LSIRD NAPLIO CONFERENCE PAPERS

Location of optimal areas for the development of an alternative livestock species: the cashmere goat

Kate Corcoran, Barry Dent, Julian Smith & Pablo Lara

SUMMARY

Optimal areas of location for the development of livestock enterprises within the European Union can be identified based on a classification methodology using ranking and weighting of physical, climatic and socio-economic characteristics for areas being evaluated. Two less Favoured Areas (LFA) in hill and upland situations in Andalucia in Spain and Scotland were classified sub-regionally in order to identify optimal areas for the location and development of cashmere goat production.

ARC/INFO Geographical Information System (GIS) facilitated the ranking and ordering of spatially variable characteristics and was an integral part of the overall enterprise location 'decision support system'. The technique of spatial overlay allowed areas coincident for optimal characteristics to be identified and mapped in sequence to produce a final map of optimal areas. The order of coverage overlay (thematic layers of information) was significant with areas where characteristics were judged to be most limiting, eliminated first. The order of coverage overlay was as follows:- LFA status (Sub-group 'Areas at Risk of Depopulation') overlaid with areas satisfying altitude > 400 ms. Areas satisfying both these criteria were then overlaid with climatic maps and resultant areas classified as optimal, sub-optimal or unsuitable. Administrative areas corresponding with the second overlay were identified and scored based on a number of socio-economic criteria to give a final map with a five category optimality range. The spatial units used in the analysis were the 'comarcas' in Spain and the 'region' in Scotland. Socio-economic factors considered in the administrative scoring process were: presence and level of sheep and goat subsidy; land area (SAU); goat and sheep population; milk and meat goat population; goat and sheep slaughter numbers; number and size of farms and population density.

The process was robust in firstly identifying areas of suitability and secondly in providing a methodology which could take account of changing levels in one or more characteristics.

INTRODUCTION

Structural adjustment is now firmly a feature of EU farming and a both rural and farm diversification promoted widely through EU policy. The policy context of the post Structural Funds reform period since 1988, has been: firstly to move away from production centred support instruments which, for the livestock sector has seen a high proportion of previous support decoupled from output, though this has been in part replaced by headage payments. A second post-reform measure seeks to provide the means for a wider integrated rural development policy through support measures targeted at designated Objective Areas, with Objectives 1 and 5B relating directly to disadvantaged 'Less Favoured Areas' (LFA)s. For agriculture in these areas this has resulted in funds being made available for diversification of farm enterprises where enterprise options to provide viable farm-family income are limited and where average incomes in LFAs are currently 50-70% of those achieved in non-LFA areas (Commission of the European Communities, 1993). The impact of farm diversification to supplement farm income has been disappointing and uptake low. While current policy seeks to encourage diversification, a more integrated approach of extension support, enterprise economic assessment and initial targeting of areas with a high probability of success is necessary if more diversification is to occupy a permanent place on farms. Policy makers need information or rules upon which to base decisions on improving the uptake of alternative enterprises. This study concentrated on developing a methodology for identification target optimal areas for the alternative ruminant, the cashmere goat. The production fibre and meat from cashmere goats represents a potentially sound extensive diversification option (Milne, 1993; Corcoran, 1994). Optimal locations are likely to be for those areas which provide most favourable conditions when the interaction of physical, climatic and socio-economic characteristics are evaluated. The hypothesis presented here is that 'if the optimal occurrence of characteristics are spatially located and ranked, based on the scoring and weighting of relevant factors, then areas which are 'most favourable' for a particular enterprise can be positively identified and ranked.

Geographical Information Systems (GIS) provide effective tools for this type of reasoned land evaluation processes through the development of decision rule structures and predictive modeling (Eastman, 1993). Siddiqui, Everett and Vieux, (1996) used a GIS spatial analytical hierarchy process (AHP) to take into account regulatory restrictions, area attributes and site assessment criteria provided by experts and/or users, for a preliminary landfill site assessment in Cleveland County, Oklahoma. In the present study, AHP ranking methodology using choice heuristics and selection rules, was used to locate ideal areas for the development of cashmere goat enterprises. These were Andalucia, Spain and Scotland, UK.

SPATIAL DECISION SUPPORT

Decision rules typically contain procedures for combining criteria into a single composite index together with a statement of how alternatives are to be compared using this index. They are structured in the context of a specific objective. The nature of that objective serves to guide the development of the structure used in a particular decision rule and the relative weighting given to criteria used are in turn guided by expert opinion. In the analysis of spatial or thematic information, decisions on the selection and ranking of features within each thematic layer and the order in which these layers of information are overlaid is critical for the results generated. For spatial overlay using Boolean criteria (/constraints), suitability solutions usually lie in the union or intersection of conditions. In order to evaluate areas of suitability for cashmere goats in Andalucia, the following spatial overlay sequence was used to select areas of coincident and consistent high suitability: areas of favourable LFA status/ risk of depopulation, Objective 1 or 5B status, altitude and climate (favouring goat systems). Superimposed on these areas were administrative areas (comarcas) each carrying an index of suitability based on criterion scores and weightings of those factors considered to be essential in the evaluation process. As factors were continuous, (less to more important) a weighted linear combination in which weightings were applied to factors, followed by a summation of results yielded the suitability scores (Eastman, 1993).

i.e. S = w ix i where s = suitability

wi = weight of factor I

xi - criterion score of factor I

Scales upon which criteria scores were based were subjectively standardised. The development of weights was based on a pairwise ranking process developed by Satay (1977) using AHP where the weights sum to one. Weights are derived from the comparison of the relative importance of the two criteria involved in determining suitability for the stated objective (Table 1). On a 9 point continuous scale, a value is given for the relative importance of paired factors laid out in a matrix such that a high priority rating is given a high score for the factor row (e.g. 7), whereas a low relative rating for the factor row would be given an inverse score of 1/7 (Figure 1).

Figure 1: The continuous Rating Scale

For Andalucia, the suitability score index was calculated using the single objective 'areas which offer best potential for a cashmere enterprise at the current time for each relevant

Table 1: Pairwise comparison matrix for assessing the comparative importance of factors

(Rating of the row factor relative to the column factor)

factor' (comarcas level), based on the multi-criteria evaluation of factors shown in the row-column matrix in Table 1. A similar pairwise comparison procedure was carried out for Scotland .

Table 2: Ranked Weights

Finally the procedure sums the weights within each column, divides each co-efficient by the sum and then averages over all columns to produce a best fit approximation set of weights. This procedure achieves a good approximation with Satay's (1977) best fit, produced by computing the principal eigenvector of the pairwise comparison matrix (Table 2).
Suitability maps were produced for both areas in ARCVIEW following the union of spatial and weighted attribute data. Spatial polygon data coverages were digitised using ARC/INFO ADS at a scale of 1:1,000,000. At this scale discrete farm polygons are lost and mapped output reflects potential at a low resolution. The overlay procedure followed to compile the 'optimal location' map is given in the APPENDIX.

RESULTS

The coverage overlay order in the UNION process and the ITEMS chosen for reselection influenced the polygons satisfying the selection criteria and the spatial distribution of selected polygons on the final suitability map. The inclusion of 'Mountainous Areas' within the LFA selection would have taken in additional tracts in Huelva, Cordoba, Jaen, Granada, Almeria and the northern part of Malaga. However the addition of the constraint of areas with an altitude greater than 400 metres was felt to allow the 'theme' of depopulation to be modeled at altitudes where hill farms exist. The climate ranges selected were: 'cold and hot steppe' and 'temperate warm and hot summer', these having to occur in areas above 400 metres. No upper constraining limit of altitude was included in the evaluation process although this constraint is valid and would have eliminated further areas from the final suitability map.

An evaluation of land areas suitable for cashmere goat enterprises in Andalucia., was carried out at the University of Cordoba in 1996. This produced a high degree of consistency with areas identified and provided validation of the AHP methodology (Lara, 1996). This ground-truthing was essential when the weighting and ranking of factors was carried out outside Andalucia.
Map 1

Map 1, includes areas in the following comarcas: Alhama, Alto Almazora, Anavalo Occidental, Bajo Almazora, Baza, Campina Alta, Campina del Norte, Campina del Sur, Campo Tabernas, El Condado, Guadix, Huescar, Iznalloz, La Loma, La Sierra, La Sierra Norte, Pedroches, Penibetica, Sierra, Sierra Cazorla and Sierra Morena. The factor 'common grazing' not relevant in Andalucia, would serve to modify index scores in Northern Spain.

Map 2 indicates areas where goats might targeted under the present 'no subsidy' regime in Scotland. In the absence of subsidy paid on goats, LFA 'specialist sheep' was taken as a proxy of land use suitability for fibre goat systems. Precipitation was then overlaid on this 'composite' coverage with the rainfall range divided into six suitability categories falling between 500 and 3000 mm. Areas with a rainfall above 1500 mm were considered least suitable with those between 500 - 1000 mm suitable on a sliding scale. This product coverage was then overlaid with the administrative regional boundaries which carried the suitability index score for attribute agricultural data. The area with highest 'suitability' lies in the north around Inverness where rainfall is lowest at 500 mm per annum. Moving to the next suitable zone - 'fair to moderate suitability', more extensive tracts of Highland Region, upland Tayside and the southern part of Grampian Region were selected. These areas have a low to moderate rainfall with high levels of extensive sheep systems. Areas in the west in Highland Region, Argyll, Strathclyde and Ayrshire and in the north east in Grampian Region, Fife, Central Region Lothian and borders have 'moderate to poor' suitability ranking. This rank reflects a combination of factors which scored low for goat production in these areas: high rainfall, poor access to grazing or the presence of a high level of non-LFA farms made up by arable, intensive livestock, or mixed farm types for example.

Map 3 shows areas of suitability within Scotland if subsidy were extended to goats in line with levels applied in southern European countries i.e. at 90% of sheep subsidy. LFA areas carrying Highland and Island Enterprise attracts highest subsidy and scored highly. Map 3 reflects the a situation where higher scoring, extensive specialist sheep target areas, frequently to fall within areas of higher rainfall, which carry a lower ranking. This gives an end-picture of higher suitability in northern highland areas, becoming marginally less suitable moving south through Argyll. Grampian, Central Region Tayside, Strathclyde, Dumfries and Galloway and eastwards through the Borders region to carry a 'moderate' suitability rank. The areas of Fife, East Lothian and Berwickshire have a 'least' suitability allocation being the principal cereal growing areas of Scotland. Other areas to the west in Ayrshire also carry a low suitability rank due to a high concentration of dairying in the locality.
Map 2

The Islands region varies in suitability ranking: Orkney, being lowland in type has a fair suitability based on lower levels of existing sheep production, while the opposite situation exists in the north in Shetland. The Western Isles have a high suitability rank due the high score given to common grazing in this area. The availability of common grazing and high sheep density offset the high negative rainfall score and small holding size which could reduce the suitability scoring. For Scotland expert opinion has suggested that areas with higher precipitation should be allocated lower scores at the pairwise ranking matrix stage.

Map 3

DISCUSSION

The process of spatial overlay and analytical hierarchy process for indexing factors in the attribute evaluation was robust, firstly in identifying areas of suitability and secondly in providing a methodology which could take account of changing levels in one or more characteristics by rerunning the overlay process and/or recalculating the best fit set of weights. If the location of the optimal occurrence of characteristics can be scored in this way and areas which are deemed 'most favourable' for a particular enterprise identified, it follows that the process could evaluate between competing enterprises for allocation of land resources for example. Optimal location could also be alternatively given as area where profit can be maximised from an enterprise; also fibre production and carcasse weight or the farm-family labour resource could be used most efficiently.

The methodology proved successful in identifying areas of suitability in Andalucia based on validation from within the area. The procedure is transferable in that local weighting and ranking of key factors which affect or influence area selection can be carried out for a range of livestock location exercises across Europe. Validation in Scotland was not possible at the current time due to the low goat population and number of goat enterprises.

CONCLUSIONS

Processes for evaluating information which helps decision makers and practitioners choose between alternatives are vital, if benefits derived from public monies are to maximised and unsuccessful schemes minimised. This study shows that initial target areas can be identified based on expert evaluation of current knowledge.

REFERENCES

Corcoran, K. (1994). Opportunities and Constraints for a European Fine Fibre Industry. Fine Fibre News (European Fine Fibre Network). No.3, pp 24-29, 1994.

Eastman, R.J., (1993). IDRISI GIS - Update Software User Manual, Version 4, Clarke University, Massachusetts, 01610, USA.

Commission of the European Communities, (1993). Farm Accountancy Data Network - The Agricultural Income Situation in Less Favoured Areas of the EC. Office for Official publications of the European Communities, L-2985 Luxembourg.

Environmental Systems Research Institute (ESRI), (1993). ARC/INFO User Manual.

Lara Velez, P. (1995). An internal study of areas of suitability for cashmere goat enterprises in Andalucia, University of Cordoba.

Milne, J.A. (1993). Economics of fibre Production. In Agriculture: Alternative animals for fibre production. EU Report No. EUR 14808 EN. (ed. A.J.F. Russel). pp 95-102.

Ministerio de Agricultura, Pesca y Alimentacion (1990). El Censo Agrario de 1989; Ganaderia - Ovinos y Caprinos (Livestock census of Spain - Sheep and Goats.)

Satay, T.L. (1977). A Scaling Method for Priorities in Hierarchical Structures, Journal of Mathematical Psychology, 15, 234-281.

Siddiqui, Everett and Vieux, (1996). Landfill Siting using Geographical Information Systems - A Demonstration, Journal of Environmental Engineering - ASCE, 1996, Vol. 122, No.6. pp.515-523.

The Scottish Office, 1996. Economic Report on Scottish Agriculture. Agriculture and Fisheries Department.

Appendix.

Procedure for Spatial Analysis of map and attribute data

Institute of Ecology and Resource Management, University of Edinburgh, Edinburgh, UK

University of Cordoba, Cordoba, Spain

Return to index