4 
Fishery Bulletin 1 12(1) 
for FI with additional HI evidence were categorized as 
HI-FI with other evidence noted. Mutilation was distin- 
guished from scavenger damage according to guidelines 
in Read and Murray (2000). In contrast to the consis- 
tent definitions of HI described above, the Hi-other 
designation due to human harassment of live pinnipeds 
and cetaceans on the beach was applied inconsistently 
by stranding responders through the years because 
it is not clearly defined by the NMFS Marine Mam- 
mal Health and Stranding Response Program (MMH- 
SRP), although such cases arguably fit the definition 
of harassment under the Marine Mammal Protection 
Act (MMPA; Title 16, U.S code [USC] 1361). Live seals 
are particularly vulnerable to harassment, especially 
in NC where their presence on the beach is relatively 
novel. As a result, each record of a live seal stranding 
was reviewed during the compilation of stranding data 
for this study. Seals not otherwise in need of medical 
attention (i.e., stranded) were considered stranded and 
classified as Hi-other when a high level of harassment 
occurred (e.g., when they were touched by humans or 
relocated because of repeated disturbance), but not 
when seals were frightened back into the water. Live 
stranded cetaceans that the public pushed back into 
the water before notifying stranding responders were 
not generally classified as Hi-other (i.e., harassed) ac- 
cording to the Southeast Regional MMHSRP policies. 
Whenever possible, genetic analysis was conducted 
to verify or assign species identification for animals 
not identified to species in the field owing to advanced 
decomposition and for species hard to identify (e.g., Ko- 
gia, Stenella species). Genetic analysis also was used 
to assign the coastal or offshore morphotype (Rosel et 
ah, 2009) for 185 samples from bottlenose dolphins. All 
other bottlenose dolphins were by default considered 
the coastal morphotype. 
Stranding data and analyses 
Stranding data from January 1997 through December 
2008 were compiled from the comprehensive database 
maintained at NMFS-Beaufort. Level A stranding data 
from NC are also maintained in the MMHSRP data- 
base; however, the local database includes additional 
data such as detailed HI information. Analyses were 
conducted with SAS vers. 9.3, or SAS Enterprise, vers. 
4.2 4 software (SAS Institute, Inc., Cary, NC). 
For analyses, each stranding was considered a sepa- 
rate event, except for mass strandings and mother-calf 
pairs, where individual animals are not independent. 
For mother-calf pairs, maternity was either genetically 
determined or presumed on the basis of a combination 
of sex, age class, date, and proximity of an adult fe- 
male to a calf. Mass strandings were defined as 2 or 
more individuals, excluding mother-calf pairs, of the 
4 Mention of trade names or commercial companies is for iden- 
tification purposes only and does not imply endorsement by 
the National Marine Fisheries Service, NOAA. 
same species at the same location (within 5 km from 
each other) on the same day. Two species were con- 
sidered exceptions to this definition. Harbor porpoises 
(Phocoena phocoena ) meeting these conditions were not 
considered a mass stranding because they do not travel 
in tight social groups (e.g., Raum-Suryan and Harvey, 
1998) and, therefore, one animal beaching is unlikely to 
influence another animal beaching (Geraci and Loun- 
sbury, 2005). Coastal bottlenose dolphins meeting the 
defined conditions also were not generally considered 
a mass stranding because of their high frequency of 
stranding and coastal abundance, except for one event 
where 3 animals stranded together alive. 
For analyses of coastal bottlenose dolphins, strand- 
ing events were stratified by dolphin length and HI cat- 
egories. There is a tendency for seasonal calving (Hohn, 
1980; Thayer et al., 2003) and a concomitant high mor- 
tality of neonates (Fernandez and Hohn, 1998). The re- 
sulting preponderance of neonatal strandings can mask 
stranding patterns of older animals when all strand- 
ings are combined. Because documentation of neonatal 
characteristics (see Thayer et al., 2003) has not been 
consistent for all strandings, length was used as a 
proxy. In NC, Thayer et al. (2003) documented the max- 
imum length of true neonates as 125 cm (mean=108.2 
cm). Thus, all strandings <125 cm were categorized as 
perinates, although we recognize that this category 
could comprise most or all of the neonates, as well as 
some specimens up to 3 months of age (Fernandez and 
Hohn, 1998). Nonperinatal (>125 cm) coastal bottlenose 
dolphin strandings were stratified into HI categories: 
HI-FI, Hl-no, HI-CBD, and Hi-other. For mother-calf 
pairs, the HI category of the mother was used. The 
sample size of the category Hi-other was too small for 
analyses (n= 26 events). Only 4 perinatal bottlenose 
dolphin strandings were positive for HI (HI-FI=2, Hl- 
other=2); therefore, perinates were not stratified by 
HI category. Animals with an estimated or minimum 
length (such as that due to severed flukes) of <125 cm 
were excluded (91 individuals from 89 events). 
Temporal patterns 
Yearly and monthly patterns of stranding events were 
evaluated for taxonomic groups with sufficient sample 
sizes. When necessary to achieve adequate sample 
sizes, species with similar habitats were combined. An- 
nual trends were assessed by using a simple linear re- 
gression (SAS PROC REG) for the following taxonomic 
groups: balaenopterids, pygmy sperm whales ( Kogia 
breviceps)-, dwarf sperm whales (Kogia sima ); harbor 
porpoises; harbor seals (Phoca vitulina ); non -Phoca 
pinnipeds; non-Kogia pelagic odontocetes (delphinids 
except coastal bottlenose dolphins, ziphiids, and sperm 
whales [Physeter macrocephalus]), and the coastal 
bottlenose dolphin categories mentioned previously, in- 
cluding coastal and inshore strandings. For the simple 
regression only, data (plus 0.5 due to zeros) were natu- 
ral log transformed. Monthly stranding patterns were 
