During the study (Jan 2018 through Dec 2019), the average Temp and RH of the Sulaibiya area was 27.6 ± 2.8˚C and 38.1 ± 5.1%, respectively. The average THI was 72.1 ± 2.7 and the average THI_adj during the summer season was 44.5 ± 1.7. The average monthly Temp (˚C), RH (%), PPT (mm), THI and THI_adj are shown in Figure 1. During the month of July in summer season, the maximum Temp was 40.1˚C, THI 82.1 and THI_adj 50.3, while during the short winter season (Dec to April) Temp reach 14.5˚C and THI 58.0. The average monthly PPT in Kuwait varied from 0.0 to 133.7 mm.

The average daily WI per animal category was measured for lactating cows and estimated for all other categories. The average WI including dietary water was 100.6 ± 1.8 L·d⁻¹ for lactating cows, 66.1 ± 0.5 L·d⁻¹ for dry cow, 40.3 ± 0.3 L·d⁻¹ for heifers (1 year), and 17.9 ± 0.1 L·d⁻¹ for calves (Table 4). The WI for lactating cows is comparable with Shapasand et al. [35] who estimated the WI under average THI of 82.0 to be 118.2 L·d⁻¹ and Solomon et al. [36] who found that WI in a region with an average Temp of 37.5˚C to be 128.0 L·d⁻¹. In addition, the WI values for other animal categories agree with the estimated values as shown in Table 4. The high rate of water consumption is critical for the dairy cows to cope with the heat stress in Kuwait. In summer, the average Temp and THI in Kuwait were 34.0˚C ± 0.70˚C, and 78.4 ± 0.6, respectively which are beyond the thermal comfort zone of dairy cows (26˚C) [36]. Under this high heat stress cows consume more water to compensate for high dehydration rates [37]. Dairy cattle to acclimatize with Temp of 32˚C compared with 2˚C - 10˚C need to drink two to four times more water [12]. Studies have shown that WI is positively correlated with Temp [38], given the preference of cows for warmer water [39] [40].

^aMeasured WI of Holstein dairy cows (n = 689) in Kuwaiti dairy farms during the period of 24 month (mean ± SE). ^b Computed by the use multiplicative adjustment factor to the average WI values of Le Riche et al. (2017) and Ward and McKague (2019) which are 10.52, 23.67 and 38.75 L cow⁻¹ d⁻¹ for calves, heifers and dry cows, respectively. ^c WI is estimated by increasing the average WI values of Le Riche et al. (2017) and Ward and McKague (2019) by 53.20% based on Equation (1) [15]. ^d WI is estimated by increasing the average WI values of Le Riche et al., (2017) and Ward and McKague (2019) by 87.30% based on Equation (2) [16]. ^eDietary Water based on the feed rations of dairy farms in Kuwait. ^f Estimated WI values and DW.

1) Milk Production

During the study, the average number of lactating cows was 83.0 ± 2.0, 477.0 ± 0.4 and 123.0 ± 1.5 for farms A, B and C, respectively. The average number of non-lactating animals (calves, heifers and dry cows) was 59.0 ± 12.7, 118.0 ± 22.4 and 92.0 ± 21.5 for farm A, B and C, respectively as shown in Table 1. Throughout the study, the average milk production was 17.0 ± 0.4 L cow⁻¹ d⁻¹ (Table 5) and ranged from 15.5 ± 0.6 L cow⁻¹ d⁻¹ in the summer to 19.3 ± 0.4 L cow⁻¹ d⁻¹ in the winter. The effects of farm (P < 0.0001) and season (P < 0.001) were significant. These results agree with Razzaque et al. [41] who studied the MY of 25 locally born and adapted Holstein lactating cows and found that it was 16.9 L cow⁻¹ d⁻¹. Milk production of imported pregnant Holstein heifers in Kuwait ranged from 8.2 to 14.8 L cow⁻¹ d⁻¹ [42] [43]. Another investigation evaluated the performance of dairy heifers in Kuwait over 8 years and found that the average MY of lactating cows was 10.7 L cow⁻¹ d⁻¹ [44], which revealed an adaptation problem to Kuwait’s climate when compared with the performance of these cows in their country of origin [45].

The low rate of MY in summer compared to winter is consistent with the recognized seasonal pattern of MY [46]. Geers et al. [47] reported that milk production decreased when Temp was >16˚C. In the current study, the average fat and protein contents of milk produced in Kuwait was 2.7% ± 0.0% and 3.9% ± 0.0% with similar values between seasons. Our milk fat content was lower than that reported by Razzaque et al. [41] which was 3.1%, while the milk protein content was higher compared with 3.1% reported by the same study. These differences could be due to the effect of Temp on milk composition [48] along with other reasons including the genetic factors, farm management practices, the components of feed rations or water quality and its availability to the herd. The overall average of FPCM was calculated to be 15.3 ± 0.4 kg cow⁻¹ d⁻¹ over the 24-month period (Table 5).

2) Whole Farm Water Utilization

The overall average water utilized on the confined dairy farms was 505.2 ± 22.6 L cow⁻¹ d⁻¹ of which 218.9 ± 2.5 L cow⁻¹ d⁻¹ (43.3%) was for drinking, 194.7 ± 22.9 L cow⁻¹ d⁻¹ (38.5%) was for spray and 58.9 L cow⁻¹ d⁻¹ (11.7%) was for cleaning of milk house (Table 5). The farm management (P < 0.001), season (P < 0.0001) and THI_adj (P < 0.0001) had a significant effect on the total amount of water used on the farm which was higher in summer (636.8 ± 31.1 L cow⁻¹ d⁻¹) compared to the winter (314.3 ± 4.4 L cow⁻¹ d⁻¹).

In winter, the main components of water use were associated with WI (69.5%) followed by washing of milking parlour (19.4%), dietary and milk water (11.8%). However, during the summer, most of the water used on the farm was associated with the cooling system (either in the barns or milking parlour). In the summer, the cooling (spray) water accounted for about 51.7% of the daily water use, followed by WI (34.4%) and wash water (9.0%). The volume of the wash water was varied based on the farm (P < 0.0001) and THI_adj (P < 0.001), which is consistent with the conclusion of Harner et al. [19] and Janni et al. [49] who reported that water used on 14 dairy farms in Minnesota was between 8.6 to 35.3 L cow⁻¹ d⁻¹ and was varied from 140% to 400% due to several reasons including animal performance and weather conditions.

Average WI of dairy livestock ranged from 218.4 ± 2.6 L cow⁻¹ d⁻¹ in winter to 219.2 ± 3.8 L cow⁻¹ d⁻¹ in summer which is consistent with the results of the NRC [50] and with the findings of Murphy et al. [51] and West [52]. Additionally, Gorniak et al. [53] found that as THI increased over 30, the WI by lactating cows increased. The high WI of cows in Kuwait is likely due to the high Temp combined with the THI value (78.4 ± 0.6), which was adjusted to 48.5 ± 0.3 based on the use of cooling pads and spray water systems to reduce the effect of heat stress on the cows. Similarly, VanderZaag et al. [9] reported a high rate of WI by lactating cows in summer compared to winter. However, in the current study, the water used in washing the milking parlour was 58.9 ± 2.3 L cow⁻¹ d⁻¹ (ranged from 57.5 ± 3.0 in summer to 60.9 ± 3.4 in winter), which is higher than the average values in temperate conditions. Robinson et al. [54] reported that the average wash water for tie stall barns in ON was 30.2 L cow⁻¹ d⁻¹. VanderZaag et al. [9] reported an average value of 28.1 L cow⁻¹ d⁻¹ for washing the milking system. The high rate of water utilized for washing in Kuwait could be due to the use of the high volume hoses (without a trigger valve) to clean the milking parlour, equipment and milk tank throughout the milking event, coupled with the harsh weather conditions and the design of the housing system that governs the cleaning process.

3) Blue WF

The average whole farm blue WF (L·kg⁻¹ FPCM) was estimated to be 40.1 ± 2.8 L·kg⁻¹. On a seasonal basis, the blue WF (L·kg⁻¹) was 19.2 ± 0.8 in the winter and 54.5 ± 4.0 in the summer. The effect of farm, season and THI_adj (P < 0.0001) was significant. The blue WF was affected by changes in water utilization and milk production. The blue WF was negatively correlated with milk production (r = −0.73) and positively associated with the total farm water utilization (r = +0.74).

Our findings showed that the whole farm water utilization was higher in summer as water spray and WI increased. Armstrong [45] and Fuquay [55] concluded that hot weather conditions especially during summer depressed the performance and milk production of lactating cows. However, to reduce the effect of heat stress a protocol of cooling by ventilation, shade, spray and fans were implemented on Kuwait dairy farms in Kuwait. This cooling program aims to reduce the effects of heat stress that were conveyed in the value of THI_adj, which proved to be successful in alleviating the effect of high ambient THI.

The increase in WI during the summer was not accompanied by an increase in milk production. Murphy et al. [51], found a positive correlation between milk production, WI and the minimum Temp. Likewise, the same relationship was found by Cardot et al. [56]. However, for Kuwait’s dairy farms, the quality of water and the level of total dissolved solid (TDS) could negatively affect the rate of milk production and the effectiveness of WI under heat stress condition as identified by Shapasand et al. [35] and Challis et al. [57].

^aWinter from December to April (151 day), Summer from May to November (214 day). ^bmeans ± SE (n = 49 observation for each farm, 689 total number of cows). Note: Total or average value could differ due to rounding. ^cCalculated based on IDF (2015). ^dTHI_adj for summer season only. ^eMeasured WI values (L cow⁻¹ d⁻¹) of lactating cows along with extrapolated values for the dry cows, heifers and calves. ^fUnit: L cow⁻¹ d⁻¹. ^gL water kg⁻¹ FPCM; L of water = kg of water. ^*Total spray includes the water used to spray the cows in barns and milking parlour to reduce heat stress during the summer season only. ^**Wash water includes water used for washing of cows, milking parlour, and equipment in milk house.

Effluent water on dairy farms consists mainly of rinse water from the milking parlour, spillage of milk residue, urine and manure runoff [58] [59]. The operations in the milking house require systematic cleaning of the dairy cows, floor area and equipment, which generates wastewater. The calculation of the grey WF for dairy farms is based on the assumption that if the actual nutrient concentration is ≤MAC discharge, then no further dilution or treatment for the effluent water is required, and the grey WF is zero.

NO₃-N and PO₄ Concentrations of Milk House Wastewater

The concentration of NO₃-N in effluent water ranged from 4.2 ± 0.7 mg·L⁻¹ to 12.3 ± 1.7 mg·L⁻¹ with an overall average of 8.8 ± 1.2 mg·L⁻¹. However, it is not a serious problem as its concentration was <10 mg·L⁻¹ [30]. Mantovi et al. [60] and Morin et al. [61] similarly reported that the wastewater generated from the milk house was characterized to have 0.3 to 6.5 mg·L⁻¹ NO₃-N, respectively. However, the concentration of PO₄ was found to be a problem in Kuwait’s dairy farms. Based on KEPA guidelines, the maximum acceptable level of PO₄ in effluent water is 30 mg·L⁻¹.

The analysis of milk house wastewater showed that the PO₄ concentration was 75.2 ± 25.6 mg·L⁻¹ for farm (C), 66.2 ± 17.3 mg·L⁻¹ for farm (B) and 105.2 ± 31.5 mg·L⁻¹ for farm (A) with an overall average of 82.2 ± 14.3 mg·L⁻¹. The PO₄ content was affected significantly by farm (P < 0.05) and the month of sample collection (P < 0.0001).

The high concentration of PO₄ is due to the use of different detergents as well as the presence of cow urine, manure and milk in the wash water [62] [63]. Mantovi et al. [60] and Morin et al. [61] stated that wash water generated from the milk house has a PO₄ content of 61.3 to 306.5 mg·L⁻¹. In addition, Singh et al. [24] found different concentrations of PO₄ based on farm size which was 413.3 mg·L⁻¹, 341.1 mg·L⁻¹ and 192.9 mg·L⁻¹ for large, medium and small farms, respectively. The PO₄ content in milk house wash water in Kuwait was higher than Janni et al. [49] for 14 dairy farms, which was 56 mg·L⁻¹ due to different weather conditions and management practices.

To estimate the grey WF, the PO₄ concentration in effluent water must be diluted to be within the MAC. The daily effluent water for farms (C), (B) and (A) was 24,501.9 ± 9,503.3 L·d⁻¹, 14,217.3 ± 1,692.6 L·d⁻¹, and 5,648.8 ± 404.8 L·d⁻¹, respectively. Accordingly, the grey WF was calculated to be 50,774.7 ± 20,066.0 L·d⁻¹ for farm (C), 34,054.9 ± 11,340.1 L·d⁻¹ for farm (B) and 21,126.7 ± 6,820.1 L·d⁻¹ for farm (A). Effluent water and grey WF per day were affected significantly by farm (effluent water (P < 0.0001); grey WF (P < 0.01)) and the period of sample collection (P < 0.01).

The effluent water per lactating cow was 208.9 ± 81.4 L cow⁻¹ d⁻¹, 29.7 ± 3.5 L cow⁻¹ d⁻¹ and 55.5 ± 5.8 L cow⁻¹ d⁻¹ for farm (C), farm (B) and farm (A), respectively. Consequently, the grey WF was calculated to be 432.4 ± 172.0 L cow⁻¹ d⁻¹ for farm (C), 71.2 ± 23.7 L cow⁻¹ d⁻¹ for farm (B), and 215.2 ± 76.8 L cow⁻¹ d⁻¹ for farm (A). Grey WF per animal basis was affected significantly by farm (P < 0.0001) and month of sample collection (P < 0.1).

The grey WF per FPCM (L·kg⁻¹ FPCM d⁻¹) was estimated to be 52.1 ± 22.4 for farm (C), 3.4 ± 1.3 for farm (B) and 13.6 ± 5.1 for farm (A). The grey WF per FPCM was affected significantly by farm (P < 0.0001) and the period of sample collection (P < 0.01).

In Kuwait, the variations between farms in the volume and nutrient contents of effluent water (PO₄ concentration was 75.2 ± 25.6 mg·L⁻¹, 66.2 ± 17.3, 105.2 ± 31.5 mg·L⁻¹ for farm (C), farm (B) and farm (A), respectively) could be due to the variation in the effluent volume and nutrient status between farms along with the type and amount of pipeline chemicals used, washing duration and frequency [64] that influenced by the amount of animal excretions (urine/feces), PPT level and the balance between pipeline wash water and manure runoff. Commonly, as the herd size increases, the volume of effluent water per animal decreases [65]. Nevertheless, sometimes there is no correlation between herd size and the volume of washing water, suggesting that the capacity of milking parlour and PPT level can have a major effect [66].

In general, the grey WF of dairy farms contributes to water depletion and deterioration. The dilution of water to reach the acceptable level is a burden that should be considered by finding alternative solutions to manage water. This can be achieved using various treatment methods including constructed wetlands that are considered to be an effective in allowing for water re-use and treatment as they remove >95% of pollutants [67].

The determination of the compositions of effluent water for dairy farms is critical as there are no previous assessments in Kuwait for the volume and nutrient contents of milk house effluent water. This will enable farmers to plan for strategies to manage wastewater on their farms to reduce the cost and increase the efficiency of water utilization.