Using qualitative behavior assessment to investigate the effect of tourist presence on the welfare in captive tigers (Panthera tigris) in three tourism facilities in Thailand

Numerous facilities around the world offer tourists interactive experiences with captive tigers. Yet, the animal welfare implications of this practice have not been widely studied. This study aimed to investigate whether qualitative behavioral assessment (QBA) could: (i) provide a valid indicator of tiger's emotional state and (ii) be applied to assess whether unfamiliar human presence with hand-raised captive tigers had an impact on the emotional state of those tigers. To investigate this, QBA was applied to video clips of hand-raised captive tigers from three sites (two offering unfamiliar human interaction, Sites A and C, and one retirement site with no direct interactions, Site B) in Thailand. QBA allows inferences to be made about animal emotion on the basis of descriptions of behavioral expression. Analysis, using a free choice profiling methodology, was provided by observers (N = 38) split between three groups; tiger keepers and vets from the Thai venues (n = 12), UK-based animal behavior MSc and vet students (n = 16), and international tiger keepers (n = 10). Tigers (N = 35) were split between Sites A (n = 7), B (n = 18), and C (n = 10) and filmed at three time points; morning (0800-0930 h); midday, (1130-1230 h); and evening, (1630-1830h) totaling 105 clips. Using generalized procrustes analysis, a consensus profile was calculated for each observer group. Two meaningful dimensions of behavioral expression, explaining 75.0% of the variation, were observed across these groups: Dimension 1 (D1: "active"/"interested"/"agitated" to "relaxed"/"calm"/"chilled-out") and Dimension 2 (D2: "bored"/"stressed"/"frustrated" to "relaxed"/"curious"/"interested"). There was clear agreement between the three observer groups in terms of tiger emotional expression along D1. However, agreement was more variable on D2. The behavioral expression on D1 was not significantly affected by site but was significantly affected by an interaction between age and time of day. Time of day also affected scores on D2, with the Thai observer group also showing an effect of site. During the midday period, when unfamiliar humans were present, all tiger age groups showed more positive behavioral expressions on D1 (lower scores: "relaxed"/"calm"/"chilled-out") and more negative behavioral expressions on D2 (higher scores: "bored"/"stressed"/"frustrated"), which could indicate that the presence of unfamiliar humans was a stressor. However, tigers in the retirement Site C also displayed similar behavioral expressions, which could indicate a deeper welfare issue. With further development, QBA could be used as part of a valid tool for long-term measurement of behavioral expression in captive tigers.