Seasonal climatology maps of the ADT (1993-2013) show the cyclonic eddy of Southern California clearly visible in winter, summer and autumn (Fig. 2A, C, D). There is also an increase in the number of mesoscale structures related to an anticyclonic flow, with an elongated meander being more pronounced during summer and autumn (Fig. 2D). Sea surface circulation is mostly towards the equator during spring and summer. This flow, which corresponds to the California Current, is driven by the prevailing southward winds. In spring, when winds along the coast are stronger, upwelled waters near the coast have lower temperature and consequently display the lowest elevation along the entire peninsula (Fig. 2B). Upwelled water is evident in the sea level minimums (ADT) and is reduced significantly in summer and autumn, especially in the region south of 28°N (Fig. 2C, D). Also during spring, mesoscale structures are reduced and the cyclonic gyre of Southern California is non-existent.

In the southern region, a coastal flow towards the pole is evident in summer and increases in autumn (Fig. 2C, D), as indicated by counter-currents. During summer-autumn, connectivity may be established between waters of tropical origin from the southern margin of the peninsula and the Pacific anticyclonic gyre (subtropical waters) from the eastern margin, as indicated by current lines (Fig. 2C, D).

Performances of the best-fitted models are shown in Table 2. Effect plots of the best-fitted models (Fig. 3) were used to describe the effect of environmental, temporal and spatial factors on selected demersal fish larvae on the west coast of the Baja California Peninsula.

M. productus, Sebastes sp., and C. xanthostigma belong to the so-called “Northern Complex”, distributed in the north-south plane across the area of influence of the California Current, but probably close to the coast when their distribution reaches the southern waters off the peninsula (Moser et al. 1987). C. stigmaeus and C. xanthostigma display a particular affinity for warm water, but are included in the “Northern Complex” due to the high affinity with the northern anchovy, Engraulis mordax; C. stigmaeus is the most temperate taxon in this group (Moser et al. 1987). The distribution of the slope/transitional larval fish assemblage is primarily related to the bathymetric range of adults.

Larvae of M. productus and Sebastes sp. show the Southern California Bight as their preferential distribution centre, while larvae of Sebastes sp. and Citharichtys stigmaeus are distributed across the California Current (Moser et al. 1993, Bautista-Romero et al. 2018). In contrast, C. xanthostigma, Synodus lucioceps, and Symphurus atricaudus are associated with fauna in warm to temperate areas (Moser et al. 1987). Their distribution is limited to the Baja California peninsula, mainly in the vicinity of Punta Eugenia (28°N), although Symphurus spp. and C. xanthostigma also occur with a low frequency in southern California (Moser et al. 1993).

The highest larval abundances of M. productus (Bailey et al. 1981, Funes-Rodríguez et al. 2009) and Sebastes spp. (Moser and Boehlert 1991, Moser et al. 1993) are consistent with those reported for coastal and slope waters north of the peninsula, closely related to the habitat used by adults and larvae. A second increase in M. productus to the south of the peninsula was also observed, evidenced as latitudinal differences in abundance, which seemed to break at Punta Eugenia (28°N). This distribution is probably related to the stock of dwarf hake, similar to the findings reported by Funes-Rodríguez et al. (2009). Additionally, the abundance of M. productus decreases off Punta Baja (~30°N), apparently bordered by waters transported towards the coast from the outer margin of the California Current around Isla Guadalupe. This finding could be related to an anticyclonic flow around an elongated meander, such as the one described by Durazo and Baumgartner (2002) and Soto-Mardones et al. (2004). These observations also explain the transport of mesopelagic larvae to the coast (Funes-Rodríguez et al. 2011) and the partition of the breeding area of coastal species (Aceves-Medina et al. 2018).

M. productus and Sebastes sp. occur mainly in winter-spring and are scarce in other seasons of the year (Moser and Boehlert 1991, Moser et al. 1993, Funes-Rodríguez et al. 2009). In this study, their larvae occurred principally in spring under sub-Arctic waters characteristic of the California Current. The current is defined by more homogeneous flow off Baja California, intense winds that favour coastal upwelling, low temperature and salinity from the surface to 100 m depth off California, and maximum coast-to-ocean slope at the sea surface, mainly in spring (Lynn and Simpson 1987, Soto-Mardones et al. 2004, Durazo 2009, 2015).

The most significant environmental predictors of both species were similar and defined mostly by the seasonal variation in the strength of the California Current southward flow along the west coast of the Baja California Peninsula. Although larvae of M. productus are positively related to relatively lower SST, other studies (Bjorkstedt et al. 2002, Bautista-Romero et al. 2018) recorded a preference for low salinity, low dynamic height and unstratified water in a wide mixed layer depth. This preference explains the broad distribution of M. productus larvae in waters south of the peninsula. Similarly, larvae of Sebastes sp. are related to unstratified waters and low temperatures, these parameters being the most significant predictors of its larval distribution, as evidenced in localities with low SST, high geostrophic flow or lower ADT values in other studies (Weber and McClatchie 2010, Ash and Checkley 2013, Kosllow et al. 2013). This would explain the relationship with other species that take advantage of inshore areas in spring, characterized by highly productive and relatively stable conditions in the Southern California Bight (Weber and McClatchie 2010).

Upwelling areas may exert major effects in the planktonic stages, derived from the advection of eggs and larvae outside of breeding areas and consequently affecting the recruitment of stocks (Bailey 1981, Sakuma and Ralston 1997). Similarly, the scarcity of larvae of coastal species is associated with the region of upwelling and offshore Ekman transport, suggesting that adults avoid spawning in upwelling areas, thus avoiding offshore transport of their larvae into the California Current (Doyle et al. 1993). Although coastal species with benthic eggs tend to be relatively scarce during the upwelling season in Angola Bay, three species of flatfishes showed positive responses to upwelling events. These were associated with the highest microzooplankton biomass, suggesting that the spawning of these species is timed to maximize food availability (Pattrick and Strydom 2014).

To compensate for the effect of the offshore Ekman transport, the spawning season in M. productus and Sebastes sp. occurs in early winter in California (Moser et al. 1993), when offshore transport shows a seasonal decrease (36-27°N, Bailey 1981). Additionally, most late eggs and larvae of M. productus are located near or within the thermocline (Ahlstrom 1959, Cass-Calay 1997, Moser et al. 1997) limited by the mixing layer, in a vertical distribution range below the photic layer (50-100m depth) (Sakuma et al. 2007). Larvae of Sebastes sp. are captured mainly within the mixed layer (Sakuma et al. 2007) and increase abruptly in abundance near upwelling fronts (Moser and Boehlert 1991, Bjorkstedt et al. 2002). This finding is likely associated with surface convergence related to the increased thermocline slope, so the distribution of fish larvae may change along the pycnocline slope associated with fronts (Bjorkstedt et al. 2002). In other regions, larvae of flatfishes showed positive responses to upwelling events, suggesting that spawning is timed to maximize food availability for larvae in Algoa Bay. However, larvae hatching from benthic eggs tend to be larger and more developed upon hatching than those from pelagic eggs, and would therefore have some enhanced ability to resist transport (Pattrick and Strydom 2014).

Larvae of C. xanthostigma and C. stigmaeus occur mainly in the second half of the year, the latter throughout winter (Moser et al. 1993, Bautista-Romero et al. 2018). During the second half of the year, a reduced core of sub-Arctic water occurs as far south as the tip of the peninsula (28°N). This leads to a progressive narrowing of the California Current near the surface as a result of mixing with warmer and more saline waters along its southerly path (Durazo 2015), governed by a warm countercurrent on the southern margin of the peninsula (Valle-Rodríguez and Trasviña-Castro 2017). Both species exhibited the best fit to the model at low salinities and a mixed layer depth. Dynamic height had little or no effect on these species, which occur mainly associated with the eastern margin of the Pacific gyre in a poorly stratified water column. Similarly, other studies have reported higher catches of Coryphaena hippurus larvae near fronts and eddies in the northern Gulf of Mexico (Kitchens and Rooker 2014).

Citharichthys spp. larvae may migrate vertically through the thermocline and pycnocline, which could favour recruitment to settlement areas near the coast (Sakuma and Ralston 1997, Sakuma et al. 1999). In this study, flatfish larvae preferred coastal shelf waters, with some species (e.g. S. atricaudus and S. lucioceps) preferentially distributed across a wide continental shelf in Bahía Sebastián Vizcaíno. This finding is consistent with other regions where the narrow continental shelf reduces the potential spawning habitat (e.g. sardine), with clear limits in terms of temperature and bottom depth in European Atlantic waters (Bernal et al. 2007). Additional examples are the high larval density and number of coastal and shelf species on the coastal side, apparently limited by the shelf-slope front off the northern Catalan coast (Sabatés and Olivar 1996). However, in other areas the largest number of early stages occur in shelf waters of high productivity influenced by large rivers, such as the Bay of Biscay (Ibaibarriaga et al. 2007) and the Columbia River plume (Doyle et al. 1993). This is not the case of the subtropical Baja California Peninsula, a region with negligible freshwater input, favourable winds and upwelling events over the whole year, which peak from spring to early summer (Durazo 2015).

S. atricaudus and S. lucioceps are part of the so-called “Southern Coastal Complex”, distributed mainly in the coastal zone around Punta Eugenia (28°N) (Moser et al. 1987). Despite their similar geographic distribution, the preferred habitat of each of these two species differs. S. lucioceps thrives mainly in relatively homogeneous waters with high dynamic height and salinity, even when their larvae occur throughout the year with peak abundance in autumn (Moser et al. 1993). For its part, S. atricaudus occurs in areas with shallow mixed layer depth and unstratified waters. This is consistent with its peak seasonal abundance (summer-autumn), as it remains within cyclonic eddies around Punta Eugenia and thus avoiding offshore transport, while settlement occurs in late autumn and winter (Charter and Moser 1996).

In summary, oceanic circulation and environmental forcing have significant consequences (extension and delimitation) in the distribution of demersal fish larvae in the southern part of the California Current. Using GAMs to explore the preferred habitat allowed us to identify common environmental variables that influence larval distribution (temperature, salinity, dynamic height and degree of column stratification). This is possible since larval associations reveal that different species share similar life cycles despite the dispersive environment. The fact that these occur sympatrically (eggs and larvae) may contribute to their being affected by the same mechanisms (Moser and Smith 1993). However, the set of variables and ranges where habitat preference is maximized is usually species-specific, and this allows a better description of the most important environmental characteristics that are closely related to the habitat used by adults and larvae.