2.5.1. Misting Procedure

Misting the inside of a stationary trailer and/or access to external fans when available is suggested as an option by the TQA [20] for summer conditions, although the placement and the operating pressure of nozzles are not clearly addressed and can be customized by transport companies. Misting in this study indicated spraying water into the air or onto the back of pigs and/or onto the bedding materials. For the trailer used in this study, 20 misting nozzles (TX-V626, Teejet Technologies, 2 in Zone 1, 6 in Zone 2, 2 in Zone 3, 1 in Zone 4, 6 in Zone 5, and 3 in Zone 6) were located along the middle length on the bottom level, and the right-side length on the top level. While loading pigs at commercial pig barns, two methods of misting were observed based on the water availability, including misting during the process of loading the pigs onto the trailer, or misting briefly after the trailer was fully loaded. Fan banks were located in the abattoir and were available only when the trailer was parked by the fan banks prior to unloading. The duration of misting varied among practices and usually lasted about 5 to 15 min.

#### 2.5.2. Effects of Trailer Management Methods for Hot Weather

Temperature measurements from all of the thermistors in each zone were averaged for the segmen<sup>t</sup> during transport (depart from pig barn where pigs were loaded until the arrival at the abattoir). A temperature and humidity index (THI) was computed (Equation (1)) using the center-zone temperature and RH data [14,25]. The average and maximum THI inside the trailer, and the average THI for outside condition during the transport segmen<sup>t</sup> were obtained.

$$THI = 0.8T\_{db} + RH(T\_{db} - 14.4) + 46.4\tag{1}$$

Four response variables were tested individually to assess the spatial thermal variability within the trailer during transport: (1) average temperature difference between the trailer zones and the outside condition; (2) average THI; (3) maximum THI; and (4) difference between the maximum THI recorded in the trailer and the average THI for outside condition. Each response variable was analyzed for 20 hot weather trips in thermal categories including *Warm*, *Mild*, and *Very Hot* by analysis of variance (ANOVA) for effects of thermal category, trailer zones, misting methods at loading (no misting, misting during loading process or misting after all pigs were loaded), and zone x misting interaction. The analyses were done by PROC MIXED in SAS (version 9.4, [35]). PROC UNIVARIATE was used to verify normality of the dependent variable and accepted at *p* > 0.05. The Tukey–Kramer test for differences of least square means was used to determine significant differences between variable means (*p* < 0.05) due to unequal sample sizes.

#### 2.5.3. Temporal Thermal Profile inside the Trailer for Hot Weather

Center-zone temperatures in the six trailer zones, the outside temperature and center-zone THI conditions were plotted against elapsed time for two representative monitoring trips, including a morning trip in the *Very Hot* category and an afternoon trip in the *Warm* category. Important segments, including arriving at a pig barn, loading pigs onto the trailer, misting inside trailer, road transport, arrival at abattoir, access to fans and misting during waiting, and unloading pigs are identified. The temporal temperature profile was investigated for both hot and cold weather conditions, and the temporal THI profile was conducted for the hot weather conditions only, with one representative sample trip from each data set.

#### 2.5.4. Spatiotemporal Visualization of Variability inside the Trailer for Hot Weather

Temperature distribution patterns on a trailer deck basis (top or bottom) provide visualization of multidimensional temporal variability within the trailer and insight into ventilation patterns during transport. Based on our understanding of the ventilation patterns, cooler regions indicate proximity to an air inlet and hotter regions indicate air outlets in the trailer, except in the case of misting or sprinkling. This is expected from the understanding of pressure distributions on the outside of a moving trailer, with lower pressure inside the trailer that drove air in.

Data from 84 pig-level temperature sensors were linearly interpolated in Matlab® to develop a series of animations that represent the dynamic spatiotemporal profile across the trailer inside [14]. The animations are interpreted as follows: red indicates warmer temperatures and blue indicates cooler temperatures; green circles represent pig-level thermistors and their locations within the trailer; and a colored text box indicates critical events which occurred during the monitoring trip. The animations subjectively describe the ventilation patterns within the trailer, areas receiving benefits from the cooling methods (including misting onto pigs and access to external fans), and the evaporative cooling persistence into the transport segment.

The spatiotemporal visualization of trailer interior temperatures was developed for both hot and cold weather conditions. Representative trips monitored during *Very Hot* (2 trips), *Warm* (1 trip), and *Mild* (1 trip) categories were selected to create the animations. Table A1 (Appendix A) provides the descriptive information for these four trips, with viewable animations available in the Supplemental Materials accompanying this article.

#### *2.6. Evaluation of Cold Weather Trips*

#### 2.6.1. Boarding and Bedding Procedures

Trailer boarding (covering of trailer openings) was recommended in the TQA for winter conditions. The TQA guidelines outlined 25%, 50%, 75%, and 90% boarding coverage for specific winter outside temperature ranges. Three variations in boarding patterns (uniformly along the trailer sides, boarding gradually more towards rear, and boarding all at the back) are not addressed in the previous TQA guidelines and are evaluated in this study. The industry often applies boarding uniformly along the trailer side.

Bags of conventional kiln-dried pine shavings were used as bedding in this study (0.06 m<sup>3</sup> each). The use of bedding was characterized by the number of bags placed onto the trailer prior to each monitoring trip and was designated for specific outside temperature ranges: light bedding (1 or 2 bags); medium bedding (3 bags); and heavy bedding (4 to 6 bags). According to the TQA guidelines, the trailer operator had the flexibility to slightly adjust the number of bags of bedding within each designation [20].

#### 2.6.2. Effects of Trailer Management Methods for Cold Weather

Trailer managemen<sup>t</sup> methods for cold weather investigated were trailer boarding percentage, bedding level, and boarding position. A boarding-bedding combination was created for the analysis, where LM indicates light boarding (25%) and medium bedding (3 bags); MM indicates medium boarding (50%) and medium bedding; and MH indicates medium boarding and heaving bedding (4–6 bags).

The average temperature difference between the trailer zones and the outside condition was analyzed for 16 trips under Tout thermal category *Cold* by ANOVA for effects of trailer zones, combination of bedding and boarding percentage, boarding position (as a nesting factor in the boarding and bedding combination, including boarded evenly, more towards the rear, or all at the rear), and zone x boarding position interaction. The analysis was performed by PROC MIXED statement in SAS (version 9.4, [35]). PROC UNIVARIATE was used to verify normality of the dependent variable and accepted at *p* > 0.05. The Tukey–Kramer test for differences of least square means was used to determine significant differences between factor means (*p* < 0.05) due to unequal sample sizes.

For the four trips monitored under the *Very Cold* category, the analysis was simplified due to the lack of combinations of boarding-bedding placement and boarding position in trips monitored, and only the effect of trailer zones on trailer interior temperature rise was analyzed. The boarding-bedding combination and boarding position were not tested. A representative trip in the *Very Cold* category was selected to create the spatiotemporal animation (Table A1 (Appendix A)).

## **3. Results and Discussion**

For all 40 trips analyzed, the average duration of a complete trip was 3.5 ± 0.8 h, ranging from a minimum of 0.8 h to a maximum of 4.9 h [14]. The segmen<sup>t</sup> during road transport had an average of 2.4 ± 0.8 h, ranging from 0.9 to 4.2 h. A total of four pigs were found DOD for all trips monitored, out of approximately 7000 market-weight pigs transported.

#### *3.1. Evaluation of Hot Weather Trips*

#### 3.1.1. Effects of Trailer Management Methods for Hot Weather

Table 2 lists descriptive statistics for the 20 transport trips categorized in thermal categories *Mild*, *Warm*, and *Very Hot*. Variables summarized include number of trips, the mean ( ±standard deviation) of outside temperature, trip duration, and waiting time are provided. The range in average outside temperature varied from 16.7 to 35.3 ◦C. Mean trip duration averaged 2.1 to 2.7 h, and average waiting times before unloading varied from 4 to 28 min.


**Table 2.** Descriptive statistics for 20 monitored transport trips categorized as thermal categories *Mild*, *Warm*, and *Very Hot*.

> 1 N/A: there were no instances of this combination occurring.

Table 3 provides results of effects of trailer misting methods for hot weather managemen<sup>t</sup> for the corresponding trips included in Table 2. There was no effect of zone on any of the measured variables for these hot weather trips. Results for the main effects (thermal category, trailer zones, and misting methods) are described in the following paragraphs.



means for the maximum THI for the during transport segment. 7 Numbers indicate the least square means for the difference between the maximum THI documented for inside the trailer and the average THI for the outside condition for the during transport segment. A–C Different superscripts within a row under the analyzed factor indicate the means differ significantly (*p* < 0.05) using the Tukey–Kramer test for differences of least squares means.

#### *Animals* **2018**, *8*, 203

**Thermal category.** Results from Table 3 show that the thermal category had a significant effect on the trailer interior temperature rise. For *Mild* and *Warm* categories, trailer interior was warmer than Tout but cooler for the *Very Hot* trips, presumably from evaporative cooling of mist water applied prior to transport. When compared to the *Mild* category, average THI was higher for the *Warm* and *Very Hot* categories. THI levels between 78 and 84 are considered dangerous for livestock animals, and THI levels greater than 84 constitute an emergency condition [14,15,25]. We observed dangerous average THI conditions during transport for all trips in the *Very Hot* category, and both dangerous and emergency maximum THI for *Warm* and *Very Hot* categories. The difference between maximum THI recorded in the trailer and the average outside THI was notably different between thermal categories (*p* < 0.001). The difference of maximum THI decreases when the outside temperature became higher.

**Trailer section.** Trailer sections (front: Zones 1 and 4; middle: Zones 2 and 5; and rear: Zones 3 and 6) did not affect the average temperature rise, average THI, or maximum THI. The trailer thermal environment was uniformly hot during each of these trips. The uniform distribution of temperature rises during hot weather trips agree with the results reported by [10], where no difference in temperature among compartments of a double-decked pot belly trailer for weaned piglets was observed for trips with 29 ◦C average ambient temperature. They did not report the THI conditions in their study.

**Misting.** Misting before the start of trip slightly cooled the trailer for hot weather categories, as measured by average temperature rise in the trailer (*p* < 0.001), maximum THI recorded (*p* < 0.01), and the difference between maximum THI and average outside THI (*p* < 0.001). However, the variation in average temperature rise between no misting, misting during loading, and misting after loaded was less than ±1 ◦C. Only misting during loading resulted in an average interior temperature cooler than outside. One explanation may be that the average temperature difference over the entire transport segmen<sup>t</sup> may mask any relatively shorter-term cooling benefits from misting, which usually only lasted 5–15 min. This is further confirmed by average THI which was also not different between misting at loading categories (*p* > 0.1). A previous study [36] that assessed thermal environment in a goose-neck horse trailer during summer conditions reported that the THI was not uniform in the trailer, and more extreme conditions were found toward the front, suggesting that the front section openings served as outlets. The THI in their trailer was more affected by the ambient condition rather than different trailer positions [36], which matches with our results in which the maximum THI difference was found in the top front zone ( ΔTHImax = 9.2), although our study did not note any statistical difference in any THI responses for different zones. The effects of misting after loading, was a higher average temperature rise than misting during loading (*p* < 0.05). The effects of no misting were similar for mean THI to both misting conditions, and a lower maximum THI (*p* < 0.01) than either misting condition, and a lower ΔTHImax (*p* < 0.001) than misting after loading. These results show that one can expect misting during loading to result in similar or reduced interior temperature, similar average THI, and higher maximum THI in the trailer during transport compared to no misting. The use of misting pushed the thermal environment to a dangerous condition for at least some portion of the transport segment, although it did not have a lasting effect.

#### 3.1.2. Temporal Thermal Profile inside Trailer during Hot Weather

Obtaining temperature and THI profile of the trailer during the entire course of the transport trip is helpful to understand the nature and to assess the variability of the interior thermal conditions that are encountered by pigs. The spatiotemporal thermal profile was investigated for both hot and cold weather conditions. Figure 2 and Figure 3 show the change of center-zone temperatures and the THI respectively from the six trailer zones with elapsed time from the time the trailer left the home base until the trailer was completely unloaded at the abattoir. The two example trips during hot weather shown here are: (a) a morning trip in the *Very Hot* category that was completed from 5:12 a.m. to 1:24 p.m. and (b) an afternoon trip in the *Warm* category, from 9:45 a.m. to 5:20 p.m. The outside

temperature is also plotted on Figure 2, and the trip segmen<sup>t</sup> and trailer managemen<sup>t</sup> were numbered with explanation provided.

**Figure 2.** Ceiling-centered temperature profile representing 6 zones inside a pig trailer during: (**a**) a *Very Hot* morning trip; and (**b**) a *Warm* afternoon trip. Zones 1 to 3 represent trailer top deck, and Zones 4 to 6 represent bottom deck. Events occurred during the trip are numbered on the figure as follows (if present): **1.** Arrival at a commercial pig barn; **2.** Loading; **3.** Misting applied inside trailer; **4.** En route; **5.** Arrival at abattoir; **6.** Access to fans and misting during waiting; **7.** Unloading.

**Figure 3.** Ceiling-centered THI profile representing 6 zones inside a pig trailer during: (**a**) a *Very Hot* morning trip; and (**b**) a *Warm* afternoon trip. Zones 1 to 3 represent trailer top deck, and Zones 4 to 6 represent bottom deck. Events occurred during the trip are numbered on the figure as follows (if present): **1.** Arrival at a commercial pig barn; **2.** Loading; **3.** Misting applied inside trailer; **4.** En route; **5.** Arrival at abattoir; **6.** Access to fans and misting during waiting; **7.** Unloading.

Figure 2 and Figure 3a,b demonstrate how trailer interior zone temperatures and THI paralleled the outside temperature during summer conditions, except after arrival at the abattoir when misting with fans was used. In Figure 2, all six zones followed Tout, with some zones showing more noticeable changes after cooling was applied, and to a lesser extent, prior to the start of road transport as the trailer heated up. The six zone temperatures are nearly identical to the outside temperature when the trailer was empty driving on the road without pigs. Once the trailer stopped for loading, the zone temperatures began to diverge, and then became more uniform again during the transport segment. During loading and cooling, Zones 1 and 2 (top front and middle) were the warmest, although zone

effect on average temperature rise was not significant (Table 3). The center-zone THI conditions (Figure 3) paralleled the temperature profile over time (Figure 2), except that the THI substantially increased to approximately 83–85 after misting was applied in the trailer for the trip shown in Figure 2 and Figure 3a, suggesting that the pigs experienced a temporarily dangerous thermal comfort condition for most zones, and an emergency condition for Zone 1. However, this change in thermal comfort condition was not illustrated by the temperature history profile. In the other trip that did not receive misting at loading, the THI in six zones increased when pig loading started, but to a much lesser extent than that with misting applied simultaneously. The visualization of the THI history profile supports our statistical analysis results for the ΔTHImax, where an average increase in THI of 4.9 is noted for no misting applied (same case as represented by the trip in Figure 3b), and an increase of 11.1 in THI for misting after loaded (same case as represented by the trip in Figure 3a). In Figure 3a, the discomfort THI condition lasted about 30 min from the onset the misting, and approximately 10–15 mins into the transport segment, but not the entire transport duration. This supports the results that no variation in trailer zones was found for the average THI. For these two representative trips, Zones 1 and 4 (trailer front section) had the most extreme thermal conditions with pigs present for both temperature and the THI. However, this variation between zones were likely masked by the statistical analysis over the entire transport segment.

With regard to misting with fans before unloading used in both trips, substantial non-uniformity in zone temperature and THI was observed. In Figure 2a, interior trailer temperatures in Zones 4–6, representing the trailer bottom deck, rapidly decreased soon after access to misting following fans, resulting in a minimum 4 ◦C difference between the trailer top and bottom decks; the magnitude of difference was even greater for the trip of Figure 2b, with Zones 5 and 6 showing a temperature reduction of 10 ◦C cooler than the other zones. In Figure 3a, the THI in the same zones (4–6) decreased similarly to that of the temperatures, but the variation between zones was smaller than that in Figure 2a. In Figure 3b, the THI in both Zones 5 and 6 paralleled the temperature decrease. In addition, the THI in Zone 2 dropped substantially when receiving fans and misting, although the temperature profile in Zone 2 did not show such trend. This further evident that the cooling effect was not uniform in all trailer zones. One explanation for the different zone thermal responses for these two is the uncontrolled trailer parking position, providing different airflow coverage for different parts of the trailer. The uneven benefits from cooling methods were also seen by the pig surface temperature analysis from the same study [14,15]. In a study conducted on broiler transport [37] with external fans at the side of the trailer, simulated ventilation patterns and velocities using a Computational Fluid Dynamics (CFD) model were not uniform for all locations. They concluded that little of the air flow generated by the fans entered the trailer. By adjusting fan heights, they observed a 41% difference in air flow rate in two adjacent top trailer rows, which supports our observation that only the two to three zones on the bottom deck showed any positive effect of fans at the abattoir.

#### 3.1.3. Spatiotemporal Variability inside the Trailer during Hot Weather

While the temporal thermal history plots in Figure 2 and Figure 3 illustrate variations in zone temperature and THI over time, the animations of spatial variability in the top and bottom sections of the trailer provide a more detailed view of the thermal variation within and between zones. Figure 4a–c provide animation screenshots for three procedures (misting during loading, as transport is started, and waiting by fans after misting was applied at the abattoir) during a representative *Very Hot* monitoring trip (average Tout > 32 ◦C). The first two procedures, misting during loading and at the start of a trip show substantial spatial variability in both bottom and top decks associated with the misting coverage. During misting, the center zones in the bottom deck were as much as 8 ◦C cooler than the middle and rear of the top deck. Analysis using only the average temperature of the thermistors, such as depicted in Figure 2, will likely mask this large temperature variation. About 15 min after misting was stopped, the trailer was about to embark on a 2 h journey (Figure 4b), and there were still residual cooling effects noted especially on the lower deck. Upon arrival at the abattoir, misting was applied, and the

truck was then moved to the fan bank. The resulting upper-deck temperature is uniform, but extreme (36–38 ◦C) as was most of the lower deck except in the very center section where some evaporative cooling was still taking place.

**Figure 4.** Trailer interior temperature distribution for three events during a hot summer monitoring trip (Tout > 32 ◦C). Lighter color areas represent cooler temperatures inside the trailer, and potentially the cooling effects of the misting lasted into the transport segment. The three events were: (**a**) trailer was stationary, and misting was applied inside onto the pigs and bedding; (**b**) trailer was moving; and (**c**) stationary trailer loaded with pigs at abattoir waiting besides external fans with misting previously applied. The green circles indicate position of air temperature sensors at the pig level.

These screenshots, and the complete animations provided in the Supplemental Material to this article, illustrate the lack of uniformity in spatial temperature. In the example in Figure 4, this is likely from unequally distributed misting nozzles across the trailer, and the limited effectiveness of the fans at the abattoir.

Misting and fan operation for cooling during hot weather created large temperature variation within the trailer and was not applied uniformly to all pigs, which can be observed in the sample data visualizations (Figure 2, Figure 3 and Figure 4). Misting at loading showed cooling benefits that lasted into the road transport segment, but increased values of THI for at least some portion of the trip. Similar cooling effects that lasted into the road transport segmen<sup>t</sup> were observed for misting either during loading or after loaded, but no conclusion can be derived regarding which misting method results in greater temperature depression. The efficacy of misting is affected by the location and direction of the misting nozzles such that the coverage area is optimized, and critically, requires high and uniform velocity distribution to ensure evaporation during transport.

#### *3.2. Evaluation of Cold Weather Trips*

#### 3.2.1. Effects of Trailer Management Methods for Cold Weather

Results of the effects of trailer boarding and bedding on average temperature difference for *Cold* and *Very Cold* thermal categories is shown in Table 4. Statistical significance between the means for different levels of the analyzed factors is indicated by superscripts in the same row. Boarding position x trailer zone interaction was not significant, thus not discussed.

**Trailer Section.** The trailer zones have a significant effect on the average temperature rise between trailer interior and the outside, indicating the thermal environments between trailer zones were significantly different during transport (*p* < 0.001). For the 16 trips in the *Cold* category, the front section of the trailer (Zones 1 and 4) was warmer than the middle and rear sections, while the middle section (Zones 2 and 5) and rear section (Zones 3 and 6) were not different from one another. For the *Cold* category, the front section was the warmest area of the trailer, while the rear section was the coolest, which indicates the trailer rear section was the air inlet and the front section as the outlet. This result is not consistently the same as that of the hot weather analysis, which reveals that different trailer sections responded to the environment differently between cold weather and hot weather conditions.

**Boarding—Bedding Combination and Boarding Position.** Although Table 4 shows that the main effect of the boarding and bedding combination was significant for the average temperature rise between trailer interior and Tout, heavier boarding and bedding combination did not show any benefits for increasing trailer zone temperatures. Similarly, for the 16 trips analyzed, none of the boarding positions (as a nesting factor in the boarding—bedding combination) yielded warmer trailer interior temperatures than another. Furthermore, the average temperature rise between trailer interior and the outside was not different based on boarding distribution. More boarding could reduce the air exchange rate by limiting the air inlets and exhaust areas and, hence, the air circulation patterns as well. Since air circulation patterns are affected, air velocity over the animals can also be affected, which in turn theoretically could change convective heat loss. However, our earlier analysis on pig surface temperature as a function of 25% vs. 50% boarding showed no significant difference [15]. Our results agree with [30–32], who reported no difference between trailer boarding levels for pig mortality or morbidity for temperature ranging from 5.1 to 23.3 ◦C; and bedding level did not affect the mortality or morbidity rate at the abattoir.



*p*-value < 0.001, and NS is not significant (*p* > 0.05). 4 Numbers in the parentheses indicate the number of monitoring trips completed and analyzed for this combination. 5 NA indicates insufficient observations across levels of main effect for testing. A–C Different superscripts within a row under the analyzed factor indicate the means differ significantly (*p* < 0.05) from the Tukey–Kramer test for differences of least squares means.

#### *Animals* **2018**, *8*, 203

#### 3.2.2. Temporal Thermal Profile inside Trailer during Cold Weather

Figure 5 show the change of center-zone temperatures from the six trailer zones with elapsed time for a complete transport trip in the *Very Cold* category that was conducted during the evening, from 6:00 p.m. to 1:45 a.m. the next day.

**Figure 5.** Ceiling-centered temperature profile representing 6 zones inside a pig trailer during a *Very Cold* evening trip. Zones 1 to 3 representing trailer top deck, and Zones 4 to 6 representing bottom deck. Events occurred during the trip are numbered on the figure as follows: **1.** Arrival at a commercial pig barn; **2.** Loading; **3.** En route; **4.** Arrival at abattoir; **5.** Unloading.

For this winter monitoring trip, variations of up to 8 ◦C between zone temperatures were observed. The temperature rise added by the pigs' heat production was more obvious than that of the summer trips, reaching an 18 ◦C temperature difference on average between trailer interior and Tout which was about −14 to −12 ◦C for most of the trip. This trend was similar to that found by [10,11], where a magnitude of 15–20 ◦C temperature rise in the trailer was noted for winter trips conducted for weaned piglets transport in Illinois and Iowa. Trailer temperatures during the initial part of the road trip were similar to outside, until about 40 min into the trip when the trailer began to warm up. Zones 3 and 6 (rear of trailer) were consistently colder after the end of the trip, but during transport no clear trends in differences between trailer zones or trailer decks can be seen in this plot, although the front was warmer on average for all cold trips (*p* < 0.05). However, the maximum difference between zone temperatures was 12 to 15 ◦C. Reasons for this large variability are not clear, but it is likely the amount and placement of boarding affected the relative position of inlets and outlets for the trailer.
