Dahlheim et ai.: Temporal changes in abundance of Phocoena phocoena inhabiting the inland waters of Southeast Alaska 
245 
porpoise densities derived from the results of the sur- 
veys conducted in 1991-1993. Regions with higher por- 
poise density in the early 1990s than other regions with- 
in the same period were given greater trackline effort in 
the later surveys. As in the early years, both zig-zag and 
straight-line paths were used for tracklines in an effort 
to include as many different habitats as possible. 
Between 2010 and 2012, the following 4 charter ves- 
sels were used to conduct our surveys: July 2010, the 
FV Steller (21.3-m commercial fishing vessel with a 
combined bridge and observer height of 4.8 m); Sep- 
tember 2010, the FV Northwest Explorer (43.8-m re- 
search and fishing vessel with a combined bridge and 
observer height of 5.6 m); June and September 2011, 
the Alaska Department of Fish and Game’s RV Medeia 
(33.5-m research vessel with a combined bridge height 
and observer height of 7.4 m); and July 2012, the RV 
Aquila (50.0-m commercial fishing vessel with a com- 
bined bridge and observer height of 7.2 m). Different 
vessels were used because the John N. Cobb was de- 
commissioned in 2008. 
During all line-transect surveys, a team of 3 re- 
searchers (1 recorder and 2 observers) were on effort 
at any given time. However, the total number of biolo- 
gists participating in a survey varied and determined 
the amount of time that observers spent off effort. The 
surveys conducted in 2006 and 2007 comprised 4 biolo- 
gists, allowing 1 researcher to be off effort while the 
others manned stations. A full observer rotation took 
1.5 h, during which each observer spending 30 min at 
each station. In this case, the observer only had a rest 
period of 30 min between watches. To minimize fatigue, 
we also went off effort for meals, a schedule that pro- 
vided each observer with an additional rest period. In 
2010, 5 biologists participated in the survey. As with 
the 2006 and 2007 surveys, a full observer rotation 
took 1.5 h with a 1-h rest period between watches. In 
2011 and 2012, a full complement of 6 biologists par- 
ticipated in the surveys, allowing observers to spend 30 
min at each station and have a 90-min rest period. As 
noted earlier for the surveys conducted during 1991- 
1993, schedules of observer rotations were selected 
randomly each day. 
To gather positional and navigational information, 
the computer used for data collection was either in- 
terfaced directly to the ship’s GPS system (2006 and 
2007) or connected to a portable GPS unit (2010-2012). 
The computer program WinCruz (R. Holland, NOAA 
Southwest Fisheries Science Center) was used to re- 
cord all sighting and environmental data (e.g., cloud 
cover, wind strength and direction, and sea conditions). 
All other field methods for data collection (e.g., scan- 
ning techniques and field equipment) were the same 
as those used for surveys conducted in the early 1990s. 
Analysis for corrected distances 
For the surveys conducted during 2006-2012, distance 
from the vessel to the animal was originally recorded 
in WinCruz under the inherent assumption that the 
reticle given by the observer was taken from the hori- 
zon. However, within the inland waterways of South- 
east Alaska, many of the sightings occurred against the 
land versus the true horizon. In these instances, the 
observer used the shoreline as their “horizon,” and the 
distance to the sighting needed to be recalculated. 
Sightings were recorded in WinCruz as the position 
of the observer (vessel) at the time of a sighting. All 
sightings were plotted in ArcMap, vers. 10.1 (Esri, Red- 
lands, CA), and a line representing the distance to the 
real horizon, at the correct sighting angle, was drawn. 
The horizon line was truncated wherever it crossed 
land. The length of this new line, representing the dis- 
tance from the observer to the shore at the angle of the 
sighting, was then converted to “reticles to land” with 
the DistRet function in Geofunc (National Marine Mam- 
mal Laboratory, http://www.afsc.noaa.gov/nmml/soft- 
ware/excelgeo.php, accessed January 2013), a Microsoft 
Excel add-in that performs trigonometric calculations 
for plane and spherical geometry pertinent to marine 
mammal survey sighting methods (the appropriate for- 
mulas are described in Lerczak and Hobbs [1998]), to 
account for the corresponding observer height and radi- 
ans per reticle for 7x50 binoculars. The resulting value 
of reticles to land was added to the original observer 
reticles to calculate the actual reticles of the animal 
from the vessel. Final distance from the observer to the 
animal was calculated from this new reticle value with 
the DistRet function. During the 1991-1993 surveys, 
observers obtained a distance to shore at the angle of 
the sighting from the ship’s radar; however, to be con- 
sistent with distances calculated for sightings in the 
later surveys, the above protocol also was applied for 
all harbor porpoise sightings for those years. 
Line-transect estimation of density and abundance 
The density and abundance of harbor porpoise was es- 
timated for 3 survey periods, 1991-1993, 2006-2007, 
and 2010-2012, and effort was allocated in 6 strata 
that were defined on the basis of distribution of harbor 
porpoise throughout Southeast Alaska and the configu- 
ration of different water features (such as large straits 
and bays). Data were pooled across sequential years 
as a strategy to minimize variability due to the spa- 
tial distribution of harbor porpoise and the allocation 
of survey effort, therefore, providing more robust esti- 
mates of abundance for the 3 different periods. 
Previous studies reported no evidence of seasonal 
changes in distribution for harbor porpoise occupying 
the inland waters of Southeast Alaska (Dahlheim et 
ah, 2009). Therefore, given the broader spatial coverage 
and greater number of surveys conducted in summer, 
we chose to derive abundance estimates from the sum- 
mer season only. 
Estimation of detection probability 
Often, when the collection of sighting data is consis- 
tent across years or strata in visual line-transect sur- 
