Reilly and Fiedler: Interannual variability of dolphin habitats 



437 



30 



20 



3 



10 - 



- 



-10 - 



-20 



Cruise Tracks 



1986 

 1987 

 1988 

 1989 

 1990 



160 150 



— i — 

 140 



— I — 

 130 



— I ' 1 — 



120 110 



Longitude 



Figure 3 



Cruise tracks of RV David Starr Jordan and McArthur during August-November, 

 1986-90. 



Ocean Model 541, or Seabird SEACAT Model 21). 

 Vertical structure was sampled in two ways. Expend- 

 able bathythermographs (XBTs) were deployed four 

 to six times daily (every 55 to 110 km). Twice daily, 

 just before dawn and just after dusk, the ships 

 stopped to deploy conductivity-temperature-depth 

 instruments (CTDs) to 1000 m. Sea water samples 

 were collected with Niskin rosettes for chlorophyll 

 and nutrient analysis. Surface chlorophyll was moni- 

 tored continuously with a Turner Designs Fluorom- 

 eter calibrated from discrete samples drawn at the 

 surface at least six times per day (see Fiedler et al., 

 1990 for details). 



The research vessels searched a total of 140,597 

 km (Fig. 3), and recorded a total of 2,014 sightings of 

 dolphin schools of the seven types considered here, 

 during five annual cruises, 1986 through 1990 (Table 

 1). The environmental data used include continuous 

 temperature, salinity and fluorescence measure- 

 ments along most of the trackline, 17,303 surface 

 chlorophyll measurements, 4,726 XBT deployments 

 and 1,596 CTD stations (Table 2). 



Analytic methods 



We estimated dolphin relative abundance as the 

 number of schools sighted per unit distance searched, 

 for each day, for seven pod-type categories (Table 1). 

 These are the most frequently sighted types of dol- 

 phin schools in the region. They are also the types 



captured by the purse seine fishery. Only days with 

 at least two hours of sighting effort (approximately 

 37 km) during periods of fair or better sighting con- 

 ditions (Beaufort 5 or less) were used in the analy- 

 ses. The distance searched in one day varied between 

 37 and 222 km. We analyzed daily sightings in rela- 

 tion to environmental conditions measured during 

 that day. Our use of Beaufort 5 as the cut-off follows 

 results from Holt ( 1987) who analyzed the effects of 

 sea state on dolphin school sightability from ETP ship 

 surveys. 



Reilly (1990) found that dolphin habitats in the 

 ETP could be defined statistically by a combination 

 of thermocline depth, surface temperature, and sa- 

 linity (where temperature and salinity were com- 

 bined into sigma-t, an index of surface water den- 

 sity, by using a simple linear function described by 

 Pickard and Emery, 1982). We used thermocline 

 depth, surface temperature, salinity, and sigma-t, 

 plus surface chlorophyll (log-transformed), and a 

 measure of thermocline strength (the difference in 

 depth between the 20° and 15° isotherms). Surface 

 temperature, salinity, and chlorophyll were averaged 

 from the day's continuous sampling during sighting 

 effort. Thermocline depth (represented by the depth 

 of the 20°C isotherm: Donguy and Meyers, 1987) and 

 strength were estimated from XBT casts made dur- 

 ing or within a few hours of sighting effort. We in- 

 cluded sigma-t, in addition to both temperature and 

 salinity, in the multivariate analyses as a form of 



