Limnol. Oceanogr., 44(2), 1999, 466–4699

Recruitment of intertidal barnacles was markedly higher in May–September 1997, just after the apparent onset of El Niño conditions in the waters off California, than over the same period during the previous year. This increase was geographically broad, spanning five degrees of latitude, and was unusually large, relative to interannual differences in barnacle recruitment previously documented for this region. Increased onshore transport associated with El Niño events probably caused high recruitment in 1997. This explanation accords with the findings of previous, smaller scale studies in central and northern California and indicates that oceanographic transport anomalies associated with El Niño can strongly affect benthic communities in the northeast Pacific. The catastrophic effects of the El Niño–Southern Oscillation on the productivity of marine ecosystems in South America are well known (Cushing 1971; Barber et al. 1985). Although El Niño events are largely tropical, they have also been linked to substantial changes in temperature, salinity, and circulation patterns in the California Current System of the northeast Pacific (Enfield and Allen 1980; Chelton and Davis 1982; Simpson 1983, 1984; Emery and Hamilton 1985; Huyer and Smith 1985; Norton et al. 1985; Simpson 1992; Jacobs et al. 1994; Lynn et al. 1995; Ramp et al. 1997; Liu et al. 1998). These large and widespread changes affect planktonic communities at multiple trophic levels (Chelton et al. 1982; Graybill and Hodder 1985; Miller et al. 1985; Mysak 1986; Pearcy and Schoener 1987; Brodeur et al. 1992; Ainley et al. 1995). Because many animals that inhabit benthic communities have a planktonic larval phase, one might expect these changes in planktonic communities to affect recruitment rates to benthic populations. However, evidence for such effects is mixed. In some cases, recruitment is higher than normal (Paine 1986; Roughgarden et al. 1988; Ebert et al. 1994), but, in other cases, recruitment is below levels seen in prior or subsequent years (Tegner and Dayton 1987; Ebert et al. 1994; Shkedy and Roughgarden 1997). If large scale environmental changes induced by El Niño cause these changes in recruitment rates, as is often proposed (e.g., Paine 1986; Tegner and Dayton 1987; Roughgarden et al. 1988; Ebert et al. 1994), then El Niño events should have geographically broad effects on recruitment. However, most of these studies were confined to one or a few sites in close proximity to one another, so it is difficult to verify whether the observed changes in recruitment were due to El Niñorelated regional environmental changes or to changes in more local factors. The exception, a recruitment study at several sites in northern California and the Southern California Bight, did not find a significant change in recruitment across sites in either region (Ebert et al. 1994). Thus, while large changes in planktonic communities occur during El Niño years, evidence for effects on recruitment to benthic populations at the regional scale is lacking. From 3 May 1996–18 September 1997, we monitored recruitment of the intertidal barnacles Balanus glandula, Chthamalus dalli, and Chthamalus fissus at nine sites spanning approximately five degrees of latitude in central and northern California. These sites extended from Cape Mendocino (408249N) to San Simeon (358359N) (Fig. 1). Serendipitously, this period included the onset of warm sea surface temperatures, anomalously weak upwelling, and a strengthening of the Aleutian low-pressure system in May 1997, presumably associated with the 1997–1998 El Niño (Liu et al. 1998; CoastWatch, West Coast Regional Node, National Marine Fisheries Service, La Jolla, California; http:// cwatchwc.ucsd.edu/cwatch.html). To investigate a possible effect of El Niño on intertidal barnacle populations in the northeast Pacific, we compared recruitment during May– September 1997 with recruitment during the same months in 1996. Because the sites span a large region, they provide an opportunity to determine whether changes in recruitment during an El Niño year occur over a broad scale. Recruitment was monitored by bolting five 50-cm2 styrene plates covered with safety walk tape (3M) to the substrate at each site (Farrell et al. 1991). Plates were arranged haphazardly in the midintertidal (mussel) zone on the most wave-exposed rocks accessible. Approximately once a month, we collected plates and deployed fresh ones in the same locations. Sampling intervals varied in length from 26 to 31 d because collection dates were constrained to occur when there were daytime negative tides for several days in a row (allowing us to visit all sites during consecutive days). C. fissus is visually indistinguishable from C. dalli, so recruits were identified to the genus level only (cf. Farrell et al. 1991). From these data, we calculated mean monthly recruitment for May–September 1996 and May–September 1997. Recruitment data for August 1996 were lost due to equipment failure, so we calculated mean monthly recruitment rates for 1997 both with and without August 1997. All differences reported as significant in this paper were significant in both cases. Exact recruitment rates and P-values reported here correspond to the first case described above. Because the data did not meet parametric assumptions, we tested for significant differences between years with a nonparametric Friedman’s test modified for within-site replication (Gibbons 1985). Overall, recruitment increased highly significantly across the sampled region from 1996 to 1997 for Balanus (Q 5 17.7, P , 0.001) and Chthamalus (Q 5 18.3, P , 0.001). When sites were tested individually, recruitment was significantly higher at a majority of the sites for both Balanus and Chthamalus (Fig. 2). However, at Pescadero Beach and Soberanes Point, Balanus recruitment actually decreased in 1997 (Fig. 2), suggesting that El Niño studies restricted to one or a few sites will not necessarily detect a change in recruitment that occurs over a geographically broad scale, even when that change is highly significant. Earlier studies indicate that the presence of a statistically