TY - JOUR
AU - Kamykowski, Daniel
AB - Abstract The Atlantic dipole phosphate utilization (ADPU) index, derived through statistical conversion of 20th century Atlantic basin subpolar sea surface temperatures, is used as a fingerprint of Atlantic meridional overturning circulation (AMOC) variability and as an indicator of global Meridional Overturing Circulation (MOC) variability. ADPU index correlations with differences in sea level anomalies (SLAs) between Canada and the UK and across the Isthmus of Panama demonstrate intrabasin and interbasin associations with MOC variability. Cross-correlation analyses of ADPU index, SLAs, and sardine (S) and anchovy (A) catch differences [S −A] (normalized sardine catch minus normalized anchovy catch) confirm strong correlations between ADPU and [S −A] off Japan, California, Peru and Southwest Africa (Benguela). Statistically significant cross correlations also exist between the ADPU index and SLAs for Japan, California, Peru and Benguela, and for SLAs and [S − A] for Japan, California and Peru, but the short time-series lengths compared with the length of the multidecadal cycle limit the interpretation of the observed lead-lags. Though correlation is not causality, the correlation analyses developed here are useful in support of hypothesis generation. The proposed hypothesis to explain the observed small pelagic fishery synchronies asserts: (i) ocean bathymetry and continental distributions interact with multidecadal variations in MOC strength that occur along the conceptual global conveyor belt to generate changes in global oceanic planetary waves and mesoscale eddies that propagate through the world ocean; (ii) each small pelagic fishery region has a unique spatial relationship with pertinent oceanic planetary wave and mesoscale eddy source regions that affect the timing and strength of the waves and eddies that influence the nearby boundary current; (iii) synchronous changes or phasing among global fisheries depend on how and when MOC variability mediated by oceanic planetary waves and mesoscale eddies reaches each fishery region; (iv) oceanic planetary waves and/or mesoscale eddies influence the strength or meandering of the boundary current adjacent to a small pelagic fishery region to change local SLAs and environmental conditions to favour sardine or anchovy populations at different times. Introduction Sardine and anchovy populations off Japan, California, Peru and Southwest Africa (Benguela) exhibit apparently synchronous multidecadal fluctuations (Kawasaki, 1983; Lluch-Belda et al., 1989, 1992; Checkley et al., 2009). California, Peru and Benguela are eastern boundary current (EBC) regimes characterized by coastal upwelling dynamics with offshore extensions due to squirts and jets (MacCall, 2002, 2009; Dottori and Clarke, 2009). Japan is a western boundary current (WBC) regime characterized by current meandering with eddy-related turbulent mixing and frontal upwelling dynamics (Taguchi et al., 2005, 2007; MacCall, 2009; Kaneko et al., 2013). Sardine (anchovy) regimes for California, Peru and Benguela (EBC) generally occur when sea surface temperature (SST) is warmer (cooler) than average and for Japan (WBC) when SST is cooler (warmer) than average (Chavez et al., 2003; Tourre et al., 2007). A flow-related framework suggests that sardines are favoured when planktonic sardine larvae are given time to develop in a weaker EBC or meandering WBC, while anchovies are always coastal and are favoured by increased nearshore productivity, as when EBC coastal upwelling strengthens (MacCall, 2009). Changes in North Pacific EBC and WBC fish populations likely result from large-scale, simple and direct forcing of the ocean basin (Chavez et al., 2003). Proposed interbasin synchronizing mechanisms involve forcing associated with climate variability (Klyashtorin, 1998; Alheit et al., 2009) that propagates through the atmosphere and ocean via complementary teleconnections characterized as atmospheric bridges or oceanic tunnels (Liu and Alexander, 2007; Alheit and Bakun, 2010). Interdecadal variability, driven by stochastic forcing, appears to be associated with changes in the wind-driven, upper-ocean circulation in the Pacific, and with changes in the Atlantic meridional overturning circulation (AMOC) in the North Atlantic (Liu and Alexander, 2007). The time-scale of interdecadal variability likely is determined mainly by Rossby wave propagation in the Pacific and Atlantic extratropics, but the Atlantic time-scale can also be determined by the advection of the returning branch of AMOC (Liu and Alexander, 2007). To more fully examine ecosystem implications of oceanic tunnels associated with AMOC, Kamykowski (2010) calculated an Atlantic Dipole Phosphate Utilization (ADPU) index derived from an SST-based index suggested and justified by Latif et al. (2004, 2006) and Latif and Keenlyside (2010). The phosphate-based ADPU index provided more biological relevance in terms of seasonal patterns and physiological impacts than SST alone, since the latitudinally variable phosphate depletion temperatures (PDTs; Kamykowski and Zentara, 1986), used in the ADPU index calculation, approximated regional seasonal transitions between new nitrate-based and regenerated ammonium-based production regimes (Dugdale and Goering, 1967). Though nitrogen-limitation of oceanic phytoplankton was identified on ecological time-scales, phosphate was preferred here, because world ocean coverage was more complete for mapping purposes (Kamykowski and Zentara, 1986), while a persistent oceanic nitrate:phosphate ∼ 16:1 ratio meant phosphate still provided a nitrate proxy (Weber and Deutsch, 2010). Furthermore, the ADPU index magnified AMOC oscillations through the 20th century to better fit selected ecological time-series (Kamykowski, 2010, 2012). The 20th century ADPU index suggested stronger AMOC centred on 1890, 1935, 1959 and 2007, and weaker AMOC centred on 1918, 1948 and 1984. ADPU index application to the historic Atlantic basin SST record identified an alternating, latitudinal banded pattern of differential surface phosphate utilization (SPU) between times of weak versus strong AMOC. Atlantic ecosystem variability was related to AMOC's temporal and spatial patterns through generalized bottom-up control (Greene and Pershing, 2007; Greene, 2012). Kamykowski (2012) extended these Atlantic basin patterns to the world ocean under the assumption that AMOC variability was symptomatic of simultaneous global Meridional Overturning Circulation (MOC) variability including both thermohaline circulation (THC) and wind-driven circulation variability. The alternating, latitudinal banded SPU pattern observed in the Atlantic (Kamykowski, 2010) also occurred in the Indo-Pacific, suggesting a global coherency, and the AMOC temporal pattern could be related to Atlantic and Pacific sardine or anchovy fisheries synchrony suggesting global ecosystem impacts again through generalized bottom-up controls (Greene and Pershing, 2007; Greene, 2012). Oceanic planetary waves and mesoscale eddies, generated along the conceptual global MOC conveyor belt (Chelton et al., 2007; Huang, 2010) and modulated under strong versus weak AMOC conditions (Senjyu et al., 1999; Dottori and Clarke, 2009), could favour sardines or anchovies at different times (MacCall, 2009). Propagation times from the pertinent oceanic planetary wave and mesoscale eddy source areas for different regional fisheries might set the clock for the phase relationships among the different fisheries. Though caution is required (Woodworth, 2006), sea level variability provides a way to monitor multidecadal basin-scale and regional impacts of MOC variability. Measured local sea level variability responds to tides, atmospheric pressure, wind and buoyancy at different time-scales (Kjerve et al., 1978). Buoyancy forcing is often ignored but can exert significant influence, especially over longer time-scales (Piecuch and Ponte, 2012). Kelvin and Rossby waves contribute to buoyancy-driven, de-trended, decadal sea level patterns (Kjerve et al., 1978; Dottori and Clarke, 2009; Holbrook et al., 2011) that correlate with locally applicable climatic indices (Papadapoulos and Tsimplis, 2006; Zhang and Church, 2012). For example, sea level anomalies (SLAs) are influenced by the North Atlantic (Arctic) Oscillation and Gulf Stream latitudinal position in the Northeast Atlantic; by the North Atlantic (Arctic) Oscillation in the Northwest Atlantic; by the El Niño–Southern Oscillation (ENSO) in the Pacific and Southwest Atlantic; and by the Pacific Decadal Oscillation in the Northeast Pacific (Papadopoulos and Tsimplis, 2006; Zhang and Church, 2012). Low frequency changes in sea surface height are also associated with changes in overturning circulation (Hakkinen, 2001; Ezer et al., 2013) through basin-scale sea surface height adjustments, but future thermal forcing influences due to CO2 increase and ocean warming are more complex (Kienart and Ramstorf, 2012). The present paper revisits Kamykowski's (2012) examination of interbasin small pelagic fishery synchronies with the added consideration of how ADPU-indexed MOC variability influences selected SLAs. First, the ADPU index is compared to SLA variability representative of North Atlantic intrabasin and Isthmus of Panama interbasin ocean dynamics. Then the relationships between the ADPU index and the small pelagic fisheries off Japan, California, Peru and Benguela are reexamined using the regional multidecadal SLAs to monitor baroclinic influences on those regions. Finally, the hypothesis outlined in Kamykowski (2012) is expanded to further consider how regional SLAs and phasing among the four small pelagic fisheries may result from the influence of MOC-related planetary waves and mesoscale eddies on EBC and WBC regimes. Methods ADPU index The AMOC-related ADPU index described in Kamykowski (2010, 2012) served as an indicator (Zhang, 2008) of multidecadal MOC variability through the 20th century. Global, monthly, 2° latitude–longitude resolution SST data between 70°N and 70°S from 1890 to 2007 were downloaded from the Extended Reconstructed SST, version 2 (ERSSTv2; Smith and Reynolds, 2004). Related ERSSTv3 data (Smith et al., 2008) favourably compared with other available 20th century global SST datasets (Yasunaka and Hanawa, 2011). Monthly SST data were entered into the first worksheet of an MSExcel workbook, hereafter identified as the ERSSTv2 workbook. These monthly ERSSTv2 temperatures were converted into monthly surface phosphate concentration ([PO4] for [PO4–3]) according to Kamykowski (2008), based on a global statistical analyses of [PO4] versus temperature scatterplots. The geographically variable PDT were derived as the x-intercepts of linear regressions calculated using all the National Oceanographic Data Center (NODC) [PO4] and temperature profile data (multiseasonal, multidecadal) within the upper kilometre available through the last century in each 10° latitude–longitude quadrangle between 70°N and 70°S (Kamykowski and Zentara, 1986). PDT identified the local temperature above which [PO4] was no longer detectable by traditional colorimetric analyses (Strickland and Parsons, 1968), and thus represented regional seasonal transitions between new and regenerated production regimes (Dugdale and Goering, 1967). These NODC-derived PDT were designated N:PDT. This 10°-resolution matrix was processed using linear interpolation to 2°-resolution to match the ERSSTv2 grid and then entered in the second worksheet in the ERSSTv2 workbook. In order to support estimates of [PO4], upper ocean temperature–phosphate linear regression fits were calculated from the Geochemical Ocean Sections Study (GEOSECS) dataset collected through the 1970s. Separate northern hemisphere and southern hemisphere ocean basin equations were calculated for the GEOSECS slope (G:TPS) versus GEOSECS x-intercept (G:PDT). The hemispheres were treated separately because of the improved non-linear fits compared with the global treatment likely due to the separation of the more oceanic southern hemisphere from the more continental northern hemisphere (Kamykowski, 2008). To better represent longitudinal variability in the x-intercepts than provided by the spatially limited GEOSECS transects, 2° latitude–longitude resolution N:PDT were substituted for G:PDT in the appropriate hemispheric G:TPS versus G:PDT equation. Hybrid G:TPS and N:PDT linear regression relationships for estimating [PO4] as: $$\eqalign{ \hbox{Northern Hemisphere: [PO}_{4}]=\ &(1.3541 \, (1+\hbox{N:PDT})^{- 0.7500})\cr & \times (\hbox{SST} - \hbox{N:PDT})}$$(1) or $$\eqalign{\hbox{Southern Hemisphere: [PO}_{4}]=\ & (3.6562 \, (1+(\hbox{N:PDT})^{- 1.0801})\cr & \times(\hbox{SST} - \hbox{N:PDT})}$$(2) were entered into the same worksheet as the N:PDT to support the calculation of [PO4] from SST at each 2° latitude–longitude gridpoint for January. The N:PDT and the hybrid equation matrices were then vertically copied 11 more times in the second worksheet to match the SST data for February through December in the first worksheet of the ERSSTv2 workbook. Negative nutrient concentrations calculated at SST > N:PDT were assigned a minimum detectable concentration of [PO4] =0.03 µM (Strickland and Parsons, 1968) in a third worksheet in the ERSSTv2 workbook that accommodated 12 months per year. The fourth and fifth worksheets in the ERSSTv2 workbook selected the maximum and minimum surface [PO4] values, respectively, from among the 12-monthly values for each year between 1890 and 2007 at each 2° latitude–longitude gridpoint in the third worksheet. Annual maximum minus minimum differences representing SPU for each 2° latitude–longitude gridpoint were calculated in the sixth worksheet of the ERSSTv2 workbook (Louanchi and Najjar, 2000). SPU was reported as “μM” to emphasize that the annual surface [PO4] range, used as a measure of differential interannual water column fertility, was not a quantitative annual water column [PO4] utilization rate commonly reported as “μM m−2 y−1”. The annual ADPU index was calculated as a ratio of average SPU for the North Atlantic (40–70°N, 65°W–20°E) divided by average SPU for South Atlantic (40–70°S, 65°W–20°E). A sample of a related ERSSTv2 workbook applied to Argo drifter data was provided as supplementary material in Kamykowski (2008). The application of profile-based, climatological temperature–phosphate relationships, the assumed temporal stability of PDT, and the interpretation of SPU were more fully discussed in Kamykowski (2008, 2010, 2012). The segments of the 1890–2007 ADPU index time-series applied here were modified from Kamykowski (2012) due to the variable lengths of regional sea level time-series. Sea level anomalies Monthly sea level data in millimetres were obtained from the World Meteorological Association's Global Sea Level Observing System (http://www.gloss-sealevel.org/) and from the University of Hawaii Sea Level Center (http://uhslc.soest.hawaii.edu/). Gauge sites considered here included four Atlantic stations: Newlyn, United Kingdom (UK) (1920–2009); Halifax, Canada (CAN) (1920–2009); Cristobal, Panama (PAN) (1907–2008); Simons Bay, South Africa (1958–2007) and four Pacific stations: Ayukawa, Japan (1959–2007); La Jolla, California (1950–2007); Balboa, PAN (1907–2008); and Antafagasta2, Chile (1961–2007). Annual sea level averages were calculated from the monthly data. Linear regressions then were calculated from scatterplots of annual sea level versus year for each gauge site. Subtraction of linear regression sea level estimates from the actual data for each year for each site de-trended the time-series from long-term climatic changes and provided annual SLA time-series that were then smoothed using running average to emphasize multidecadal trends. Differences between the de-trended smoothed annual SLAs were calculated for intrabasin [Halifax – Newlyn] and interbasin [Cristobal – Balboa] comparisons. Small pelagic fisheries analysis Time-series of small pelagic fish catch from all countries were compiled between 1950–2007 for all Food and Agriculture Organization (FAO) areas and categories. Small pelagic fish were considered because they were generally categorized as omnivorous microphagists that fed at the base of the foodweb (van der Lingen et al., 2009) and because they exhibited drastic population/biomass fluctuations (Alheit et al., 2009). The present analysis considered Japanese pilchard, Japanese anchovy, California pilchard, California anchovy, South American pilchard, Anchoveta (=Peruvian anchovy), South African pilchard and South African anchovy. These fisheries were referred to here as Japan sardine (JS), Japan anchovy (JA), California sardine (CS), California anchovy (CA), Peru sardine (PS), Peru anchovy (PA), Benguela sardine (BS) and Benguela anchovy (BA) from FAO areas 61 for Japan (J), 77 for California (C), 87 for Peru (P), and 47 for Benguela (B) as in Kamykowski (2012). The great advantages of FAO's catch statistics were global coverage, complete time-series since ∼1950, and regular updating (Watson et al., 2004; FAO, 2005–2013). Kleisner et al. (2012) acknowledged that fish catch statistics were not a silver bullet when it came to evaluating stock status, but that catch did represent the only method of obtaining a global picture of stock status when analysed with an understanding of the scenarios which might cause misinterpretations. Froese et al. (2012) showed that trends in fish catch data were consistent with trends in biomass data of fully assessed stocks. Similarly, Kamykowski (2012) related FAO catch data for the eight small pelagic fisheries considered here to biomass data for these fisheries (Barange et al., 2009). Six of the eight sardine and anchovy fisheries (JS, CA CS, PA, PS, and BS) exhibited significant correlations between catch and biomass (representing temporal coherence), while the other two (JA and BA) exhibited related temporal patterns. The standard Regime Indicator Series (RIS) index (Lluch-Cota et al., 1997): $$\hbox{RIS}=(\hbox{JS} - \hbox{JA})+(\hbox{CS} - \hbox{CA})+(\hbox{PS} - \hbox{PA}) - (\hbox{BS} - \hbox{BA})$$(3) provided a global synthesis of small pelagic fish catch of sardine and anchovy from the four regions considered here based on ∼1925 through 1990 catch data. The RIS index was based on catch values normalized to the maximum catch within the time-series to eliminate sensitivity to absolute sardine and anchovy population abundances. This series was offered as a good indicator of decadal-scale global changes since it emphasized common long-term variability of the stocks versus individual year-to-year variations. In the original RIS formulation, Japan, California and Peru were treated as similarly phased (plus signs), while Benguela was in opposite phase (negative sign). The RIS3 index formulation (Kamykowski, 2012) modified the original RIS index by changing the sign of (CS −CA) from positive to negative based on a reanalysis of the phasing among the four target fisheries in the available FAO catch data 1950 through 2007. The RIS3 index provided the basis for conveniently representing the time-series for the four small pelagic fishery regions as [S −A] (normalized sardine catch minus normalized anchovy catch) differences (Kamykowski, 2012). Software Microsoft Excel was used for data manipulation. FAO FishStat Plus ver. 2.32, universal software for fishery statistical time-series and the associated FAO Capture Production 1950–2009 dataset (FAO, 2011) were used to compile pelagic fisheries time-series. Systat Software's SigmaPlot was used to smooth data (running average with sampling proportion equal to 0.1 that standardized the number of datapoints in each time-series to n =101) to emphasize the multidecadal signal for subsequent regression and other statistical analyses. The Gnu Regression, Econometrics and Time-series Library (Gretl; Cottrell and Lucchetti, 2007) and ClimLab2000 (Tanco and Berri, 2000) were used for time-series analysis including cross correlation. The effective number of datapoints was reduced in each time-series due to autocorrelation following Garrett and Toulany (1981). Table 1 provided the statistical significance applied to the correlations considered next. Table 1. The original number of observations in all cases equal 101 after smoothing. . n* . r p = 0.1 . r p = 0.05 . r p = 0.02 . r p = 0.01 . DCor . p . XCor . p . HSL & NSL 35 0.275 0.325 0.381 0.418 −0.27 ∼0.1 −0.32 =0.05 ADPU & (H-N)SL 36 0.275 0.325 0.381 0.418 0.55 <0.01 0.59 <0.01 CSL & BSL 35 0.275 0.325 0.381 0.418 −0.48 <0.01 −0.49 <0.01 ADPU & (C-B)SL 37 0.275 0.325 0.381 0.418 0.77 <0.01 0.77 <0.01 ADPU & J[S − A] 34 0.275 0.325 0.381 0.418 0.81 <0.01 0.81 <0.01 ADPU & JSLA 36 0.275 0.325 0.381 0.418 −0.01 >0.1 −0.32 =0.05 JSLA & J[S − A] 36 0.275 0.325 0.381 0.418 0.02 >0.1 −0.47 <0.01 ADPU & C[S − A] 34 0.275 0.325 0.381 0.418 −0.54 <0.01 −0.70 <0.01 ADPU & CSLA 36 0.275 0.325 0.381 0.418 0.41 =0.01 0.41 =0.01 CSLA & C[S − A] 36 0.275 0.325 0.381 0.418 0.06 >0.1 0.52 <0.01 ADPU & P[S − A] 34 0.275 0.325 0.381 0.418 0.63 <0.01 0.63 <0.01 ADPU & PSLA 36 0.275 0.325 0.381 0.418 −0.16 >0.1 0.33 <0.05 PSLA & P[S − A] 36 0.275 0.325 0.381 0.418 0.41 =0.01 0.46 <0.01 ADPU & B[S − A] 35 0.275 0.325 0.381 0.418 −0.38 =0.02 −0.39 <0.02 ADPU & BSLA 38 0.275 0.325 0.381 0.418 0.07 >0.1 −0.35 <0.05 BSLA & B[S − A] 42 0.257 0.304 0.358 0.393 0.10 >0.1 0.23 >0.1 . n* . r p = 0.1 . r p = 0.05 . r p = 0.02 . r p = 0.01 . DCor . p . XCor . p . HSL & NSL 35 0.275 0.325 0.381 0.418 −0.27 ∼0.1 −0.32 =0.05 ADPU & (H-N)SL 36 0.275 0.325 0.381 0.418 0.55 <0.01 0.59 <0.01 CSL & BSL 35 0.275 0.325 0.381 0.418 −0.48 <0.01 −0.49 <0.01 ADPU & (C-B)SL 37 0.275 0.325 0.381 0.418 0.77 <0.01 0.77 <0.01 ADPU & J[S − A] 34 0.275 0.325 0.381 0.418 0.81 <0.01 0.81 <0.01 ADPU & JSLA 36 0.275 0.325 0.381 0.418 −0.01 >0.1 −0.32 =0.05 JSLA & J[S − A] 36 0.275 0.325 0.381 0.418 0.02 >0.1 −0.47 <0.01 ADPU & C[S − A] 34 0.275 0.325 0.381 0.418 −0.54 <0.01 −0.70 <0.01 ADPU & CSLA 36 0.275 0.325 0.381 0.418 0.41 =0.01 0.41 =0.01 CSLA & C[S − A] 36 0.275 0.325 0.381 0.418 0.06 >0.1 0.52 <0.01 ADPU & P[S − A] 34 0.275 0.325 0.381 0.418 0.63 <0.01 0.63 <0.01 ADPU & PSLA 36 0.275 0.325 0.381 0.418 −0.16 >0.1 0.33 <0.05 PSLA & P[S − A] 36 0.275 0.325 0.381 0.418 0.41 =0.01 0.46 <0.01 ADPU & B[S − A] 35 0.275 0.325 0.381 0.418 −0.38 =0.02 −0.39 <0.02 ADPU & BSLA 38 0.275 0.325 0.381 0.418 0.07 >0.1 −0.35 <0.05 BSLA & B[S − A] 42 0.257 0.304 0.358 0.393 0.10 >0.1 0.23 >0.1 A corrected number of observations (n*) was calculated for all correlations to compensate for autocorrelation based on the equation (1/n* = 1/n + (2/n2)*(n − 1)*rp(1)) adapted from Garrett and Toulany (1981) using the actual number of observations (n) and the partial autocorrelation function at lag 1 of the product of the two variables minus the mean of the product. In all cases, partial autocorrelation function decreased rapidly to 0 after lag 1. The correlation coefficients (r) significant at p = 0.1, 0.05, 0.02, 0.01 for n* are provided to set the significance level (p) of the direct correlation (DCor) and cross-correlation (XCor) coefficients calculated here (www.gifted.uconn.edu/siegle/research/correlation/corrchrt.htm). Open in new tab Table 1. The original number of observations in all cases equal 101 after smoothing. . n* . r p = 0.1 . r p = 0.05 . r p = 0.02 . r p = 0.01 . DCor . p . XCor . p . HSL & NSL 35 0.275 0.325 0.381 0.418 −0.27 ∼0.1 −0.32 =0.05 ADPU & (H-N)SL 36 0.275 0.325 0.381 0.418 0.55 <0.01 0.59 <0.01 CSL & BSL 35 0.275 0.325 0.381 0.418 −0.48 <0.01 −0.49 <0.01 ADPU & (C-B)SL 37 0.275 0.325 0.381 0.418 0.77 <0.01 0.77 <0.01 ADPU & J[S − A] 34 0.275 0.325 0.381 0.418 0.81 <0.01 0.81 <0.01 ADPU & JSLA 36 0.275 0.325 0.381 0.418 −0.01 >0.1 −0.32 =0.05 JSLA & J[S − A] 36 0.275 0.325 0.381 0.418 0.02 >0.1 −0.47 <0.01 ADPU & C[S − A] 34 0.275 0.325 0.381 0.418 −0.54 <0.01 −0.70 <0.01 ADPU & CSLA 36 0.275 0.325 0.381 0.418 0.41 =0.01 0.41 =0.01 CSLA & C[S − A] 36 0.275 0.325 0.381 0.418 0.06 >0.1 0.52 <0.01 ADPU & P[S − A] 34 0.275 0.325 0.381 0.418 0.63 <0.01 0.63 <0.01 ADPU & PSLA 36 0.275 0.325 0.381 0.418 −0.16 >0.1 0.33 <0.05 PSLA & P[S − A] 36 0.275 0.325 0.381 0.418 0.41 =0.01 0.46 <0.01 ADPU & B[S − A] 35 0.275 0.325 0.381 0.418 −0.38 =0.02 −0.39 <0.02 ADPU & BSLA 38 0.275 0.325 0.381 0.418 0.07 >0.1 −0.35 <0.05 BSLA & B[S − A] 42 0.257 0.304 0.358 0.393 0.10 >0.1 0.23 >0.1 . n* . r p = 0.1 . r p = 0.05 . r p = 0.02 . r p = 0.01 . DCor . p . XCor . p . HSL & NSL 35 0.275 0.325 0.381 0.418 −0.27 ∼0.1 −0.32 =0.05 ADPU & (H-N)SL 36 0.275 0.325 0.381 0.418 0.55 <0.01 0.59 <0.01 CSL & BSL 35 0.275 0.325 0.381 0.418 −0.48 <0.01 −0.49 <0.01 ADPU & (C-B)SL 37 0.275 0.325 0.381 0.418 0.77 <0.01 0.77 <0.01 ADPU & J[S − A] 34 0.275 0.325 0.381 0.418 0.81 <0.01 0.81 <0.01 ADPU & JSLA 36 0.275 0.325 0.381 0.418 −0.01 >0.1 −0.32 =0.05 JSLA & J[S − A] 36 0.275 0.325 0.381 0.418 0.02 >0.1 −0.47 <0.01 ADPU & C[S − A] 34 0.275 0.325 0.381 0.418 −0.54 <0.01 −0.70 <0.01 ADPU & CSLA 36 0.275 0.325 0.381 0.418 0.41 =0.01 0.41 =0.01 CSLA & C[S − A] 36 0.275 0.325 0.381 0.418 0.06 >0.1 0.52 <0.01 ADPU & P[S − A] 34 0.275 0.325 0.381 0.418 0.63 <0.01 0.63 <0.01 ADPU & PSLA 36 0.275 0.325 0.381 0.418 −0.16 >0.1 0.33 <0.05 PSLA & P[S − A] 36 0.275 0.325 0.381 0.418 0.41 =0.01 0.46 <0.01 ADPU & B[S − A] 35 0.275 0.325 0.381 0.418 −0.38 =0.02 −0.39 <0.02 ADPU & BSLA 38 0.275 0.325 0.381 0.418 0.07 >0.1 −0.35 <0.05 BSLA & B[S − A] 42 0.257 0.304 0.358 0.393 0.10 >0.1 0.23 >0.1 A corrected number of observations (n*) was calculated for all correlations to compensate for autocorrelation based on the equation (1/n* = 1/n + (2/n2)*(n − 1)*rp(1)) adapted from Garrett and Toulany (1981) using the actual number of observations (n) and the partial autocorrelation function at lag 1 of the product of the two variables minus the mean of the product. In all cases, partial autocorrelation function decreased rapidly to 0 after lag 1. The correlation coefficients (r) significant at p = 0.1, 0.05, 0.02, 0.01 for n* are provided to set the significance level (p) of the direct correlation (DCor) and cross-correlation (XCor) coefficients calculated here (www.gifted.uconn.edu/siegle/research/correlation/corrchrt.htm). Open in new tab Results Intrabasin and interbasin sea level anomalies Based on direct correlation (DCor, Table 1), smoothed SLAs measured at Halifax, CAN and Newlyn, UK (Figure 1a) are weakly negatively related (r = −0.27, p ∼ 0.1). Nevertheless, the ADPU index, is positively correlated (r =0.55, p < 0.01) with SLA differences calculated as (Halifax −Newlyn) (Figure 1b). Based on direct correlation (DCor, Table 1), smoothed SLAs measured at Cristobal, PAN and Balboa, PAN (Figure 2a) are negatively correlated (r = −0.48, p < 0.01) and the ADPU index is positively correlated (r =0.77, p < 0.01) with SLA differences calculated as (Cristobal −Balboa) (Figure 2b). Cross correlations (XCor, Table 1) for the North Atlantic SLAs improve the correlations with Halifax leading Newlyn (r = −0.32, p =0.05) by 2 years and the ADPU index lagging (Halifax −Newlyn) (r =0.59, p < 0.01) by 4 years. Cross correlations (XCor, Table 1) for Panama SLAs slightly improve the cross isthmus correlation with Crystobal leading Balboa (r = −0.49, p < 0.01) by 4 years but leave the relationship between the ADPU index and (Cristobal −Balboa) unchanged (r =0.77, p < 0.01). Figure 1. Open in new tabDownload slide (a). Plot of the smoothed sea level anomalies (SLAs) for Halifax, CAN (black) and Newlyn, UK (white) from 1920 to 2010. (b). Plot of the smoothed ADPU index (black) and the SLA difference [Halifax – Newlyn] (white) from 1920 to 2010. Figure 2. Open in new tabDownload slide (a). Plot of the smoothed sea level anomalies (SLAs) for Cristobal, PAN (black) and Balboa, PAN (white) from 1907 to 2010. (b). Plot of the smoothed ADPU index (black) and the SLA difference [Cristobal – Balboa] (white) from 1907 to 2010. Sardine and anchovy synchrony re-examined Region-specific relationships among the ADPU index, SLAs and [S − A] exhibit significant direct correlations (Table 2; DCor, Table 1) between ADPU and [S −A] for Japan (positive), California (negative), Peru (positive) and Benguela (negative) and significant positive correlations between SLAs and [S − A] for Peru and between ADPU and SLAs for California. All other direct correlation relationships in Table 2 are not significant (DCor, Table 1). Plots of contemporaneous ADPU index, SLAs and [S −A] off Japan (Figure 3a), Peru (Figure 3c), California (Figure 3b) and Benguela (Figure 3d), constrained by the length of the available SLAs and [S −A] time-series, display the relationships among the time-series that provided the correlations reported in Table 2. The vertical dashed lines during the mid-1980s in each plot separate an increasing trend in the ADPU index (weakening AMOC) starting ∼1960 from a decreasing trend in the ADPU index (strengthening AMOC) continuing through 2007. Visual comparisons of the three time-series for each region in Figure 3 demonstrate that the increasing ADPU index (weakening AMOC) from ∼1960 to the mid-1980s coincides with a tendency for decreasing SLAs and increasing [S −A] off Japan, increasing SLAs and [S −A] off Peru, and decreasing SLAs and [S −A] off California and Benguela. A decreasing ADPU index (strengthening AMOC) after the mid 1980s, coincides with a tendency for increasing SLAs and decreasing [S − A] off Japan, and decreasing SLAs and [S − A] off Peru, and increasing SLAs and [S − A] off California and Benguela. Table 2. Selected correlation coefficients for Japan, California, Peru and South Africa (Benguela) among running average smoothed ADPU, SLA and [S − A]. . ADPU & [S − A] . SLA & [S − A] . ADPU & SLA . JAPAN r = 0.81*, p ≪ 0.01 r = 0.02, p = 0.83 r = −0.01, p = 0.93 CALIFORNIA r = −0.54*, p ≪ 0.01 r = 0.06, p = 0.54 r = 0.41*, p ≪ 0.01 PERU r = 0.63*, p ≪ 0.01 r = 0.41*, p ≪ 0.01 r = −0.16, p = 0.12 BENGUELA r = −0.38*, p ≪ 0.01 r = 0.10, p = 0.34 r = 0.07, p = 0.49 . ADPU & [S − A] . SLA & [S − A] . ADPU & SLA . JAPAN r = 0.81*, p ≪ 0.01 r = 0.02, p = 0.83 r = −0.01, p = 0.93 CALIFORNIA r = −0.54*, p ≪ 0.01 r = 0.06, p = 0.54 r = 0.41*, p ≪ 0.01 PERU r = 0.63*, p ≪ 0.01 r = 0.41*, p ≪ 0.01 r = −0.16, p = 0.12 BENGUELA r = −0.38*, p ≪ 0.01 r = 0.10, p = 0.34 r = 0.07, p = 0.49 According to Table 1, all r > 0.30 in this table are significant at p ≤ 0.05. Those values are marked with asterisks. Open in new tab Table 2. Selected correlation coefficients for Japan, California, Peru and South Africa (Benguela) among running average smoothed ADPU, SLA and [S − A]. . ADPU & [S − A] . SLA & [S − A] . ADPU & SLA . JAPAN r = 0.81*, p ≪ 0.01 r = 0.02, p = 0.83 r = −0.01, p = 0.93 CALIFORNIA r = −0.54*, p ≪ 0.01 r = 0.06, p = 0.54 r = 0.41*, p ≪ 0.01 PERU r = 0.63*, p ≪ 0.01 r = 0.41*, p ≪ 0.01 r = −0.16, p = 0.12 BENGUELA r = −0.38*, p ≪ 0.01 r = 0.10, p = 0.34 r = 0.07, p = 0.49 . ADPU & [S − A] . SLA & [S − A] . ADPU & SLA . JAPAN r = 0.81*, p ≪ 0.01 r = 0.02, p = 0.83 r = −0.01, p = 0.93 CALIFORNIA r = −0.54*, p ≪ 0.01 r = 0.06, p = 0.54 r = 0.41*, p ≪ 0.01 PERU r = 0.63*, p ≪ 0.01 r = 0.41*, p ≪ 0.01 r = −0.16, p = 0.12 BENGUELA r = −0.38*, p ≪ 0.01 r = 0.10, p = 0.34 r = 0.07, p = 0.49 According to Table 1, all r > 0.30 in this table are significant at p ≤ 0.05. Those values are marked with asterisks. Open in new tab The complex, region-specific relationships among the ADPU index, SLAs and [S − A] for Japan, California, Peru and Benguela (Figure 3) are examined further by cross correlation (Table 3). Table 3 (using n =101) lists selected high correlations within ±25 years, the associated lead-lag years for that correlation, and the range of years with correlations generally with p ≤ 0.01 around that correlation. All selected cross correlations in Table 3 are significant (p ≤ 0.05) based on Table 1 (XCor, using n* ≈ 35) except SLAs and [S −A] for Benguela (p > 0.1). For ADPU and [S − A], the ±25-year cross correlations for Japan, California, Peru and Benguela have single peaks (not shown). The ADPU index has a 0 year (−8:8 year) lead-lag for positively correlated Japan, a 5 year (−3:15 year) lag for negatively correlated California, a 1 year (−8:6 years) lead for positively correlated Peru, and a 1 year (−6:1 year) lead for negatively correlated Benguela. The full plots for SLAs and [S − A] ±25-year cross correlations for Japan, California, Peru and Benguela (Figure 4) are complex, often with multiple peaks. Based on the association of JS with colder regional temperature and thus lower regional SLAs and of CS, PS and BS with warmer regional temperature and thus higher regional SLAs (Table 3), SLAs have a 9-year (3–12 year) lag for negatively correlated Japan, an 8-year (−15:−2-year) lead for positively correlated California, a 2-year (−3:6-year) lag for positively correlated Peru, and a 4-year (4:5-year) lag for positively trending Benguela. Regression analysis of [S − A] versus SLA without time adjustment (Figure 5a) for combined Japan, California, Peru and Benguela fisheries has a significant positive slope (r2 =0.005, p ≪ 0.01) with the trend dominated by Peru. Regression analysis of [S − A] versus SLA adjusted for lead-lag offsets (Figure 5b) yields a significant positive slope (r2 =0.37, p ≪ 0.01) for combined California, Peru and Benguela, representing EBC and a significant negative slope (r2 =0.20, p ≪ 0.01) for Japan, representing WBC. The full plots for the ADPU and SLAs ±25-year cross correlations for Japan, California, Peru and Benguela (Figure 6) are also complex, often with multiple peaks suggesting both leads and lags. The relationship between the ADPU index and SLAs is assumed to be negative off WBC Japan and positive relationships off EBC California, Peru and Benguela (Table 3) based on the regional SLAs and related temperature relationships with [S − A] (Figure 5b) and on documented opposing cross-ocean effects (Chavez et al., 2003). The ADPU index consequently has a 6-year (−9; −5-year) lead for negatively correlated Japan, a 0-year (−13:3-year) lead-lag for positively correlated California, an 8-year (−11: −6-year) lead for positively correlated Peru, and a 4-year (2:8 year) lag for positively correlated Benguela. Other relationships between the ADPU index and regional SLAs are possible based on the complex cross-correlation patterns (Figure 6) and the lack of information on the actual leads and lags affecting the different regions. Table 3. The first term is most significant cross-correlation coefficient (r) for Japan, California, Peru and South Africa (Benguela) among running average smoothed ADPU, SLA and [S − A]. . ADPU and [S − A] . SLA and [S − A] . ADPU and SLA . Japan r = 0.81*, 0 year, −8:8 year r = −0.47*, lg 9 year, 3:12 year r = −0.32*, ld 6 year, −9: −5 year California r = −0.70*, lg 5 year, −3:15 year r = 0.52*, ld 8 year, −15: −2 year r = 0.41*, 0 year, −13:3 year Peru r = 0.63*, ld 1 year, −8:6 year r = 0.46*, lg 2 year, −3:6 year r = 0.33*, ld 8 year, −11: −6 year Benguela r = −0.39*, ld 1 year, −6:1 year r = 0.23, lg 4 year, 4:5 year r = 0.39*, lg 4 year, 2:8 year . ADPU and [S − A] . SLA and [S − A] . ADPU and SLA . Japan r = 0.81*, 0 year, −8:8 year r = −0.47*, lg 9 year, 3:12 year r = −0.32*, ld 6 year, −9: −5 year California r = −0.70*, lg 5 year, −3:15 year r = 0.52*, ld 8 year, −15: −2 year r = 0.41*, 0 year, −13:3 year Peru r = 0.63*, ld 1 year, −8:6 year r = 0.46*, lg 2 year, −3:6 year r = 0.33*, ld 8 year, −11: −6 year Benguela r = −0.39*, ld 1 year, −6:1 year r = 0.23, lg 4 year, 4:5 year r = 0.39*, lg 4 year, 2:8 year The second term is the lead (ld) or lag (lg) for the most significant r of the first element in the label (for California ADPU and [S − A]: ADPU lags [S − A] by 5 year). The third term is the range of lead-lag years of r values generally at the p = 0.01 based on the original 101 number of observations after smoothing (Figure 4 and 5). According to Table 1, all r > 0.30 in this table are significant at p = 0.05. Those values are marked with asterisks. Open in new tab Table 3. The first term is most significant cross-correlation coefficient (r) for Japan, California, Peru and South Africa (Benguela) among running average smoothed ADPU, SLA and [S − A]. . ADPU and [S − A] . SLA and [S − A] . ADPU and SLA . Japan r = 0.81*, 0 year, −8:8 year r = −0.47*, lg 9 year, 3:12 year r = −0.32*, ld 6 year, −9: −5 year California r = −0.70*, lg 5 year, −3:15 year r = 0.52*, ld 8 year, −15: −2 year r = 0.41*, 0 year, −13:3 year Peru r = 0.63*, ld 1 year, −8:6 year r = 0.46*, lg 2 year, −3:6 year r = 0.33*, ld 8 year, −11: −6 year Benguela r = −0.39*, ld 1 year, −6:1 year r = 0.23, lg 4 year, 4:5 year r = 0.39*, lg 4 year, 2:8 year . ADPU and [S − A] . SLA and [S − A] . ADPU and SLA . Japan r = 0.81*, 0 year, −8:8 year r = −0.47*, lg 9 year, 3:12 year r = −0.32*, ld 6 year, −9: −5 year California r = −0.70*, lg 5 year, −3:15 year r = 0.52*, ld 8 year, −15: −2 year r = 0.41*, 0 year, −13:3 year Peru r = 0.63*, ld 1 year, −8:6 year r = 0.46*, lg 2 year, −3:6 year r = 0.33*, ld 8 year, −11: −6 year Benguela r = −0.39*, ld 1 year, −6:1 year r = 0.23, lg 4 year, 4:5 year r = 0.39*, lg 4 year, 2:8 year The second term is the lead (ld) or lag (lg) for the most significant r of the first element in the label (for California ADPU and [S − A]: ADPU lags [S − A] by 5 year). The third term is the range of lead-lag years of r values generally at the p = 0.01 based on the original 101 number of observations after smoothing (Figure 4 and 5). According to Table 1, all r > 0.30 in this table are significant at p = 0.05. Those values are marked with asterisks. Open in new tab Figure 3. Open in new tabDownload slide Plots of the smoothed ADPU index (dot line), sea level anomalies (SLAs; grey), and [sardine minus anchovy] ([S − A]; black) for (a). Japan (J), (b). California (C), (c). Peru (P), and (d). South Africa (Benguela) (B). The vertical grey line in each plot divides the ADPU index between weakening AMOC between 1960 and the mid 1980s and strengthening AMOC after the mid 1980s. Figure 4. Open in new tabDownload slide Cross correlation results for sea level anomalies (SLAs) and [sardine minus anchovy] ([S − A]) for (a). Japan (J), (b). California (C), (c). Peru (P) and (d). South Africa (Benguela) (B). SLA leads [S − A] for negative units and lags for positive units. x-axis unit values are J = 0.48, C = 0.57, P = 0.46 and B = 0.44 years due to smoothing. The two horizontal lines in each plot mark the baseline positive or negative correlation levels where p = 0.05 for n = 101. Figure 5. Open in new tabDownload slide Plots of [sardine minus anchovy] ([S − A]) versus sea level anomalies (SLAs) for Japan (J), California (C), Peru (P) and South Africa (Benguela) (B) for (a) direct time-series comparisons of smoothed data as depicted in Figure 3 and column 3 of Table 2, and (b) time-series comparisons after adjustment for SLA cross-correlation lead-lags as given in column 3 of Table 3. The regression lines, confidence intervals and prediction intervals are calculated separately for C, P and B (positive slope) and for J (negative slope). Figure 6. Open in new tabDownload slide Cross-correlation results for ADPU index and sea level anomalies (SLAs) for (a). Japan (J), (b). California (C), (c). Peru (P) and (d). South Africa (Benguela (B)). ADPU index leads SLA for negative units and lags for positive units. x-axis unit values are J = 0.48, C = 0.57, P = 0.46 and B = 0.44 years due to smoothing. The two horizontal lines in each plot mark the baseline positive or negative correlation levels where p = 0.05 for n = 101. Discussion Intrabasin and interbasin sea level anomalies Intrabasin patterns of opposite cross-ocean phases were highlighted in previous considerations of boundary current intensity and North Pacific fisheries (Chavez et al., 2003; Di Lorenzo et al., 2008; Powell and Xu, 2011) and of multidecadal SPU patterns in the Atlantic and Indo-Pacific basins (Kamykowski, 2010; 2012). The weak negative correlation between North Atlantic Halifax, CAN and Newlyn, UK SLAs and the positive correlation of ADPU with the (Halifax −Newlyn) SLA difference further document intrabasin cross-ocean differences. The pattern is consistent with increased penetration of the warmer North Atlantic current with associated higher sea level in the eastern subarctic Atlantic during stronger AMOC (Medhaug and Furevik, 2011). The strong negative correlation of SLAs between Cristobal, PAN on the Atlantic side and Balboa, PAN on the Pacific side of the Isthmus of Panama is consistent with the influence of the northeast trade winds. These winds blow across the isthmus at ∼10°N from January through April (Black et al., 1999; Dong and Sutton, 2005) with an onshore velocity in the Atlantic and an offshore velocity in the Pacific. The SLA difference across the Isthmus of Panama (Cristobal −Balboa), that tends to be positive when the ADPU index is higher (weaker AMOC) and negative when the ADPU index is lower (stronger AMOC), demonstrates an interbasin continuity at multidecadal time-scales that likely is related to AMOC influence on the intertropical convergence zone (ITCZ). During a weaker (stronger) AMOC, the average annual ITCZ position migrates south (north) and the northeast trade winds strengthen (weaken) (Zhang and Delworth, 2005). Encouraged by these intrabasin and interbasin relationships, AMOC-related SLA sensitivity is considered next in three Pacific and one Atlantic small pelagic fishery regions. Sardine and anchovy synchrony re-examined The observed patterns among the ADPU index, SLAs and [S − A] have constraints. The relatively short time-series lengths that start ∼1950 provide only one cycle in a multidecadal oscillation, and a relatively large 2° latitude–longitude grid size is used in the SPU calculation to accommodate global representation. These temporal and spatial constraints limit the statistical strength of the correlation patterns and frustrate detailed speculation on the regional ecological controls of the different small pelagic fisheries, but do support speculation on larger scale boundary current processes. Though extrapolation into the future is possible, changing global conditions due to anthropogenic influences (IPCC AR4 WG1, 2007) and to complex hydrographic and small pelagic fish population responses to those changes (Burrows et al., 2011) limit predictive confidence. In a previous analysis (Kamykowski, 2012), the ADPU index was compared with contemporaneous [S − A] off Japan, California, Peru and Benguela using direct correlation. Despite the shorter time-series used here due to SLA time-series constraints, the ADPU and [S − A] (Table 2) remain significantly positively correlated for Japan and Peru and significantly negatively correlated for California and Benguela. Significant cross correlations between ADPU and [S − A] span 0-year lead-lag for all regions (Table 3) in agreement with the significant direct correlations. Of the small pelagic fisheries considered here, three (California, Peru and Benguela) represent the broader, slower and shallower EBC regime, and the fourth (Japan) represents the narrower, faster and deeper WBC regime (Garrison, 2013). As an example of EBC processes (Dottori and Clarke, 2009), the California Current exhibits low frequency variation off the California coast originating at the equator. Large-scale interannual Rossby waves, mostly generated remotely by equatorial winds, propagate westward from the California coast and influenced sea level, the dynamic height, temperature, geostrophic flow variability, and the salinity beneath the surface layer. Because both the anomalous vertical and alongshore currents are proportional to the time derivative of the interannual sea level, anomalous currents associated with Rossby waves induce temperature fluctuations proportional to the anomalous dynamic height. The alongshore and vertical advections contribute to the temperature fluctuations in the same sense, a higher-than-normal sea level results in downward and poleward displacement of warmer water and a local higher-than-normal temperature. Sardine regimes (Table 4) for California, Peru and Benguela (EBC) generally occur when SST is warmer than average (Tourre et al., 2007; Chavez et al., 2003) and thus when SLA is higher. As an example of WBC processes (Taguchi et al., 2005, 2007), the Kuroshio Current exhibits the effect of trans-Pacific westward Rossby wave propagation on the Kuroshio Current Extension. Rossby wave arrival at the western boundary causes the eastward current to accelerate, leading to enhanced differential advection of subtropical or subarctic origin along the western boundary layer at multidecadal intervals. Kuroshio Current meanders are associated with decadal-scale changes in sea level in the same sense all along the Japanese coast (Senjyu et al., 1999), with higher SLAs associated with higher temperatures (Sasaki and Schneider, 2011). Sardine regimes (Table 4) for Japan (WBC) generally occur when SST is cooler than average (Tourre et al., 2007; Chavez et al., 2003) and thus when SLA is lower than average. Table 4. Hypothetical scenarios for the sardine regimes in the four fisheries discussed in this paper Sardine regime characteristics . Japan . California . Peru . Benguela . Boundary currenta WBC EBC EBC EBC Observed relative temperatureb Colder Warmer Warmer Warmer Observed relative SLAc Lower Higher Higher Higher Expected relative current statea More meandering Weaker Weaker Weaker Expected AMOC strength relationshipc Stronger Stronger Stronger Stronger Observed concurrent AMOC strengthc Weaker Stronger Weaker Stronger Hypothesized reason for discrepancy Planetary Wave Travel Time None Planetary Wave Travel Time None Oceanic planetary wave & meso-eddy source regionsc Bering Straits, ITCZ, Indonesian Throughflow ITCZ, Indonesian Throughflow, Southern Ocean Southern Ocean, Indonesian Throughflow Agulhas Extension, ITCZ, Southern Ocean Oceanic planetary wave & meso-eddy signald,e Years to Decades Years to Decades Years to Decades Years to Decades Sardine regime characteristics . Japan . California . Peru . Benguela . Boundary currenta WBC EBC EBC EBC Observed relative temperatureb Colder Warmer Warmer Warmer Observed relative SLAc Lower Higher Higher Higher Expected relative current statea More meandering Weaker Weaker Weaker Expected AMOC strength relationshipc Stronger Stronger Stronger Stronger Observed concurrent AMOC strengthc Weaker Stronger Weaker Stronger Hypothesized reason for discrepancy Planetary Wave Travel Time None Planetary Wave Travel Time None Oceanic planetary wave & meso-eddy source regionsc Bering Straits, ITCZ, Indonesian Throughflow ITCZ, Indonesian Throughflow, Southern Ocean Southern Ocean, Indonesian Throughflow Agulhas Extension, ITCZ, Southern Ocean Oceanic planetary wave & meso-eddy signald,e Years to Decades Years to Decades Years to Decades Years to Decades aMacCall, 2009. bChavez et al., 2003. cThis paper. dWang, 2002. eChelton et al., 2007. Open in new tab Table 4. Hypothetical scenarios for the sardine regimes in the four fisheries discussed in this paper Sardine regime characteristics . Japan . California . Peru . Benguela . Boundary currenta WBC EBC EBC EBC Observed relative temperatureb Colder Warmer Warmer Warmer Observed relative SLAc Lower Higher Higher Higher Expected relative current statea More meandering Weaker Weaker Weaker Expected AMOC strength relationshipc Stronger Stronger Stronger Stronger Observed concurrent AMOC strengthc Weaker Stronger Weaker Stronger Hypothesized reason for discrepancy Planetary Wave Travel Time None Planetary Wave Travel Time None Oceanic planetary wave & meso-eddy source regionsc Bering Straits, ITCZ, Indonesian Throughflow ITCZ, Indonesian Throughflow, Southern Ocean Southern Ocean, Indonesian Throughflow Agulhas Extension, ITCZ, Southern Ocean Oceanic planetary wave & meso-eddy signald,e Years to Decades Years to Decades Years to Decades Years to Decades Sardine regime characteristics . Japan . California . Peru . Benguela . Boundary currenta WBC EBC EBC EBC Observed relative temperatureb Colder Warmer Warmer Warmer Observed relative SLAc Lower Higher Higher Higher Expected relative current statea More meandering Weaker Weaker Weaker Expected AMOC strength relationshipc Stronger Stronger Stronger Stronger Observed concurrent AMOC strengthc Weaker Stronger Weaker Stronger Hypothesized reason for discrepancy Planetary Wave Travel Time None Planetary Wave Travel Time None Oceanic planetary wave & meso-eddy source regionsc Bering Straits, ITCZ, Indonesian Throughflow ITCZ, Indonesian Throughflow, Southern Ocean Southern Ocean, Indonesian Throughflow Agulhas Extension, ITCZ, Southern Ocean Oceanic planetary wave & meso-eddy signald,e Years to Decades Years to Decades Years to Decades Years to Decades aMacCall, 2009. bChavez et al., 2003. cThis paper. dWang, 2002. eChelton et al., 2007. Open in new tab With these EBC and WBC characterizations in mind, and assuming that any change in the ADPU index causes an SLA response, SLA lags the ADPU index for the three Pacific small pelagic fishery regions with a peak range of 0–8 years but a span of 16 years and leads the ADPU index for the one Atlantic small pelagic fishery region with a peak range of 4 years but a span of 6 years (Table 3). These regional ADPU and SLA lead-lags may have significance in terms of how much time is required for the MOC signal mediated by Kelvin waves, Rossby waves or mesoscale eddies to reach a small pelagic fisheries region, but the single expression of a multidecadal cycle frustrates rigorous interpretation. Higher than average SLAs associated with relatively warmer regional conditions lead sardine population peaks off California by an 8-year (−15: −2-year) lag, off Peru by 2 years (−3:6 years), and tend to lag off Benguela by 4 years (4:5 years) (Table 3). Lower than average SLAs associated with relatively colder regional conditions lag sardine population peaks off Japan by 9 years (3:12 years) (Table 3). These lead-lags may have regional significance in terms of what forcing and how much time is required to set up sardine-favourable conditions and how long it takes for the fish populations to respond, but again the single expression of a multidecadal cycle frustrates rigorous interpretation. Combining the three EBC regions, a scatter plot of [S − A] versus SLA adjusted for lead-lags has a positive linear regression slope. The [S − A] versus SLA scatterplot for Japan, a WBC region, has a negative linear regression slope. Again, the specific regional lead-lags between SLAs and [S − A] may contain information on the development of favourable environmental conditions for the growth of either sardine or anchovy populations with appropriate development times (Herrick et al., 2007). Nevertheless, the available time-series are too short and the available 2° latitude–longitude grid size is too coarse to discern multidecadal regional environmental differences with confidence. Hypothesis generation based on the previous observations The patterns presented here are correlation, not causality. Correlations in fishery science generally have a history of low retest verification (Myers, 1998). Nevertheless, correlation studies have a useful place in support of hypothesis generation. Though less than ideal, the time-series patterns and the correlation and regression analyses presented here support speculation on the ocean processes that may contribute to the observed phase relationships among the small pelagic fisheries. The speculation begins with previously stated trends. Sardine regimes (Table 4) for California, Peru and Benguela (EBC) generally occur when SST is warmer than average and for Japan (WBC) when SST is cooler than average (Chavez et al., 2003; Tourre et al., 2007). SLAs higher than average are associated with higher temperatures both off California (Dottori and Clark, 2009) and off Japan (Sasaki and Schneider, 2011). These relationships exist in the context that flow (MacCall, 2009) not temperature may be a unifying feature controlling the phasing among the four small pelagic fishery regions. Weaker EBCs may result from stronger poleward coastal Kelvin waves and complementary westward Rossby waves or mesoscale eddies for California, Peru and Benguela, and greater WBC meanders that entrain Oyashio water for Japan may result from stronger westward-propagating Rossby waves (Taguchi et al., 2005, 2007; Dottori and Clarke, 2009). In an ideal, instantly responsive ocean, the sardine regimes in each area could be expected to be directly related to stronger MOC due to more energetic opposing oceanic planetary waves and mesoscale eddies that interfere with current dynamics. Though this expected relationship might generally have held in the present dataset for California and Benguela since 1950, it is reversed for Japan and Peru (Figure 3). The hypothesis presented here proposes that Japan and Peru receive a delayed signal associated with MOC variability based on their greater distances from influential MOC-related oceanic planetary wave and mesoscale eddy source regions (Figure 7a). Source regions are specific geographic locations where MOC variability contributes to changes in atmospheric or oceanic circulation and complementary changes in the oceanic planetary waves or mesoscale eddies. An atmospheric example is a different migration pattern for the ITCZ, while an oceanic example is a different current velocity over a bathymetric feature or through a continental constriction. Eastward propagating equatorial Kelvin waves and coastal Kelvin waves that propagate counterclockwise in the northern hemisphere and clockwise in the southern hemisphere travel at about 2 m s−1 (Wang, 2002). Westward-propagating Rossby waves travel about a third as fast near the equator but quickly decrease in speed to a few cm s−1 at 20° latitude (Chelton et al., 2007). At >20° latitude, nonlinear mesoscale eddies propagate nearly due west at about the phase speed of nondispersive baroclinic Rossby waves with slight poleward (equatorward) deflection of cyclonic (anticyclonic) eddies (Chelton et al., 2007). Figure 7. Open in new tabDownload slide (a). The ocean background shows bathymetry at 2900–3100 m depth (small squares). Conceptual plot of proposed Kelvin waves (solid arrow lines), Rossby waves (<20° latitude; dotted arrow lines) and mesoscale eddies (>20° latitude; dotted arrow lines; cyclones move poleward and anticyclones equatorward). Also depicted are representative source regions [deep water formation zones (darker oval “O”) and deep MOC limb bathymetry interactions (dark oval “B”); ITCZ (lighter oval “I”) and surface MOC limb chokepoint interactions (lighter oval “S”)] responsive to MOC variability that can impact sea level and fisheries in the Japan, California, Peru and South Africa (Benguela) fisheries regions (fish). (b). The numbers in the rectangles refer to representative references [1) Bertrand et al., 2008; 2) Biastoch et al., 2008; 3) Dottori and Clarke, 2009; 4) Drushka et al., 2010; 5) Garzoli and Bianchi, 1987; 6) Huang, 2010; 7) Ivchenko et al., 2006; 8) Johnson and Marshall, 2004; 9) Rao et al., 2010; 10) Sugimatsu and Isobe, 2010; 11) Taguchi et al., 2007; 12) Timmerman et al., 2005; 13) Valsala et al., 2010; 14) Zhang, 2007] that discuss oceanic planetary waves and mesoscale eddies in the different world ocean regions. Huang (2010) provides an overview (top left) of oceanic planetary wave propagation induced by deep-water formation in the world ocean in his Figure 5.187. Japan, California, Peru and South Africa (Benguela) fisheries are marked by fish. Information from these references contributed to Figure 7a. Using adjoint modelling, Heimbach et al. (2011) found that the time-scales of influence on the subtropical North Atlantic were short (months) from local disturbances but extended back to a decade and longer as the region of influence expanded to occupy much of the Atlantic basin and significant areas of the global ocean. The present situation is complicated by proposed multiple oceanic planetary wave and mesoscale eddy source regions associated with poorly defined MOC-derived variability. One version of the proposed hypothesis is that regional SLAs are responding to oceanic planetary waves and mesoscale eddies simultaneously generated by perfectly phased global MOC variability with simultaneous weakening or strengthening throughout the world ocean. MOC variability may be in phase in the AMOC region since North Brazil Current variability is related to the thickness of deep Labrador Sea water (Kieke et al., 2009; Zhang et al., 2011) and with the Agulhas leakage (Biastoch et al., 2008). On the other hand, interocean phasing differences in MOC strength are probable because of the complexity of the global MOC (Schmitz, 1995) and the reported 20th century bipolar seesaw between the Arctic and Antarctic (Chylek et al., 2010). In the latter case, the oceanic planetary wave and mesoscale eddy signals impacting different parts of the world ocean may be much more complex due to the MOC simultaneously weakening in some areas, while strengthening in other areas based on regional circulation loops. Either way, the variable MOC may influence oceanic planetary wave and mesoscale eddy generation (Figure 7a) at: (i) the ITCZ migration zone in the eastern tropical Pacific (lighter oval “I”; Flores-Morales et al., 2012); (ii) the MOC deep limb origination zones in the North Atlantic and Southern Oceans (darker oval “O”; Huang, 2010), (iii) MOC deep limb (darker oval “B”; Huang, 2010) and surface limb (darker oval “B”; Boyer et al., 1993) perturbation zones related to deep currents and complex ocean bathymetry, and (iv) continental chokepoints like the Bering Straits, the Indonesian Throughflow, the southern tip of Africa, and the Drake Passage (lighter oval “S”; van Aken, 2007; Huang, 2010; Garzoli and Matano, 2011). Though ITCZ influences and oceanic tunnels are emphasized here (Figure 7a), other atmospheric bridges apparently also transmit MOC variability globally (Liu and Alexander, 2007) and thus likely would add complexity to a more comprehensive source region map. Future longer time-series of the instantaneous AMOC state as measured by the ADPU index may show positive or negative relationships with SLAs in a given region depending on: (i) the global pattern of how MOC strength actually varies relative to AMOC strength; (ii) proximity to the AMOC-sensitive ITCZ or to one or more MOC ocean bathymetry or continental chokepoint source areas for oceanic planetary wave and mesoscale eddy generation (Figure 7a), and (iii) propagation velocities and residual energies from these planetary wave-source areas for equatorial and coastal Kelvin waves and for latitude-dependent Rossby waves and mesoscale eddies. For example, the balance and timing of multidecadal physical forcing of California waters (Figure 7a; Table 4) may be variously influenced by (i) more proximate ITCZ-related coastal Kelvin waves (Flores-Morales et al., 2012), possibly restricted to the northern hemisphere due to an equatorial filter (Johnson and Marshall, 2004), (ii) more distant eastward equatorial Kelvin waves originating in the Indian Ocean that excite coastal Kelvin waves (coast on right in the northern hemisphere) off California (Huang, 2010), or (iii) the even more distant sequence of westward Rossby waves or mesoscale eddies originating in the Southern Ocean (Ivchenko et al., 2006) exciting in sequence coastal Kelvin waves off Australia (coast on left in the southern hemisphere), eastward equatorial Kelvin waves, and coastal Kelvin waves off California (Huang, 2010). The available post-1950 data suggest that California and Benguela require years (Table 4: underlined) but Japan and Peru require decades (Table 4: underlined) to be influenced by the instantaneous MOC state. Changing influence from different wave and eddy source regions through a century may contribute to the way that the four fishery phase with each other in different multidecadal cycles as suggested by the difference between the RIS (∼1925–1990) and RIS 3 (1950–2007) formulations based on different time-spans. Representative literature (Figure 7b) demonstrates that oceanic planetary wave and mesoscale eddy activity are implicated in global ocean variability. Once the oceanic planetary wave or mesoscale eddy signals arrive, several precedents for impacts on water column structure and regional ecosystems are available. ENSO variability on ecosystems were associated with eastward equatorial Kelvin wave and westward Rossby wave propagation in the equatorial Pacific (Wang and Fiedler, 2006). Coastally trapped Kelvin waves were identified as a main driver of climatic variability off western South America and were implicated in the spatial reorganizations of living organisms including Peru anchovy (Bertrand et al., 2008). Off California, the logarithm of the zooplankton population averaged over the region lagged coastal SLA by ∼2 months, and this pattern was consistent with the westward Rossby wave propagation (Clarke and Dottori, 2007). Mid-ocean mesocale eddies were implicated in oceanic chlorophyll patterns (Chelton et al., 2011) and higher trophic level aggregations (Godø et al., 2012). The full exposition of the influence of MOC variability on small pelagic fish synchrony in the world ocean may best be approached with an appropriately formulated atmosphere–ocean general circulation model that can integrate the complexities of MOC global variability and of oceanic planetary wave and mesoscale eddy generation and propagation on time- and space-scales applicable to the multidecadal cycling of regional small pelagic fisheries. Conclusions Analysis of the relationships between the ADPU index and SLAs in the North Atlantic and across the Isthmus of Panama supports intrabasin and interbasin scale influences of AMOC-related MOC variability. Analysis of the relationships between the ADPU index, SLAs and [S − A] off Japan, California, Peru and Benguela, suggests a multidecadal effect of MOC variability on regional SLAs sand small pelagic fisheries. The mechanisms contributing to the patterns among the ADPU index, SLAs and [S − A] require more information on the exact timing and strength of multiple AMOC-related atmospheric bridges and oceanic tunnels that propagate the signal of MOC variability to Japan, California, Peru and Benguela. Based on these observations, a hypothesis is proposed to explain the observed small pelagic fishery synchronies: (i) ocean bathymetry and continental distributions interact with multidecadal variations in MOC strength that occur along the conceptual global conveyor belt to generate changes in global oceanic planetary waves and mesoscale eddies that propagate through the world ocean; (ii) each small pelagic fishery region has a unique spatial relationship with pertinent oceanic planetary wave and mesoscale eddy-source regions that affect the timing and strength of the waves and eddies that influence the nearby boundary current; (iii) synchronous changes or phasing among global fisheries depend on how and when MOC variability mediated by oceanic planetary waves and mesoscale eddies reaches each fishery region; (iv) oceanic planetary waves and/or mesoscale eddies influence the strength or meandering of the boundary current adjacent to a small pelagic fishery region to change local SLAs and environmental conditions to favour sardine or anchovy populations at different times. Funding This material is based in part upon work supported by the National Science Foundation under Grant Number OCE-0726271 (PI: D. Kamykowski). Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author and do not necessarily reflect the views of the National Science Foundation. Acknowledgements I thank S-J. Zentara for contributions to the NODC-based global PDT and comments on the manuscript. I also particularly thank colleagues for public access to the ERSSTv2 dataset, ClimLab 2000, Gretl, FAO software and data, and reviewers for comments that helped with the evolving conceptualization of the issues involved. References Alheit J. , Bakun A. . Population synchronies within and between ocean basins: apparent teleconnections and implications as to physical–biological linkage mechanisms , Journal of Marine Systems , 2010 , vol. 79 (pg. 267 - 285 ) Google Scholar Crossref Search ADS WorldCat Alheit M. , Roy C. , Kifani S. . Checkley D. , Alheit J. , Oozeki Y. , Roy C. . Decadal-scale variability in populations , In Climate Change and Small Pelagic Fisheries , 2009 Cambridge Cambridge University Press (pg. 64 - 87 ) 372 Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Barange M. , Bernal M. , Cergole M. C. , Cubillos L. A. , Daskalov G. M. , de Moor C. L. , De Oliveira J. A. A. , et al. Checkley D. , Alheit J. , Oozeki Y. , Roy C. . Current trends in assessment and management of stocks , In Climate Change and Small Pelagic Fisheries , 2009 Cambridge Cambridge University Press (pg. 191 - 255 ) 372 Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Bertrand S. , Dewitte B. , Tamc J. , Diaz E. , Bertrand A. . Impacts of Kelvin wave forcing in the Peru Humboldt current system: scenarios of spatial reorganizations from physics to fishers , Progress in Oceanography , 2008 , vol. 79 (pg. 278 - 289 ) Google Scholar Crossref Search ADS WorldCat Biastoch A. , Boning C. W. , Lutjeharms J. R. E. . Agulhas leakage dynamics affects decadal variability in Atlantic overturning circulation , Nature , 2008 , vol. 456 (pg. 489 - 492 ) Google Scholar Crossref Search ADS PubMed WorldCat Black D. E. , Peterson L. C. , Overpeck J. T. , Kaplan A. , Evans M. N. , Kashgarian M. . Eight centuries of North Atlantic ocean atmosphere variability , Science , 1999 , vol. 286 (pg. 1709 - 1713 ) Google Scholar Crossref Search ADS PubMed WorldCat Boyer D. L. , Chen R-R. , Tao L. , Davies P. A. . Physical model of bathymetric effects on the Antarctic circumpolar current , Journal of Geophysical Research , 1993 , vol. 98 (pg. 2587 - 2608 ) Google Scholar Crossref Search ADS WorldCat Burrows M. T , Schoeman D. S. , Buckley L. B. , Moore P. , Poloczanska E. S. , Brander K. M. , Brown C. , et al. The pace of shifting climate in marine and terrestrial ecosystems , Science , 2011 , vol. 334 (pg. 652 - 655 ) Google Scholar Crossref Search ADS PubMed WorldCat Chavez F. P. , Ryan J. , Lluch-Cota S. E. , Niquen C. M. . From anchovies to sardines and back: multidecadal change in the Pacific Ocean , Science , 2003 , vol. 299 (pg. 217 - 221 ) 10.1126/science.1075880 Google Scholar Crossref Search ADS PubMed WorldCat Checkley D. M. Jr , Ayon P. , Baumgartner T. R. , Bernal M. , Coetzee J. C. , Emmett R. , Guevara-Carrasco R. , et al. Checkley D. , Alheit J. , Oozeki Y. , Roy C. . Habitats , In Climate Change and Small Pelagic Fisheries , 2009 Cambridge Cambridge University Press (pg. 12 - 44 ) 372 Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Chelton D. B. , Gaube P. , Schlax M. G. , Early J. J. , Samelson R. M. . The influence of nonlinear mesoscale eddies on near-surface oceanic chlorophyll , Science , 2011 , vol. 334 pg. 328 10.1126/science.1208897 Google Scholar Crossref Search ADS PubMed Google Preview WorldCat COPAC Chelton D. B. , Schlax M. G. , Samelson R. M. , de Szoeke R. A. . Global observations of large oceanic eddies , Geophysical Research Letters , 2007 , vol. 34 L15606 2 10.1029/2007GL03081 Google Scholar OpenURL Placeholder Text WorldCat Crossref Chylek P. , Folland C. K. , Lesins G. , Dubey M. K. . Twentieth century bipolar seesaw of the Arctic and Antarctic surface air temperatures , Geophysical Research Letters , 2010 , vol. 37 pg. L08703 10.1029/2010GL042793 Google Scholar Crossref Search ADS WorldCat Clarke A. J. , Dottori M. . Planetary wave propagation off California and its effect on zooplankton , Journal of Physical Oceanography , 2007 , vol. 38 (pg. 702 - 714 ) Google Scholar Crossref Search ADS WorldCat Cottrell A. , Lucchetti R. J. . Department of Economics and Wake Forest University and Dipartimento di Economia Universit_a Politecnica delle Marche , Gretl User's Guide , 2007 http://ricardo.ecn.wfu.edu/pub//gretl/manual/PDF/gretl-guide.pdf (Last accessed 7 October 2013) Google Scholar OpenURL Placeholder Text WorldCat Di Lorenzo N. , Schneider N. , Cobb K. M. , Franks P. J. S. , Chhak K. , Miller A. J. , McWilliams J. C. , et al. North Pacific gyre oscillation links ocean climate and ecosystem change , Geophysical Research Letters , 2008 , vol. 35 10.1029/2007GL032838 Google Scholar OpenURL Placeholder Text WorldCat Crossref Dong B. , Sutton R. T. . Mechanism of interdecadal thermohaline circulation variability in a coupled ocean–atmosphere GCM , Journal of Climate , 2005 , vol. 18 (pg. 1117 - 1135 ) Google Scholar Crossref Search ADS WorldCat Dottori M. , Clarke A. J. . Rossby waves and the interannual and interdecadal variability of temperature and salinity off California , Journal of Physical Oceanography , 2009 , vol. 39 (pg. 2543 - 2561 ) Google Scholar Crossref Search ADS WorldCat Drushka K. , Sprintall J. , Gille S. T. , Brodjonegoro I. . Vertical structure of Kelvin waves in the Indonesian throughflow exit passages , Journal of Physical Oceanography , 2010 , vol. 40 (pg. 1965 - 1987 ) Google Scholar Crossref Search ADS WorldCat Dugdale R. C. , Goering J. J. . Uptake of new and regenerated forms of nitrogen in primary productivity , Limnology and Oceanography , 1967 , vol. 12 (pg. 196 - 206 ) Google Scholar Crossref Search ADS WorldCat Ezer T. , Atkinson L. P. , Corlett W. B. , Blanco J. L. . Gulf Stream's induced sea level rise and variability along the US mid-Atlantic coast , Journal of Geophysical Research: Oceans , 2013 , vol. 118 (pg. 685 - 697 ) Google Scholar Crossref Search ADS WorldCat FAO Fisheries and Aquaculture topics. Fishery statistics and FAO role. Topics Fact Sheets. , 2005–2013 Text by Richard Grainger. In: FAO Fisheries and Aquaculture Department [online]. Rome. Updated 27 May 2005. [Cited 7 October 2013]. http://www.fao.org/fishery/topic/14790/en. (Last Accessed 16 August 2012) Google Scholar FAO Statistics and Information Service of the Fisheries and Aquaculture Department. Capture production 1950–2009. FISHSTAT Plus – Universal software for fishery statistical time series (online or CD-ROM) , 2011 Food and Agriculture Organization of the United Nations. http://www.fao.org/fishery/statistics/software/fishstat/en. (Last accessed 7 October 2013) OpenURL Placeholder Text WorldCat Flores-Morales A. L. , Parés-Sierra A. , Gómez-Valdivia F. . Semiannual Kelvin waves in the northeastern tropical Pacific , Journal of Coastal Research , 2012 , vol. 28 (pg. 1068 - 1072 ) Google Scholar Crossref Search ADS WorldCat Froese R. , Zeller D. , Kleisner K. , Pauly D. . What catch data can tell us about the status of global fisheries , Marine Biology , 2012 , vol. 159 (pg. 1283 - 1292 ) Google Scholar Crossref Search ADS WorldCat Garrett C. , Toulany B. . Variability of the flow through the Strait of Belle Isle , Journal of Marine Research , 1981 , vol. 39 (pg. 163 - 189 ) Google Scholar OpenURL Placeholder Text WorldCat Garrison T. . Oceanography: An Invitation to Marine Science , 2013 8th edn Belmont CA Brooks/Cole-Thomson Learning pg. 605 Google Scholar Garzoli S. L. , Bianchi A. . Time-space variability of the local dynamics of the Malvinas–Brazil confluence as revealed by inverted echo sounders , Journal of Geophysical Research: Oceans , 1987 , vol. 92 (pg. 1914 - 1922 ) Google Scholar Crossref Search ADS WorldCat Garzoli S. L. , Matano R. . The South Atlantic and the Atlantic meridional overturning circulation , Deep-Sea Research II , 2011 , vol. 58 (pg. 1837 - 1847 ) Google Scholar Crossref Search ADS WorldCat Godø O. R. , Samuelsen A. , Macaulay G. J. , Patel R. , Hjøllo S. S. , Horne J. , Kaartvedt S. , et al. Mesoscale eddies are oases for higher trophic marine life , PLOS ONE , 2012 , vol. 7 e30161. 10.1371/journal.pone.0030161 Google Scholar OpenURL Placeholder Text WorldCat Crossref Greene C. H. . Towards a more balanced view of marine ecosystems , Fisheries Oceanography , 2012 , vol. 22 (pg. 140 - 142 ) Google Scholar Crossref Search ADS WorldCat Greene C. H. , Pershing A. J. . Climate drives sea change , Science , 2007 , vol. 315 (pg. 1084 - 1085 ) Google Scholar Crossref Search ADS PubMed WorldCat Hakkinen S. . Variability in sea surface height: a qualitative measure for the meridional overturning in the North Atlantic , Journal of Geophysical Research-Oceans , 2001 , vol. 106 (pg. 13837 - 13848 ) Google Scholar Crossref Search ADS WorldCat Heimbach P. , Wunsch C. , Ponte R. M. , Forget G. , Hill C. , Utke J. . Timescales and regions of the sensitivity of Atlantic meridional volume and heat transport: toward observing system design , Deep-Sea Research II , 2011 , vol. 58 (pg. 1858 - 1879 ) Google Scholar Crossref Search ADS WorldCat Herrick S. F. Jr , Norton J. G. , Mason J. E. , Bessey C. . Management application of an empirical model of sardine-climate regime shifts , Marine Policy , 2007 , vol. 31 pg. 71080 Google Scholar Crossref Search ADS WorldCat Holbrook N. J , Goodwin I. D. , McGregor S. , Molina E. , Power S. B. . ENSO to multi-decadal time scale changes in East Australian current transports and Fort Denison sea level: oceanic Rossby waves as the connecting mechanism , Deep-Sea Research II , 2011 , vol. 58 (pg. 547 - 558 ) Google Scholar Crossref Search ADS WorldCat Huang R. X. . Ocean Circulation: Wind-Driven and Thermohaline Processes , 2010 Cambridge Cambridge University Press pg. 791 Google Scholar Solomon S. , Qin D. , Manning M. , Chen Z. , Marquis M. , Averyt K. B. , Tignor M. , Miller H. L. , H. L. . IPCC AR4 WG1 Climate Change 2007: The Physical Science Basis , 2007 UK and New York, NY Cambridge University Press Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change Google Scholar Ivchenko V. O. , Zalesny V. B. , Drinkwater M. R. , Schroter J. . A quick response of the equatorial ocean to Antarctic sea ice/salinity anomalies , Journal of Geophysical Research , 2006 , vol. 111 C10018. http://dx.doi.org/10.1029/2005JC003061 Google Scholar OpenURL Placeholder Text WorldCat Johnson H. L. , Marshall D. P. . Global teleconnections of meridional overturning circulation anomalies , Journal of Physical Oceanography , 2004 , vol. 34 (pg. 1702 - 1722 ) Google Scholar Crossref Search ADS WorldCat Kamykowski D. . Estimating upper ocean phosphate concentrations using ARGO float temperature profiles , Deep-Sea Research I , 2008 , vol. 55 (pg. 1580 - 1589 ) Google Scholar Crossref Search ADS WorldCat Kamykowski D. . Atlantic meridional overturning circulation and phosphate classified bottom-up control of Atlantic pelagic ecosystems through the 20th century , Deep-Sea Research I , 2010 , vol. 57 (pg. 1266 - 1277 ) Google Scholar Crossref Search ADS WorldCat Kamykowski D. . 20th Century variability of Atlantic meridional overturning circulation: planetary wave influences on world ocean surface phosphate utilization and synchrony of small pelagic fisheries , Deep-Sea Research I , 2012 , vol. 65 (pg. 85 - 99 ) Google Scholar Crossref Search ADS WorldCat Kamykowski D. , Zentara S-J. . Predicting plant nutrient concentrations from temperature and sigma-t in the world ocean , Deep-Sea Research , 1986 , vol. 33 (pg. 89 - 105 ) Google Scholar Crossref Search ADS WorldCat Kaneko H. , Yasuda I. , Komatsu K. , Itoh S. . Observations of vertical turbulent nitrate flux across the Kuroshio , Geophysical Research Letters , 2013 , vol. 40 10.1002/grl.50613 Google Scholar OpenURL Placeholder Text WorldCat Crossref Kawasaki T. . Why do some pelagic fishes have wide fluctuations in their numbers? Biological basis of fluctuation from the viewpoint of evolutionary ecology , FAO Fisheries Report , 1983 , vol. 291 (pg. 1065 - 1080 ) Google Scholar OpenURL Placeholder Text WorldCat Kieke D. , Klein B. , Strammac L. , Rhein M. , Koltermann K. P. . Variability and propagation of Labrador Sea water in the southern subpolar North Atlantic , Deep-Sea Research I , 2009 , vol. 56 (pg. 1656 - 1674 ) Google Scholar Crossref Search ADS WorldCat Kienart H. , Ramstorf S. . On the relation between meridional overturning circulation and sea-level gradients in the Atlantic , Earth System Dynamics Discussion , 2012 , vol. 3 (pg. 325 - 356 ) Google Scholar Crossref Search ADS WorldCat Kjerve B. , Greer J. E. , Crout R. L. . Wiley M. L. . Low frequency response of estuarine sea level to non-local forcing , In Estuarine Interactions , 1978 New York Academic Press (pg. 497 - 513 ) 603 Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Kleisner K. , Zeller D. , Froese R. , Pauly D. . Using global catch data for inferences on the world's marine fisheries , Fish and Fisheries , 2012 10.1111/j.1467-2979.2012.00469.x Google Scholar OpenURL Placeholder Text WorldCat Crossref Klyashtorin L. B. . Long-term climate change and main commercial fish production in the Atlantic and Pacific , Fisheries Research , 1998 , vol. 37 (pg. 115 - 125 ) Google Scholar Crossref Search ADS WorldCat Latif M. , Böning C. W. , Willebrand J. , Biastoch A. , Dengg J. , Keenlyside N. , Schweckendiek U. , et al. Is the thermohaline circulation changing? , Journal of Climate , 2006 , vol. 19 (pg. 4631 - 4637 ) Google Scholar Crossref Search ADS WorldCat Latif M. , Keenlyside N. S. . A perspective on decadal climate variability and predictability , Deep–Sea Research II , 2010 , vol. 58 (pg. 1880 - 1894 ) Google Scholar Crossref Search ADS WorldCat Latif M. , Roeckner E. , Botzet M. , Esch M. , Haak H. , Hagemann S. , Jungclaus J. , et al. Reconstructing, monitoring, and predicting multidecadal-scale changes in the North Atlantic thermohaline circulation with sea surface temperature , Journal of Climate , 2004 , vol. 17 (pg. 1605 - 1614 ) Google Scholar Crossref Search ADS WorldCat Liu Z. , Alexander M. . Atmospheric bridge, oceanic tunnel, and global climatic teleconnections , Reviews of Geophysics , 2007 , vol. 45 RG2005. http://dx.doi.org/10.1029/2005RG000172 (Last accessed 7 October 2013) Google Scholar OpenURL Placeholder Text WorldCat Lluch-Belda D. , Crawford R. J. M. , Kawasaki T. , MacCall A. D. , Parrish R. H. , Schwartzlose R. A. , Smith P. E. . World-wide fluctuations of sardine and anchovy stocks: the regime problem , South African Journal of Marine Science , 1989 , vol. 8 (pg. 195 - 205 ) Google Scholar Crossref Search ADS WorldCat Lluch-Belda D. , Schwartzlose R. A. , Serra R. , Parrish R. , Kawasaki T. , Hedgecock S. D. , Crawford R. J. M. . Sardine and anchovy regime fluctuations of abundance in four regions of the world oceans: a workshop report , Fisheries Oceanography , 1992 , vol. 1 (pg. 339 - 347 ) Google Scholar Crossref Search ADS WorldCat Lluch-Cota D. B. , Hernández-Vázquez S. , Lluch-Cota S. E. . Empirical investigation on the relationship between climate and small pelagic global regimes and El Niño-Southern Oscillation (ENSO) , FAO Fisheries Circular No. 934 , 1997 Rome FAO pg. 48 Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Louanchi F. , Najjar R. G. . A global monthly climatology of phosphate, nitrate and silicate in the upper ocean: export production and shallow remineralization , Global Biogeochemical Cycles , 2000 , vol. 14 (pg. 957 - 977 ) Google Scholar Crossref Search ADS WorldCat MacCall A. D. . Bakun A. , Broad K. . Sardine regimes and mesoscale flow structure , In Climate and Fisheries: Interacting Paradigms, Scales and Policy Approaches. International Research Institute for Climate Prediction , 2002 Palisades, New York Columbia University (pg. 39 - 42 ) 67 Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC MacCall A. D. . Checkley D. , Alheit J. , Oozeki Y. , Roy C. . Mechanisms of low-frequency fluctuations in sardine and anchovy populations , In Climate Change and Small Pelagic Fisheries , 2009 Cambridge Cambridge University Press (pg. 285 - 299 ) 372 Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Medhaug I. , Furevik T. . North Atlantic 20th century multidecadal variability in coupled climate models: sea surface temperature and ocean overturning circulation , Ocean Science , 2011 , vol. 7 (pg. 389 - 404 ) Google Scholar Crossref Search ADS WorldCat Myers R. A. . When do environment-recruitment correlations work? , Reviews in Fish Biology and Fisheries , 1998 , vol. 8 (pg. 285 - 305 ) Google Scholar Crossref Search ADS WorldCat Papadopoulos A. , Tsimplis M. N. . Coherent coastal sea-level variability at interdecadal and interannual scales from tide gauges , Journal of Coastal Research , 2006 , vol. 22 (pg. 625 - 639 ) Google Scholar Crossref Search ADS WorldCat Piecuch C. G. , Ponte R. M. . Buoyancy-driven interannual sea level changes in the southeast tropical Pacific , Geophysical Research Letters , 2012 , vol. 39 L05607, 10.1029/2012GL051130 Google Scholar OpenURL Placeholder Text WorldCat Crossref Powell A. M. Jr , Xu J. . Abrupt climate regime shifts, their potential forcing and fisheries impacts , Atmospheric and Climate Sciences , 2011 , vol. 1 (pg. 33 - 47 ) Google Scholar Crossref Search ADS WorldCat Rao R. R. , Girish Kumar M. S. , Ravichandran M. , Rao A. R. , Gopalakrishna V. V. , Thadathil P. . Interannual variability of Kelvin wave propagation in the wave guides of the equatorial Indian Ocean, the coastal Bay of Bengal and the southeastern Arabian Sea during , Deep-Sea Research I , 2010 , vol. 57 (pg. 1993 - 2006 ) 1–13 Google Scholar Crossref Search ADS WorldCat Sasaki Y. N. , Schneider N. . Decadal shifts of the Kuroshio extension: application of thin-jet theory , Journal of Physical Oceanography , 2011 , vol. 41 (pg. 979 - 993 ) Google Scholar Crossref Search ADS WorldCat Schmitz W. J. . On the interbasin scale thermohaline circulation , Reviews of Geophysics , 1995 , vol. 33 (pg. 151 - 173 ) Google Scholar Crossref Search ADS WorldCat Senjyu T. , Matsuyama M. , Matsubara N. . Interannual and decadal sea-level variations along the Japanese coast , Journal of Oceanography , 1999 , vol. 55 (pg. 619 - 633 ) Google Scholar Crossref Search ADS WorldCat Smith T. M. , Reynolds R. W. . Improved extended reconstruction of SST [1854–1997] , Journal of Climate , 2004 , vol. 17 (pg. 2466 - 2477 ) Google Scholar Crossref Search ADS WorldCat Smith T. M. , Reynolds R. W. , Peterson T. C , Lawrimore J. . Improvements to NOAA's historical merged land–ocean surface temperature analysis (1880–2006) , Journal of Climate , 2008 , vol. 21 (pg. 2283 - 2296 ) Google Scholar Crossref Search ADS WorldCat Strickland J. D. H. , Parsons T. R. . A practical handbook of seawater analysis , Fisheries Research Board of Canada Bulletin , 1968 , vol. 167 pg. 311 Google Scholar OpenURL Placeholder Text WorldCat Sugimatsu K. , Isobe A. . Intrusion of Kuroshio-derived bottom cold water into Osaka Bay and its possible cause , Journal of Oceanogaphy , 2010 , vol. 66 (pg. 191 - 199 ) Google Scholar Crossref Search ADS WorldCat Taguchi B. , Xie S-P. , Mitsudera H. , Kubokawa A. . Response of the Kuroshio Extension to Rossby waves associated with the 1970s climate regime shift in a high-resolution ocean model , Journal of Climate , 2005 , vol. 18 (pg. 2979 - 2995 ) Google Scholar Crossref Search ADS WorldCat Taguchi B. , Xie S-P. , Schneider N. , Nonaka M. , Sasaki H. , Sasai Y. . Decadal variability of the Kuroshio Extension: observations and an eddy-resolving model hindcast , Journal of Climate , 2007 , vol. 20 (pg. 2357 - 2377 ) Google Scholar Crossref Search ADS WorldCat Tanco R. A. , Berri G. J. . CLIMLAB2000 , 2000 Version 1.1.0, A Statistical Software Package for Climate Applications. IRI-TR-00/1, International Research Institute (IRI) for Climate Prediction, LDEO, Palisades, NY, 10964 USA OpenURL Placeholder Text WorldCat Timmerman A. , An S-I. , Krebs U. , Goose H. . ENSO suppression due to weakening of the North Atlantic thermohaline circulation , Journal of Climate , 2005 , vol. 18 (pg. 3122 - 3139 ) Google Scholar Crossref Search ADS WorldCat Tourre Y. M. , Lluch-Cota S. E. , White W. B. . Global multi-decadal ocean climate and small-pelagic fish population , Environmental Research Letters , 2007 , vol. 2 pg. 9 10.1088/1748-9326/2/3/034005 Google Scholar Crossref Search ADS WorldCat Valsala V. , Maksyutov S. , Murtugudde R. . Possible interannual to interdecadal variabilities of the Indonesian throughflow water pathways in the Indian Ocean , Journal of Geophysical Research , 2010 , vol. 115 C10016. http://dx.doi.org/10.1029/2009JC005735. (Last accessed 7 October 2013) Google Scholar OpenURL Placeholder Text WorldCat van Aken H. M. . The Oceanic Thermohaline Circulation: An Introduction , 2007 New York, NY Springer pg. 326 Google Scholar van der Lingen C. D. , Bertrand A. , Bode A. , Brodeur R. , Cubillos L. , Espinoza P. , Friedland K. , et al. Checkley D. , Alheit J. , Oozeki Y. , Roy C. . Trophic dynamics , In Climate Change and Small Pelagic Fisheries , 2009 Cambridge Cambridge University Press (pg. 112 - 157 ) 372 Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Wang B. . Curry J. A , Pyle J.A. . Kelvin Waves , In Encyclopedia of Atmospheric Science , 2002 Elsevier (pg. 1062 - 1068 ) 2623 Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Wang C. , Fiedler P. C. . ENSO variability and the eastern tropical Pacific: a review , Progress in Oceanography , 2006 , vol. 69 (pg. 239 - 266 ) Google Scholar Crossref Search ADS WorldCat Watson R. , Kitchingman A. , Gelchu A. , Pauly D. . Mapping global fisheries: sharpening our focus , Fish and Fisheries , 2004 , vol. 5 (pg. 168 - 177 ) Google Scholar Crossref Search ADS WorldCat Weber T. S. , Deutsch C. . Ocean nutrient ratios governed by plankton biogeography , Nature , 2010 , vol. 267 (pg. 550 - 554 ) Google Scholar Crossref Search ADS WorldCat Woodworth P. L. . Some important issues to do with long-term sea level change , Philosophical Transactions of the Royal Society of London , 2006 , vol. 364 (pg. 787 - 803 ) Series A, Mathematical, physical, and engineering sciences Google Scholar Crossref Search ADS WorldCat Yasunaka S. , Hanawa K. . Intercomparison of historical sea surface temperature datasets , International Journal of Climatology , 2011 , vol. 31 (pg. 1056 - 1073 ) Google Scholar Crossref Search ADS WorldCat Zhang D. , Msadek R. , McPhaden M. J. , Delworth T . Multidecadal variability of the North Brazil Current and its connection to the Atlantic meridional overturning circulation , Journal of Geophysical Research , 2011 , vol. 116 C04012, 10.1029/2010JC006812 Google Scholar OpenURL Placeholder Text WorldCat Crossref Zhang R. . Anticorrelated multidecadal variations between surface and subsurface tropical North Atlantic , Geophysical Research Letters , 2007 , vol. 34 L12713, doi.org/10.1029/2007GL030225 Google Scholar OpenURL Placeholder Text WorldCat Crossref Zhang R. . Coherent surface–subsurface fingerprint of the Atlantic meridional overturning circulation , Geophysical Research Letters , 2008 , vol. 35 L20705, 10.1029/2008GL035463 Google Scholar OpenURL Placeholder Text WorldCat Crossref Zhang R. , Delworth T. L. . Simulated tropical response to a substantial weakening of the Atlantic thermohaline circulation , Journal of Climate , 2005 , vol. 18 (pg. 1853 - 1860 ) Google Scholar Crossref Search ADS WorldCat Zhang X. , Church J. A. . Sea level trends, interannual and decadal variability in the Pacific Ocean , Geophysical Research Letters , 2012 , vol. 39 pg. L21701 10.1029/2012GL053240 Google Scholar OpenURL Placeholder Text WorldCat Crossref Author notes Handling editor: Howard Browman © 2013 International Council for the Exploration of the Sea. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com © 2013 International Council for the Exploration of the Sea. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com
TI - Twentieth century Atlantic meridional overturning circulation as an indicator of global ocean multidecadal variability: influences on sea level anomalies and small pelagic fishery synchronies
JO - ICES Journal of Marine Science
DO - 10.1093/icesjms/fst165
DA - 2014-04-01
UR - https://www.deepdyve.com/lp/oxford-university-press/twentieth-century-atlantic-meridional-overturning-circulation-as-an-oiueCiELe4
SP - 455
EP - 468
VL - 71
IS - 3
DP - DeepDyve
ER -