MACCALL: LOW-FREQUENCY VARIABILITY IN FISH POPULATIONS CalCOFl Rep., Vol. 37, 1996

PAllERNS OF LOW-FREQUENCY VARIABILITY IN FISH POPULATIONS OF THE CALIFORNIA CURRENT ALEC I). MACCALL N~tion~lManne Fi5herim Service South\\ect Fi5heries Science Chter 31 50 Paradise Ilr. Tihuroii, C::ilifwtiia 94c120

ABSTRACT kinds of conditions. Rather, there are probably a great Long-term data sets include sedimentary fish-scale number of possible regimes and abrupt discontinu- paleochronologies, faunal surveys froiii the 1SOOs, and ities connecting them, flip-flops froin one regime to the records of game fish catches maintained by the Tuna another; niultifarious regimes involving biology or Club at Avalon since 1898. These data suggest a inode climate, or oceanography, or migrations, temperature, of low-frequency biological variability over a cycle of or weather, or combinations of these . . . . My main .5O to 70 years that is associated with alternation of warm point is that there are no simple statistics in the ordi- and cold physical regimes. The very warm conditions nary sense. There are internal, interactive episodes in the northeastern Pacific Ocean since 1976 seein sim- locked into persistence, and one is entirely fooled if ilar to conditions experienced from about 1850 to 1870. one takes one of these short intervals of a decade or Coastal pelagic fish abundances since 1930 indicate an so and decides there is some sort of simple probabil- orderly sequence of four or five dominant species through ity associated with it. . . . organisms must respond to these cycles that is very similar to the order and timing more than just fluctuations around some optimum of fluctuations in catches of siniilar pelagic fishes in condition. Actually, many of their characteristics and Japanese waters. The consistency arid predictability of fluctuations of populations must be related to these this rotation and the biological relationship to low- very large alternations of conditions. frequency fluctuations in physical conditions are iin- Isaacs coined the term wyim to describe distinct en- portant areas for further research. These patterns of low- vironmental or climatic states, and yqime SI$ has recently frequency variability result in boom-and-bust fisheries, gained acceptaiice as a term for the abrupt transitions and pose serious probleiiis for “sustainable development.” between regimes. It is notable that publication of lsaacs’ regime concept in 1976 itself coincided with a profound INTRODUCTION reginie shift in the northeastern Pacific. That shift was With recognition of the worldwide decline in fish not apparent at the time, and it required another decade stocks, the guideline of “sustainable development” is re- for the oceanic warming (figure 1) to be recognized as ceiving international support as an explicit goal of fish- a new and persistent condition (MacCall and Prager ery and ecosystem management. Ironically, this conies 1 988; Trenberth 1990). at a time when the biological basis of sustainability in In ths paper I exanine variability in southern Cahfornia some of the world’s niajor fisheries is itself being ques- tioned because of natural patterns of variability (e.g., Lluch-Belda et al. 1989; Beaniish 1995). In a broader 19 , context, the science of ecology is moving away froni Fta- tic and equilibrium views of ecosystems. Rather, scien- tists are coining to perceive ecosystems as pulsing and fluctuating entities, driven by external sources of vari- ability and by internal nonlinear dynamics. Odurii et al. (1995) state the possibility that “what is sustainable in ecosystems, is a repeating oscillation that is often poised on the edge of chaos.” At the 1973 CalCOFI Symposium, John Isaacs (1976) took issue with the common implicit assumption in fish- eries science that there is “some steady state . . . around which there is a normal distribution of perturbations.” 15 ’ , He continued, 1890 1910 1930 1950 1970 1990 YEAR

This certainly does not seein to be the case. The as- Figure 1. Mean annual temperatures (“C) at the Scripps Institution of sumption is that there are some normal statistics to all Oceanography Pier, La Jolla, California.

1 00 MACCALL: LOW-FREQUENCY VARIABILITY IN FISH POPULATIONS CalCOFl Rep., Vol. 37, 1996

fish populations over the longest time scales supported The major warming that began about 1976 is clearly by the qualitative and quantitative information available. visible in the Scripps Pier temperature record. This event Even the quantitative time series are niuch too short for was widespread and occurred over the entire northeast- rigorous statistical analysis. My approach to these data ern Pacific Ocean, from Mexico to Alaska (Cole and is that of “historical science” (Gould 2989; Francis and McLain 1 989). However, the geographic extent of ear- Hare 1994); consequently, some of my interpretations lier temperature regimes and regime shifts is less clear. are speculative. An earlier regime shift occurred about 1940, begin- Perhaps because fish are an important commodity, ning a long period of cold temperatures. Consequently better long-term records exist for fishes than for any the Scripps Pier record consists of three conspicuous other biological component of the California Current temperature regimes: a moderate and highly variable pe- System. Paleochronologies of fish debris preserved in riod fiom the beginning of the time series to about 1940, anaerobic sediments provide invaluable information on a cold period from 1940 to 1976, and a warm period prehistoric fluctuations and very-low-frequency vari- since 1976. Ware (1995) describes similar alternations of ability. The earliest historical records indicative of fish temperature regimes off British Columbia. abundance in the California Current are relatively qual- It is not clear how these low-frequency temperature itative and begin with the Pacific Railroad Survey of regimes are related to the El Nifio/Southern Oscillation. 1856. Quantitative big game fishing records have been However, brief warm events related to both El Niiio and maintained since the turn of the century by the Tuna variability in the Aleutian Low (Wooster and Hollowed Club at Avalon, on Santa Catalina Island. Although Tuna 1995) are superimposed on these underlying tenipera- Club data have seen little use in scientific contexts, these ture regimes. Major warm events have occurred at a tran- records constitute the longest quantitative and unbroken sition (1940), in the midst of a cool period (1958-.59), series of biological observations in this region. The State and in the midst of a warm period (1982-83 and of California began keeping systematic records of fish 1992-93). landings in 191 6, and began comprehensive biological sampling of coastal pelagic fish landings in the 1930s. FISH ABUNDANCES Good information on ichthyoplankton exists for niuch The anaerobic sediments of the Santa Barbara Basin of the last 50 years due to the California Cooperative provide a remarkable paleochronology of fish scale-de- Oceanic Fisheries Investigations (CalCOFI) and other position rates (Soutar 1967). Soutar and Isaacs (1974) fishery and oceanographic programs that were initiated developed a 160-year time series of Pacific sardine following World War 11. (Savdirzopr sqax) and northern anchovy (Engraulis ~OY- For the purpose of this paper, I use ocean tempera- dux) scale-deposition rates that indicated large natural ture as a simple proxy for the more complicated and long-term fluctuations in both of those species. poorly understood suite of oceanographic variables that Baumgartiier et al. (1‘992) extended and refined these influences biological populations. Sea-surface tenipera- studies, and developed a paleochronology spanning the tures have been monitored continuously at the Scripps last 1,700 years at a temporal resolution of 30 years Institution of Oceanography Pier (Scripps Pier) in La (figure 2). Power spectra of these long time series show Jolla, beginning in 1916 (figure I), forming one of the high variability in both sardine and anchovy scale-de- longest time series of sea-surface temperatures on the position rates at a period of approximately GO years, Pacific coast. Earlier observations of sea-surfxe tem- and at longer periods that differ for the two species. perature from ships-of-opportunity are provided by the Ware’s (1995) analysis of a much shorter physical and bi- COADS data set (Mendelssohn and Roy l996), but there ological data set also indicated a spectral peak in the range are insufficient observations from southern California of 50 to 75 years. Neither Baunigartner et al. (1992) nor waters before 1920 to provide annual resolution of teni- Ware (1995) was able to identify a physical mechanism perature fluctuations (K. Parrish, PFEG, NMFS, Pacific associated with this low-frequency variability. Grove, CA, pers. conini. 9/13/95). However, Parrish’s examination of the post-1920 COADS data did denion- strate that low-frequency temperature fluctuations in Anecdotal Information southern California do not necessarily parallel temper- Biological surveys of the Pacific coast began in the ature fluctuations off central or northern California. niid-l80Os, and provide useful indications of presence Thus, annual average Scripps Pier temperatures may pro- or absence of species but have poor resolution for the vide a siniplified representation of southern California purpose of estimating the timing and duration of ,envi- ocean conditions but may not always be a reliable indi- ronmental regimes. However, in comparison with con- cator of conditions in adjacent Pacific coastal regions ditions during most of the twentieth century, some of such as central California. the early survey information is so peculiar that Hubbs

101 MACCALL LOW-FREQUENCY VARIABILITY IN FISH POPULATIONS CalCOFl Rep., Vol. 37, 1996

PACIFIC SARDINE

NORTHERN ANCHOVY 56 -- J(1 -- T 15

20

15

10

5

0 xx) xg a Ma 700 800 900 loo0 1100 1200 1JM) 1400 1Mo 16W 17W 1800 1803 2000 YEAR Figure 2. Sardine and anchovy scale-deposition rates in the Santa Barbara Basin (reproduced from Baumgartner et al. 1992)

(1948) was able to infer the existence of an unusually California until a small population reappeared in San warm ocean climate during the mid 1800s. Diego Bay about 1984 (Jones et al. 1988). The San Diego According to Hubbs, the Pacific Railroad Survey sam- Bay population has remained viable over the past decade pled fish fauna in the San Diego area during the period (R. Burhans, SIO, UCSD, La Jolla, CA, pers. comm. 1853-57. This superficial survey encountered many 2/15/96). Hubbs’ inference of warm oceanic conditions “southerly” fish species that no longer occur in south- in the mid-nineteenth century is consistent with reap- ern California. Hubbs reported that when Jordan and pearance of the Pacific seahorse at San Diego during the Gilbert conducted a thorough sampling of the California warming since 1976. coast in about 1880, the fish fauna from Monterey Bay and southern California still had an unusually strong rep- Tuna Club Records resentation of subtropical species. These anomalies led The Tuna Club was formed at Avalon, Cathna Island, Hubbs to infer that southern California experienced ex- in the summer of 1898, shortly following the discovery ceptionally warm oceanic conditions from about 1850 of the sport of angling for large ocean game fishes by to 1870; the beginning and ending dates are imprecise. its founder, Dr. C. E Holder (Macrate 3 948). The club’s In 1857, the Pacific Railroad Survey encountered five records, which now span nearly 100 years, document specimens of Pacific seahorse (Hippocutnpus iqens) in San sizes and numbers of several fish species taken in annual Diego Bay (figure 3). This species is normally found tournaments. From 1909 to 1919, a summer tourna- much farther to the south and was exceedingly rare in ment was held from May 1 to September 30, and a

102 MACCALL: LOW-FREQUENCY VARIABILITY IN FISH POPULATIONS CalCOFl Rep., Vol. 37, 1996

15 I . . 12 =. ::11 I+ 10 1 . I + I -I L!? I =I sg LI . July 1 I %8 .== I kv) .= r7 I E .. I. LL .- . $6 Y -8 0 $5 +w 8 4 .= -I I 3 n 2 . I 1 . n - 1850 1900 1950 21 YEAR Jan 1 . Figure 3. Number of Pacific seahorses reported by various sources (data 1 0 1910 1930 1950 1970 1990 from Jones et al. 1988). YEAR

Figure 5. Dates at which first tuna were caught in Tuna Club annual separate winter tournament was held from October 1 to tournaments. April 30. From 1920 on, annual tournaments corre- sponded to calendar years. during the extended cold regime from 1940 to 1976. The nature of the Tuna Club data dlffei-s among groups Moderate numbers of tuna reappeared in the annual tour- of fish. Naturally, tunas were of greatest interest, and the naments following the 1976 warming. The two periods club yearbooks provide information on the largest fish of abundance reflect different species of tuna: pre-1940 caught each year, catch by month, average weight of all catches consisted mainly of northern bluefin tuna, while fish reported, and date of first catch. Unfortunately the post-1976 catches have been mostly yellowfin and big- club did not distinguish aniong several species of large eye tunas. Interpretation is also complicated by the growth tunas, which include northern bluefin tuna (’Thunnus of commercial fishing for tunas in southern California thynnus), yellowfin tuna (Thunnus albacaves), and bigeye during the 1920s (Bayliff 1992; Wild 1992), with con- tuna (Tlzunnus obesus). The Tuna Club considered striped sequent declines in abundance and size. marlin (Tetvaptuvus audux) and swordfish (Xiphim glad- The dates on which the first tuna was taken in the ius) to be worthy alternatives to large tuna and kept annual tournaments also coincide with the large shifts separate records for those species. Club yearbooks also in Scripps Pier temperatures (figure 5). The post-1976 provide data on the largest fish caught in each annual pattern is similar to that prior to 1940, when the earli- tournament for a number of “lesser” species, including est catches were often made in April or May. During the albacore (Thunnus alalunga), white seabass (Atvactoscion long cold regime, the few tuna that were caught first ap- nobilis), yellowtail (Seviola lalandi), and giant sea bass peared in July or August. Remarkably, the earliest an- (Steveolepis g&m). nual appearance of tuna in club records is March 4, and Tunas were landed in large numbers in many tour- first appearances were in March of three of the four years naments through the 1920s (figure 4), but were very rare 1902-5. Sea-surface temperatures are normally quite cold in March, and immigration of bluefin tuna to south- 1000 ern California would be unlikely at that time of year, 900 suggesting that there may originally have been a small

800 resident population of tuna in southern California. A re- cent anomaly in figure is a tuna that was caught on 700 5 January 3, 1987, in waters far to the south of the usual 600 5 fishing grounds near Catalina Island. 500 Catches of striped marlin increased sharply in the z2 400 1930s, providing an alternative to tuna, which were 300 becoming scarce. Although there is little long-term pat-

200 tern to annual marlin catches (figure 6), the size of the largest fish dropped conspicuously about 1940. In con- 100 trast to many changes associated with the 1940 regime 0 IO 1910 1930 1950 1970 1990 shift, large marlin did not reappear after 1976 (figure YEARS 7). And unlike the dates for tuna, the dates of the first Figure 4. Number of tuna reported in Tuna Club annual tournaments marlin caught each year have been virtually constant,

103 MACCALL LOW-FREQUENCY VARIABILITY IN FISH POPULATIONS CalCOFl Rep., Vol. 37, 1996

800

700

600

500

wLT 2 400 3

300

200

100

0 1890 1910 1930 1950 1970 1990 an 1 - YEAR ) 1910 1930 1950 1970 1990 Figure 6. Number of marlin reported in Tuna Club annual tournaments. YEAR Figure 8. Dates at which first marlin were caught in Tuna Club annual tournaments

200

I 2 I- 150 0 4 NOV-DEC L 0 I 2 100 z Y I- t- 4 OCT P 0 $ 50 w d SEPT

0 3 1910 1930 1950 1970 1990 YEAR

Figure 7. Largest (points) and annual average (solid line) weight of marlin I 1910 1930 1950 1970 1990 taken in Tuna Club annual tournaments. YEAR

Figure 9 Month in which the last marlin was caught in Tuna Club annual tournaments. except for one fish caught in southern waters on January 1, 1983 (figure 8). However, marlin appear to be de- parting southern California waters later in the warm pe- riod since the mid-l 970s (figure 9). Tuna Club catches suggest that albacore were abun- 30 t mmm dant in southern California during the first two decades of this century, but catches declined severely in the late 1920s and again declined in the late 198Os, paralleling difficult periods for California’s commercial albacore fish- ery (Laurs and Dotson 1992). From 1910 to 1919, when winter tournanient records were kept separately, alba- core continued to be caught after October 3, and the largest “winter” fish was consistently larger than the largest “suninier” fish (figure 10). This phenomenon “Winter” iOct 1 to April 301 I cannot have lasted long after the winter tournaments 1 were discontinued, because large albacore disappeared -, ~, , , , , , ~, , 1890 1910 1930 1950 1970 1990 altogether from the annual tournaments shortly after YEAR 1920. Large late-season albacore reappeared in southern Figure 10. Largest albacore taken in Tuna Club annual tournaments. Winter California waters about 70 years later, and in the mid- records were not maintained after 1920.

1 04 MACCALL: LOW-FREQUENCY VARIABILITY IN FISH POPULATIONS CalCOFl Rep., Vol. 37, 1996

30 I and that the initial biomass of giant sea bass was at least 1,300 tons in southern California waters. . I Like the Pacific seahorse described earlier, the north- ern populations of these species are probably established or strengthened during prolonged warm periods, such as occurred during the mid to late 1800s. During cooler regimes, these populations would no longer be self- sustaining, and would slowly decline due to lack of re- cruitment (cf. Jacobson and MacCall 1995). In the ab- sence of fishing, adults of these species have low mortality rates, and a substantial population could remain resident for decades. However, during cold regimes even a mod-

1890 1900 1910 1920 1930 1940 1950 erate fishery would deplete the population very rapidly. YEAR

Figure 11. Association of commercial harvesting (solid line) with decrease Small Pelagic Fishes in size of largest yellowtail (points) taken annually by Tuna Club members. Patterns and causes of fluctuations in pelagic fish pop- ulations of the California Current have long been fun- 1980s once again were encountered very late in the year damental issues of study by CalCOFI. Reviews include (T. Foreman, IATTC, La Jolla, pers. conini., 1987). In MacCall (1986) and the Pacific Fishery Management both the 1920s and 1980s, the appearance of large late- Council (1996). Historical fluctuations in several species season albacore was followed by a severe decline in south- of small pelagic fishes are summarized very briefly here. ern California catches. The links to climate variability Scientific monitoring of California’s pelagic fish re- are unclear, and may be related to environmental con- sources began in the 1920s, and useful estimates of sar- ditions in distant waters. These albacore fluctuations dine landings by age group were provided by the show a 60-year pattern that seem out of phase with the California Division of Fish and Game’s Bureau of temperature regimes. An intriguing possibility is that the Commercial Fisheries beginning in 1932. But it wasn’t albacore patterns seen in the 1920s and 1980s consis- until Murphy (1966) developed a formal technique now tently lead a regime shift to colder conditions by ten to known as virtual population analysis (VPA) that con- fifteen years. sensus could be reached on the detailed historical pat- Tuna Club records of yellowtail, white seabass, and tern of sardine population abundance. The Pacific sar- giant sea bass caught in the annual tournaments do not dine reached a peak abundance in 1934, declined sharply, show clear long-term patterns other than a severe in- and recovered to a secondary peak in 1941 before de- pact from coniniercial fishing during the period 1916 to clining to unmeasurably low levels over the following 1930. Southern California’s giant sea bass were con- 30 years. The recent sardine recovery was first detected niercially depleted by 1930, after which catches by men- in the early 1980s (Watson 1992; Wolf 1992), and the bers of the Tuna Club were rare. For yellowtail, the population has grown steadily since then (Barnes et al. coincidence of intense commercial harvesting and a 1992), not yet having shown evidence of a peak in abun- decline in size is striking (figure 11). After the 1920s, dance. Jacobson and MacCall (1995) have described a southern California’s catches of these species were main- temperature-dependent relationship between recruit- tained by seasonal migration of younger fish from dis- ment and parental stock for sardines that links the long- tant waters. term cycle of sardine abundance to the low-frequency There may have been resident California populations variability of sea-surface temperatures. of several species of fish, including bluefin tuna, yel- Pacific mackerel (elsewhere known as chub mackerel, lowtail, and giant sea bass. The strongest evidence for a ScoMzbcr.ja~oizicus)experienced two major peaks in abun- resident population exists for white seabass, which sup- dance, in 1932 and again in 1982 (Parrish and MacCall ported continuous harvests in the San Francisco area be- 1978; MacCkll et al. 1985; Pacific Fishery Management tween 1880 and 1916 (Skogpberg 1939). Ragen (1990)- Council 1996). The northern anchovy (Erzg~aulismor- summarized by Dayton and MacCall (1992)-developed dax) appears to have been at a very low abundance when a population model that relates niaxiinuiii size of fish the CalCOFI ichthyoplankton surveys began in the early caught each year to the rate of exploitation, and used 1950s, but increased in the mid-1960s and experienced the Tuna Club and California commercial catch data to a brief period of high abundance in the mid-1970s estimate preexploitation abundances of three species. (Methot 1989; Pacific Fishery Management Council Ragen concluded that initial biomasses of white seabass 1996). The abundance of northern anchovy is unclear and yellowtail would have been each about 20,000 tons, before 1954, but MacCall (1986) interpreted ichthy-

105 MACCALL: LOW-FREQUENCY VARIABILITY IN FISH POPULATIONS CalCOFl Rep., Vol. 37, 1996

of these pelagic fishes are strongly separated, so that in recent years no two of these species have simultaneously been highly abundant. The sequence of the four well- documented species alternates between piscivores (bonito and Pacific mackerel) and planktivores (anchovy and sar- dine). An implication of this pattern is that biological interactions may play a substantial role in determining the patterns of peak abundances. Yet the timing of these peaks also appears to be strongly conditional on regime- scale physical fluctuations such as the 1976 transition from cold to warm conditions. Fishing pressure undoubtedly influences these fluc- tuations, but its effect is difficult to isolate. In the ab- 1930 1940 1950 1960 1970 1980 1990 sence of intense sardine fishing, it is likely that the par- YEARS tial sardine recovery, reaching 2.5 million tons (age 2 and Figure 12. Abundances of major pelagic fishes in southern California, with each species scaled to unit maximum. older; Murphy 1966) in 1941, would have produced a much higher abundance, probably exceeding that of 1934 (3.6 million tons). The potential magnitude ofthe 1941 oplankton information from 1939-40 as indicating a peak has been estimated quantitatively by MacCall (1979), moderate level of abundance, somewhat higher than was who simulated historical sardine abundances under al- seen during the early 1950s. Pacific bonito (Sarda chilien- ternative fishng pressures. MacCall’s simulated 1941 peak six) has long been harvested commercially in southern was higher than the 1934 peak in scenarios where the California waters, but recruitment of young-of-the-year simulated fishing intensity was less than about half the was rare before a major increase in southern California actual historical rate. Thus it is reasonable to speculate abundance that began in the mid-1950s (Collins et al. that a valid sardine peak occurred in 1941, nine years 1980; MacCall 1986). after the Pacific mackerel peak in 1932. Thus the early Jack mackerel (Zackuvus symmetricus) abundance has mackerel-sardine sequence is not necessarily inconsis- not been monitored, and the history of fishery landings tent with the sequence seen in the 1980s and 1990s, and has been strongly influenced by the availability of alter- the coincidence of historical peaks in the early 1930s native fishing targets such as sardine, anchovy, and Pacific may have been an artifact of fishery development. mackerel. But length distributions reported by MacCall and Stauffer (1983) suggest that there may have been a Relationship to the Northeastern Pacific recruitment pulse in the mid to late 1940s, at about the The southerly portion of the Cahfornia Current treated same time the southern California fishery suddenly in- in this paper bears a special relationship to the contigu- creased. The evidence for a large jack mackerel popu- ous ecosystems to the north, extending from central lation around 1950 is much weaker than the evidence California to southwestern Alaska. A large symposium for other species. (Beamish 1995) recently addressed climatic effects on Pelagic fish fluctuations appear to be unpredictable fishes in this northern area. Many pelagic fishes (e.g., when they are considered individually, but a pattern sardines, Pacific mackerel, jack mackerel, and Pacific emerges when the time series of historical abundances whiting, Merluccius pvoductus) seem to move to the south- of these fishes are viewed simultaneously (figure 12). In ern portion of the California Current for reproduction order to emphasize the temporal pattern, I have scaled in late winter through early summer. As older adults, abundances to a unit maximum for each species peak these fishes migrate in the summer and fall to northerly (Le., where two widely separated peaks occur, each is waters, especially under conditions of high abundance rescaled separately). Sardine and Pacific mackerel were and warm ocean temperatures. Low-frequency fluctua- simultaneously abundant in the early 1930s; unfortu- tions in these species are more apparent at the northern nately, pre-1950 information on other species is weak and southern ends of the range. or lacking. In the late 1940s and early 1950s no species Sardines were abundant off British Columbia in the appears to have been abundant, with the possible ex- 1930s, when they supported a substantial fishery (Murphy ception ofjack mackerel. 1966). Under intense harvesting and a regime shift to Since the 1950s, peak abundances have occurred se- colder conditions, they subsequently disappeared from quentially at regular intervals of about a decade. Currently, that area for over 40 years. In the 1990s sardines are once Pacific mackerel is declining in southern California, and again becoming abundant from the Columbia River to Pacific sardine continues to increase. Peak abundances British Columbia (Hargreaves et al. 1994).

1 06 MACCALL: LOW-FREQUENCY VARIABILITY IN FISH POPULATIONS CalCOFl Rep., Vol. 37, 1996

It is unclear whether Scorizber was abundant in British 5 Columbia in the 1930s, but in the El Niiio events of 1982-83 and 1992-93 it was common enough in the Sardinops Pacific Northwest to be considered a potentially im- fl portant predator of young salmon (D. Ware, Pacific Biological Station, DFO, Nanainio, pers. comm., 1996). Scomber spp. Brodeur and Ware (1995) showed that jack mackerel were abundant in the southern Gulf of Alaska during 1955-58, but were almost entirely absent when the area was resampled in 1980-89. Salmon (Oncovhynchus spp.) in the northeastern Pacific have shown a strong response to regime shifts. Francis and Hare (1994) show that Alaskan salmon catches dropped abruptly with the shift to a cold regime, which 1890 1910 1930 1950 1970 1990 YEARS they identify as having occurred in 1946, and rose Figure 13. Annual catches of several groups of pelagic fishes in Japanese abruptly with the 1976 shift to a warmer regime. Beamish waters (redrawn from Matsuda et al. 1992). and Bouillon (1995) also examined long-term fluctua- tions in fish catches from the northeastern Pacific, and found major shifts around 1940 and 1976 for both salmon The waters off Japan have supported harvests of sev- and nonsalmon species. They conclude that these fluc- eral small pelagic fish species for centuries. Qualitative tuations in aggregate catch are closely related to low- analysis of historical fish market records kept since 1550 frequency fluctuations in the Aleutian Low. indicates strong low-frequency fluctuations in catches of Archaeological evidence shows that giant bluefin tuna sardine (Surdinups mclanosticta), with major peaks occur- have been taken by artisanal fisheries off Vancouver Island ring every 100 to 120 years (Ito 1961). Reliable pelagic and the Queen Charlotte Islands on many occasions dur- fish catch records exist for most of the twentieth cen- ing the last 5,000 years (Crockford 1994). Although tury, but Japanese fishery statistics are maintained sepa- bluefin tuna have been absent from that area during the rately by many independent agencies, and it is difficult twentieth century, Crockford reports an oral account to obtain comprehensive summary statistics. In a study of artisanal harpoon fishing that must have taken place of alternations in pelagic fish abundances off Japan, during the decade of the 2880s. The tuna were said to Matsuda et al. (1991, 1992) proposed a cyclic dominance have appeared during brief, exceptionally warm periods model (A defeats B, B defeats C, and C defeats A) to in August. Unfortunately, dating of archaeological re- explain the temporal pattern of pelagic fish catches off mains is too imprecise to establish a statistically useful Japan (figure 13).Matsuda et al. identified three pelagic time series, but this pattern is consistent with other ev- fish groups, consisting of chub mackerels (mostly Scomber idence of low-frequency variability. japonicus) , Japanese sardine (Savdinops melanostictus), and a mixed-species group composed of anchovy (Engvaulis Relationships to Other Ecosystems japonica), horse mackerels (Trachurusjaponica and Decapterus Throughout the world, sardine and anchovy stocks muvoadsi), and Pacific saury (Cololabis saira). exhibit low-frequency fluctuations (Lluch-Belda et al. There are remarkable sirmlarities between the sequence 1989), many of which appear to be in phase. A re- of pelagic fish species groups in Japanese waters and the markable hemispheric coincidence of Savdinops spp. fluc- sequence of similar groups off California. In both cases, tuations in Japanese waters and off the west coasts of Savdinops is followed by Engvaulis, then by Scombev japon- both North and South America has been described by icus, and then returns. Matsuda et al. combined Engvaulis Kawasaki (1991). In the Peru-Chile system, a large sar- and Trachuvus into a single group; these two genera may dine (Savdinops sagax) fishery developed following the have formed separate peaks in California, but if they collapse of the anchoveta (Engvaulis ringens) fishery in the were combined, they would occupy the same sequen- early 1970s. In the 1990s there is evidence that the sar- tial position as in Japan. The Japanese system does not dine resource is declining off Peru and Chile, and that appear to have an analog of California’s Pacific bonito, the anchoveta is achieving higher levels of abundance which intervened between California’s tentative Eachuvus and productivity (R. Parrish, PFEG, Pacific Grove, pers. peak and the Engvaulis peak. In California, sequential comm., 1996). Other pelagic fish species in the Peru- peaks in pelagic fish abundance have occurred at near- Chile system have not been monitored closely enough regular intervals of slightly over a decade (figure 12). In to discern their relationships to the fluctuations in sar- Japan, the peaks in catches of pelagic fish groups are dine and anchoveta. less distinct, but appear to occur at intervals of about

107 MACCALL LOW-FREQUENCY VARIABILITY IN FISH POPULATIONS CalCOFl Rep., Vol. 37, 1996

TABLE 1 in doniinaiit pelagic fish species. In contrast, an iiiter- Dates of Peak Abundance for Similar Pelagic Fishes in mediate and prolonged physical regime (such as pre- Japan and California 1940) could result in less temporal separation of domi- Species group Japan California ~~______..__~~____~ nant populations, perhaps consistent with the coincidence si I~lllbri. 1‘132 of high sardine and iiiackerel biomasses around 1930. 1030 1934 (19411) IJtC 1940\; tolllblnrii, 1955-65 S

108 MACCALL: LOW-FREQUENCY VARIABILITY IN FISH POPULATIONS CalCOFl Rep., Vol. 37, 1996

by Jordan and Gilbert and by others around 1880 en- other ecologically similar locations experience low- countered remarkable abundances of barracuda (Sphyraena frequency variability that is incompatible with a simple nqcnfea) , Pacific bonito, and even Spaniqh iiiackerel interpretation of “sustainable developiiient.” However, (Sconzberor?iorzis corrrolor). In thiq century, barracuda and an analogous principle is vitally needed to guide fishing bonito have ddom reached Monterey Bay, and catcheq industries and their management through inevitable of Scowzberoirzcirus have been restricted to southern Baja boom-and-bust cycles. It is first necessary to recognize California. the importance of low-frequency variability or regime behavior of ecosystem in contrast to standard steady- CONCLUSION state assumptions. This step has taken nearly 20 years (cf. The low-frequency fluctuations of pelagic fish pop- Isaacs 197h), but perhaps some of the delay has been due ulations in the southern California region are somewhat to the need for a clear object lesson. That lesson was cyclic, and are related to the alternation of warm and provided by the sharp regime shift in the northeastern cold temperature regimes on periods of 50 to 70 years. Pacific Ocean about 1976. The fluctuations in pelagic fishes are also correlated with We are now beginning to understand this low-fre- fluctuations in other geographical areas such 3s Peru- quency variability and the corresponding biological re- Chile and Japan, a phenonienon cliniatologists call tele- sponses (e.g., Jacobson and MacCall 1995). Fortunately, connections. The remarkable similarity of temperate there appears to be substantial pattern and statistical pre- pelagic fish assemblages in these distant areas indicates a dictability in the long-term physical and biological vari- larger suite of physical and biological similarities that ability of the California Current ecosystem. The scien- favor persistence of these particular species types (Bakun tific problems now faced by CalCOFI are perhaps inore and Parrish 1980). Given general physical and biologi- complicated than was originally envisioned, and are of cal similarities, we should also expect there to be siin- fundamental importance to fields such as climatology, larities in the population dynamics of these species and oceanography, and ecosystem nianagement. It niay be in the interactions among them. From this viewpoint, taken as a sign of institutional and intellectual health that it is reasonable to hypothesize that the worldwide sin- as our view of the California Current ecosystem ma- ilarity of fluctuations in sardine and anchovy stocks tures, the need for and value of CalCOFI monitoring (Lluch-Belda et al. 1989; Kawasaki 1991) may extend and research continues to grow. to the full assemblage of major pelagic fishes in those systems. In the near future, the only hope of testing that ACKNOWLEDGMENTS hypothesis must come &om paleochronologies or from The ideas in this paper have evolved over the last long-term data that have already been collected froiii decade or more, and I would like to acknowledge many other comparable ecosystems; within any single system, helpful people without naming all of them individually. we otherwise have little hope of gaining further insight Several members of the Tuna Club generously helped during our lifetimes. me obtain recent yearbooks. I thank Kenichi Tatsukawa, It is difficult to resist speculating on what the near fu- Yoshiharu Matsuniiya, and the University of Tokyo’s ture may hold. The present warin regime is now 20 years Ocean Research Institute for their gracious support and old, and it has been 56 years since the 1940 shift to a hospitality. Daniel Lluch-Belda, chairiiian of SCOR WG- cold regime (figure 1). The appearance of large late-sea- 98 and director of the Centro de Investigaciones Bio16,’oicas son albacore and the subsequent decline in southern del Norte, has been a continual source of inspiration and California catches in the mid-I 980s resembles events in support. Over the years, Richard l’arrish has contributed the mid-1920s that preceded the regime shift by 10 to many useful thoughts on the behavior of oceans and 15 years. Also, Japanese sardine catches peaked in 1988 ecosystems. Larry Jacobson generously provided useful and have since declined sharply. If the recent pattern of review and coinment on short notice. I ani especially time lags remains consistent (table I), we should expect grateful to Michael Mullin for his patience and for giv- California’s peak sardine abundance about a decade later, ing me the stimulus to finally commit these ideas to around 1998. These biological indicators suggest that a paper. This paper is a product of SCOR WG-98. transition to a cooler regime is likely in the next decade or so. On the basis of spectral patterns, Ware (1995) has boldly predicted that the transition could occur around the year 2001. Unfortunately, we do not know whether it will be intense or moderate. Even with the limited understanding provided by his- torical science, it is clear that the ecosystems and pelagic fish resources of California, Peru-Chile, Japan, and many

109 MACCALL: LOW-FREQUENCY VARIABILITY IN FISH POPULATIONS CalCOFl Rep., Vol. 37, 1996

Uaumgartner, T., A. Soutar, and V. Ferreira-Uartrina. 1992. Reconstruction -. 1986. Changes in the bioniass of the California Current ecosystem. of the history of Pacific sardine and northern anchovy populations over In Variability and management of large niarine ecosystenis, K. Sherman the past hvo millennia from sedinientt of the Sant.i Barbara Uasin, California. and L. M. Alexander, eds. Amer. Assoc. Adv. Sci., Sel. Symp. 99. Boulder: Calif. Coop. Oceanic Fish. Invest. Rep. 3324-40. Westview Press, pp. 33-54. Bayliff, W. H. 1992. Northern bluefin tuna. 117 California’s living marine re- MacCall, A. I]., and M. H. I’ragcr. 1988. Historical changes in abundance sources md their utilization, W. S. Leet, C. M. Dewees, and C. W. Haugen, of six fish species off southern California, based on egg and larva samples. eds. Calif. Sea Grant Extension, Publication UCSGEP-92-12, pp. 138-1 40. Chf Coop. Oceanic Fish. Invest. Rep. 29:91-101. Beamish, R. J., ed. 1995. Climate change and northern fish populat~ons.Can. MacCall, A. Ll., and G. I). Stauffer. 198 Biology and fishery potential of Spec. I’ubl. Fish. Aquat. Sci. 121, 739 pp. jack n~ackerel(Trdiur~r. dlif Coop. Oceanic Fish. Invest. Uearnish, R. J., and L?. R. Bou~llon.1005. Marine fish production trends IXep. 24:46-56. off the Pacific coast of Canada and the United States. Irr Climate change MacCall, A,, I<. A. Klnigbeil, and I<. D. Methot. 198.5. Recent increased and northern fish populations, K. J. Beamish, ed. Can. Spec. Publ. Fish. abundance and potenti.11 productivity of Pacific mackerel (Smrirhcrja~Jo,ii- Aquat. Sci. 121, pp. 585-591. (us). Calif. Coop. Ocemic Fish. Invest. Rep. 21:1 19-129. Brodeur, R. I)., and D. M. Ware. 1 905.Interdecadal variability 111 distribu- Macrate. A. N., Jr. 1948. The history ofthr Tuna Club. Avalon, Calif.: Tuna tion and catch rates of epipelagic nekton in the northeast Pacific Ocean. Club. 197 pp. Irr Climate change and northern fish populations, R. J. Uearnish, ed. Matsuda, H., T. Wada, Y. Takeuchi, and Y. Matsuiiiiya. 1991. Alternative Can. Spec. Publ. Fish. Aquat. Sci. 121, pp. 329-356. models for species replacement of pelagic fishes. Res. Popul. Ecol. 33:41-56. Cole, L). A,,and D. R. McLam. 1989. Interannual variability of tenipera- -, 1992. Model analysis of the effect ofenv~ronmentalfluctuation on ture in the upper layer of the North Pacific eastern boundary region, the species replacement pattern of pelagic fishes under interspecific coni- 1971-1987. U. S. Dep. Comnier., NOAA Tech. Memo., NOAA-TM- petition. Kes. Popiil. Ecol. 34:309-319. NMFS-SWFC-12.5, 20 pp. Mendelrsohn, K., and C. Koy. 1996. Comprehensive ocean dataset extrac- Collins, IX., 1). Huppert, A. MacCall, J. Kadovich, and G. Swuffer. 1980. tion user’s guide. NOAA Tech. Memo., NOAA-TM-NMFS-SWFSG Pacific bonito niansgenient ~nforniationdocument. Calif Dep. Fish Game, 228, 67 pp. Mar. Kes. Tech. Rep. 44, 94 pp. Methot, R. I). 1989 Synthetic estimates of historical abundance and nior- Crockford. S. 1094. New archaeological and ethnogr~pliicevidence of an tality for northern anchovy. Amer. Fish. Soc. Symp. 6:66-82. extinct fishery for giant bluefin tuna (T/iuiiriw t/i)itiirr{s orirrrtdis) 011 the Murphy, G. I. 1966. Population biology of the Pacific sardine (Sardinops Pacific Northwest coast of North America. !,I Fish exploitatioii in the past. rarwrlra). Proc. Calif. Acad. Sci., 4th Ser. 34:1-84. Proceedings of the 7th meeting of the ICAZ Fish liemains Working Odum, W. E., E. P. Oduni, and H. T. Odurii. 1995. Nature’s pulsating par- Group, W. Van Nrer, ed. Annales du MusCe Royal de I’Afrique Centrale, adigm. Estuaries 18:547-555. Sciences Zoologiques 274, pp. 163-168. Pacific Fishery Management Council. 1996. Aniendiiient 7 to the Northern Dayton, 1’. K., and A. D. MacCall. 1992. Pre-exploitation dbundmces of Anchovy Fishery Maiiagenient Plan. Pacific Fishel-); Management Council, important large recreatiotlal and coniniercial fishrs off southern California. 2130 S.W. Fifth Ave., Suite 224, Portland, Ore. 97201. Irr California Sea Grant biennial report of conipleted projects 1988-00. I’arrish, K. H., and A. D. MacCa11. 1978. Climatic variation and exploita- Calif. Sea Grant, Publication K-CSGCI’-033, pp. 91-9t1. tion 111 the Pacific mackerel fishery. Calif. Dept. Fish Game. Fish Bull. Francis, I<.C., and S. R. Hare. 1994. Uecadal-scde regime shifts in the large 167, 110 pp. marine ecosystems of the North-east l’dcific: a case fin historical science. Kagen, T. J. 1990. Pre-exploitation abundances for the white seabass Fish. Oceanogr. 3:279-291. (Amirtosiion mildis), yellowtail (Swida lolorirki), and giant sea bars (Sterrcokypis Gould, S. J. 1989. Wonderful life: the Burgess Shale and the nature of hi\- gigas) off southern California. Irr The estimation of theoretical population tory. New York: W. W. Norton and Co. 347 pp. levels for natural populations. Univ. Calif., San Diego, 1’h.D. diss., pp. Hargreaves, N. B., U. M. Ware, and G. A. McFarlane. 1994. Return of 12-50. Pacific sardine (Snrdiriops sqax) to the Uritirh Columbia coat In 1992. Skogberg, T. 1939. The fishes of the family Sciaenidae (croakers) of Califorma. Can. J. Fish. Aquat. Sci. 51:460-463. Calif. Dept. Fish Game, Fish Uull. .54, 62 pp. Hubbs, C. L. 1948. Changes in the fish fauna ofwestern North America cor- Soutar, A. 1967. The accumulation of fish debris in certain California coastal related with changes in ocean temperature. J. Mar. Res. 7:459-482. sediments. Calif Coop. Oceanic Fish. Invest. Rep. 11:136-139. Isaacs, J. D. 1976. Some ideas and frustrations about fishery SC~CIICC. Calif. Soutar, A., and J. D. Isaacs. 1974. Abundance of pelagic fish during the Coop. Oceanic Fish. Invest. Kep. 18:34-43. 19th and 20th centuries as recorded in anaerobic sedmients of the Califomias. Ito, S. 1961. Fishel-). biology of the sardine, Sadimp r~r/m~~s~im(T &. S), in Fish. Bull., U. S. 72257-273. the waters around Japan B~ill.Jpn. Sea Keg. Fish. Res. Lab. 9:l-227. Trenberth, K. E. 1990. Recent observed interdrcadal climate changes in Jacobson, L. D., and A. D. MacCall. 199.5. Stock-recruitment models for the Northern Hemisphere. Bull. Amer. Meteorol. Soc. 71 :988-993. Pacific sardine (Snrdiliops sqix). Can. J. Fish. Aquat. Sci. 32506-577. Ware, D. M. 1995. A century and a half of change in the climate of the NE Jones, A. T., 1’. Ilutton, and R. E. Snodgrass. 1988. Reoccurrence of the Pacific. Fish. Ckeanogr. 4:267-277. Pacific srahorse, Hippocilrripus irTypu, in San rliego Bay. Calif Flsh Game Watson, W. 1992. I>istribiition of larval Pacific sardine, Sardiriop sqox, In 74236-238. shallow coastal waters between Oceanside and San Onofre, California: Kawasaki, 7.1991. Long-term variability In the pelagic fish populations. 111 1978-1986. Calif. Coop. Oceanic Fish. Invest. Rep. 33:89-99. Long-term variability of pelagic fish populations and their environntent, Wild, A. 1992. Yellowfin and skipjack tunas. III California’s living marine T. Kawasaki, S. Tanaka, Y. Toba, and A. Tanibachi, e&. Oxfbrd: I’erganion resources and their util~ration,W. S. Leet, C. M. Drivers, and C. W. Press, pp. 47-60. Haugen, eds. California Sea Grant Extension, Publication UCSGEP-92- Laurs, K. M., and I<. C. llotson. 1992. Albacore. Jrr California’s living ma- 12, pp. 140-143. rine resources and thcir utilization, w. S. Leet, C. M. De\vees, and (:. w. Wolf, 1’. 1992. Kecovery of the l’aclfic sardine and the California sardine Haugen, ed5. Calif. Sea Grant Extension, I’ublication UCSGEP-92-12, fishery. Calif Coop. Oceanic Fish. Invest. Rep. 33:7(1-86. pp. 236-138. Wooster, W. S., and A. B. Hollowed. 1995. Decadal-scale variations in the Lluch-Uelda, I)., R. J. M. Crawf~ird,T. Kawasaki, A. 1). MacCall, R. H. eastern subarctic Pacific I. Winter ocean conditions. Irr Climate change I’arrish, K. A. Schwartzlose, and 1’. E. Smith. 1989. World-wide fluctu- and northern fish populations, K. J. Beaniish, ed. Can. Spec. I’ubl. Fish. ations of sardine and anchovy stocks: the repie problem. S. Afr. J. Mar. Aquat. SCI. 121, pp. 81-85. SCI.8:195-205. MacCall, A. 1>. 1979. Population estimates for the waning years of the Pacific sardine fishery. Calif Coop. Oceanic Fish. Invest. Rep. 20:72-82.

110