Effects Of Heat Stress Of Different Duration On The Production Performance Of Cows And The Establishment Of Warning Model Of Heat Stress

Heat stress(HS)of dairy cows is one of central problems in dairy farming.HS decreases milk yield and compromises milk compositions of dairy cows,which have been traditionally attributed to the reduction in DMI during HS.However,a number of studies in recent years have demonstrated that reduced DMI only explains 35%to 61%of the decrease of milk yield during HS.To explore other reasons of milk yield decrease in HS,this study compare the effects of short-term heat stress(STHS)and long-term heat stress(LTHS)on cows in climate-controlled chambers in terms of the production.We determined the physiological index,production performance,rumen fermentation parameters,ruminal microbiome,blood parameters and rumination behavioristics of cows during STHS and LTHS to study the negative effect of HS on dairy cows.In order to accurately predict the occurrence of HS in dairy cows,this study explored multiple environmental factors and 9 most common thermal index thresholds when dairy cows are in heat stress,and studied their accuracy.Because HS will lead to a decrease in the rumination time(RT)of dairy cows,so a new warning model for HS of dairy cows was constructed in this study based on the changes of RT by machine learning technology.These main results are summarized as follows:1.Compared with STHS,LTHS decreased milk yield by 17.2%and dry matter intake by 12.6%,indicating that LTHS caused a more severe decline in milk production and feed intake.In addition,LTHS decreased milk protein concentration by 6.8%and milk protein yield by 22.4%.In comparison with STHS,LTHS decreased rumen liquor volatile fatty acid(VFA,29.7%),blood glucose(11.6%),and nonesterified fatty acid(NEFA,13.6%)concentrations,but increased milk urea nitrogen(MUN)by 15.1%,blood urea nitrogen(BUN)by 8.6%,and creatine concentrations by 15.4%.The production of casein is significantly positively correlated with the concentrations of Ile,Met and Tyr in blood,and the concentration of casein in milk is significantly positively correlated with the concentration of Ile in the blood.Concentrations of Ile,Leu,Val,Met and Tyr in blood are significantly positively correlated with DMI.In addition,LTHS causes decreases in the relative abundance of Bacteroidetes,Prevotellaceae and Prevotella in the rumen.Our results suggest that while reduced feed intake may be mainly responsible for reduced milk production during STHS,impaired rumen metabolism and suppressed mobilization of adipose tissue could be main reasons for further reduction in milk yield during LTHS.2.13741 sets of total ruminant time(TRT)and corresponding meteorological factor data were analysed through broken-stick regression,the results showed that TRT was decreased in HS.The decrease of TRT can be used as a marker of the occurrence of HS.The results of broken-stick regression analysis showed that when the average daily temperature in summer exceeds 23~oC,or the average daily relative humidity exceeds 67%,the average daily temperature of the black bulb exceeds 24~oC,the maximum daily radiation intensity exceeds 500 W/m~2,or the cumulative daily radiation value exceeds 19 MJ/m~2,TRT of cows will decrease.In addition,it was found that the temperature humidity index(THI)has the highest correlation with TRT.When THI is higher than 71,cows will suffer HS.Compared with other thermal indexes,THI is more suitable as an evaluation index for HS of dairy cow in northern China.In addition,the decrease of DRT in high temperature environment is the reason for the decrease of TRT,and NRT is not affected by HS.3.8436 sets of TRT and meteorological data collected from dairy farms were used in the K nearest neighbor(KNN)algorithm of machine learning technology by MATLAB for prediction models of TRT,NRT and DRT.K values of TRT,NRT and DRT were calibrated to 58,48 and 45 by cross-validation.The KNN accuracy of TRT,NRT and DRT was 0.6171,0.2798 and 0.8138,respectively.KNN algorithm is suitable for TRT and DRT prediction.According to the degree of TRT decrease in HS,HS can be classified into the following four categories:TRT?534 min/day,no HS;506 min/day?TRT<534 min/day,mild HS;469min/day?TRT<506 min/day,moderate HS;TRT<469 min/day,severe HS.The predicted results of TRT can be used as a warning of HS.