Strong and others 



Chapter 32 



Distribution and Population Estimates in Oregon 



of prey there, such as smelt species (Strong and others 

 1993). With the exception of Surf Scoters (Melanitta 

 perspicallata), other seabirds were more scattered and farther 

 offshore as well in 1993. In 1993, Common Murres (Uria 

 aalge) and Pigeon Guillemots (Cepphus columbd) largely 

 abandoned nest sites in June, and very few (murres) or no 

 (guillemots) fledglings were seen at the end of the nesting 

 period. It is probable that very low prey availability caused 

 the reproductive failure for these alcid species, and likely 

 that Marbled Murrelets were also impacted. Both 1992 and 

 1993 were cited as 'El Nino years', but waters off Oregon 

 were warmer and upwelling weaker in 1993 (NO A A Coastal 

 Ocean Programs 1992-1993). 



Regional Characteristics 



We did not attempt to extrapolate from the central 

 region's offshore distribution where we lacked data on 

 offshore distribution for the north and south regions. In the 

 northern region, inshore densities were much lower. 

 Assumption of a proportionate dispersal offshore as for the 

 central region would probably be invalid as it would result 

 in extremely scattered birds. Other data show Marbled 

 Murrelets to have a very clumped distribution (Nelson and 

 Hardin 1993a, Strong and others 1993). Low overall densities 

 on the north coast was characteristic of all survey methods 

 and years, with the exception of one vessel transect on 2 1 

 July 1993. On that day, Murrelets were concentrated in the 

 vicinity of Netarts Bay, and the average density (24.2 birds/ 

 km 2 ) was far higher than any other records for the region. 

 This 'outlier' was interpreted as a movement of non-nesting 

 birds from the central region. It is possible these birds failed 

 or did not attempt to nest due to low prey availability in that 

 year (see above). 



The southern region has very different physical 

 characteristics than the rest of the state, with many offshore 

 rocks, rocky shorelines, and variable bathymetry. Coastline 

 densities here were most variable (C.V., table 3), though our 

 survey effort was small and, in 1993, took place under 

 largely fair to poor conditions (Beaufort state 3 to 4). Because 

 of these considerations, we have lower confidence in our 

 density estimates for this region. It may be appropriate to 

 further divide the region north and south of Cape Blanco, 

 based on physical characteristics and recorded murrelet 

 densities. Near the California border (south of Goat Island), 

 murrelets from nesting areas in California's protected redwood 

 parks may forage in Oregon waters, thereby confusing 

 measures of the state population. 



Interpreting results was problematic in the center-south 

 subregion. The single survey of the region in 1992 generated 

 the highest daily average densities recorded, but four surveys 

 of the area in 1993 each recorded densities well below the 

 rest of the central region (table 3). Aerial surveys in 1993, 

 however, again produced relatively high densities, although 

 this may have resulted from vagaries in aerial surveying. To 

 account for the different offshore distribution between years 

 in this area, and bring the estimates into closer agreement, 



we only extrapolated to a 500-m wide block of area in 

 computing the 1992 density estimate. 



Other Adjustments to the Estimate 



While not including a factor for birds beyond 1 km in 

 northern and southern Oregon may be seen to cause 

 underestimation, other considerations of distribution and 

 sampling may compensate for this. The surf zone off Oregon's 

 beaches typically ranges from 100 to 400 m out to sea 

 depending on swell size. While we did observe Marblec 

 Murrelets within the surf zone, particularly in 1992, the> 

 occurred at lower densities than beyond the breakers. If we 

 were to assume, as an approximation, that the inshore 100 1 

 was without murrelets, the effect would be to reduce the 

 estimate by 1 percent. 



A proportion of the birds that flew in response to the 

 vessel went in the direction of vessel travel where they could 

 have been double-counted if they landed in the transect' s 

 path. In 1993, we quantified this and found that 21.9 percent 

 of the birds which flew went in the vessel's direction of 

 travel. This was far less than 50 percent since murrelet 

 usually flew against the wind, and we usually ran transect 

 with the wind (birds rarely departed east or west). Of 10." 

 percent of birds which flew in avoidance (fig. 7), 22 percen 

 flew in direction of travel. If each were double-counted once, 

 the adjustment would be 0.107 x 0.22 = 2.3 percent of the 

 estimate. This, for example, would amount to 350 birds double 

 counted in the 1993 strip transect state estimate, a relatively 

 minor difference. It is possible that many birds may relocate 

 independently of vessel movement during the course of ou 

 transects, which last 2-9 hours. But because there is equa 

 probability of birds either relocating into our path or moving 

 out of it, no error was anticipated from this behavior. 



Offshore sampling in central Oregon accounted for 

 relatively small proportion of the total survey effort, but the 

 contribution to the total estimate from those data was large, 

 particularly in 1993 (table 2). Selection of offshore sampling 

 locations took place prior to each day's sampling, and wer 

 where murrelets were found to be consistently present during 

 coastline transects. This has the potential for bias to areas of 

 higher density within the whole region, although the effect is 

 probably slight. Specific areas of abundance were virtually 

 impossible to predict, since the clumped distribution of birds 

 shifted daily on a scale of 10's of kilometers (Strong and 

 others 1993). 



In 1992 there was a significant correlation between 

 observation conditions and number of birds sighted (r 

 0.112, P < 0.001), but not in 1993. We did not detect a 

 difference in the average distance at which birds were seen 

 between excellent and good conditions; it only decreased ; 

 fair or poor conditions (ANOVA, P < 0.001). This suggests 

 that our observations had consistency of detections with 

 respect to weather at Beaufort states less than 3. 



In addition to the above considerations, other aspects of 

 Marbled Murrelet biology and behavior may affect the results 

 of marine transects for population estimation. Birds tending 



350 



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



