Moving geometric average or raw milk classification based on single measurements?
2006
Buehlmann, A. (Agroscope Liebefeld-Posieux (ALP), Posieux (Switzerland)) | Eichenberger, M.
In many countries, especially in EU, the moving geometric average, for several samples per month over a series of months, is taken into account for raw milk classification, whereas in Switzerland single measurements are carried out. In order to compare the systems, the Swiss Testing Laboratory results from the years 2001 to 2004 were analysed with the different methods. The average rates of non-conformity based on single results are 1.53 % due to an excessively high total microbial count, and 4.49 % because the cell count tolerance limits are exceeded. When the three-month moving geometric average is taken into account and the current Swiss limits are applied, the percentage of non-conformity is only 0.77 % for the total microbial count and 1.81 % for the cell count. This is a reduction of 50 % and 40 % respectively compared with single measurements. The clear seasonal periodicity of the rates of non-conformity is evened out and slightly delayed in time with the moving geometric average. Classification based on single measurements is stricter and more sensitive to short-term changes. This depends of course on the rate of sampling. In order to achieve equivalent rates of non-conformity, the limit for the total microbial count would have to be lowered from the present 200 imp/ìl (BactoScan-8000-count units) to ca. 130 imp/ìl. In the case of the cell count, where 350 SCC/ìl (Somatic Cell Counts) is laid down, this would have to be lowered to ca. 280 SCC/ìl. The use of the moving geometric average gives producers a certain protection, in that it sof-tens the weight of possibly undeserved extreme values. However, in view of the reliability of present laboratory equipment and the optimised logistics of sampling for routine assessment of raw milk, it is justifiable to retain single value classification. This method is obviously stricter and gives good security to milk products manufacturers and consumers. Moreover, there is a graduated price reduction system in Switzerland, which comes into play in case of repeated non-conformity of individual suppliers. However, if the moving geometric average were used, a greater frequency of sampling would be advisable.
显示更多 [+] 显示较少 [-]