The Scottish Government

« Previous | Contents | Next »

Listen

2. Materials and methods

2.1 Data sources and acknowledgements

SAMS records of the years 2003-2007 were received from SG. The authors thank Gary Ferguson for continued assistance, and Kevin Duffy and other SG personnel for work on the SAMS extracts. CTS records of the years 2003-2007 were received from Defra. The authors thank Steve Holdship for continued assistance, Simon Roberts for the agreements, and the IBM team for work on the data extracts. Both the CTS and SAMS extracts were firstly processed by Dr Nick Savill to a consistent form for all the three livestock species, as described below in Section 2.2.

Results of the June 2007 Agriculture and Horticulture Census of Scotland (Agri-census) were received from RERAD. The data are derived from an annual survey, and are not a complete census of the livestock populations in each year. In each annual census, for two thirds of minor agricultural holdings in Scotland the data are estimates based on previous years' census returns. For mapping, boundaries of the Scottish Civil parishes from 1991 were used, as available from the UKBorders service at EDINA.

A list of the registered livestock markets, showgrounds, collection centres and abattoirs in Scotland was received from RERAD, and further cross-checked through a joint effort between EPIC and the Animal Health and Welfare Division of RERAD. The authors are especially thankful for the information provided by Derek Wilson and Paul Honeyman, and assistance from Nick Ambrose and Mike Lamont. However, this list is not entirely complete at present because: i) seasonal livestock shows can be held on different farms from year to year; ii) farms can operate as livestock dealers; iii) farms can operate as collection centres, for example, for livestock moved off the islands; and iv) other reasons such as inconsistencies in CPHIDs.

Scottish parish numbers within each of the four island archipelagos were received from Andrew Urquhart at the Animal Health and Welfare Division of RERAD. This classification has been developed for the purpose of regionalisation during livestock infectious disease control efforts.

2.2 Data processing

From CTS single move records ( i.e. normal ON, normal OFF, birth, death, etc.) individual cattle movement histories were derived by pairing OFF and ON records allowing for the date of movement. Then, cattle moved from an individual holding to another individual holding on the same calendar dates were combined into batches.

Next, data from CTS and SAMS were processed to a consistent table form for all the three livestock species. This table included for each moved batch the CPHIDs of the departure and destination holdings, date of movement (arrival at the final destination rather than market date was used unless indicated), total number of animals in the batch, if the final destination was an abattoir, and for movements via a livestock market - the market's CPHID and market date. Then, using (C) county component of CPHIDs the holdings were classified into those located in Scotland and cross-border in England and Wales. Next, using the (P) component of CPHIDs, the Scottish holdings were classified into those located in each of the four island archipelagos and the mainland. As mentioned, holdings on Skye were included in the mainland. For cattle only, the types of departure and destination holdings as given in CTS, animal production usage, and departure dates were additionally specified in the batch movement table. Cattle were classified into animals in different production practices based on purpose of the animals' breed as specified by RADAR and provided in CTS. Movements of dairy cattle, beef cattle and those of dual and unknown purpose breeds were extracted from CTS separately. However, when all the batch movements in 2003-2007 of beef cattle and cattle of dual and unknown purpose breeds were combined, beef animals constituted >99.8% of the total batches moved, and, correspondingly, batches of dual and unknown purpose cattle <0.2%. The three groups were combined into a single category 'beef cattle' for further analysis. It should be noted that classifications used to account for dairy versus beef cattle by Defra/ BCMS and in the annual Agri-census of Scotland are different. Therefore, care has to be exercised when co-using data from these two information sources.

Abattoirs processing sheep and pigs were uniquely identified in SAMS by the EC-registration numbers and CPHIDs. However, for many cattle movements onto Scottish abattoirs only partial information was available for a given abattoir in the CTS database: either the CPHID, or EC-number or geo-reference. Multiple data sources were utilized to locate the cattle slaughter moves. In particular, the abattoirs with missing CPHIDs but known EC-numbers were checked against the EC-number- CPHID key tables: one derived from SAMS, and the other received from the Livestock Traceability Unit of RERAD. If only the geo-reference was available for an abattoir, the county and parish were identified using the boundaries of the Scottish Civil parishes 1991, as available from the UKBorders service at EDINA.

In the developed sheep and pig batch movement datasets, all movements recorded as 'to slaughter' were defined as slaughter-moves. Next, CPHIDs were cross-checked against the list of registered livestock markets, showgrounds, collection centres and abattoirs in Scotland to correct any missed slaughter-moves and market-moves, and classify the show-moves. Sheep and pig movements to incinerators, cutting rooms and knackers' yards were added to slaughter-moves. Occasional movements off slaughter-move destinations were observed. Such movements constitute <0.1% of sheep batches moved off in Scotland in 2004-2006. Movements to and from known lairages near markets, and to and from collection centres were added to the market-moves. Movements through premises with a single CPHID operating both as a showground and a market were treated as market-moves. As mentioned, the farm-showground movements were counted and the return moves excluded. Landless keepers (holdings with 7000's (H) component of the CPHIDs) and known sheep dealers registered in each county and parish were combined into a distinct category. A few recorded movements from a wildlife park, to and from veterinary clinics, to embryo transfer units and to a laboratory were excluded.

In sheep movement records in SAMS in 2003-2007, the end of year 'edge effect' was rarely observed. For example, 30 sheep batches were recorded through a market on 30 Dec 2003, but reported arriving on destination farms on 1 Jan 2004. For sheep, such movements were counted in the year of arrival. The edge effect was larger for 2002-2003 when the SAMS system was starting. For <0.1% of sheep batches moved in 2004-2006 the final destination was not recorded, however, all such batches went to a market, and therefore were counted in market-move categories, either they were or were not indicated as those ultimately going to slaughter. A few other SAMS records were excluded as anomalous, for example, if the recorded year of market was after the arrival year. Such anomalous records constituted <0.1% of the sheep batch move records in 2004-2006 otherwise eligible for analysis.

In pig movement records in SAMS in 2003-2007 incomplete records with missing holding of destination were rarely observed. If encountered in a given year, such records were considered anomalous and excluded from the analysis. During 2004-2007, the pigs were not reported to SAMS to be moved off Scottish holdings to other holdings or for slaughter via markets. Limited (<4.0% of the total pig batches) movement via markets was recorded in 2003. Therefore, from SAMS alone the market trade with pigs in Scotland appeared to be uncommon. However, when we cross-checked the CPHIDs of the holdings moving pigs in Scotland in 2003-2007 against the list of livestock markets and other registered livestock dealers a certain, though very limited, amount of pig movement to the markets was detected, with the overall increasing trend from 3.1% of the total pig batches moved in Scotland in 2003 recorded being moved to a market, to 7.1% in 2006. In 2007 such proportion was 6.8% of the total pig batches moved. All pig batches recorded being moved to the markets in 2003-2007 were not reported going to slaughter as the final destination. However, pig batches moved from the markets (<2.0% of pig batches moved off annually in 2003-2007 had a market as the holding of departure) were reported both going to abattoirs and other holdings. As with sheep, occasional movements off slaughter-move destinations were observed for pig batches moved in 2003-2007.

When classifying the sheep and pig movement types, movements via a Scottish market between two other Scottish holdings were combined with similar direct movements without a market, except the farm-market-farm and farm-market-abattoir categories. For example, movements from a Scottish landless keeper via a Scottish market to another Scottish landless keeper were combined into a single category with direct movements from one Scottish landless keeper to another. Cross-border movements were processed similarly if the departure and destination holdings were on the different sides of the border. However, if the departure and destination holdings were on one side of the border and the intermediate market on the other, this was specified explicitly. For example, a movement from an English landless keeper to an English abattoir via a Scottish market was fully specified. The domestic show-moves category included all movements on and off registered Scottish showgrounds, including the few movements off the showgrounds to Scottish markets or abattoirs. Similarly, the cross-border show-moves category included all the cross-border moves with either holding of departure or destination being a registered showground.

For cattle movements, the CTS location type classification was used to define the movement types. Farm-moves were defined as those to or from agricultural holdings and common land. Movements through calf collection centres and known lairages near markets were added to the market-moves. Movements with known departure and market but unknown destination holding, or known market and destination but unknown departure holding were added to the market-moves, similar to that done in sheep movement analysis. This could include some movements with the final destination being an abattoir. Movements to cutting rooms, cold stores, head boning plants, incinerators, knackers' yards, hunt kennels, meat products plants, protein processing plants and centres for collection of BSE surveillance materials were treated as movements to slaughter. A few movements off premises classified as destinations of slaughter-moves were recorded. Such movements accounted, for example, for less than <0.1% of beef cattle batches moved off holdings in Scotland in 2004-2006. Cattle movements to or from artificial insemination centres, semen shops and embryo transfer units were excluded. Show-moves were defined as movements to or from a holding classified as a showground in CTS. As for sheep and pigs, cattle farm-showground movements were counted for the analysis, and the return moves were excluded. Cattle movements to or from holdings of unknown business were excluded (none for beef animals in 2003-2007). Landless keepers were treated as a separate category, which included the businesses classified as landless keepers in CTS and any other holdings with 7000's (H) component of the CPHIDs.

After the OFF-ON pairing batch move protocol for CTS records and geo-referencing of cattle moves to slaughter on Scottish parish level as applied in this analysis, few records still contained only information on the holding of departure in Scotland, but insufficient information on the market and destination. This constituted, for example, <0.1% of dairy and beef cattle batches moved off Scottish holdings in 2006. Such records were excluded from the analysis. Similarly for the movements when both the departure and market information were missing. Occasionally, anomalous records with the departure year being more than one year behind the arrival year were encountered. This constituted, for example, <0.1% of the total beef cattle batches moved off Scottish holdings in 2005. Such batches were excluded from the analysis. If either the departure year, or both the departure and market years, were one year prior to the arrival year this was considered a normal edge of the year effect. Such movements, for example, constituted <0.1% of the total dairy and beef cattle batches reported off Scottish holdings in 2005 and arriving at the destination premises in 2006. In the analysis, such cattle movements were counted in the year of departure.

Data derived from the CTS records have been recommended as the most accurate for estimating the current cattle population size in GB by RADAR ( RADAR, 2006). However, good agreement between the estimates of total cattle population in GB between the GB Agri-censuses and CTS in 2003 has been reported (Green and Kao, 2007). In this report we considered all the three major livestock species in Scotland. To maintain consistency between the data for different species, cattle population in Scotland in 2003-2007 was referenced as available from the annual June Agri-censuses of Scotland.

When mapping the livestock spatial distribution, the outlines of the Scottish Civil parishes 1991 were linked to the parish (P) component of the CPHIDs of agricultural holdings in the June 2007 Agri-census. The data for mapping were further cleaned to accommodate slight discrepancies in parish names and boundaries. Mapping colour panels were adjusted to achieve visual consistency so that the relative differences in densities of animals of different livestock species could be appreciated.

2.3 IT, statistical and mapping tools

Processing of extracts from CTS and SAMS was done in the Python programming environment and then with SAS® 9.1.3 for Windows, SAS Institute Inc., Cary, NC, USA. Analysis was performed with the SAS 9.1.3. The results were graphed with Microsoft® Office Excel 2003. Maps were produced with the ArcView® 9.2, ESRI, Redlands, CA, USA.

2.4 Regionalization and dictionary of movement types

Livestock movement was analysed from the standpoint of regionalisation of Scotland into the four island archipelagos and mainland, with additional separation of the cross-border movement. Allocation of Scottish parishes within each island archipelago was as used by RERAD for livestock infectious disease control, in particular, as used for the livestock movement control during the FMD 2007 outbreak in England, with the rest of Scotland considered as the mainland. For all the three investigated livestock species, consistent regionalisation was used and consistent movement types were defined.

The movements were investigated for the following regions and directions:

Cross-border into mainland Scotland
Cross-border off mainland Scotland
From Scottish islands to mainland Scotland
From mainland Scotland to Scottish islands
Onto and off the Shetland Isles
Onto and off the Orkney Isles
Onto and off the Western Isles
Onto and off islands under the jurisdiction of Argyll and Bute

Short dictionary for the movement types definitions:

F - farm
Ab - abattoir
Mt - market
LK - landless keeper
GB - located in GB outside Scotland, i.e. in England or Wales

Table 1. Complete list of livestock movement types on and off Scottish holdings and abattoirs recorded for sheep, cattle or pigs in 2003-2007

^a 'no market' identifies a direct move not via a market

« Previous | Contents | Next »