Strong and others 



Chapter 32 



Distribution and Population Estimates in Oregon 



was a compromise between losing birds to avoidance at 

 narrower strips and not detecting birds in wider strips. 

 Both of these effects are present with a 100 m strip and, 

 combined, could result in as much as 10 percent under 

 estimation. This may explain some of the difference 

 between strip and line method results. 



Line transects may err either high or low, depending on 

 how well the detection curve model represents the true 

 detection distribution. Because birds avoided the vessel and 

 we adjusted for this in the data to model detection curves, 

 fits to any model are necessarily approximations. In spite of 

 these factors, the general agreement between the two methods 

 suggests we are in range of an accurate population estimate. 



Averages Versus Weighted Averages 



Because transect length and humber of shore observations 

 varied by day, we were able to compare estimates weighted 

 by effort with direct averages of each day (table 1). Estimates 

 weighted by transect length were quite consistently higher 

 for vessel transects, slightly higher for aerial transects, and 

 lower for shore counts (table I). There was no significant 

 correlation of transect length to densities, however, and no 

 significant differences between regional estimates were found 

 (Mests). Some vessel transects in each region were aborted 

 when fair conditions degraded to poor, resulting in shorter 

 transects under worse conditions, which may have resulted in 

 lower densities (see 'observation conditions'). In the central 

 region, the two most frequently taken transects were 72 km 

 (Newport to Florence) and 27 km (Depoe Bay to Newport) in 

 length; approximately 10 km of the shorter route was off 

 rocky shore (Boiler Bay to Otter Crest) which always had 



very low murrelet densities and would make a smaller 

 contribution in weighted data. This probably explains the 

 consistency of higher estimates for weighted vessel data. 



Year Comparisons 



Densities averaged far higher in 1992 for all methods 

 and regions except shore counts (table 3). This was due to 

 extremely high concentrations of birds very close to shore in 

 1992 (fig. 5). The inshore concentration was most pronounced 

 before mid - July 1992 (fig. 6), but data for the whole year 

 were averaged for analyses here. 



Overall population estimates were significantly different 

 between the two years for aerial and both line and strip 

 vessel estimates (Mests, P < 0.01). Differences between 

 years likely reflects their different distribution offshore and 

 some error in the assumption of equal densities within 500- 

 m (central region) and 1,000-m (north and south regions) 

 increments of distance from shore. For example, due to the 

 scarcity of birds offshore in 1992, the assumption of equal 

 density in 1-km squares and truncation beyond this may 

 have caused over estimation in that year. From the offshore 

 transect data, only 45 percent of the birds occurred over 500 

 m from shore in 1992, compared with 82 percent in 1993 

 (fig. 5, table 1). Using the same logic, the 1993 estimates in 

 other than the central region may have been under estimates, 

 since well over half of the observations occurred beyond 500 

 m and many birds were present beyond 1 km. This 

 consideration would bring the overall state totals closer 

 together in the two years. 



Murrelets and other seabird species were concentrated 

 close to shore in 1992 because of an apparent high availability 



60 



50- 



w 40 

 O 



LLJ 



30 



LU 



O 20 



D 



a. 



10- 



0)(2) (1) 



(1) 



i i i i 1 1 1 1 1 1 1 r 



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 

 NUMBER OF BIRDS 



H 1992 (n = 4918) 1 993 (n = 3368) 



Figure 8 Group size of Marbled Murrelets seen during vessel surveys. Groups of 

 over 8 birds were not recorded in 1 993. The numbers in parenthesis indicate number 

 of groups that were too few to show up on a bar graph. 



USDA Forest Service Gen. Tech. Rep. PSW-152. 1995. 



349 



