Drinking water intake in feedlot cattle: A meta-analysis

High-quality drinking water is critical for efficient feedlot production. We performed a systematic review and meta-analysis of drinking water intake (DWI) of cattle in feedlots and developed and evaluated models that predict DWI against industry standard predictions. Three search engines were used to identify experiments that included DWI in feedlot cattle. Results of experiments were evaluated to ensure that these were: from peer-reviewed journals or theses published in English; in vivo and evaluated use of DWI in feedlot cattle; had appropriate study designs; had appropriate analysis of data; and contained data to determine the effect size for outcomes. Environmental and production measures were extracted, and temperature humidity index (THI) and heat load index (HLI) were calculated. A mixed model meta-regression with the random effects of experiment within study was used to evaluate the univariate effects of temperature, THI, HLI, body weight, dry matter intake, average daily gain, dry matter percentage of the diet, sex, and shade on DWI. Multivariable, multi-level models to predict DMI were constructed by backwards stepping separately for temperature and THI when average daily temperature is > 12 °C. Existing equations for DWI for feedlot cattle were evaluated and compared to the models developed in this study by Lin's concordance correlation coefficient rho_c, Pearson's correlation, and Bland and Altman's (1986) limits of agreement. Thirty-five studies with up to 131 experimental comparisons were included in the meta-analysis. The mean DWI was 31.2 L/d (SD = 4.31), mean temperature was 23.1 °C (SD = 24.10), and temperature humidity index (THI) was 69.1 (SD = 9.62). Two models to predict DWI for feedlot cattle were developed, the first with temperature and body weight both independently predicting DWI [Root mean square error (RMSE) = 2.40 L/d] based on 20 studies and 75 experiments. The second model had THI and dry matter intake independently predicting DWI (RMSE = 2.83 L/d) based on 14 studies and 57 experiments. These models are consistent with estimates from other individual studies, but lower than those of the NRC (2016). Drinking water intake decreased by 0.2 L for every 100 mg/L increase in water sulfate (SO4). Shade reduced DWI by 3.4 L/d. We provide updated robust estimates of changes in DWI for feedlot cattle with increases in environmental temperature, THI, body weight, and feed intake to guide industry in water resource allocation.