35% of residents), the nearest CA site was outside the district. Overlay techniques were employed to calculate the numbers of house- holds in each refuse collection round that fell within differen t bring-site catchments. Separate analyses were carried out for minirecycling centers and CA sites. From the resulting allocation tables it was straightforward to Estimated time in minutes <5 Sites Main roads 0 2.5 5 7.5 10 12.5 km >25 20–25 15–20 10–15 5–10 FIGURE 13.5 Estimated travel time by car to nearest CA site. (ß Crown Copyright=database right 2007. An Ordnance Survey=EDINA supplied service.) ß 2007 by Taylor & Francis Group, LLC. distribute the site weight data among collection rounds on the basis of household numbers and so reconstruct the average weekly elements of the domestic waste stream on a local area level. Statistical analyses were conducted to examine associations between socioeconomic variables (e.g., percentage of households in Council Tax bands A and B) and the relative use of different outlets for household waste across the 40 round days. A hierarchical cluster analysis was then performed which divided the round days into five distinct clusters using three classificatory variables based on the proportion of total household waste arising from (1) refuse collection rounds, (2) CA sites, and (3) mini- recycling centers. In the second phase of the research, the classification results were used to select a sample of six round days, and a door-to-door questionnaire survey was conducted with over 100 hou seholds in each area. Each of the five round-day clusters was represented in the sample. and a total of 615 ques- tionnaire responses were obtained in August and September 1999. The questionnaire survey elicited information on householders’ attitudes toward recycling activities and their use of local waste management facil- ities. In addition, an assessment of home composting activity was conducted in which bins or heaps found at 200 of the sampled households were examined, and the basic relationships betwe en garden characteristics and composting activity were investigated. Findings from this research are described elsewhere (Wheeler and Parfitt, 2002). Berg Apton Snetterton CA site 0 2.5 5 7.5 10 12.5 km Main roads Eye/Brome Norwich Beccles Wymondham Morningthorpe Ketteringham FIGURE 13.6 Predicted catchment areas (based on shortest travel time) of CA sites. (ß Crown Copy- right=database right 2007. An Ordnance Survey=EDINA supplied service.) ß 2007 by Taylor & Francis Group, LLC. The final phase of the pr oject used the datab ase generated by the local reco nstructio n of the hou sehold waste stream to exami ne the extent to wh ich recy cling perform ance in South No rfolk could be impro ved by esta b- lishin g new bring sites or introdu cing kerbside- collectio n scheme s for dry recyclabl es or green waste in part or all of the district. Thi s sce nario work require d a number of assum ptions about level s of partici pation in kerb side collections and the mater ial co mposition of the hou sehold waste stream. The former were based on da ta from a DETR sur vey of local autho rity recycli ng perform ance (Parf itt et al., 2001), wh ile the latte r used resu lts from resea rch condu cted for Project Integra in Ham pshi re (see http: == ww w.integr a.org.uk = about =res earch) and the Environ ment Agenc y (Parf itt et a l., 1999; Parfitt, 2000). Predic tions from the sc enario work were subse quently compared agains t the statuto ry recycling targets for Sou th Norf olk as sho wn in Ta ble 13.1 . 13.3 Results Consi derable variabil ity was found in the average weights of refuse co l- lect ed fo r in dividual round da ys. Figure 13.7 maps the average kilogram per hou sehold per week weights (herea fter kg=hh =wk) for the year March 1999– Febr uary 2000 and sho ws a rather mixe d pattern with, for examp le, no sim ple tendenc y for val ues to be higher in mo re urbanized nei ghborh oods. Equa lly, correl ations with the percenta ge of hou seholds in particul ar Coun- cil Ta x bands were relative ly weak. These resu lts may we ll reflect the comp lex bounda ries and diver se socioecono mic comp osition of some rou nd days (see Section 13 .2). However , clear er trends were appare nt in the use of bri ng sites, especial ly when the round days were grou ped into the five clusters rather than analyze d individually. For instance, greater TABLE 13.1 South Norfolk Recycling Performance and Future Targets Year Household Waste Recycling Rate 1998–1999 (base year) 10% 1999–2000 11% 2003–2004 20% 2005–2006 30% Source: From the Department for Environment, Food and Rural Affairs, in A Consultation Document on the Distribution of the £140 Million Waste Minimisation and Recycling Fund in England, DEFRA, London, 2001 (Annexe B). ß 2007 by Taylor & Francis Group, LLC. quantiti es of waste were taken to CA sites by re sidents in the northe rn half of the district where average travel times to suc h facili ties were low est. By contr ast, househo lders res ident in the most southe rly cluster of rou nd days (Diss, Harleston , and Long Stratton ) had the lowest level s of accessi - bility to CA facili ties and the highest me an quantitie s of refuse collectio n round waste. Results from the questi onnaire sur vey indicated that man y of these residen ts were driving at least 15 mi n to thei r nearest CA site, whic h conf irmed the results of the GIS-based calcul ations sho wn in Figure 13.5. There was also a good level of agree ment bet ween the GIS-predi cted site catchme nts (e.g., Figure 13.6) and the facilities that questio nnaire respon d- ents rep orted using. The main dif ference betwe en actual and pred icted behavi ors was that some res pondents used sites on the way to (or from ) Norwi ch or other large r town s rather than thos e that they resided neares t. Neverth eless, the extent of such discrep ancies was suffici ently small to provide confiden ce in the procedur e used to allocate the bring site da ta to round days . Figure 13.8 summarizes the results of the waste stream reconstruction exer- cise for the five clusters of round days. The boundaries on this map are those of the 40 refuse collection round days grouped into the clusters, whereas the grayscale shading depicts the average total recycling rate each week. For each cluster there is also a proportional symbol whose size represents the total Norwich Bungay Harleston Diss Wymondham Beccles 8.0–9.99 Incomplete data 2.50 5 7.5 10 12.5 km 15.0–17.00 12.5–14.99 10.0–12.49 FIGURE 13.7 Average refuse collection weights (kg=hh=wk) for round days, March 1999–February 2000. (ß Crown Copyright=database right 2007. An Ordnance Survey=EDINA supplied service.) ß 2007 by Taylor & Francis Group, LLC. weight of waste per household each week and the subdivisions show the proportions of the waste stream going to different outlets. A feature of Figure 13.8 is that some of the clusters are quite geographic- ally concentrated, even though there was no explicitly spatial factor in the definition of groups. Such an outcome, however, can be attributed to the contrasts in the accessibility and use of bring sites mentioned above. The highest recycling rates were associated with round days where resi- dents had both CA sites and minirecycling centers easily accessible, but there was considerable variability in performance with the best areas achieving over 30% and the lowest less than 15% (these figures include estimates of materials taken to outside the district). Analysis of the questionnaire responses showed that the amount of material put out for the refuse collection increased with household size and declined with involvement in recycling activities or composting. When asked what would encourage them to recycle more, around 50% of respondents mentioned the provision of kerbside collections, whereas another 20% suggested addi- tional local bring sites. These options were investigated further in a series of scenarios regarding additional recycling facilities. The results of this exercise soon indicated that further bring-site provision was unlikely to be of major Norwich Bungay Beccles Harleston Proportion of total kg/hh/wk Green waste Collection round Residuals at CA Dry recyclables Diss Cluster label Diss/Harleston/Long Stratton Incomplete data 6.3 3.3 Total recycling kg/hh/wk 4.0 4.8 5.3 Rural with high CA and minirecycling site use Urban north Rural with high minirecycling site use Wymondham FIGURE 13.8 Variations in waste generation and recycling activity by round-day cluster. (ß Crown Copy- right=database right 2007. An Ordnance Survey=EDINA supplied service.) ß 2007 by Taylor & Francis Group, LLC. [...]... Strategy 2000 for England and Wales Parts 1 and 2 (London: The Stationary Office) ß 2007 by Taylor & Francis Group, LLC Department for Environment, Food and Rural Affairs, 2001, A Consultation Document on the Distribution of the £140 Million Waste Minimisation and Recycling Fund in England (London: Department for Environment, Food and Rural Affairs) Department for Environment, Food and Rural Affairs,... Sunnenberg, G., and Haynes, R.M., 2003a, Accessibility to GP surgeries in South Norfolk: a GIS- based assessment of the changing situation 1997–2000 In Socio-Economic Applications of Geographic Information Science: Innovations in GIS 9, edited by Kidner, D., Higgs, G., and White, S., pp 181–198 (London: Taylor & Francis) ¨ Lovett, A.A., Parfitt, J.P., Hummel, J., Bone, A., Sunnenberg, G., and Pearce, S.,... high participation rates If this condition was not achieved, then a district-wide kerbside collection of green waste (Scenario 7) would be required to meet the 2005–2006 target 13. 4 Conclusions The GIS- based methodology developed in this research provides a means of integrating site and round weight data to develop a better understanding of household recycling performance at a subdistrict level Reconstructing... Pearce, S., 2003b, Research for Monitoring and Evaluation of the Peterborough Recycling Cell (Norwich: School of Environmental Sciences, University of East Anglia) ¨ Lovett, A.A., Sunnenberg, G., and Haynes, R.M., 2004, Using GIS to assess accessibility to primary health care services In GIS in Public Health Practice: Opportunities and Pitfalls, edited by Maheswaran, R and Craglia, M (London: Taylor & Francis)... waste collection and recycling strategies in England and Wales Resources, Conservation and Recycling 32, 239–257 Perrin, D and Barton, J., 2001, Issues associated with transforming household attitudes and opinions into materials recovery: a review of two kerbside recycling schemes Resources, Conservation and Recycling 33, 61–74 Tucker, P., Murney, G., and Lamont, J., 1998, Predicting recycling scheme... and Pocock, R., 1999, A Review of the United Kingdom Household Waste Arisings and Compositional Data, Final report prepared under contract No EPG 7=10=21 CLO 0201 for the Wastes Technical Division, Department of the Environment (Bristol: Environment Agency) ¨ Parfitt, J.P., Lovett, A.A., and Sunnenberg, G., 2001, A classification of local authority waste collection and recycling strategies in England... both local and national levels in the future (Cabinet Office Strategy Unit, 2002; Norfolk Waste Partnership, 2002) In these circumstances, the role of GIS in the planning and monitoring of household waste management services could be invaluable At present, such applications are relatively rare in local authorities and it is likely to require improvements in several aspects of data quality and storage... Environment, Food and Rural Affairs) Eastham, L.S., 1993, Council Tax (London: Citizen Advice Notes Service Trust) EnCams, 2002, Waste Segmentation Research 2002 (Wigan: EnCams) (http:= =www encams.org, accessed on November 2002) Lovett, A.A., 2000, GIS and environmental management In Environmental Science for Environmental Management, Second Edition, edited by O’Riordan, T., pp 267–285 (Harlow: Prentice-Hall)... helping South Norfolk to meet its statutory targets and therefore attention was focused on the impact of new kerbside schemes Figure 13. 9 illustrates the predicted outcomes of a series of scenarios concerning the introduction of new kerbside schemes The GIS database was used as part of this exercise to help determine where the introduction of kerbside-collection schemes would have the greatest impact... such as that presented in this chapter demonstrates the contribution that the use of GIS can make to the provision of information from which effective household waste management policies can be formulated Acknowledgments This research was funded as an ENTRUST (Landfill Tax Credits Scheme) project through UEAEA with support from NEWS (Norfolk Environmental Waste Services) and South Norfolk Council We . 13 A GIS- Based Methodology to Support the Development of Local Recycling Strategies Andrew Lovett, Julian Parfitt, and Gilla Su ¨ nnenberg CONTENTS 13. 1 Introduction 299 13. 2 Research. the results of the GIS- based calcul ations sho wn in Figure 13. 5. There was also a good level of agree ment bet ween the GIS- predi cted site catchme nts (e.g., Figure 13. 6) and the facilities. strategies in England and Wales. Resources, Con- servation and Recycling 32, 239–257. Perrin, D. and Barton, J., 2001, Issues associated with transforming household atti- tudes and opinions into