Management application of an empirical model of sardine–climate regime shifts

ARTICLE IN PRESS

Marine Policy 31 (2007) 71–80 www.elsevier.com/locate/marpol

Management application of an empirical model of sardine–climate regime shifts Samuel. F. Herrick Jr.a,, Jerrold. G. Nortonb, Janet E. Masonb, Cindy Besseyc a

National Oceanographic and Atmospheric Administration, National Marine Fisheries Service, Southwest Fisheries Science Center, 8604 La Jolla Shores Drive, La Jolla, CA 92037-1508, USA b Southwest Fisheries Science Center, Pacific Fisheries Environmental Laboratory, 1352 Lighthouse Avenue, Pacific Grove, CA 93950-2097, USA c Joint Institute of Marine and Atmospheric Research, Pacific Fisheries Environmental Laboratory, 1352 Lighthouse Avenue, Pacific Grove, CA 93950-2097, USA

Abstract In previous work, we show that accumulated anomalies of physical indices are proportional to California sardine landings and that the accumulated anomaly curves change the sign of their slope, showing maxima (minima) when climate is favorable (unfavorable) to successful completion of the sardine life cycle change. Here, we find unique time series characteristics of the periods when the climate changed for sardines in the 1930–2004 period. Only one 50–70 year cycle is examined but the consistency of the dominant signals in measurements taken independently within the sardine’s environment at locations separated by thousands of kilometers, supports the argument that the events affecting the ocean-climate of the California Current region and consequently sardine life-cycle are large-scale and persist over multi-decadal periods. Year-to-year monitoring of the climate regime-state in the physical environment and its accumulating effects on sardine populations is also described. The ability to analyze climate shifts and monitor their effects on the sardine populations can reduce uncertainty in making resource management, social and business decisions. Possible effects of management decisions affecting transboundary fisheries issues within United States (US) and between the US and its Pacific neighbors are clarified. The methods presented will add an analysis of low-frequency events to the current management oriented analyses of interannual events, which are part of the existing Pacific Fishery Management Council sardine management plan. r 2006 Elsevier Ltd. All rights reserved. Keywords: Pacific sardine; Sardine–climate regime; Transboundary management

1. Introduction Pacific sardine (Sardinops sagax) supported the largest fishery in the eastern Pacific Ocean during the 1930s and 1940s [1–4]. Sardines were taken along the coast of British Columbia, Washington and Oregon (Pacific Northwest), California, and Mexico, with the bulk of the catch off California. With a biomass estimated between 3 and 4 million metric tons (t) [2], landings along the entire coast peaked at more than 700,000 t during the 1936–1937 season and exceeded 400,000 t in each year from 1937 to 1944. At the time biologists warned that the biomass was incapable of sustaining such large harvests. Nevertheless, there was Corresponding author. Tel.: +1 858 546 7111; fax: +1 858 546 7003.

E-mail address: [email protected] (S.F. Herrick Jr.). 0308-597X/$ - see front matter r 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.marpol.2006.05.005

no mechanism in place to limit catches in accord with what the resource could sustain. Not surprisingly, the fishery began a southward range contraction in the late 1940s when landings ceased in the Pacific Northwest (PNW) and Canada. By the 1960s Pacific sardine landings were extremely low in California, which resulted in a moratorium on its directed fishery, starting in 1967. By the early 1970s incidental sardine landings were less than 10 t [5] and biomass is thought to have dropped below 5000 t [2,6]. Initially, the collapse of the Pacific sardine fishery was viewed as a classic case of overfishing: an overcapitalized industry, employing advanced technology to harvest a fragile resource. More recently, the collapse of the Pacific sardine fishery has been attributed to a combination of overfishing and an unfavorable environment [1,7,8]. When the fishery peaked, ocean waters were beginning to cool, a

ARTICLE IN PRESS 72

S.F. Herrick Jr. et al. / Marine Policy 31 (2007) 71–80

condition associated with lower sardine life-cycle success [6]. This suggests that the collapse was partly due to a change from warm to cold-water regimes off the California coast [9–11]. The general cooling and environmental shift was not immediately evident, because of the natural yearto-year variability of the California Current System. In the late 1970s and 1980s ocean temperature began to rise and improved environmental conditions together with the conservation and management measures that had been put in place by California were contributing to an increase in the harvestable sardine population [2]. Evidence of a vigorously recovering spawning biomass led California to lift its moratorium on Pacific sardine harvesting in 1986, and by the 1990s, extremely favorable environmental conditions were fueling a rapid resurgence in Pacific sardine availability along the west coasts of the US, Mexico and Canada. To avoid the experience of the earlier fishery, the US Pacific Fishery Management Council (PFMC) responded to the situation by instituting a fishery management plan (FMP) for coastal pelagic species [Pacific sardine, Pacific mackerel (Scomber japonicus), jack mackerel (Trachurus symmetricus), northern anchovy (Engraulis mordax), and market squid (Loligo opalescens)] in 1998. A major component of the FMP was a sardine harvest policy that accounts for environmental conditions, the biological status of the resource, and harvesting capacity of the fishing fleets—factors which contributed to the demise of the earlier fishery—in establishing a coastwide (Canada, US, Mexico) acceptable biological catch (ABC).1 The US share of the ABC is based on the average proportion of the total Pacific sardine biomass occurring within the US exclusive economic zone [12]. In this paper, we extend the work of Norton and Mason [3,4,13] relating low-frequency (interdecadal) ocean regime change, on a global and regional scale, to long-term variation in the distribution of the sardine population off the North American west coast. The objective is to identify climatic features which can be quantitatively indexed and incorporated in the US harvest policy as an indicator of a Pacific sardine–climate regime shift, particularly with regard to the distribution of the harvestable population, and its availability to domestic and foreign fisheries. Identification of a sardine–climate regime shift indicator will benefit resource managers and the fishing industry by enhancing their ability to reduce uncertainty, prepare for and manage the risk associated with a sardine–climate regime shift, and in devising appropriate response strategies [14]. Since from a US perspective we appear to be in a favorable (warm water) sardine–climate regime currently, we would anticipate a low-frequency change in climate and the environment to result in an unfavorable (cold water) sardine–climate regime relative to the current situation; i.e., 1

In the coastal pelagic species FMP, a maximum sustainable yield (MSY) control rule is established to calculate the coastwide ABC for Pacific sardine [12].

a reduced sardine population, and its contraction southward. In the case of the domestic Pacific sardine fishery the potential for significant, but imperfectly predictable, sardine–climate regime shifts underscores the need to integrate such knowledge into the domestic harvest policy in an attempt to minimize the disruptions and related economic impacts to the sardine industry, and other sardine interests, of potentially large and long-term adjustments to ABCs, and their geographic or seasonal allocation among US northern and southern subarea fishery sectors [12,15]. Our work provides an indicator of the onset of a sardine population retreat, enabling the PFMC and US fishery sectors to make provident conservation-management and business decisions, resulting in less adverse impacts in terms of both economic efficiency, and industry and community stability. Beyond the domestic concerns, are those relating to the transboundary2 distribution of the resource and how the harvestable population might be shared between the US and Mexico, if the harvestable population contracts southward with sardine–climate regime shifts unfavorable to sardine growth and reproduction. In this regard allocation of the coastwide ABC could be a major issue which would seem to make the case for cooperative international management of Pacific sardine—with built in adjustment mechanisms to account for sardine–climate regime, as well as other anticipated change—all the more pressing [14,16]. In the next section, we investigate the variability of known large-scale oceanic environmental indicators for the 1930–2004 period, and compare these widely spaced measurements to variations in California sardine landings over the same period. This is done to derive a relationship between the physical record and persistence in the size and availability of the sardine resource, and in turn, develop a low-frequency environmental change indicator for a sardine–climate regime shift. We then discuss how the sardine–climate regime shift indicator can be incorporated into the existing harvest control rule to account for lowfrequency environmental effects on the coastwide distribution of the harvestable sardine population. Finally, we develop several options to discuss the implications of early detection and year-to-year monitoring of sardine–climate regime change, and how we might deal with these options from a conservation and management policy standpoint. 2. Methods Variation in Pacific sardine abundance and distribution, as reflected by sardine landings along the coast of North America, is frequently associated with environmental fluctuations that are indexed by the sea surface temperature (SST) obtained daily at the Scripps Institution of 2

Here, we interpret transboundary according to the FAO definition of a transboundary resource: fish stocks that occur within the exclusive economic zones of two or more coastal nations [16].

ARTICLE IN PRESS S.F. Herrick Jr. et al. / Marine Policy 31 (2007) 71–80

Fig. 1. Annual landings (1916–1968) in Oregon, Washington and British Columbia (Pacific Northwest, dashed), California (solid) and Ensenada, Mexico (Baja California, squares) for the early sardine fishery. Annual sea surface temperature (SST) at La Jolla, California is shown in the lower panel. Landings weight is given as the natural log (loge) of the weight in metric tons. Landings were near zero or not available where they are not plotted. Regional fisheries had beginnings and ends during this period, when all the fisheries were regulated by economic and environmental factors independent of management. The Pacific Northwest (PNW) and Ensenada, Baja California fisheries did not occur simultaneously.

Oceanography pier at La Jolla, California (La Jolla SST) (Fig. 1). This shore-based ocean temperature measurement has been shown to be an index for temperature changes in large areas off the west coast of North America [17,18]. Although the rate of sardine egg and larval development is sensitive to sea temperature [19], it is likely that there are a wide variety of environmental conditions, often occurring together with elevated SST, that favor successful completion of the sardine life cycle. Jacobson and MacCall [2] found significant relationships between the natural logarithm (loge) of sardine reproductive success and La Jolla SST averaged over the previous 2–5 years. This relationship is currently incorporated into the maximum sustainable yield (MSY) control rule that establishes the ABC for sardine fisheries off the west coast of the US. In this manner, the control rule contains a biomass safeguard that accounts for changing environmental conditions, via La Jolla SST [12]. The MSY control rule formula [20] is HGyþ1 ¼ ðBy  CÞ  F  D,

73

where HG is the US harvest guideline for year, y+1; B is the estimated biomass of the sardine population at the beginning of the y+1 year; cutoff, C, is the lowest level of estimated biomass at which directed harvest is allowed; F is the fraction of the biomass above cutoff that can be taken by the fishery3; and D is the proportion of total stock biomass in US PNW+CA (California) EEZ waters. PNW represents the fishery north of California. In the current specification of the MSY control rule formula F ranges between 0.05 and 0.15 based on sea temperature, while C and D are currently fixed at 150,000 mt and 87%,4 respectively. Herein we focus on the dynamics of the variable D; how and when it is likely to be affected given the prospect of a sardine–climate regime change. Our methodological approach is to investigate the relationship between sardine landings along the west coast of North America and environmental changes during period 1930–2004, and to investigate what we can learn in terms of being able to foretell the impact of future sardine–climate regime shifts. From a conservation and management standpoint our interest is in how this knowledge can be used to enhance US harvest policy as it pertains to the distribution of the harvestable population of Pacific sardine between fishery sectors within the US EEZ, and between the Canadian, US and Mexican EEZs in response to sardine–climate regime changes. Previous studies [4] found that the loge of the landings weight (W) is equal to a scaling function, f, of the accumulation or integration of a suitable environmental index anomaly (A) and a term, t, representing factors not well indexed by A X  log e ðW Þ ¼ f ðAÞ þ t . (1) When this relationship was tested for the 1930–2000 period (Fig. 2) it was found that the accumulated La Jolla SST anomaly (Eq. (1), righthand side) explained about 75% of the variance in the loge transformed California landings series. It was also found that the anomaly indices of annual sea level pressure at Darwin, Australia (DSLP) and an eastern equatorial Pacific SST (EqSST) index were reasonable proxies for the La Jolla SST anomaly index. These three index series are highly correlated (nominal correlation coefficient, r0 40:8) for temporal patterns of 5–60 years [4]. The following tests are based on these and other results [21,22], which imply that major modes of physical variation in the northeastern temperate Pacific are forced from the equatorial Pacific air–ocean system and that these modes influence the sardine’s environment and consequently influence their reproductive success and harvest availability. 3 The value for F is a proxy for FMSY (the fishing mortality rate that achieves MSY under long-term equilibrium assumptions) and is based on the Jacobson and MacCall [2] relationship between sardine reproductive success and La Jolla SST averaged over the previous 3 years [20]. 4 The value of D is based on aerial fish spotter data [12].

ARTICLE IN PRESS 74

S.F. Herrick Jr. et al. / Marine Policy 31 (2007) 71–80

3. Time series development Anomalies computed with a progressive mean are derived for the three physical data series, " # y X X b ðyÞ ¼ V ðyÞ  V ðiÞ=n , (2) i¼fy

where Xb(y) is the back-anomaly, V(y) is the measured value for year y; n is the number of years averaged (yfy+1). The second term on the right is the progressive mean. Eq. (2) provides the dashed curve in Fig. 2. Primary considerations for devising an analysis scheme are to focus on climate variability and to analyze the changes as quickly as possible. The physical series derived with Eq. (2) show considerable interannual variability, in the sign of the slope of the accumulated anomaly (sign of the anomaly). In an attempt to remove or greatly attenuate the effects of the interannual variability while maintaining early detection, an average slope of the accumulated anomaly curve is computed by averaging the anomaly of the four previous years and the current year, y, Sb ðyÞ ¼ Fig. 2. Sum of the SST anomaly (broken) compared to the loge transformed shore-based sardine landings (solid), illustrating Eq. (1). Changes in the trend or slope of the SST accumulation appears to precede large changes in the slope of the landings curve, then the two series have parallel trends for long periods. The nominal correlation coefficient (r0 ) between the two series is 0.87. SST anomaly is from Eq. (2) and both series are standardized and may be offset for comparison. The landings range from about 500,000 t in the late 1930s and early 1940s to 10 t in the late 1970s and early 1980s.

Eq. (1) incorporates a temperature integration relationship, first noted by Marr [1], and suggests that long lasting anomalous conditions may have similar or greater effects than shorter, extremely anomalous interannual events. Anomalous conditions, persisting longer than 5 years may be considered climate scale [2,4], but the focus of this paper is on climate events of decadal scale during the 1930–2004 interval. The three physical indicators presented below are proxies for large-scale events and are measured independently in areas separated by thousands of kilometers (see below). These events have a spatial scale, which includes the tropical and temperate Pacific Ocean and they have greater than interannual persistence [23,24]. In the analyses, annual means are used to describe unique conditions that indicate decadal scale change in the sign of the slope of the accumulated environmental anomaly indicators during 1930–2000. The slope of the accumulated SST anomaly curve is given by the SST anomaly value. For the physical series annual mean values represent many daily measurements made throughout the year. Annual sardine landings data for the period are from sales receipts completed by dozens of licensed fish dealers for landings received from hundreds of fishers [1,3,25,26].

y X

X b ðiÞ=5,

(3)

i¼y4

where Sb is the mean of the anomaly or the backward looking slope of the accumulated anomaly curve for the current year and the previous 4 years. Series derived in this way are termed, ‘‘S-series’’. Back-anomaly means (Eq. (3)) were derived for the three physical series (Fig. 3). First, the annual La Jolla SST anomaly is derived from monthly mean SST taken at the Scripps Institution of Oceanography Pier in La Jolla, California (32.91N, 117.31W), beginning in 1917 (fy ¼ 1917) and available from shorestation.ucsd.edu. The back-anomaly means or accumulated anomaly slope is the La Jolla S-SST index. Second, EqSST indices were computed from monthly average equatorial SST anomaly for the area defined by 41N–41S, 1501W–901W. This series is also known as the JMA ENSO index and is available at www.coaps.fsu.edu/ products/jma_index.php, where fy ¼ 1870 with reconstruction [27]. Eq. (3) gives S-EqSST. Third, sea level atmospheric pressure at Darwin, Australia (12.41S, 130.91E) is related to El Nin˜o-like physical events in the equatorial and California Current regions [28–32]. Darwin SLP is available from Kousky [33] at (www.cpc.ncep.noaa.gov), where fy ¼ 1882. S-DSLP is from Eq. (3). For comparison, all time series are graphed as standardized variables. Following from the results of Norton and Mason [3,4] that the slope of the accumulated physical anomaly changes sign at apparent sardine–climate shifts, we hypothesize that if the anomaly mean has the same sign for three consecutive annual values and that if this sign is different than the previous interval, it will be an indicator of sardine climate change. This three consecutive S-series

ARTICLE IN PRESS S.F. Herrick Jr. et al. / Marine Policy 31 (2007) 71–80

75

4. Results The 3yS tests show that shifts to predominately negative average slope in the accumulated curves occurred in 1946, 1944, and 1951 for La Jolla S-SST, S-EqSST and S-DSLP, respectively (Fig. 3). In 1946, 5 years after the maximum California catch [5], there is substantial indication that a climate shift took place and the persistence of the negative S-DSLP values in 1951, completes the unanimity of indicators. Peak sardine landings were in the early1940 s, but the late 1940s brought a dwindling of the harvestable population, and consequently reduced landings in the PNW and Canada (Fig. 4, Northern fishery), further indicating a change in sardine–climate regime and confirming the S-series (3yS) analysis. Two of the physical indicators reversed sign by the 3yS test during the long period of negative sardine climate between the late 1940s and the late 1970s, however the accumulated S-series values remained strongly negative during the 1944–1978 period (Fig. 4). Shifts to predominately positive slopes and

Fig. 3. Application of the 3yS test to the three physical S-series. Change from one regime to the next through persistent change in the sign of the averaged anomalies (3yS test) is shown. Comparison of average backslope anomaly, S-series (Eq. (3)) for La Jolla SST, top; EqSST, middle; and DSLP. The sardine series is the between year difference of the loge transformed values so that it will have the same form as the physical series (Eq. (1)). The horizontal line indicates zero anomaly or zero slope in the accumulated anomaly curve. Years (-1900) when climate change is diagnosed, by the 3yS test (see text), are indicated and followed by a sign indicating the climate regime expected to follow. All series have been standardized for comparison. A 5-year mean backslope, which bridges interannual frequency, was chosen to reduce interannual variability while maintaining analysis timeliness. Values of S-series depend only on the values available at the time of computation. The physical time series and the sardine landings series show many of the same features. Because the progressive mean is less stable before 1940 earlier applications of the 3yStest may be less reliable.

(3yS) test uses the current year and the previous 6 years. All other sequences of 3yS test values are taken to indicate persistence of the ongoing climate mode: sardine favorable environments are indicated by positive values and sardine unfavorable environments are indicated by negative values. We test the La Jolla S-SST, S-EqSST and S-DSLP indices (Fig. 3) to find if these 3yS tests indicate unique characteristics of climate change events.

Fig. 4. History of the west coast sardine fishery compared to the history of the physical environment as indexed by accumulated La Jolla S-SST. The vertical is loge landings in metric tons and the horizontal is accumulated La Jolla S-SST. The time sequence, indicated by year (-1900) begins at the upper right showing the largest landings and descends to the lower left (dashed line), then ascends to the upper right (solid line). The fisheries are indicated to end or to begin based on a 1000 mt division. All La Jolla SSST values depend only on the values available at the time of computation. Eq. (1) is illustrated by the relationship between the variables. Dependence of the relationship on initial conditions, management and economic factors is shown by the difference between the ascending and descending curves. When the composite, apparently sigmoid curves are fit by a logistic equation, y ¼ a/(1.0+bexp (-cx))+d, a ¼ 1.21, b ¼ 1.76, c ¼ 2.91 and d ¼ 6.32; and r0 2 ¼ 0.80. The shaded rectangles show loge-transformed biomass estimates for 1933–1965, during the unregulated fishery [2]. The relation of S-SST to biomass is nearly linear in this region (r0 2 ¼ 0.81). Trends of the logetransformed biomass and landings values are similar (r0 2 ¼ 0.59); Norton and Mason [4] give additional comparisons.

ARTICLE IN PRESS 76

S.F. Herrick Jr. et al. / Marine Policy 31 (2007) 71–80

positive sardine climate in all S-series occurred during 1982 (Fig. 3). Using the 3yS test, there are no sardine climate shifts in mean anomaly after 1983. Negative S-series values happened between 1997 and 2004, but these have not triggered persisting negative anomaly according to the 3yS test (Fig. 3). La Jolla S-SST and S-DSLP indices also indicated shorter-term climate transients in the 1960s (Fig. 3). In these cases the La Jolla S-SST and S-DSLP analyze different events that are recorded in all three S-series. Comparison of Figs. 3 and 4 shows that only the 1961 positive transient in the La Jolla S-SST is apparent in the landings (Fig. 4, loop in broken line). The change that took place in the S-DSLP series in 1967 occurred after the California fishery ended in 1965. Overall, the three physical indices show consistent climate changes and each supports the longer-term inferences provided by the others. Three years of back-slope measurements (3yS) of climate change appear adequate to analyze the sardine–climate changes that occurred in the late 1940s and early 1980s. Generally, a 4yS test gives the same results as the 3yS test. However, there are false positive 3yS indications; consequently, it is important to monitor the quality and extent of sardine-favorable environments on a year-to-year basis. The geographical extent and economic potential of harvestable sardine populations can be monitored using the accumulations of the S-series. As the accumulated value decreases (increases) the harvestable populations apparently decrease (increase) as shown by Eq. (1) and illustrated by the history of the California sardine fishery (Fig. 4). An increase in the harvestable sardine population within the US EEZ means that its range will increase northward into the PNW and Canada. In Fig. 4, the decline of the historic fishery and accumulation of a negative La Jolla S-SST anomaly is shown by the broken line, and the fishery rebound and accumulation of a positive anomaly after 1982 is shown by the solid line. Catch was greater than 300,000 t annually in the late 1920s through the early 1940s (Fig. 4, upper right), and then in 1946 the 3yS test on the La Jolla S-SST series indicated a sardine–climate change. The harvestable sardine population started to disappear north of California and by 1950 the fisheries in the PNW and Canada had begun a 50-year hiatus. During 1957–1960 accumulated anomalies increased and a loop is seen in the descending catch curve as positive anomalies replaced negative anomalies in the accumulation (Figs. 3 and 4). Sardine landings increased briefly then declined precipitously as negative anomalies accumulated after a short reversal. In 1965, the California sardine landings fell below 1000 t and the commercial fishery was lost, except for small, mostly incidental catches [34]. Landings reported in California were less than one ton in 1982 and 1983 [5]. However, in 1982 the 3yS test indicated a sardine climate shift to more favorable conditions and a sardine favorable environment spread northward. A fishery out of Ensenada, Mexico began in 1985 with 3722 t and increased to more than 35,000 t in

1995. In 1988 the Southern California catch exceeded 1000 t, and by 1998 had increased to 32,462 t (Fig. 2). The PNW fishery exceeded 1000 t in 2000 and 2001 as positive anomaly values continued to accumulate (Figs. 2 and 4). A reduction in the growth rate of California landings, and a lower level of landings compared to the early fishery, between 1996 and 2004 is partially due to management of the US fishery between 1986 and 2004. The results presented here depend on mean index values and the restraints of anomaly accumulation. If the indicator value is less than (more then) mean, then an unfavorable (favorable) environment is assumed. The mean is an objective regulator of these tests, but other objective regulators might also be used. These will be explored as the study progresses. The accumulation process constrains the endpoints of the accumulated series and makes the interpretation more difficult in these areas, but there is less constraint with the progressive mean used here. Jacobson and McCall [2] scaled the observed recruitment response to the population level by subtracting the loge (spawning biomass) from the loge (recruitment estimates). This subtraction in their model substitutes for the accumulation of environmental suitability or unsuitability summarized by Norton and Mason [4] in Eq. (1). Otherwise the summarizing equations of the two studies are similar. Jacobson and McCall [2] concluded that longterm fluctuations in the environment may cause long-term fluctuations in sardine productivity, and little or no sardine harvest may be sustainable during periods of adverse environmental conditions, such as those of the late 1960s and early 1970s (Fig. 4). Previous work by Norton and Mason [3,4,13] supports this conclusion (Eq. (1)). 5. Discussion Our results suggest an enhanced ability to predict the persistence of a climate change event after it has occurred. This will greatly benefit the PFMC, the sardine fishing industry and fishery dependent communities in their longterm decision-making. Although only one 50–70 year cycle can be realized from the 1930–2004 period, multiple independently measured indices tend to support the conclusions drawn from sardine catch off California and the La Jolla S-SST measurements (Figs. 3 and 4). Periods of high sardine abundance in the 1930s, 1940s, 1990s and early 2000s and lack of abundance in the 1960s and 1970s define a 50–70 year cycle (Figs. 1 and 2) which is similar in length to cycles that appear in fisheries and physical variables throughout the northern hemisphere [3,4,22,35,36]. Sixty year cycles are also a dominant mode of temporal variability in the centuries long sardine scale deposition records from the anoxic sediments of the Santa Barbara basin in the upper Southern California Bight [9]. It is likely that much of the California Current ecosystem is changing composition in 50–70 year temporal modes and sardine abundance and northerly range extension are cycle mode indicators. During the 1930–2004 interval, these

ARTICLE IN PRESS S.F. Herrick Jr. et al. / Marine Policy 31 (2007) 71–80

sardine climate phases have appeared to shift over periods of about 5 years and the shifts have lasted more than 20 years. We have shown that transition periods can be analyzed over 7-year intervals and that a correct analysis may lead to prediction if the change persists as it did in the 1930–2004 period. However, we have detail on only one cycle and have made some inferences about part of the previous cycle [4]. We do not know if all the cycles have beginnings and ends that may be defined in the same way as the 1930–2004 cycle, but we have the possibility of monitoring ongoing sardine–climate regime shifts with the accumulated anomaly and providing a context for the 3yS test and the 3 year analysis that is currently used [12]. The ability to monitor ongoing variability in the harvestable population of Pacific sardine allows us to say something about its sequential geographic availability in response to a sardine–climate regime change. The intent is to incorporate this capability into the US MSY harvest control rule for Pacific sardine so that the D term (proportion of the total stock biomass in the US EEZ) might be made dynamic in this regard. Our analysis, leading to the 3yS test, is an important step towards achieving this objective. From Fig. 4 we derive Fig. 5, which generalizes the full sardine–climate regime cycle. Following the pattern of the historical fishery, a shift to an unfavorable sardine–climate regime is likely to result in a rather abrupt contraction and relocation southward of the harvestable population as negative anomalies persist (down slope in Fig. 5). This is consistent with observations of other coastal pelagic fisheries where rapid changes in fish populations were

Fig. 5. Diagram of expected changes the fraction, D, described in the Harvest Guideline Control Rule. The harvestable population increases (lower left to upper right) and decreases (upper right to lower left, heavy arrows) with climate persistence (accumulating physical anomaly). The vertical lines show the fisheries involved as the harvestable population increases and declines: left, at lowest abundance, is a Mexico (MX) only fishery; middle is a combined MX and California (CA) fishery and the right panel, which includes maximum abundance, accommodates the MX, CA, Pacific Northwest (PNW) and Canadian fisheries. Small arrows indicate periods when management decisions may have significant economic and social results.

77

triggered when ocean temperatures or other environmental variables surpassed certain threshold values [37]. We use Fig. 5 to illustrate several conservation and management options for the distribution term, D, in the MSY harvest control rule as we progress through a full sardine–climate regime cycle. Beginning with the most productive, fully expanded US fisheries (upper right in Fig. 5), and descending to lower yielding US fisheries, the first management decision (arrow) comes when it is found likely that the harvestable population is decreasing and that the PNW and Canadian fisheries will be lost (1946 in Fig. 4). From a US conservation and management standpoint there are a number of options, but the one that is most probable is to decrease D, based on smaller, southward shifting sardine populations in the US EEZ, while at the same time providing the PNW fishery the opportunity to harvest as much sardine as possible under the PFMC’s newly adopted long-term sardine harvest guideline allocation framework.5 This would achieve optimum economic yield by maximizing the value of the harvest guideline6 and allowing for a smooth transition of the PNW sector out of the disappearing fishery. With continued negative anomaly accumulation (right to left in Fig. 5), the next decision arrow on the down slope of the cycle deals with the transition of the harvestable population into Mexico’s EEZ—which corresponds to about 1965 in Fig. 4. The options at this point depend on whether or not there is cooperative transboundary management of the harvestable population between the US and Mexico. Accumulating knowledge concerning the movement of harvestable population centers [38] provide opportunities for international cooperation between the US and Mexico in providing dynamically located, moving refuges for sardine populations that would insure the preservation of seed stocks sufficient to rapidly expand and fill the entire range with harvestable populations when environmental conditions are favorable. Under its current harvest policy in the EEZ, the US would decrease D and invest in the resource, as California did when it declared a moratorium on the directed sardine fishery in 1967. This would allow a greater portion of harvestable population to move into Mexico’s EEZ. By this point (lower right to left arrow in Fig. 5), virtually all the 5 Under Amendment 11 to the PFMC’s Coastal Pelagic Species FMP, 35% of the US Pacific sardine harvest guideline would be released coastwide on January 1; on July 1, 40% of the harvest guideline, plus any unharvested portion from the initial allocation, would be released coastwide; and, on September 15, the remaining 25% of the harvest guideline would be released coastwide plus any unharvested portion from the first reallocation [39]. 6 The Pacific Northwest fishery is built on the availability of freshly harvested and frozen large sardines. California tagging studies [40] and data from the fisheries indicate that the larger fish move farther north destined for high value, highly discriminating export markets for sardines as bait and for human consumption, whereas the California fisheries in recent years have been harvesting smaller, relatively lower-valued sardines [15].

ARTICLE IN PRESS 78

S.F. Herrick Jr. et al. / Marine Policy 31 (2007) 71–80

harvestable population would be in Mexico’s waters and there would be continued depletion of residual sardine populations off southern California through incidental catch, as occurred during the late 1960s and early 1970s. If Mexico cooperates with the US, it may also invest in the resource, so that there is greater assurance that the sardine population will recover and expand northward into the PNW and Canada given a change to favorable sardine– climate regimes. On the other hand, non-cooperation on the part of Mexico might result in a harvest rate in Mexico’s EEZ that is sufficient to retard the harvestable population’s expansion into US waters. If there is no investment in the sardine resource by Mexico, it is likely to result in a long recovery period, after a favorable sardine–climate regime change (left to right in Fig. 5). If the recovery period exceeds the temporal length of the cycle’s sardine-favorable period, then a resource loss to PNW and Canadian fisheries may be substantial. In the hypothetical case of a decrease in the southern California sardine fishery and non-cooperation with Mexico, the US might undertake an aggressive strategy by effectively increasing D and increasing its sardine harvest rate, in a reversal of current policy. This would reduce the harvest population to low levels so that less of the harvestable population enters Mexican waters, and even in the absence of a Mexican fishery, a rapid recovery of the population after a favorable sardine–climate regime change might be less likely. Management decisions related to the transition of the harvestable population centers between the US and Mexico (left arrows in Fig. 5) relate to economic and social impacts as well as to conservation investment. In the absence of cooperative transboundary resource utilization, exhaustive harvest of a sardine population which is moving out of a nations EEZ may result in less immediate social and financial impacts and allow for an easier transition of fishery-dependent citizens into other fisheries or industries. These orderly, phased, changes might not be possible if harvestable populations are being managed from a conservation-only standpoint. When environmental conditions become favorable for an increasing harvestable population, the cycle would start to upslope with increases in weight and latitudinal extent of harvestable populations and D-values would be adjusted to accommodate an expanded harvestable population in the US EEZ (left to right in Fig. 5). With an extreme persistent positive anomaly and increasing harvestable populations, D might be increased to 100% under current US harvest policy. However, it is conceivable that given a resurgence of the resource, Mexico could increase its harvest rate so that less of the harvestable population enters US waters, in which case a formal cooperative conservation and management agreement with Mexico would seemingly be desirable. If the next unfavorable–favorable sardine cycle follows the same pattern as the one observed during 1930–2004 a formal cooperative conservation and management agreement with Canada may not be as pressing, since the

Canadians have only been taking a small fringe of the coast-wide harvestable population. However, if the atmosphere and oceans continue to warm under the influence of anthropogenic forces, then the sardine populations may shift farther north than in the present cycle. It may become possible for Canadian fishers to harvest a substantial fraction of the larger, more fecund, sardines when they migrate into waters off British Columbia in late summer and early fall. If the Canadian harvest is extensive, it might shift conservation investment onto US fishers and make a cooperative conservation and management agreement with Canada more desirable for the US. 6. Conclusions Unlike the past when there was no conservation and management in place to alleviate fishing pressure during a sardine–climate regime change, today the US has a conservation and management plan in place for Pacific sardine which can be adjusted to accommodate sardine– climate regime changes in terms of managing the US fisheries and conserving, or investing in, the sardine resource. On the domestic front, the studies cited in this report suggest that when the harvestable population decreases in the US EEZ, its range to the north is also decreasing and that loss of the Pacific Northwest Fishery must be anticipated, if informed fishery, social and business decisions are to be pursued. By incorporating the lowfrequency climate change signal into the Harvest Control Rule there would be ample forewarning to provide for the anticipated loss of US Pacific sardine fisheries. Reduced uncertainty in this regard for fishery managers and the US sardine industry would preclude ill-timed investment in conservation-management infrastructure and industry expansion. If warning of population decline is timely, more effective and economically efficient conservation and management can be implemented, industry over capitalization can be avoided, and a smoother transition to other fisheries or other industries is possible, lessening adverse welfare impacts on the US public and socioeconomic disruption of fishery dependent communities. On the international front, as a negative climate anomaly persists the largest part of harvestable population may shift to the Mexican EEZ. The harvestable population apparently shifted to the Mexican EEZ in the 1960s then back to the US in the 1980s. If the sardine population centers are moving from the EEZ of one country into the EEZ of another, then exploitive harvesting of the population by the country currently harvesting the population center may reduce the social impacts of eventual fishery reduction and loss. However, given the natural southward progression of the harvestable population under a persistent negative anomaly it would seem desirable for the US to establish cooperative transboundary management of the harvestable sardine populations with Mexico. This would create an opportunity to provide a dynamically located moving refuge for the sardine population that would facilitate

ARTICLE IN PRESS S.F. Herrick Jr. et al. / Marine Policy 31 (2007) 71–80

rapid expansion of the population when environmental conditions are favorable. Under a favorable sardine–climate regime and the harvestable population fully extended into the Canadian EEZ, a cooperative transboundary management agreement with Canada may gain importance. Finally, based on the most recent environmental and fishery conditions preliminary analyses using our 3yS test suggest that positive anomaly values and favorable sardine climate conditions will hold through 2006 and current US harvest rates will remain sustainable.

References [1] Marr JC. The causes of major variations in the catch of Pacific sardine. In: Rosa H, Murphy G, editors. Proceedings of the world scientific meeting on the biology of sardines and related species. Rome: Food and Agriculture Organization of the United Nations; 1960. p. 667–791. [2] Jacobson LD, MacCall AD. Stock-recruitment models for Pacific sardine (Sardinops sagax). Canadian Journal of Fisheries and Aquatic Sciences 1995;52:566–77. [3] Norton JG, Mason JE. Locally and remotely forced environmental influences on California commercial fish and invertebrate landings. California Cooperative Oceanic Fisheries Investigations Reports 2004;45:136–45. [4] Norton JG, Mason JE. Relationship of California sardine (Sardinops sagax) abundance to climate-scale ecological changes in the California Current system. California Cooperative Oceanic Fisheries Investigations Reports 2005;46:83–92. [5] Mason JE. Historical patterns from 74 years of commercial landings from California waters. California Cooperative Oceanic Fisheries Investigations Reports 2004;45:180–90. [6] Barnes JT, Jacobson LD, MacCall AD, Wolf P. Recent population trends and abundance estimates for the Pacific sardine (Sardinops sagax). California Cooperative Oceanic Fisheries Investigations Reports 1992;33:60–75. [7] Uber E, MacCall A. The collapse of California’s sardine fishery. In: Glantz M, Feingold L, editors. Climate variability, climate change and fisheries. Boulder, CO, USA: National Center for Atmospheric Research; 1990. p. 17–23. [8] Wolf P. Recovery of the Pacific sardine and the California sardine fishery. California Cooperative Oceanic Fisheries Investigations Reports 1992;33:76–86. [9] Baumgartner TR, Soutar A, Ferreira-Bartrina V. Reconstruction of the history of Pacific sardine and northern anchovy populations over the past two millennia from sediments of the Santa Barbara Basin, California. California Cooperative Oceanic Fisheries Investigations Reports 1992;33:24–40. [10] Chavez FP, Ryan J, Lluch-Cota SE, Niquen M. From anchovies to sardines and back: multidecadal change in the Pacific Ocean. Science 2003;299:217–21. [11] Lluch-Belda D, Crawford RJM, Kawasaki T, MacCall AD, Parrish RH, Schwartzlose A, et al. World wide fluctuations of sardine and anchovy stocks: the regime problem. South African Journal of Marine Science 1989;8:195–205. [12] PFMC. Amendment 8 (to the northern anchovy fishery management plan) incorporating a name change to: the Coastal Pelagic Species Fishery Management Plan. Portland, OR, USA: Pacific Fishery Management Council; 1998 (335p). [13] Norton JG, Mason JE. Environmental influences on species composition of the commercial harvest of finfish and invertebrates off California. California Cooperative Oceanic Fisheries Investigations Reports 2003;44:123–33.

79

[14] Miller KA, Munro G. Cilmate and cooperation: a new perspective on the management of shared fish stocks. Marine Resource Economics 2004;19:367–93. [15] Herrick SF, Hill K, Reiss C. An optimal harvest policy for the recently renewed United States Pacific sardine fishery. In: Hannesson R, Barange M, Herrick S, editors. Climate change and the economics of the world’s fisheries: examples of small pelagic stocks. Glos, UK: Edward Elgar; 2006. p. 126–50. [16] FAO. Report of the Norway-FAO expert consultation on the management of shared fish stocks, Bergen, Norway, 7–10 October, 2002. FAO Fisheries Report No. 695, Rome. [17] McGowan JA, Cayan DR, Dorman LM. Climate–ocean variability and ecosystem response in the northeast Pacific. Science 1998;281: 210–7. [18] Norton JG. Apparent habitat extensions of dolphinfish (Coryphaena hippurus) in response to climate transients in the California Current. Scientia Marina 1999;63:239–60. [19] Butler JL, Smith PE, Lo NCH. The effect of natural variability of the life-history parameters on anchovy and sardine population growth. California Cooperative Oceanic Fisheries Investigations Report 1993;34:104–11. [20] PFMC. 2005 Status of the Pacific Coast coastal pelagic species fishery and recommended acceptable biological catches. Stock assessment and fishery evaluation—2005. Portland, OR, USA: Pacific Fishery Management Council; 2005 (51p). [21] Mestas-Nun˜ez AM, Enfield DB. Eastern equatorial Pacific SST variability: ENSO and non-ENSO components and their climatic associations. Journal of Climate 2001;14:391–402. [22] Schneider N, Cornuelle BD. The forcing of the Pacific decadal oscillation. Journal of Climate 2005;18:4355–73. [23] McGowan JA. Climate and change in oceanic ecosystems: the value of time-series data. Trends in Ecology and Evolution 1990;5:293–9. [24] Lluch-Belda D, Hernandez-Vazquez S, Lluch-Cota DB, SalinasZavala CA, Schwarzlose RA. The recovery of the California Sardine as related to global change. California Cooperative Oceanic Fisheries Investigations Reports 1992;33:50–9. [25] Radovich J. The collapse of the California sardine fishery: what have we learned? California Cooperative Oceanic Fisheries Investigations Reports 1982;23:56–78. [26] Conser R, Hill K, Crone P, Lo N, Felix-Uraga R. Assessment of the Pacific sardine stock for US management in 2005: Appendix 2 to the Status of the Pacific Coast coastal pelagic species fishery and recommended acceptable biological catches. Stock assessment and fishery evaluation—2005. Portland, OR, USA: Pacific Fishery Management Council; 2005 (149p). [27] Hanley DE, Bourassa ME, O’Brien JJ, Smith SR, Spade ER. A quantitative evaluation of ENSO indexes. Journal of Climate 2002;16:1249–58. [28] Huyer A, Smith RL. The signature of El Nin˜o off Oregon, 1982–1983. Journal of Geophysical Research 1985;90:7133–42. [29] Mysak LA. El Nin˜o, interannual variability and fisheries in the northeast Pacific Ocean. Canadian Journal of Fisheries and Aquatic Sciences 1986;43:464–97. [30] Norton JG, McLain DR. Diagnostic patterns of seasonal and interannual temperature variation off the west coast of the United States: local and remote large scale forcing. Journal of Geophysical Research 1994;9(16):19–30. [31] Alexander MA, Blade´ I, Newman M, Lanzante JR, Lau NG, Scott JD. The atmospheric bridge: the influence of ENSO teleconnections on air–sea interaction over the global oceans. Journal of Climate 2002;15:2205–31. [32] Fu LL, Qui B. Low frequency variability of the North Pacific Ocean: the roles of boundry- and wind-driven baroclinic Rossby waves. Journal of Geophysical Research 2002;107:3220–35. [33] Kousky VE, editor. Climate diagnostic bulletin. Camp Springs, MD, USA: National Oceanic and Atmospheric Administration, National Weather Service, National Centers for Environmental Prediction; 2005 (92p).

ARTICLE IN PRESS 80

S.F. Herrick Jr. et al. / Marine Policy 31 (2007) 71–80

[34] Wolf P, Smith PE, Bergen DR. Pacific sardine. In: Leet WS, Dewees CM, Klingbeil R, Larson EJ, editors. California’s living marine resources: a status report. CA, USA: University of California Agriculture and Natural Resources SG01-11; 2001. p. 299–302. [35] MacCall AD. Patterns of low-frequency variability in the California Current. California Cooperative Oceanic Fisheries Investigations Reports 1996;37:100–10. [36] Klyashtorin LB. Climate change and long-term fluctuations of commercial catches. Fisheries Technical Paper 410. Rome: Food and Agriculture Organization of the United Nations; 2001. 86p. [37] Stenseth NC, Mysterud A, Ottersen G, Hurrell JW, Chan KSik, Lima M. Ecological effects of climate fluctuations. Science 2002;297:1292–6.

[38] Jacobson LD, Bograd SJ, Parrish RH, Mendelssohn R, Schwing FB. An ecosystem-based hypothesis for climatic effects on surplus production in California sardine (Sardinops sagax) and environmental dependent surplus production models. Canadian Journal of Fisheries and Aquatic Sciences 2005;62:1782–96. [39] PFMC. Allocation of the Pacific sardine harvest guideline: Amendment 11 to the Coastal Pelagic Species Fishery Management Plan. Portland, OR, USA: Pacific Fishery Management Council; 2005 (94p). [40] Clark FN, Janssen Jr JF. Movements and abundance of the sardine as measured by tag returns. California Department of Fish and Game: Fisheries Bulletin 1945;61:7–42.