The Phenology of Coastal Upwelling in the California Current Steven J. Bograd and Daniel M. Palacios NOAA Southwest Fisheries Science Center, Environmental Research Division William J. Sydeman PRBO Conservation Science Scott R. Benson NOAA Southwest Fisheries Science Center, Protected Resources Division Mark H. Carr and Mark D. Readdie Long Marine Laboratory, University of California, Santa Cruz Letter of Intent for FY2007 funding under Fisheries and the Environment (FATE) BACKGROUND Many marine organisms have life histories adapted to seasonal events in the environment. Changes in the amplitude or phase of seasonal events can therefore significantly affect the productivity and community structure of marine ecosystems, from primary producers to fish stocks to apex predators (Cushing, 1990; Beare and McKenzie, 1999; Bograd et al., 2002; Logerwell et al., 2003). Such phenological effects are potentially more disruptive than those associated with interannual climate events and decadal climate shifts. Thus, any set of indicators of ecosystem state would need to include those that quantify changes in the seasonal cycle of dominant physical processes. Phenology plays a particularly critical role in the California Current System (CCS), in which ecosystem productivity and structure is driven largely by the seasonal cycle of coastal upwelling. The impact of an anomalous seasonal cycle was particularly evident in 2005, when the onset of coastal upwelling was delayed by several weeks in the northern CCS (Schwing et al., 2006),resulting in anomalously warm sea surface temperatures (Kosro et al., 2006; Pierce et al., 2006),low surface chlorophyll levels (Thomas and Brickley, 2006), a spatial redistribution of zooplankton species (Mackas et al., 2006), record low rockfish recruitment and a lack of forage species (Brodeur et al., 2006), total breeding failure of a dominant planktivorous marine bird (Sydeman et al., 2006), and changes in California sea lion foraging strategies (Wiese et al.,2006). Our objective is to develop simple indicators that quantify the timing of the onset of coastal upwelling in the California Current (i.e., the date of the “spring transition”), the temporal evolution and overall intensity of upwelling, and the total duration of the upwelling season, as well as the spatial variations of these properties along the West Coast between Baja California and Vancouver Island. Once developed, these indicators will be applied in two contexts. First, we will validate their usefulness through correlations with historical biological time series, including time series of seabirds, salmon, rockfish, and leatherback turtles. Second, we will develop and test (in spring 2007) an operational spring transition index that could provide an early warning signal to resource managers of the probability of a disruption to the California Current ecosystem. DATA SETS Physical Data The coastal upwelling index (UI; Bakun, 1973; Schwing et al., 1996) has been a workhorse of fisheries oceanography for more than three decades, greatly contributing to our understanding of physical-biological coupling in eastern boundary current systems. The UI represents the magnitude of the offshore component of Ekman transport, which approximates the amount of water upwelled from the base of the Ekman layer. Positive values are the result of equatorward wind stress; negative values imply downwelling, the onshore advection of surface waters accompanied by a downward displacement of water. Upwelling indices are computed from the 6-hourly sea level pressure (SLP) fields obtained from the U.S. Navy Fleet Numerical Meteorology and Oceanography Center (FNMOC). We will use the 40-year time series (1967-2006) of daily UI at six locations along the U.S. West Coast, from 33°N to 48°N, separated by 3° latitude, to develop indicators of upwelling conditions. In addition, we will use complementary surface wind and coastal sea level data from NOAA buoys and shore stations, respectively, to contribute to indicator development. NOAA buoys closest to the UI locations will be used. Figure 1. Seabird reproductive phenolgy from the Farallon Islands. Biological Data The biological time series available for analysis with the phenological indices include ~35years of nesting dates for seabirds of the Farallon Islands (Figure 1a). The seabird dataset includes information on the phenology and productivity of planktivorous auklets (Ptychoramphus aleuticus) and piscivorous murres (Uria aalge). In general, the timing of nesting is predictive of murre and auklet breeding success (Abraham and Sydeman 2004). However, the factors affecting seabirds timing of breeding are not well understood. Preliminary analyses indicate that mean annual auklet timing of egg-laying varies by upwards of 60 days between years, and is related to wintertime SST in a non-linear fashion (Figure 1b). A similar pattern has been found for murres. Auklet nesting is highly synchronous in years of early upwelling and cold SST. When upwelling is delayed, egg- laying is more spread out and less peaked (Sydeman, unpublished data). The leatherback aerial survey data include information on distribution and abundance in shelf waters (<90m depth) between Point Conception and the California/Oregon border during 1990-2003, and similar fine-scale survey data between Point Piños and the Golden Gate Bridge during 2002-2006, which also included qualitative data on leatherback prey, jellyfish. Leatherbacks migrate to the west coast of North America to forage on seasonal aggregations of jellyfish prey, although the factors affecting the development and location of prey patches are poorly understood. Recent telemetry studies indicate that leatherback arrival at the California Current generally occurs during late spring through early summer, with peak densities occurring during September relaxation events when jellyfish abundance tends to be greatest; however, temporal and spatial variation in this pattern is common. Previous analyses have established a link between leatherback abundance off California and the Northern Oscillation Index (NOI) during1990-2003 (Benson et al., in press). More recently, leatherback and jellyfish densities were observed to be very low during 2005-2006, when the onset of upwelling was delayed. These patterns suggest that coastal upwelling is central to the occurrence of jellyfish and leatherbacks, but more detailed data on the timing, duration and intensity of upwelling are required to elucidate the physical processes involved. In this project, we will examine interannual patterns of leatherback arrival, distribution and abundance relative to the timing and duration of upwelling/relaxation events during 1990- 2006. Indices will be validated against aerial surveys conducted within the same regions during July-October 2007. The PISCO subtidal monitoring program conducts annual community surveys of rocky reefs along a section of the California coast from Santa Cruz to the Channel Islands. Initiated in 1999,the surveys monitor aspects of fish (density and size), invertebrate and algal community structure, rockfish recruitment and small scale physical oceanography (temperature, currents). Biological data are used to investigate large-scale biogeographic patterns and to understand long-term variability in kelp forest communities. Physical aspects of the coastline (relief, substrate, swell exposure), in addition to the derived upwelling indicators, will be used in the analysis to understand their effects on patterns of community structure. Local rockfish recruitment dynamics are analyzed in conjunction with the pelagic juvenile rockfish trawls done by SWFSC-FED. APPROACH Development of Upwelling Indices This project will build upon and expand the utility of the historical Upwelling Index (UI; Bakun, 1973).Since upwelling has a cumulative effect on ecosystem productivity and structure, we focus on the cumulative upwelling index (CUI),based on integrating the mean daily upwelling indices at six locations along the U.S. West Coast (Schwing et al., 2006). The st integration begins on January 1 , and continues to the end of the year. The climatological mean (1967- 2005) CUI at the six locations are shown in Figure 2, yielding a mean start (when CUI reaches its annual minimum) and end date (when CUI reaches its annual maximum) of the upwelling season at each latitude. Note that the climatological upwelling season is nearly year-round (though weaker) off southern California and Baja, but gets progressively shorter with northward latitude. We define the start date of the upwelling season (i.e., the date of spring transition) as the st date on which the CUI, integrated from January 1 , reaches its minimum value. This is the date after which positive UI (upwelling) prevails. Similarly, the end date of the upwelling season is defined as the date on which the CUI reaches its maximum value. That is, downwelling prevails after this date. We will determine the date of spring transition, the intensity of upwelling, and the total duration of the upwelling season from the 40-year CUI series at each location as described below: Spring Transition Index (STI): This is determined as the date at which the minimum value st of CUI is achieved, following an integration beginning on January 1 . This is the same definition used by Pierce et al. (2006) and Schwing et al. (2006). This index will be quantified as the Julian date of the spring transition (STI) and the anomaly from the ). climatological mean spring transition date (STIanom The