| 隨機小語 |
天下有大勇者，猝然臨之而不驚，無故加之而不怒。
[孟子] |
|
第二十八卷(1999)
總乳體細胞數分級與高體細胞數個體分隔集乳策略模擬
黃鈺嘉 蔡秀容 張秀鑾
行政院農業委員會畜產試驗所
本試驗的第一個目的為估計在不同的體細胞數(SCC)分級系統下，不同乳牛(或羊)族群的生乳等級分布。假設場總乳體細胞數平均值變數源自一截去單尾的常態分布，模擬族群平均為 10萬 到 250萬(SCC/毫升) ，標準偏差 (SD) 為平均值的 2/3, 2/4 ,2/5, 2/6至 2/7，來估計總乳的SCC 累積頻率。 如果總平均是 60萬而SD 是20萬，則比 100萬高而受罰的牧場估計為2.3%. 當族群平均為 40萬，而SD是13.3萬時,即使收乳標準降為 80萬，受罰的牧場仍將少於 0.2%。但是, 如果牧場規模縮小或其他原因而使場間變異增大時, 例如SD增加到總平均的 1/2，在前述情況的收乳標準，將有 8.3% 與2.3%的牧場受罰。 本試驗的第二個目的為探討區隔少數高體細胞母牛乳以符合收乳要求的因應策略。假設個體的體細胞分數為常態分布，模擬分析一萬個牧場，泌乳頭數從 30 到 250 頭，體細胞分數 (SCS)變動於 3.0 到 6.3間，SD為1.5 到 2.5間。一般而言，分隔百分之一到百分之五的高體細胞數個體，能使許多原場平均為 100~150萬的牛場降至100萬以下，以一組不同 場-年-月 的田間乳牛資料來查證這個模擬, 當分隔 3.67% 高於500萬的母牛, 能使場的月平均由 120萬降到 101萬，為協助酪農策略演練,一個以網為基礎的計算程式編寫於 http://www.angrin.tlri.gov.tw 可供參用。
關鍵語：體細胞數、體細胞分數、分級方法。
SIMULATION OF GRADING MILK BY SOMATIC CELL COUNT AND
SEPARATING HIGH SOMATIC CELL COUNT MILK FROM BULK TANK
Y. C. Huang, S. J. Tsai and H. L. Chang
Taiwan Livestock Research Institute, Council of Agriculture
The first objective of this experiment is to explore the grade distribution of bulk milk of dairy farmers (cows and goats) under different somatic cell count (SCC) grading systems. Truncated normal distribution with population means from 100,000 to 2,500,000(SCC/ml) and different standard deviations (SD) from 2/3, 2/4, 2/4,2/5,2/6 to 2/7 of means, were used to estimate the cumulated frequency of the SCC of bulk milk. If the population average of SCC is 600,000 and SD is 200,000, there were be 2.3% farms in penalty for higher than 1,000,000. As the population mean of SCC down to 400,000 and SD is 133,333, the farms in penalty will be less than 0.2% for higher than 800,000. But, as variance increased, such as reducing the herd size in all farms and the SD increased to 1/2 of mean, there will be more than 8.3% and 2.3% in penalty for the same population means and thresholds. The second objective is to understand the influence of few high somatic cows to SCC of bulk milk. Ten thousands farms with 30 to 250 milking cows with different somatic scores (SCS, 3.0 to 6.3) and different SD (1.5 to 2.5) were simulated under normal distribution assumption. In general, separating one to five percent high somatic cell cows can reduce SCC to less than 1,000,000, if herd average is higher than 1,000,000 but less than 1,500,000. A set of herd-year-month cow field data was used to verify this simulation, as 3.67% high SCC cows, higher than 5,000,000, was assembled separately which can reduce SCC from 1,198,128 to 1,006,221. For practicing the strategies, a web-based program for forecasting the possible outcomes of separately milking cows was available on http://www.angrin.tlri.gov.tw.
Key Words: Somatic cell count, Somatic cell score, Grading methods.
  
|
|
|
|
|
