198 
Fishery Bulletin 1 14(2) 
Table 2 
Predicted probabilities of a harbor seal {Phoca vitulina) entering the water when vessels of various types approached within 
4 distance classes. Confidence intervals (95%) are given in parentheses below their associated point estimates. Data were 
collected during 27 May-30 June 2001; sample sizes are listed in parentheses in the first 2 rows of this table. 
Distance class 
Cruise ship 
Tour boat 
Power boat 
Skiff 
Inflatable 
Kayak 
No. of vessels 
(5) 
(77) 
(25) 
(12) 
(14) 
(8) 
No. of ice bergs 
(103) 
(694) 
(147) 
(81) 
(110) 
(64) 
0-50 m 
0.99 
0.62 
0.47 
0.66 
0.77 
0.98 
(0.96-1.00) 
(0.50-0.73) 
(0.24-0.71) 
(0.39-0,85) 
(0.54-0.90) 
(0.92-1.00) 
51-100 m 
0.50 
0.16 
0.10 
0.01 
0.12 
0.62 
(0.23-0.76) 
(0.11-0.24) 
(0.04-0.23) 
(0.00-0.11) 
(0.05-0.26) 
(0.32-0.85) 
101-300 m 
0.22 
0.05 
0.01 
0.06 
0.02 
0.18 
(0.08-0.49) 
(0.03-0.08) 
(0.00-0.04) 
(0.02-0.22) 
(0.00-0.10) 
(0.06-0.44) 
>300 m 
0.02 
0.02 
0.00 
0.00 
0.11 
0.00 
(0.00-0.08) 
(0.01-0.04) 
(0.00-0.04) 
(0.00-0.53) 
(0.01-0.70) 
(0.00-0.57) 
entering the water increases with decreasing seal-ves- 
sel distance (Figs. 5 and 7) — a finding that agrees with 
those of other studies (Calambokidis et al. 1 ; Jezierski, 
2009; Jansen et ah, 2010; Hoover-Miller et ah, 2013, 
Young et al., 2014). However, seals in Tracy Arm dur- 
ing our study appeared to be less responsive to ves- 
sels at a given distance than seals in Disenchantment 
Bay, where only cruise ships were monitored (Jansen et 
al., 2010), and in Muir Inlet in GBNP where all vessel 
types were monitored (Calambokidis et al. 1 ), and they 
were far less sensitive than seals in Johns Hopkins In- 
let in GBNP (Young et al., 2014), even when accounting 
for vessel type. 
For our visual estimates, we tended to underesti- 
mate distances between seals and vessels, especially 
for longer distances, and we did not adjust assigned 
distance classes used in other analyses. Our underes- 
timates of distance could possibly explain some of the 
weaker distance response in Tracy Arm than that in 
other glacial fjords where seal-to-vessel distances were 
measured. 
Independent of distance, we found that seals were 
more sensitive to the presence of cruise ships than to 
other vessel types, except kayaks. Other studies in which 
multiple vessel types were monitored also reported this 
pattern (Calambokidis et al. 1 ; Calambokidis et al.,1987; 
Young et al., 2014; Blundell and Pendleton, 2015). 
Calambokidis et al. 1 found that cruise ships in GBNP 
disturbed seals on ice at an average distance of 277 m, 
Jansen et al. (2010) reported seal disturbances by cruise 
ships at distances of up to 500 m, and Young et al. (2014) 
reported disturbances by cruise ships at distances >800 
m. In contrast, we found the probability of disturbance 
by cruise ships approached 1 for ships within 50 m of 
a seal, but the response probability declined rapidly to 
0.02 when the distance between the seal and ship was 
>300 m (Fig. 7) — a more rapid decline in response than 
was observed by Jansen et al. (2010). 
For all of our analyses by vessel type, kayaks also 
had high probabilities of causing seal disturbance, with 
a disturbance response pattern similar to that caused 
by cruise ships (Table 2). Some other studies had also 
noted that kayaks disturb seals at greater distances 
than those of motorized vessel types (e.g., power boats) 
(Jezierski, 2009; Hoover-Miller et al., 2013), but Calam- 
bokidis et al. 1 found that harbor seals were equally 
sensitive to kayaks and tour boats (and less sensitive 
to pleasure boats [i.e., private vessels in our study]) 
and Young et al. (2014) found seals were less sensitive 
to kayaks than any type of power vessel. 
It is possible that harbor seals can habituate to 
the noise of power boats and can determine a boat’s 
approximate location by the sound of its engine. Kay- 
aks, on the other hand, travel more slowly and quietly 
and may go undetected by seals until they are in close 
proximity, or until they make a noise, causing the seals 
to startle and flee into the water. The lower sensitiv- 
ity to kayaks reported by Young et al. (2014) could be 
a function of the behavior of kayakers at that specific 
site, possibly a function of regulations or boater educa- 
tion, which have been shown to reduce kayak-related 
disturbance elsewhere (Hoover-Miller et al., 2013). 
Also, kayakers frequently travel in groups, which could 
account for the higher probabilities of kayaks causing 
disturbance. Further, differences among sites in the 
sizes of kayak groups could be a factor in the variabil- 
ity among sites in patterns of harbor seal response to 
kayaks. We had too few observations of kayak groups 
to estimate the effect of kayak group size on distur- 
bance probability. 
We found responses of seals to tour boats, skiffs, in- 
flatables, and power boats generally similar, although 
there was some variation in results from our study 
methods (i.e., randomized focal observations vs. vessel 
approaches) and analyses (Figs. 5-7). Analyses of the 
randomized observations indicated that seals were less 
likely to enter the water when a tour boat was pres- 
ent than when there were no tour boats present in the 
