Marco Ballin Istat, Direzione Centrale delle Statistiche economiche Strutturali Statistiche strutturali sul settore agricolo

Transcription

1 Waste in Agriculture Sector: Data Collection and Modelling Approach I rifiuti nel settore dell agricoltura: raccolta dei dati e definizione di un modello Marco Ballin Istat, Direzione Centrale delle Statistiche economiche Strutturali Statistiche strutturali sul settore agricolo ballin@istat.it Giampaola Bellini Istat, Direzione Centrale Censimento della popolazione, ambiente e territorio Progetto metodologie e statistiche ambientali bellini@istat.it Riassunto: il Regolamento Europeo delle Statistiche sui Rifiuti (2150/2002/CE) è stato recentemente adottato, obbligando gli Stati Membri a produrre statistiche sui rifiuti generati e gestiti per settore economico, tra i quali compare quello agricolo. L Istat ha sperimentato la raccolta di dati mediante l integrazione di quesiti specifici nel questionario dell Indagine sulla struttura e produzione dell azienda agricola (FSS). I quesiti aggiunti nell indagine FSS condotta nel 2003 raccolgono informazioni su alcuni rifiuti e sotto-prodotti generati e gestiti a livello aziendale. Tecniche statistiche sono state applicate per identificare gruppi omogenei di aziende rispetto alla produzione di rifiuti plastici. Keywords: agriculture plastic waste, data modelling, FSS 1. Introduction The Regulation N. 2150/2002/EC recently adopted by the European Community on waste statistics obliges Member States to produce statistical data on waste and treatment in businesses and private households in the Community. As regards the of waste, Annex I refers to waste generated by households and economic sectors including waste arising from recovery and/or disposal operations. Annex II defines the list of recovery and disposal operations for which data have to be produced, according to the Waste Framework Directive (75/442/EEC and amending acts). Agriculture, Forestry and Fishing sector needed to be investigated, among others, to identify available data sources and appropriate methodologies quantifying waste generated and treated. 205

2 2. Farm Structure Survey as a tool for data collection on waste and treatment Data production activity on waste and treatment in agriculture can be realised exploiting diverse potential sources. In this study, for the first time, data collection approach through a sample survey was attempted in order to directly estimate waste generated at farm level, or to define waste related estimator factors. Thus specific questions were added to the existing 2003 FSS questionnaire. Criteria about the possibility and the way to add questions were mainly given by the correlation of additional questions with the phenomena analysed by the FSS, the necessity to know the existence of the different typologies of waste and of some management practices, the evaluation about ability of farmer to assess and value originated and managed waste, in order to have positive results with least costs (Istat, Bellini G., Cammarrota M., 2004). In particular the added questions in Section X refer to waste/by-product production in farm (item 59) and treatment practices adopted at farm level (item 60). For all substances considered under production perspective (Metal, Plastic, Pesticide waste, Waste from olive grinding and wine-making process), we asked to fill in the questionnaire whether the farm produces the specific substance, except for Metal, Plastic and Pesticide waste, for which the amount produced is requested. Referring to treatment used at farm level, a selection on substances and modality of recovery and disposal has been realised. The following substances are thus included: by-product of vegetable production, waste from olive grinding (waste water from olive grinding, dry and wet olive residues) and wine-making process, plus the item other. Typology of treatment are Purification, Incineration for energy production, on land -, Land treatment, Disposal in water bodies, Composting, Other treatment. Table 1 summarizes the results of data collection by region. In the first column the number of the survey responding farms is quoted. The second and the third columns contain respectively the response rate to item 59.1 (farms indicating at least one waste quantity or a plastic waste amount). The fourth column contains the fraction of agriculture activity - measured in terms of ESU (European Size Unit) - represented by respondents to item

3 Table 1: Number of farms surveyed by FSS and response rate for question 59.1 on waste and fraction of agriculture activity - measured in terms of ESU (European Size Unit) - represented by respondents to question Regions Farms Response rate % of ESU Farms Response rate % of ESU surveyed for Q represented Regions surveyed for Q represented by FSS Total Only by by FSS Plastic respondents Total Only by Plastic respondents Piemonte Marche Valle d'a Lazio Lombardia Abruzzo Bolzano Molise Trento Campania Veneto Puglia Friuli Basilicata Liguria Calabria Emilia R Sicilia Toscana Sardegna Umbria Italy (a) farms resulting respondent and active in FSS 3. Modelling approach for plastic waste at farm level The plastic material generated at farm level strictly depends on ruling of specific agricultural activities such as cultivation of crops under glass or with protective cover, mulching, irrigation, nurseries cropping. Referring to things of plastic nature, the list set up in a specific case study by ANPA and ONR (2003) includes: sheet to cover greenhouses and tunnel, hard sheet for greenhouses, film for mulching, not woven fabric, geomembrane to proof, different kind of rope and strings to substain crops and trees, harvesting nets (for olives, etc.), nets for trees protection, film for silage, tubes for different irrigation techniques. The analysis of the plastic waste by farms (Istat - Ballin, 2004) has been carried out by CHAID methodology 1, using as potential predictors some of the main structural variables collected by FSS as well as those previously mentioned. The complete set of potential predictors used for this exploratory analysis is: utilized agricultural area (ha); general types of farming (8 classes accordingly with first digit of typology classification); fresh vegetables in open field (ha); fresh vegetables under glass or other accessible protective cover (ha); flowers outdoor (ha); flowers under glass or other accessible protective cover (ha); total arable land (ha); vineyards (ha); olive plantations (ha); Citrus plantations (ha); fresh fruit and berry species of temperate climate zones (ha); kiwi (ha); nurseries (m 2 ); total permanent crops (ha); total area protected (ha); number of bovines (number); number of pigs (number); number of poultry (number); mulched area (ha); irrigated area (ha); area irrigated by dripping method (ha); european size unit (1 unit=1200 of total gross standard margin). The results of the analysis are summarized in table 2. 1 Electronic text book statsoft ( 207

4 Table 2: mean by farm of plastic waste, CV, number of farms (FSS estimate), distribution of the farms among groups, total plastic waste, distribution of the total plastic waste among groups Code of the group Mean of plastic (Kg) CVs of the mean (a) Number of farms (FSS estimates) Distribution of the farms Total plastic waste (tons) Distribution of plastic (A) 443,58 19, , ,5 4,68 (B) 1050,45 20, , ,5 2,78 (C) 1150,80 16, , ,5 1,22 (D) 3202,68 23, , ,0 1,95 (E) 7740,16 26, ,00 686,0 0,74 (F) 2315,08 15, , ,5 17,22 (G) 5123,64 13, , ,0 6,73 (H) 14996,94 11, , ,5 2,93 (I) 5,25 9, , ,5 8,62 (L) 10,61 9, , ,5 3,01 (M) 26,86 29, , ,0 4,97 (N) 45,02 11, , ,5 5,99 (O) 245,47 13, , ,0 7,30 (P) 442,32 18, , ,5 5,26 (Q) 122,59 15, , ,0 4,88 (R) 1422,61 32, ,02 518,0 0,56 (S) 232,22 13, , ,0 6,16 (T) 746,41 7, , ,5 8,08 (U) 5039,04 35, ,01 786,5 0,85 (V) 170,57 18, , ,0 1,32 (W) 586,21 25, ,07 962,0 1,04 (X) 1022,81 29, , ,0 1,42 (Y) 4842,46 25, , ,5 2,29 Total , ,0 100 It should be noted that a better model could be easily obtained by changing some of the parameters chosen in fitting the tree (number of levels below the root, number of units in each parent and child node, alfa level for F test, etc). For example allowing 6 levels of the tree (instead of 5 used in the present data analysis), CHAID would find out that final group (Q) could be split using ESU and vineyards. Nevertheless the present model gives some hints on the potentiality of the method. Assuming that it fits adequately the data, we can analyse how the total plastic waste, amounting to tons, is generated by the farms. The mean of plastic waste generated by groups ranges from a minimum of 5,25 kg - group (I) - to maximum of 15 tons - group (H). The five higher mean levels of plastic waste generated by the estimated groups are: kg for group (H) including farms characterised by fresh vegetables under glass or other accessible protective cover, with mulch and ESU equal or higher than 266,9; kg for group (E) including farms characterised by fresh vegetables under glass, without mulch and ESU equal or higher than 266,9; kg for group (G) characterised by fresh vegetables under glass or other accessible protective cover, with mulch and ESU lower than 266,9 and higher than 50,7; kg for group (U) characterised by fresh vegetables under glass or other accessible protective cover, without mulch, with irrigated area equal or higher than

5 Ha and with nurseries; kg for group (Y) characterised by farms without fresh vegetables under glass, with mulch and with ESU higher than 118,9. In terms of total amount of waste the most important is the group (F). It includes farms characterized by fresh vegetables under glass or other accessible protective cover, with mulch and ESU lower than 50,7. The group dimension is about units and generates tons of waste. 5. Concluding remarks This study represents the first attempt to understand the potentiality of a sample survey, the FSS in this case, in collecting waste data and possible uses of these data in applying a modelling approach. The survey approach revealed a critical point in the direct quantification of waste by the respondents, and missing responses were quite high. Reasons for this might due to a wide list of plastic things used at farm level, and whose quantity is difficult to assess. It has to be reminded that the quantity estimated trough the farm structure survey is strictly related to agriculture, and not to forest, selviculture and fishing activities; moreover materials linked to activities pertaining to agroindustrial activities were by definition excluded. Comparing these data on waste with other data released exploiting diverse potential sources can lead to controversial conclusion, due to existing differences. It should be better understood if the sector approach and the related waste definitions effectively explain differences with respect to other results. References APAT - ONR, Rapporto Rifiuti 2003, Roma. Istat, M. Ballin, Statistics on waste management in agriculture, forestry and fisheries - Waste Statistics Regulation 2150/2002/EC. Final report for an EC funded project. Istat, Bellini G., Cammarrota M., Methodological approach for statistics on waste generated in agriculture, forestry and fishing. Final report for an EC funded project. 209