Recipient Organization
UNIVERSITY OF CALIFORNIA, DAVIS
410 MRAK HALL
DAVIS,CA 95616-8671
Performing Department
Population Health & Reproduction
Non Technical Summary
The research proposed is to design a novel temperature humidity index gauged by a dairy cows response in terms of milk production to its immediate and preceding temperature and humidity experience, hence cumulative. Current indices are based on milking and beef cows, furthermore they do not account for the cumulative effect of such environmental determinants. In addition, to date, no index has been developed based on milk production, the primary outcome of interest in dairy management systems. We anticipate that cumulative time above a threshold THI, modified by cool off periods will predict the probability of production losses due to heat stress with more accuracy than maximum daily THI. Results of this research will be used to generate a 3-dimensional matrix with the three variables temperature, humidity and duration, which can be used to replace the current 2 dimensional THI for dairy cattle heat stress.
Animal Health Component
100%
Research Effort Categories
Basic
(N/A)
Applied
100%
Developmental
(N/A)
Goals / Objectives
Currently, a combination of temperature measurement and Temperature-Humidity Indexes (THI) are utilized to determine when cooling interventions are necessary. THIs are designed to take into account the effects of humidity in amplifying heat stress due to the reduction in effectiveness of evaporative cooling via sweating and panting[6]. Bohmanova et al reviewed the correlation between milk production losses and 7 commonly used THI equations (developed approximately 40 years ago) and identified the need for new indices that reflect the modern dairy cow and environment [7]. Most of these commonly used THIs utilize only test day temperature and humidity measurements, and identify a THI greater than 68 as an environment that may lead to reduced milk production[8]. However, it has been shown that the climatic conditions in the days preceding yield measurement are more important than the conditions on the day of collection[4], [9]. All previous indexes have utilized daily maximum, minimum, and mean measurements of climatic conditions, which lack necessary detail to reflect actual time-periods of heat stress, specifically consecutive days of exposure[10]. The commonly used maximum daily THI will miss actual time above or below the heat stress threshold, which has shown to be more important than daily extremes or means [10-13]. Hourly measurements will enable the shape of daily temperature changes to be examined, allowing the effect of duration of heat stress and duration of any intervening cool-off periods to be estimated. The importance of these cool-off periods during which a cow may recover from heat stress was demonstrated by Scott et al, who showed that nighttime temperatures above 21 degrees celius prevent cows from dissipating stored daytime heat and subsequently maintain elevated rectal temperatures the following day[11]. Various environmental modifications are applied to limit production losses associated with heat stress, however these interventions require the utilization of economically and environmentally expensive resources, primarily water and electrical power. To ensure optimal utilization, cooling systems need to be applied objectively only when required to limit production losses. The dairy industry will benefit from a Cumulative THI (C-THI) that combines current environment measurements with historical accumulated heat stress exposure to more accurately predict milk production losses . OBJECTIVES. 1) Estimate the effect of temperature and humidity measures on test-day and each of the previous 30 days temperature and humidity on a herds total test-day milk production 2) Compare models with and without interaction between exposure to a THI > 68 and duration of exposure to assess whether its effect is additive or multiplicative. 3) Design an equation that predicts the C-THI cut-off at which a drop in milk yield has been observed, and subsequently validate research findings using data collected from VMTRC core herds. Inputs of this equation include a herds temperature and humidity exposure and duration of such an exposure.
Project Methods
Data collection and housing: Hourly temperature and dew point estimates collected from various weather stations located in the Tulare, Kern, Kings, and Fresno counties will be downloaded from an online database[15]. Cow level data by test-day will be obtained from Californias DHIA records processing center[16], including cow ID, birth date, parity, DIM, milk yield, fat %, protein %, SCC and used to generate the remaining herd level factors that will be included in models predicting effect of heat stress. Both environmental and production data will be housed in an SQL database and queried to generate each herds test day production, related environmental temperature and humidity at test-day (time t) and at days prior to test day (t-n, n ranging from 1 to 30) and the factors to adjust for including herd size, breed(s), inter test-day periods, production level, parity distribution (mean parity), days in milk distribution (mean herd DIM), distribution of somatic cell counts (SCC) as a proxy for herd heath, proportion of cows with inverted fat : protein ratio as a proxy for herds nutritional status and proportion of cows with double milk peaks as a proxy for bovine somatotropin (BST) use. Data analyses: The effect of temperature and humidity measures on test-day herd milk yield will be estimated using a general linear mixed model (model 1) with herd and time as random effects to account for dependencies in milk yields within the same herd and over time. Model 1 will also take the form of a herd-specific spline model with a node in a summer month when milk production is the least will be explored (based on year round observed herd milk yields) and will be used to model the shape of the herds milk yield. Model 1 will adjust for herd demographics (above), and include temperature and humidity on test-day added into the model for days t-1 to t-30 one at a time until no further significant effect of previous temperature and humidity is identified. The latter step will help identify the maximum number of days that have an association with a current test-day milk yield (example 7 days). A second general linear mixed model with a similar random effects structure will be implemented to estimate the cumulative effect of temperature and humidity on milk yield. Model 2 will be used to analyze our dataset after reclassifying each herds temperature and humidity exposure in the previous n days (number of days prior to test-day that were identified from model 1 as significantly affecting a herds current test day, in our example 7 days). The final model 2 and its coefficients for temperature, humidity and previous temperature and humidity exposure (example: for 7 days) will be used in identifying the equation that predicts the C-THI experienced by a herd on any day by solving for the product of temperature, humidity and previous temperature and humidity coefficients.