Field work

Ballenas Channel and Bahía de losÁngeles are located in the northwestern Gulf of California (Fig. 1). Due to constant upwelling and tidal mixing, this channel has the lowest sea surface temperature year-round (^{López et al. 2006}) and the highest biological productivity in the gulf (^{Álvarez-Borrego and Lara-Lara 1991}, ^{Gómez-Gutiérrez et al. 2012}, ^{Escalante et al. 2013}). Moreover, methane concentrations and sea-air methane fluxes were found to be highest in Ballenas Channel in a study of the northern gulf by ^{Macías-Zamora et al. 2013}; these authors state that this may be due to biological methane production in the water column, most likely in zooplankton guts. According to oceanographic and climatic conditions, and primary productivity, seasonal changes indicate a warm (July-October) and a cold (December-May) season, as well as two short transition periods, one in June and one in November (^{Álvarez-Borrego and Lara-Lara 1991}, ^{Hidalgo- González and Álvarez-Borrego 2004}).

Figure 1 Map of the study area (Ballenas Channel and Bahía de los Ángeles) in the northwestern Gulf of California, Mexico, showing the survey plan (line transects) used to search for cetaceans, the zooplankton sampling stations (symbols), and the bathymetry (isobath interval is 100 m). 

We visited the study area during four seasonal periods in 2003 (cold, 12 March to 10 April; cold-warm transition, 27 May to 26 June; warm, 11 July to 4 August; and late warm, 16 October to 2 December) and four periods in 2004 (cold, 6-31 March; cold-warm transition, 3 June to 1 July; warm, 2-30 July; and late warm, 7-31 October). We used small boats (6 m long) to navigate in search for fin whales and to carry out surface net tows collecting euphausiids in a fixed sampling grid (Fig. 1).

Surveys were planned to assess the distribution of fin whales in the study area and distance sampling methods were applied (^{Buckland et al. 2001}). We navigated along line transects (Fig. 1) at 14-16 km h^-1 (^{Barlow and Forney 2007}), with a team of five observers searching the areas 90° to the left and right, and in front of the boat with 7 x 50 binoculars. In order to analyze the temporal variation of fin whale abun dance, data from additional, non-systematic surveys were included. These were carried out at speeds of 16-30 km h^-1and only when one or two observers were available, when wind and sea conditions did not permit finishing a transect, or during navigations between our base station (Bahía de los Ángeles) and the start or endpoint of a transect.

Combining both survey years, we navigated for 70 days, with a mean search effort (± SD) of 3.6 ± 1.4 h d^-1 in 2003 and 3.2 ± 1.6 h d^-1 in 2004. The search effort was defined as the time spent and distance covered while actively looking for whales: 104 h (1463 km) and 91 h (1450 km) along transects in 2003 and 2004, respectively; 150 h (3193 km) and 138 h (3184 km) in non-systematic surveys in 2003 and 2004, respectively.

Euphausiids were collected by horizontal surface net tows with a conical net (diameter 50 cm, mesh size 200 |am) at previously defined stations. Two stations on each line transect were selected (Fig. 1): one in shallow waters (<200 m bottom depth) and one in deep waters (201-1400 m). Towing distance was measured with a GPS and tows were carried out at a speed of 5 km h^-1 during 5 ± 1 min (mean ± SD). This towing time was determined as being representative since longer tows often clogged the net. Net size and time of tow were used to calculate the volume of water filtered during each tow. Mean filtered water (± SD) was 79 ± 19 m³. Samples were preserved in 4% formaldehyde buffered with sodium borate. Fifty-four zooplankton samples were collected in 2003 and 77 in 2004.

At each zooplankton station sea surface temperature (SST) was recorded with a bucket thermometer. We also used satellite SST data downloaded from NASA's OceanColor Web (http://oceancolor.gsfc.nasa.gov).

Sample and data analyses

To analyze fin whale distribution we only used data from line transects since this sampling method does not bias sightings to any particular area (Fig. 1) (^{Buckland et al. 2001}). We calculated a relative abundance index (RAI, number of whales per kilometer of search effort) and compared the mean RAI for shallow (0-200 m) and deep (201-1400 m) waters. The 200-m-isobath is the edge of the continental shelf, where the zooplankton community is domi nated by few species, one of them N. simplex, and we would therefore expect to find more whales in shallow waters. There were no significant differences in the weekly RAI from both transects and non-systematic surveys (t₆₉ = -0.66, P = 0.87); therefore, we pooled these data to assess the temporal variation in the abundance of fin whales to increase sample size. We used nonparametric tests for the weekly, seasonal, and yearly statistical comparisons (^{Neave and Worthington 1988}).

Zooplankton samples were subsampled (1/16 to 1/64) with a Folsom plankton splitter depending on the amount of plankton. Euphausiid species were identified and counted according to life phases: calyptopis, furcilia, juveniles, and adults (^{Boden 1951}). Euphausiid abundance was standardized to the number of individuals per 1000 cubic meters. Again, nonparametric tests were used for the weekly, seasonal, and yearly abundance comparisons.

Since a correlation between euphausiids and whales was attempted, the population structure of euphausiids was analyzed because daytime surface net tows tend to collect more larval (calyptopis and furcilia) than postlarval stages (juveniles and subadults). Euphausiid abundance was converted to carbon biomass using an equation for Euphausiapacifica (^{Ross 1982}):

where y is carbon biomass (µg) and x is body length (mm).

This equation was applied to all life phases of N. simplex assuming the mean body length of calyptopis (2 mm), furcilia (4.5 mm), juveniles (6.5 mm), and adults (12 mm). We used the equation for E. pacifica because there is no equation available for N. simplex. To analyze the possible correlation between the weekly relative abundance of whales and the biomass of euphausiids, we applied the nonparametric Spearman correlation test.

We also compared mean SST between 2003 and 2004, both for data recorded in situ during surveys and weekly mean SST from satellite data.

Sea surface temperature (SST)

Mean SST ± SD in 2003 (25.2 ± 3.3 °C), measured in situ, was significantly higher than in 2004 (23 ± 4.1 °C; t₅₄₈ = 6.44, P< 0.001). Seasonal temperature variations were high since the difference between March and October was 2 °C, compared with an interannual difference of 1.5 °C (Fig. 2a).

Figure 2 Weekly mean sea surface temperature (± standard deviation) during 2003 and 2004 in Ballenas Channel and Bahía de los Ángeles. (a) Data collected in situ and (b) satellite data (http://oceancolor.gsfc.nasa.gov). Horizontal bars along the x axis indicate the weeks when the search effort was carried out in the study area. 

Satellite data underestimated sea-truth SST but showed a similar trend to that of in situ data, with higher SST in 2003 (23.3 ± 4.8 °C) than in 2004 (22.5 ± 4.9 °C), although there were no significant differences between years (t₈₈ = 0.755, P = 0.45). This may be due to similar SST from June to September (Fig. 2b). However, satellite data revealed a 1.0 °C difference between March 2003 and 2004, and 1.6 °C between October 2003 and 2004.

Fin whale and euphausiid distribution

The search effort for whales showed no significant differences between years for line transects (t₁₃₈ = 0.77, P = 0.44) and non-systematic surveys (t₁₃₈ = 0.86, P = 0.39) or for the combination of both (t₁₃₈ = 1.28, P = 0.203). Therefore, data from both years were pooled for further analyses.

Fin whales were more often found south of 29.2°N and near the Baja California coast than in other areas of the Ballenas Channel (Fig. 3). They were mainly observed in waters south of Coronado Island and from the eastern entrance of Bahía de los Ángeles to Bahía Las Ánimas, near the coast of Baja California. Fin whale relative abundance (RAI, whales km^-1) did not differ significantly between shallow (0-200 m) and deep waters (201-1400 m) within each survey year (Table 1). However, there were significant differences in mean RAI between years in shallow waters (Mann-Whitney U = 684, P< 0.001; Table 1) but not in deep waters.

Figure 3 Fin whale (Balaenoptera physalus) sightings in (a) 2003 and (b) 2004.Isobathinterval = 200 m. 

Table 1 Fin whale (Balaenoptera physalus) mean relative abundance index (RAI, number of whales per kilometer of search effort) in shallow (0-200 m) and deep (201-1400 m) waters of the Ballenas Channel and Bahía de los Ángeles during 2003 and 2004; n = effort days. There were significant differences between years in shallow waters. 

Most of the euphausiids found in the samples were N. simplex, mainly in the third larval phase (calyptopis; Fig. 4). This was expected because juvenile and adult euphausiids undergo diel vertical migrations (^{Lavaniegos 1996}), so during the day only larvae stay at the surface.

Figure 4. Relative biomass as organic carbon (mg C m^-3) of Nyctiphanes simplex life phases. Stations were ordered from north to south except those to the right of the dashed line, which correspond to Bahía de los Ángeles. Station numbers were omitted where no sample was collected. 

However, we assumed that the larvae are representative of juvenile and adult presence in the same area, based on previous analyses of samples collected in the Gulf of California in 1983 and 1984 (^{Lavaniegos et al. 1989}). We carried out a Pearson regression of these data between abundances (log ind/1000 m³⁾ of calyptopis larvae and postlarvae, which was significant for both day and night (R² = 0.30, P = 0.001, n = 33; Fig. 5). The resulting equation is as follows:

Figure 5 Regression of larval (calyptopes + furcilias) and postlarval (juveniles + adults) abundances (log ind/1000 m3) for diurnal and nocturnal samples collected in the Gulf of California in 1983 and 1984 (^{Lavaniegos et al. 1989}) 

Therefore, larvae at the surface during the day are representative of the presence of postlarvae (juveniles and adults). A Spearman correlation with data that were not transformed was also significant (r_S = 0.52, P = 0.002).

We identified other euphausiid species (Nematoscelisdifficilis, Euphausiadistinguenda, Euphausiaeximia),but they only made up 0.1% of the total euphausiids. Therefore, from now on we refer to N. simplex simply as euphausiids.

Annual mean abundance of euphausiids was lower in Bahía de los Ángeles (shallow waters) than in Ballenas Channel (deep waters) in both years (Fig. 6). When comparing abundances between years there were no significant differences in shallow (Mann-Whitney U = 122, P = 0.620) or deep waters (U = 665, P = 0.261), probably due to the high variability of data (see 95% confidence intervals in Table 2).

Figure 6 Distribution of annual mean abundance of euphausiids(Nyctiphanes simplex) during (a) 2003 and (b) 2004. 

Table 2 Euphausiid(Nyctiphanes simplex) annual geometric mean abundance (ind/1000 m³⁾ in shallow and deep waters of the Ballenas Channel and Bahía de los Ángeles during 2003 and 2004; n = sampling stations (see Fig. 1). Data for the 2004 cold season were excluded. 

Temporal variation in abundance of fin whales and euphausiids

Fin whales were significantly more abundant in 2004 than in 2003 (Table 3; U = 782, P< 0.001). There were also significant differences in fin whale relative abundance from season to season within each year (2003: H₃₅₇ = 10.99, P = 0.01; 2004: H₃₇₀ = 24.22, P< 0.001). Changes in seasonal abundance were similar for both years: the RAI was lowest in October and highest in July (Table 3).

Table 3 Fin whale (Balaenoptera physalus) mean relative abundance index (RAI, number of whales per hour of search effort) by year and by season in the Ballenas Channel and Bahía de los Ángeles during 2003 and 2004. There were significant differences between years and from one season to another within each year (P< 0.05). 

There were no significant differences in euphausiid abundance between 2003 and 2004 (Table 4; U = 1919, P = 0.45), with geometric means of 174 and 255 ind/1000 m³, respectively, for the whole region (Table 4). However, there were seasonal differences in 2003 (H ₂₅₄ = 16, P< 0.001), due to a strong decrease observed from the cold-warm transition period to the warm season in Ballenas Channel, while euphausiids increased in Bahía de los Ángeles. A more dramatic and significant decrease was observed in the whole region between the warm and late warm seasons in 2003 (U =95, P = 0.01). Seasonal differences also occurred in 2004 (H _3,77 = 12, P = 0.006). Pairwise comparisons indicated that euphausiid abundance increased two orders of magnitude (table 4) from the cold to the cold-warm transition seasons (U = 103, P = 0.001), but decreased again in the warm season (U = 83, P = 0.02).

Table 4  Nyctiphanes simplex geometric mean abundance (ind/1000 m³⁾ by year and by season in the Ballenas Channel and Bahía de los Ángeles during 2003 and 2004. 

In order to analyze a possible correlation between whale and euphausiid abundance on a shorter time scale, data were grouped by weekly abundance. Fin whale relative abundance seemed to increase from May to July and then drastically decreased in October (Fig. 7). Mean weekly abundance of euphausiids in 2003 showed high values in May and low variability from one week to another (Fig. 8). In 2004, euphausiid abundance started to increase from March to April, reaching maximum values in June, while the weeks in October had the lowest values. During the last week of October a single tow collected a great amount of euphausiids, so this apparent increase in abundance should be taken with caution.

Figure 7 Fin whale (Balaenoptera physalus) weekly relative abundance (whales h^-1) during 2003 and 2004. Vertical bars are weekly means and 95% confidence intervals; horizontal bars along the x axis indicate the weeks when the search effort was carried out. 

Figure 8 Euphausiid (Nyctiphanes simplex) weekly mean abundance (log ind/1000 m³⁾ during 2003 and 2004. Vertical bars are weekly means and 95% confidence intervals; horizontal bars along the x axis indicate the weeks when the search effort was carried out. Data were log-transformed to calculatemeans. 

We found no statistical correlation between the weekly abundance of fin whales and euphausiids in 2003 (Spearman's r_S = 0.079, P = 0.754) or in 2004 (r_S = 0.126, P = 0.557). It is interesting to note that there seemed to be a lag of about four weeks between euphausiid and whale maximum abundance during both sampling years (end of May to end of July 2003, and June to end of July 2004; Figs. 7, ⁸ ).

Fin whale and euphausiid distribution

Fin whale foraging occurs in regions with high biological productivity, often related to fronts, upwelling areas, and complex bottom topography (^{Laran et al. 2010}, ^{Santora et al. 2010}). The Ballenas Channel and Bahía de los Ángeles region is such a region, where ocean dynamics (^{López et al. 2006}) constantly provide nutrients to primary producers and, therefore, high biological productivity is observed year-round (^{Millán-Núñez and Yentsch 2000}). A similar pattern has been observed in Monterey Bay, California, where blue whales forage on euphausiids (Thysanoessa spinifera and E. pacifica) in a very rich coastal upwelling ecosystem (^{Croll et al. 2005}).

Even though fin whales were observed throughout the study area, they were recorded more frequently in shallow waters close to the Baja California coast and south of Coronado Island. There is a steep slope to the east of euphausiids (^{Lavaniegos 1996}, ^{Gómez-Gutiérrez et al. 2010}, ^{Tremblay et al. 2010}), may contribute to the fact that euphausids concentrate there and are available as prey for whales (^{Ladrón de Guevara et al. 2008}). ^{Brinton and Townsend (1980)} found that euphausiid abundance here and at other sites on the continental shelf of the Gulf of California was as high as 5000 ind/1000 m³ from February to June. These authors showed that even during the warmest month (August), the abundance of N. simplex remains high in the Midriff Islands region (where our study area is located), whereas it decreases in the rest of the Gulf of California. This pattern was also observed by ^{Gómez-Gutiérrez et al.(2012)}.

^{Tremblay et al. 2010}), puede contribuir al hecho de que los eufáusidos se agreguen ahí y estén disponibles como presas para los rorcuales (^{Ladrón de Guevara et al. 2008}). ^{Brinton y Townsend (1980)} documentaron que aquí y en otros sitios de la plataforma continental del golfo de California la abundancia de eufáusidos puede ser de hasta 5000 ind/1000 m³ de febrero a junio. Estos autores mostraron que aun durante el mes más cálido (agosto), la abundancia de N. simplex permanece alta en la región de las grandes islas (donde se ubica nuestra área de estudio), mientras que disminuye en el resto del golfo de California. Este patrón también coincide con lo observado por ^{Gómez-Gutiérrez et al. (2012)}.

Daytime surface tows carried out in this study certainly underestimated the abundance of juvenile and adult euphausiids and even late furcilia stages due to the circadian vertical migration of euphausiids (^{Lavaniegos 1996}, ^{Gómez- Gutiérrez et al. 2010}). Juvenile and adult N. simplex are capable of descending to a depth of 300 m during the day but usually perform short vertical migrations (150 m), and larval stages increase vertical distance during their daily vertical migrations while growing (^{Lavaniegos 1996}). Thus, calyptopis and early furcilia stages likely stay near the surface during the day and night. Even though data in this study are biased to calyptopes and small furcilias, their presence at the surface all day long is a relatively good indicator of the presence and abundance of the juvenile and adult euphausiid population in the study area, as we confirmed in the correlation analysis of data obtained in 1983 and 1984 in the Gulf of California by ^{Lavaniegos et al. (1989)}.

Undersampling in this study was inevitable due to the vertical migrations of euphausiids and also to their evasion of the net, since it is visible during the daylight hours. However, this is not necessarily the main reason for the lack of statistical correlation of weekly abundance between euphausiids and whales. Another problem with net tows is that they do not guarantee finding a correlation between whales and their prey because euphausiids form dense, heterogeneously distributed aggregations. For example, ^{Brodie et al. (1978)} conducted zooplankton tows at a depth of 300 m off Nova Scotia as well as an analysis of whale stomach contents to find a positive statistical correlation between available and ingested prey. According to these authors, estimated available krill in the study area was not sufficient to sustain the whale population. Whales foraged on dense krill aggregations at depths that were possibly not sampled during net tows. Currently, a combination of expensive sampling methods is used to associate the presence of euphausiids and whales, including acoustic backscatter, zooplankton net tows, underwater videos, and telemetry (^{Croll et al. 1998}, ^{Fiedler et al. 1998}, ^{Friedlaender et al. 2009}, ^{Laidre et al. 2010}). Moreover, mesoscale studies (hundreds of kilometers) tend to show clear spatial correlations between whales and their prey, as in Antarctic fin whales and Euphausia superba (^{Santora et al. 2010}).

Temporal variation in fin whale and euphausiid abundances

In this study, the highest abundance of fin whales occurred during the warm season, as was also observed in the Mediterranean Sea from 2001 to 2004 (^{Laran et al. 2010}). In the southwestern Gulf of California, fin whales were more abundant during the cold than during the warm season in 2004 (^{Alonso-Lozano 2014}). We found that both euphausiids and fin whales had the lowest abundances during October 2003 and 2004 in our study area. This may be related to the increase in SST during the warm season and to temperatures remaining high throughout the season. Nutrients (PO₄, NO₃, and Si(OH)₄) are scarce due to intense biological activity (phytoplankton uptake) and stratification of the water column (^{Torres-Delgado et al. 2013}). This may have had a negative effect on the reproduction and recruitment of N. simplex, particularly during 2003 (mean SST 24 °C). This species thrives in upwelling areas and although larvae may tolerate tempera tures between 13 and 30 °C, the temperature range with highest larval abundances is between 17 and 20 °C (^{Brinton and Townsend 1980}). A study showed that N. simplex avoided the upper layer during the warm season, when temperature in the mixed layer was 22-29 °C (^{Gómez-Gutiérrez et al. 2010}).

This coincides with the seasonal changes we recorded in 2003 and 2004 in the Ballenas Channel. Oxidative stress was also proposed as an ecophysiological factor in the decrease of N. simplex abundance during October due to shoaling of hypoxic conditions (^{Tremblay et al. 2010}).

In our study, when the weekly variation in the abundance of euphausiids and fin whales was analyzed, there seemed to be a relationship at this temporal scale (Figs. 7, ⁸ ). Both species were observed year-round in the Ballenas Channel and Bahía de los Ángeles, which confirms that fin whales remain at low latitudes when food is available (^{Aguilar 2009}). The lack of a significant statistical correlation between the studied species was probably due to a lag in their maximum abundance. This seems plausible, since growth rates are slower and life cycles longer the higher up organisms are in the trophic web, so there is usually a lag in maximum productivity of primary and secondary consumers. In the upwelling region south of Australia, blue whales were observed foraging on the euphausiid Nyctiphanes australis approximately two months after the upwelling pulse and the highest chlorophyll a concentration (^{Gill 2002}).

In this study, fin whales were more abundant in 2004 than in 2003 in the Ballenas Channel and Bahía de los Ángeles region, but not euphausiids. Warmer SST was observed during September and October (Fig. 2) suggesting a possible decrease in euphausiid productivity and, therefore, prey availability for whales. This apparent temperature anomaly in the Gulf of California seemed to be related to a weak El Niño event (^{McPhaden 2004}).

In a previous study in the central Gulf of California, ^{Lavaniegos et al. 1989} examined the influence of the 1983-1984 El Niño event on euphausiid populations. A difference in SST of 1-2 °C was observed between spring 1983 (El Niño) and 1984. These authors suggested that water warming during 1983 had a negative effect on calyptopis recruitment of N. simplex, since the recorded abundance was three times lower than in 1984. Furthermore, the 1986-1987 and 1992-1993 El Niño events were probably also responsible for the absence of daytime surface swarms of N. simplex in the Gulf of California during those years (^{Gendron 1992}). This coincides with our observations for whales, since the lowest abundance in our study occurred during 2003, the warmest of the two years. A lower number of daytime surface swarms were also found in the region during 2003 in the previous study by ^{Ladrón de Guevara et al. (2008)}.

Changes in the distribution and occurrence of marine organisms during one year may occur depending on oceanographic features. Consequently, long time series are necessary to better understand the variation in the abundance of euphausiids and fin whales in the Ballenas Channel and Bahía de los Ángeles region. Nonetheless, in our two-year investigation we were able to establish that both euphausiids and fin whales are more abundant in shallow waters close to the Baja California coast and south of Coronado Island, probably in association with the bottom topography that promotes high phytoplankton concentration and, hence, euphausiid abundance. In addition, euphausiids and fin whales showed the lowest abundance during October of both years, probably related to high SST and low recruitment of N. simplex. There seemed to be a lag in the maximum abundance of predators and prey, since euphausiids were most abundant late during the cold-warm transition period and fin whales in July, during the warm season. Finally, regarding the hypothesis that the Ballenas Channel may be a refuge area for fin whales during El Niño years (^{Tershy et al. 1991}), we propose that this is only true for strong El Niño events, because during our study we observed a higher abundance of fin whales during the colder year (2004) than during the warmer year (2003).

Acknowledgements

Funding was provided by the International Community Foundation (California, USA), Fundación Internacional de la Comunidad (Tijuana, Mexico), CICESE, and Mexico's National Commission for Natural Protected Areas (CONANP). The research was conducted under SEMARNAT permits No. SGPA/DGVS/00510 dated 24/1/ 2003 and No. 1564 dated 13/3/2003. We thank the staff of the natural protected area Islas delGolfo de California, in particular I Fuentes and A Zavala, for their logistic and technical support. Enriqueta Velarde (Universidad Veracruzana) also provided logistic support. Many people contributed during the field work: L Barbosa, E Bravo, L del Toro, L Enríquez, F Lafarga, O Guzón, R Mendoza, E Morteo, S Rodríguez de la Gala, A Baez, Y Schramm, C Díaz, S Jaume, and S Nigenda. We appreciate the collaboration of J Urbán (Universidad Autónoma de Baja California Sur). Zooplankton sample analysis was carried out mainly by one of the authors (PLG), but was complemented by P García, I Ambriz-Arreola, I Navarrete, and C Navarro. We received advice on data analyses from E Solana, A Jaramillo, and M López. Previous versions of the manuscript were reviewed by L Rojas, J Urbán, J Gómez-Valdez, and G Gaxiola. Two anonymous reviewers helped to improve the paper.