FISH AND WILDLIFE TECHNICAL REPORT 25 



formats that we describe can also be used for aerial 

 surveys. The reader is refered to Savard (1979) and 

 Forsell and Gould (1981) for techniques of aerial shoreline 

 surveys, and to Harrison (1982) and Briggs et al. (1985) 

 for pelagic aerial survey techniques. 



Shipboard Surveys 



Most shipboard investigators of marine bird populations 

 have relied on modifications of line- or strip-transect 

 methods and have reported their results as indexes of 

 occurrence or abundance, supplemented with anecdotal 

 information (e.g., Jespersen 1930; Wynne-Edwards 1935; 

 King and Pyle 1957; Kuroda 1960; Bailey 1968; King 

 1970; Shuntov 1972; Gould 1974; Brown et al. 1975a, 

 1975b; Ainley and Jacobs 1981; Powers 1982; Blake et al. 

 1984). These methods, however, have differed greatly, 

 especially in handling such problems as locating and 

 counting a variety of species differing in behavior and con- 

 spicuousness. Among the early pioneers, Wynne-Edwards 

 (1935) demonstrated the importance of structured obser- 

 vations related to unit of effort and repeated in the same 

 area in different seasons and years. He also recognized 

 the need for special handling of ship-following species. 

 Wiens et al. (1978) analyzed differences in detectability 

 as they affect measurement of densities of birds at sea, 

 and suggested techniques that allow for greater control 

 of specific bias-producing factors (e.g., flying birds and 

 determining the distances at which birds are first detected). 

 However, the effort needed to reduce the entire suite of 

 biases inherent in transect surveys of seabirds seriously 

 reduces the cost-effectiveness of the surveys and thus 

 limits their usefulness. This is especially true if informa- 

 tion in addition to abundance and distribution (e.g., 

 behavior or age structure) is being sought. Griffiths (1981) 

 discussed biases produced by the effect of the ship on the 

 behavior of birds at sea. Bailey and Bourne (1972) and 

 Tasker et. al. (1984) discussed problems involved in 

 counting birds at sea and called for standard techniques. 

 Bailey and Bourne (1972) stressed the need to use 10- or 

 15-min transects that could be analyzed separately or could 

 be combined, depending on local density and distribution 

 patterns. Tasker et al. (1984) reviewed the major types 

 of at-sea survey techniques and recommended three major 

 components of the system we describe a 300-m-wide 

 strip census, 10-min duration counts, and an instantaneous 

 count of flying birds. Haney (1985) and Tasker et al. 

 (1985) also discuss methods of counting birds at sea, with 

 an emphasis on standardized methods. 



Our survey method evolved from attempts to accum- 

 ulate the maximum amount of information on the distribu- 



tion and abundance of marine birds within realistic time, 

 money, logistic, and environmental constraints (Gould 

 et al. 1982). Of primary importance was the establish- 

 ment of a standardized system that would be easy to use 

 and teach, and that would provide consistent results in 

 a system useful for both management (monitoring and in- 

 ventory) and research programs. 



We use strip census techniques to develop indexes of 

 density (birds per square kilometer per transect). These 

 indexes, while not being actual counts, are consistent 

 within the data base and provide a baseline from which 

 one may define changes in the size and distribution of 

 seabird populations in time and space (Forsell and Gould 

 1981; Gould et al. 1982; Gould 1983). When conditions 

 do not permit the use of strip transects, we suggest five 

 supplemental techniques: skiff counts, station counts, 

 ship-follower counts, coastline counts, and general ob- 

 servations. These additional methods are a part of the 

 standardized system but are adaptable to a variety of geo- 

 graphic conditions. 



Sampling Design for Strip Transects 



Serious consideration and planning should be given to 

 sampling design before leaving port. Once the cruise has 

 begun, the sampling design should rarely be changed. 

 Special sightings, such as large flocks, which cannot be 

 predicted but are important to record, are handled by sup- 

 plemental techniques (see General Observations). Situa- 

 tions do arise, however, that make it worth modifying the 

 sampling design. A change in the cruise pkn would re- 

 quire a reevaluation of sampling design to accomodate 

 new areas. Encountering unexpected habitat features 

 would make it worth extending a set of 3 transects into 

 a set of 12 or more. In such a case, however, it would 

 still be correct to code the additional transects as general 

 observations, especially if the habitat change is small, very 

 localized, of short duration, or not likely to reoccur. Ex- 

 perience in both observation and data analysis makes these 

 decisions easier. 



Pelagic Areas 



There are three strategies that work well in pelagic 

 areas, depending on the mission and schedule of the 

 cruise. Single transects, or sets of transects, may be con- 

 ducted at preset times throughout the cruise, but the 

 observer should be consistent in the number of transects 

 used during each observation (e.g., one per hour). Three 

 consecutive 10-min transects every hour works well in 

 most situations. This type of sampling is useful to 



