Manatee Behavioral Response to Approaching Boats

The Florida manatee (Trichechus manatus latirostris) inhabits coastal and inland waters in the southeastern United States, and its long-term survival is threatened in part by mortality from boat collisions. This study investigated manatee behavior during boat approaches to better understand the factors that lead to boat-manatee collisions. Digital acoustic recording tags (DTAGs) and satellite-linked GPS tags were deployed on 18 Florida manatees in southwest Florida, and the boat traffic around them was mapped. Suites of manatee behavioral, environmental, boat-related, and acoustic variables were measured during opportunistic boat passes. A subset of these variables was used to create and compare numerous models to predict the occurrence, magnitude, and timing of manatee changes in behavior during boat passes. Model averaging techniques were used to discern the relative importance of each variable. Some models accounted for a relatively high proportion of variance in the occurrence (up to R2 of 0.55), magnitude (up to R2 of 0.58), and timing (up to of R2 of 0.75) of changes in manatee behavior during boat passes. Manatees clearly responded to boats: they changed their orientation (heading, roll), depth, and fluking behavior significantly more often when a boat approached closely (within 50 m) than in the absence of boats. Boats traveling at slow speeds elicited changes in behavior that were more likely to occur before the boat's closest point of approach (CPA) compared to boats traveling on a plane. These findings concur with previous evaluations that slower boats allow the manatee more time to respond; coupled with the higher risk of mortality from collisions with faster moving boats, the evidence indicates that slower boats reduce the risk of manatee mortality. Manatee activity before boat approaches was an important factor affecting probability of behavioral change during boat passes. Manatees engaged in a high fluking state (e.g., as occurs during traveling) before boat approaches were more likely to change behavior and to do so sooner (relative to the start of a boat pass) than manatees in a lower fluking state (e.g., as occurs during resting). Manatees suspected to be feeding were more likely to change heading during boat passes. Manatees spent a large proportion of time on seagrass beds, where they were less likely to change depth in response to an approaching boat. It is possible that on a seagrass bed 1) a manatee's options are limited by water depth, 2) they are more motivated to continue their current activity than to respond to a boat, 3) they are in a lower activity state and therefore less likely to change their depth, or 4) manatees cannot detect approaching boats as easily on seagrass beds because background noise is louder and transmission loss is higher. Water depth was another important feature of the habitat that influenced a manatee's response to an approaching boat. Manatees in deeper water were more likely to change fluking behavior, and their changes in depth were of larger magnitude. A manatee is more vulnerable to a boat strike while near the surface; here, manatees closer to the surface were more likely to change their behavior and with less of a time delay (relative to the start of the pass and to CPA) than manatees deeper in the water column. The acoustic properties of boat noise were important, such that louder boats were more likely to elicit a change in a behavior. Sound level, however, did not affect the magnitude or timing of changes in behavior. The rate of change in sound level before CPA was linked to the timing of behavioral changes, with steep rises in sound level associated with a shorter time delay (after the start of the pass) for the first change in behavior. The evaluation of behavioral response of tagged manatees was conducted in southwest Florida, where boat traffic is relatively high. These data were compared to that of an individual manatee in Belize, where manatees (Trichechus manatus manatus) are exposed to far less boat traffic, from a previous study that also deployed DTAGs. Although the small sample size precludes any firm conclusions, differences between the responses were striking. The Belize manatee responded to approaching boats far earlier than manatees in Florida and exhibited a sudden change in behavior, called a `jerk', in two of four passes. As a whole, this study evaluated the type and range of manatee behavioral responses to approaching boats, and identified several factors that appear to influence this response. For instance, the probability of behavior changes was linked to manatee behavior prior to a boat pass and local environmental factors such as seagrass coverage and water depth, and the speed of the boat affected the timing of these changes. Both the probability and timing of behavioral changes were influenced by other acoustic properties of boat noise, such as sound level and its rise rate. In future work, controlled approaches that maintain the high-sensitivity of DTAG-based detection of behavioral changes may be a promising approach to test specific aspects of these results, and further elucidate the environmental and behavioral factors useful for improving current management strategies to minimize boat-related manatee mortality.