Potential Feedbacks Between Paci c Ocean Ecosystems and Interdecadal Climate Variations Arthur J. Miller1*, Michael A. Alexander2, George J. Boer3, Fei Chai4, Ken Denman3 , David J. Erickson III5, Robert Frouin1, Albert J. Gabric6, Edward A. Laws7 , Marlon R. Lewis8, Zhengyu Liu9, Ragu Murtugudde10 , Sho Nakamoto11, Douglas J. Neilson1, Joel R. Norris1, J. Carter Ohlmann12, R. Ian Perry13, Niklas Schneider1, Karen M. Shell1, and Axel Timmermann14 Scripps Institution of Oceanography, La Jolla, CA USA 2 Climate Diagnostics Center, Boulder, CO USA 3 University Of Victoria, Victoria, BC Canada 4 University of Maine, Orono, ME USA 5 Oak Ridge National Laboratory, Oak Ridge, TN USA 6 Grith University, Nathan, Australia 7 University of Hawaii, Honolulu, HI USA 8 Dalhousie University, Halifax, NS Canada 9 University of Wisconsin, Madison, WI USA 10 University of Maryland, College Park, MD USA 11 Earth Science and Technology Organization, Kawasaki, Japan 12 University of California, Santa Barbara, CA USA 13 Fisheries and Oceans Canada, Nanaimo, BC, Canada 14 Institut f ur Meereskunde, Kiel, Germany 1
*Corresponding author address: Climate Research Division Scripps Institution of Oceanography La Jolla, CA 92093-0224 Phone: (858) 534-8033 Fax: (858) 534-8033 E-mail: [email protected]
Submitted to Bulletin of the American Meteorological Society October 25, 2001 1
The mechanisms responsible for interdecadal climate variations over the Paci c Ocean are unclear. While many types of feedbacks in the physical ocean-atmosphere system have been proposed to explain some aspects of these climate variations, the potential in uence of the oceanic ecosystem has not been addressed. Various mechanisms are described herein by which the oceanic biological response to Paci c interdecadal climate forcing may organize a signi cant feedback to the physical climate system. The oceanic biological processes considered in this paper appear to be able to sensitize the coupled ocean-atmosphere climate system predominantly through changes in upper-ocean absorption of radiation due to phytoplankton and to changes in dimenthylsul de (DMS) uxes to atmospheric cloud systems. The oceanic biology does not, however, appear to be able to introduce an adequately long delay in the system to set a preferred timescale in interdecadal climate variations. Modeling and observational strategies are outlined to guide future research that is needed to resolve the many unsolved issues of biological-physical feedbacks in interdecadal climate variations.
1. Introduction Oceanic ecosystems are thought to exert an in uence on the physical climate system on many time and space scales both directly (e.g., through radiative eects on the upper ocean) or indirectly (e.g., by aecting CO2 concentration in the global atmosphere). But the role of ocean ecosystems in producing and aecting the present climate, interdecadal* climate variations, and long term climate change is not well understood. The possible in uences of Paci c Ocean ecosystems in generating or modifying interdecadal timescale climate variations is of interest here. The mechanisms responsible for interdecadal climate variability in the Paci c sector are themselves not very clear. A number of possible physical feedback loops (i.e., coupled oscillations) in the ocean-atmosphere system have been proposed to explain aspects of observed and modeled climate variations (e.g., Latif 1998; Miller and Schneider 2000). However, interdecadal feedback mechanisms cannot be unambiguously identi ed in existing observations and they are not robust in the coupled general circulation models (CGCMs) whose results have been analyzed. Moreover, stochastic driving of the ocean by the atmosphere can potentially account for a considerable portion of observed and modeled interdecadal variability (Barsugli and Battisti 1998; Frankignoul et al. 1997) without recourse to physical and/or ecosystem feedback mechanisms. While physical-biological feedbacks may only account for a small fraction of the observed interdecadal climate variability, it is nevertheless of interest to ask if these exist and are of potential importance. We therefore attempt to answer a sequence of questions: * What mechanisms might allow Paci c Ocean biological responses to in uence variations of the physical climate system on interdecadal timescales? * Do these oceanic biological feedback mechanisms increase or decrease the Paci c Ocean/atmosphere sensitivity to physical climate interactions? * What are the key regions in the Paci c Ocean where these biological in uences might be active? * How can these possible biological feedbacks be tested with models and observations? The goal of this paper is to motivate a coordinated modeling and observational eort to study coupled physical-biological mechanisms of interdecadal climate variability in the Paci c.
2. Heuristic Perspective of Biology and Climate
Our current description of the climate system is usually based on a somewhat arti cial separation into an external component, which is assumed to be known or given (or to operate
* We use the term `interdecadal' to loosely refer to timescales that are longer than interannual (ENSO) and shorter than centennial (greenhouse gas forcing). 3
on a longer timescale than the components of interest), and an internal component, the behavior of which is determined by the external component and the physics (and potentially the biology) of the system. This separation is clear in the case of modern climate models in which the external system is just those quantities speci ed in the model code (e.g. the shape, size, geography and rotation rate of the earth, the composition of the atmosphere, etc.) The model climate is simulated by integrating the governing equations, given this external information, and the internal component is embodied in the prognostic variables in the model (e.g. temperatures, winds, precipitation rates, etc.) that are internally generated as the model is integrated in time. Biological processes are not part of the internal system of the collection of current global coupled climate models used to study climate, its variation and change under external forcing (e.g. Table 9.1, Ch. 9, IPCC, 2001). Biological processes are lacking as part of the external system of these models as well (except perhaps as speci ed features of the land surface such as the roughness and albedo associated with diering vegetation cover). For the oceans, current CGCMs typically do not include biological processes as part of either the internal or external system. This situation, however, is rapidly changing and simpli ed representations of some oceanic and terrestrial biology, particularly as they aect the carbon budget and radiation balances, are beginning to appear in coupled model simulations (e.g., Maier-Reimer et al. 1996; Sarmiento et al. 1998; Matear and Hirst 1999; Cox et al. 2000). A perturbation to the climate system may be generated through imposed forcing such as the solar constant or green house gas concentration of the atmosphere. Alternatively, it may be internally generated through the natural workings of the non-linear climate system. Biological feedback processes may act to enhance, transform, and/or suppress both externally forced changes and those that are internally generated. Biology can aect climate by modifying the earth's energy budget. The most direct way this may be accomplished is by altering the atmospheric or oceanic radiative balances. Modi cation of the solar radiative input to the climate system may be accomplished by changing the transparency or re ectivity of the system via aerosols and cloud eects in the atmosphere, by modi cation of the surface albedo and of the absorptivity of the ocean. Modi cation to the long wave output may be accomplished via the greenhouse gas concentration of the atmosphere and/or cloud processes. The search for important biological climate processes must also consider the nature of the response of the climate system to an arbitrary perturbation. A locally strong, biologically mediated perturbation to the ow of energy may only produce a locally weak and globally trivial eect, because the climate system rapidly distributes and dilutes local perturbations via the atmosphere. This rapid redistribution also means that local perturbations do not 4
necessarily, or even generally, lead to local responses since, to rst order, the system response is localized by the feedback processes operating in the system rather than the forcing pattern itself. [As an example, Reader and Boer (1998), Mitchell et al. (1995), Roeckner et al. (1995), Boer et al. (2000) and Chapter 12, IPCC (2001) all nd strong high latitude feedback processes localize the cooling eect of low latitude aerosols in coupled models.] In broad terms we expect direct biological perturbations to the energy stream to be potentially important where energy ows are largest, namely in the tropics where solar input is strong. Other important climate eects due to biological perturbations may be possible if they occur in conjunction with strong local feedback mechanisms based on physical (e.g., altered clouds, ice/snow albedo eects) or possibly dynamical processes (e.g. changes in baroclinicity in storm tracks). Denman et al. (1996) reviewed the main mechanisms by which marine ecosystems might respond to and generate feedbacks to a changing climate.
3. Physical and Biological Background a. Physical processes of Paci c interdecadal variability Observations clearly show enhanced variance of Paci c climate variables at interdecadal frequencies (e.g., Mann and Park, 1996; Zhang et al., 1997; En eld and Mestas-Nun~nez, 1999). The decadal variations in the physics can take the form of gradual drifts, smooth oscillations or step-like shifts such as that of 1976-77 (Miller et al., 1994) or of 1989 (Hare and Mantua, 2000). While stochastic theories can account for the basic spectral shapes, some time series analyses suggest preferred timescales of Paci c climate variability exist with periods of 10 years, 20 years and 50-80 years (Mann and Park 1996; Minobe 1999). These results, however, have only marginal statistical signi cance due to limited observations in time and space. The observed characteristic spatial patterns seen in interdecadal SST anomalies show a `canonical' structure (e.g., Tanimoto et al. 1992; Zhang et al. 1997), with central North Paci c SST near the subtropical front bracketed to the east, north and south by oppositely signed SST. A second SST pattern is centered around the subpolar front in the KuroshioOyashio Extension region (Deser and Blackmon 1995; Nakamura et al. 1997) and tends to lag the subpolar front SST anomalies by a few years on decadal timescales (Miller and Schneider 2000). The North Paci c thermocline observations show interdecadal structures linked to wind-stress curl forcing (Miller et al. 1998; Tourre et al. 1999) and to anomalous subduction from buoyancy forcing (Deser et al. 1996; Schneider et al. 1999). These observed patterns may be indicative of preferred natural modes of variability in the climate system or may be simply a passive response to the long-period atmospheric forcings. There are many ideas about how feedback mechanisms in the physical climate system 5
may control the portion of Paci c interdecadal climate variations that are not simply explained by stochastic theories (e.g., Miller and Schneider, 2000, summarize these). Many proposed mechanisms appeal to some process that introduces a delay in the system to explain key aspects of the observations. Since the proposed mechanisms cannot be tested directly because of limited observations, investigators have used full-physics coupled models to determine if the simple theories are consistent with the complicated model runs and with the available observations. One of the prominent theories to explain Paci c interdecadal climate variations is the Subduction Mode hypothesis (Gu and Philander 1997), schematically illustrated in Figure 1. This mechanism has roots in the idea that temperature anomalies generated in the central North Paci c are subducted and move towards the equator along typical pathways of thermocline waters. Depending on the exact path, the anomalies arrive in the tropical Paci c after one to several decades. There, air-sea interactions are hypothesized to amplify the exhausted temperature anomalies and to modify the atmospheric circulation in a Bjerknes teleconnection. A warm equatorial SST anomaly strengthens the atmospheric Hadley circulation and, in turn, the midlatitude westerly wind. The stronger surface westerly cools the midlatitude SST due to enhanced evaporation, Ekman transport and vertical mixing. In this way, temperature anomalies of the reversed sign could be generated within the central North Paci c, initiating the other half of the cycle. This proposed mechanism, if it works in reality, does not appear to be very ecient in triggering tropical temperature anomalies. Observations indicate that subducted temperature anomalies can be traced along their southward route up to subtropical latitudes but not further (Schneider et al. 1999). Another complication arises from the fact that temperature anomalies are not passive tracers, since they modify the subsurface density structures. This in turn can generate waves (Liu and Shin 1999) which also have the potential to alter the tropical circulation. Alternatively, decadal climate variations may result from the slow adjustment of the speed of the shallow vertical overturning cell between tropics and extratropics (Kleeman et al. 1999). Observations con rm that in the Paci c the speed of equatorward ow in the thermocline, and poleward return ow at the surface has changed over the last few decades (McPhaden and Zhang 2001). In addition, ocean hindcasts (Nonaka et al. 2001) show these to be driven by anomalies of the wind stress, and to explain part of the decadal anomalies of SST in the tropics. However, it is unclear if in observations the shallow overturning cell also caused the anomalies of the wind stress or simply responded to forcing by the atmosphere. Another in uential theory of Paci c interdecadal climate variations is the Midlatitude Gyre Mode hypothesis (Latif and Barnett 1994), schematically illustrated in Figure 1. The 6
basic idea is as follows: A warm SST anomaly (SSTA) in the western/central extratropical Paci c is argued to modify the atmospheric circulation by altering the atmospheric baroclinicity. On the one hand the atmospheric response feeds back to enhance the local SSTA, on the other hand it is associated with a wind stress curl anomaly in the central/eastern extratropical Paci c which triggers westward propagating upwelling Rossby waves. It is argued that this adjusts the circulation of the subtropical gyre after many years. The associated changes in the western boundary currents then reverse the original SSTA anomalies and commence the oppositely-sign patterns. Typical timescales for this type of coupled air-sea mode are between 10-20 years. The full-physics coupled models that have been analyzed to date, however, fail to generate either of these two broad categories of feedback loops (Schneider et al. 1999; Schneider et al. 2002). The hints of ocean to atmosphere feedbacks that have been detected do not seem to be consistent with the two leading hypotheses (at least for the Paci c Ocean). Instead, stochastic forcing of the ocean by the atmosphere can explain the bulk of interdecadal Paci c Ocean climate variability found in full-physics models. Other physical processes may contribute to the enhanced variance associated with stochastic processes at decadal timescales in the Paci c. For example, the "reemergence mechanism", where a deep mixed layer in one year is insulated from surface uxes that act to damp it when it reemerges the subsequent spring, has been documented to occur over much of the North Paci c and Atlantic (Alexander and Deser 1995; Alexander et al. 1999). This mechanism can maintain SST anomalies and their in uence on the atmosphere for several additional years in regions of deep wintertime mixed layers (Alexander et al. 2000; Watanabe and Kimoto 2000). Another physical explanation for Paci c interdecadal variability involves tropical modes that force a response in the extratropics. A signi cant fraction of the dominant pattern of long period (>10 years) SST variability in the North Paci c has been suggested to be associated with tropical forcing via the atmospheric teleconnections, the so-called "atmospheric bridge" (Lau and Nath 1996; Alexander et al. 2001). Recent fully coupled modeling studies suggest that the tropical Paci c alone can also support interdecadal climate variability (Liu et al. 2001). Jin (2001) found that decadal equatorial thermocline oscillations are possible in a linear shallow water ocean on an equatorial beta plane when the adjustments of the higher latitudes are included. Alternatively, the nonlinearities can lead to variance at decadal time scales without a classical memory timescale (such as mean advection or Rossby-wave traveling times). Interactions between ENSO and the annual cycle leads to ENSO irregularity and temperature variance on decadal timescales through the quasi-periodicity route to chaos (Jin et al. 1994; 7
Tziperman et al. 1994; Chang et al. 1996). Decadal variability in the tropics also emerges as a natural consequence of nonlinear temperature advection (Timmermann 2001; Timmermann and Jin 2001). A linearly unstable ENSO mode grows until the zonal temperature advection from the warm pool to the eastern equatorial Paci c becomes so small that the warm pool thermocline depth is readjusted to its radiative-convective equilibrium state. This resets the growing ENSO mode. Realistic parameter values yield several decades as a typical periods between warm pool resettings and explain the emergence of decadal ENSO amplitude modulations, ENSO irregularity as well as of decadal tropical variability. b. Interdecadal Paci c ecosystem changes Biological systems in the Paci c Ocean exhibit variations on interdecadal timescales, that appear to be linked to climate variability (Francis et al. 1998). In some cases, like the anchovy and sardine regimes that last 10 to 50 years (e.g., Baumgartner et al. 1992; Ware 1995; Schwartzlose et al. 1999; Field and Baumgartner 2000), most of the variance is concentrated in the interdecadal band. In other cases, like zooplankton along the North American West Coast (e.g., Bograd and Lynn 2001) and in the open northeast Paci c Ocean (Conversi and Hameed 1997), interannual variance appears to dominate the signal although this may be confounded by relatively short time series. Nevertheless, Brodeur and Ware (1992) found an increase in zooplankton biomass over the NE Paci c between the 1960s and the 1980s, and Venrick et al. (1987) documented an increase in subsurface phytoplankton chlorophyll in the Central N. Paci c over the period 1964-85. Ecosystem variations may arise from intrinsic biological interactions in the presence of a steady physical environment. On the other hand, many population variations have been linked to speci c physical changes. Often, the biological change is linked to sea-surface temperature since that is the most heavily sampled physical variable in the ocean. But, since SST is often correlated with other physical changes (such as upwelling, mixing, and horizontal currents), the speci c mechanisms linking physical and ocean ecosystem changes have remained dicult to isolate. For example, Whitney et al. (1998) have documented a dramatic decline in nitrate (an essential nutrient for phytoplankton) in the northeast Paci c that may be related to declining wind stress curl and shoaling of the ocean mixed layer. A shoaling mixed layer and warmer upper layer temperatures are also implicated in signi cant changes in the life cycle timing of the major copepod species in the northeast Paci c (Mackas et al. 1998). This, in turn, might in uence sh production. Freeland et al. (1997) have documented the shoaling of the winter mixed layer at Ocean Station P. If the present rate of shoaling continues, they present results from a simple model suggesting that the subarctic NE Paci c plankton ecosystem would on average start to experience nitrate limitation in less than 100 years, assuming no other positive feedbacks. At that time the biologically-induced 8
changes in upper layer optical transmission could lead to a feedback in mixed layer dynamics and air-sea exchanges. Interdecadal-scale variations in oceanic physical variables have many implications for marine ecosystems (Mantua et al. 1997; Sugimoto and Tadoroko 1997; McGowan et al. 1998). Widespread ecological changes associated with the 1976-77 climate shift (Miller et al. 1994) were observed throughout the North Paci c Ocean, ranging from plankton to the higher trophic level changes (Venrick et al. 1987; Polovina et al. 1994; Francis and Hare 1994). In the northwestern subtropical gyre region, chlorophyll in spring showed a steady increase from the mid 1970s to the mid 1980s (Limsakul et al. 2001). Temporal increases of phosphate and apparent oxygen utilization (AOU) in the western subarctic Paci c from 1968 to 1998 have been reported by Ono et al. (2001). Emerson et al. (2001) have reported similar increases in the northeast subtropical Paci c between 1980 and 1997. Around 197677, the small pelagic sh stocks o the coast of northern and central Peru experienced a transition from anchovy to sardine dominance (Schwartzlose et al. 1999). While there are no data to con rm this, modeling studies indicate that the tropical nutricline also shifted during the 1976-77 shift along with the thermocline shift that was noted in coral records (Guilderson and Schrag 1998). It is thus likely that the primary producers in the tropics also experienced a interdecadal shift. The likely feedbacks of this shift to the dynamics and the thermodynamics and their role in the frequency and amplitude changes in ENSO (Trenberth and Hoar 1996) need to be investigated further. To model the spring-time (March-April) phytoplankton biomass in the northwest Paci c before and after the 1976-77 climate shift, Chai and collaborators (manuscript to be submitted to GBC) have incorporated an ecosystem model with multiple nutrients and phytoplankton groups (Chai et al. 2001) into a general circulation model for the Paci c Ocean. The physical-biological model is forced with monthly averaged COADS wind stress and heat
ux from 1962 to 1992. Figure 2 shows the spring-time (March-April) vertically integrated (0-60m) total phytoplankton biomass (in units of nitrogen) for the North West Paci c (35N45N, 160E-160W). The modeled total phytoplankton biomass includes both diatoms and small phytoplankton, with the diatoms dominating the total biomass. In general, the phytoplankton spring blooms are much stronger after 1975-76 compared to the period before, which is consistent with the modeled nitrate and silicate conditions in the upper 100m. The strongest change occurs between 1962-75 and 1976-1992, with averaged spring-time integrated phytoplankton biomass increasing from 27.8 to 36 (mmol ? Nm?2 ) for these two periods. There is some interannual variability, and it seems the spring blooms are intensi ed in the northwest Paci c during El Ni~no years (e.g., 1972/73, 1976/76, 1982/83, 1986/87, and 1991/92). Haigh et al. (2001), in a model of the whole North Paci c, found post-1976 9
increased phytoplankton and zooplankton biomass generally over the subarctic gyre, due to a deepened mixed layer that entrained more nutrients. In the subarctic gyre, where light limitation also may be a factor, increased phytoplankton biomass corresponded to areas with shallower spring and summer mixed layers post-1976. Similar observed changes of phytoplankton biomass in the Northwest Paci c have been documented by many researchers, e.g., a two-fold increase in integrated chlorophyll-a was observed during summer (Venrick et al. 1987). In the subarctic North Paci c, increases in zooplankton abundance (Brodeur and Ware 1992) and a striking increase in total salmon catches (Beamish and Bouillon 1993) were reported between approximately these decades. Overall, these modeled physical-biological results seem supported by the limited observations in the North Paci c. c. Biological Feedbacks to Climate Physics The energy and chemical budgets of the atmospheric climate are modi ed and, in some instances, controlled by air-sea uxes that result from oceanic biological and physical processes. Many of the chemicals that cycle between the air and sea interact with the radiative balance of the atmosphere and ocean. The ow of radiant energy from the sun and back to space dominates the climate balance, and ocean ecosystems may modify this ow in at least two ways. The rst is through the eect of phytoplankton on upper-ocean absorption of solar radiation. The second is through the ux of dimethylsul de (DMS) to the atmosphere which aects cloud formation. Other eects may also be active. We now discuss the speci c ways these mechanisms may alter the physical climate system. i. Phytoplankton eects on upper-ocean radiation absorption A primary mechanism of biological control on upper ocean physics exists through changes in the in-water solar ux divergence, or solar transmission, brought about by varying quantities of phytoplankton biomass in the upper ocean. The absorption of solar radiation is a dominant term in the upper ocean heat equation for equatorial regions and is signi cant in summer in the higher latitudes. Solar radiation directly heats water below the air-sea interface in contrast to the other components of the air-sea heat exchange that act at the ocean surface. A fraction of the solar energy reaching the sea-surface is re ected back to the atmosphere by Fresnel re ectance or backscattered out of the ocean. The sum of Fresnel re ected and backscattered irradiance normalized by the surface-incident downwelling irradiance is termed the sea-surface albedo (or simply `albedo'). The average value of albedo is near 0.05 for equatorial regions and can vary from 0.03 to more than 0.90 as a function of the angular distribution of the incident light eld, wind speed, and the nature of particulates in the upper ocean (see Payne 1972; Sathyendranath and Platt 1988; Morel and Antoine 1994). 10
In general, however, most of the incident solar energy enters the water and is available for heating the upper ocean. In the upper ocean, solar energy decays exponentially with depth following the BeerLambert relation. This e-folding, or attenuation, scale is a function of wavelength and can be quanti ed with the diuse attenuation coecient spectrum. Both pure seawater and its constituents contribute to solar attenuation. The diuse attenuation coecient for optically pure water is considered constant. Changes in chlorophyll biomass, present in phytoplankton, and its co-varying materials (hereafter termed `chlorophyll') are primarily responsible for variations in solar attenuation, or transmission, in open ocean waters. The associated modi cation of the vertical divergence of the solar energy lead to variations in the vertical distribution of heating. This aects the upper ocean strati cation, in uences vertical mixing and can lead to changes SST, longwave, latent and sensible heat exchange with the atmosphere and oceanic and atmospheric circulation patterns. Solar energy reaches the ocean surface primarily in wavebands ranging from 250 to 2500 nm, with a peak near 500 nm. Solar energy in the red and near-infrared wavebands ( > 700 nm) is absorbed in the top few meters by the strong attenuation of pure seawater in this spectral region. Energy in the shorter (blue and green) wavelengths penetrates much deeper in pure seawater. Solar energy in the most transparent (for pure water) wavebands ( 400500 nm) has an e-folding scale near 50 m (in pure seawater; Morel and Antoine 1994). Light attenuation due to chlorophyll is strongest near 430 and 670 nm. Chlorophyll eects on solar attenuation in wavelengths between 550 and 650 nm, and beyond 700 nm, are signi cantly less. The combined pure seawater and chlorophyll attenuation spectra yield variations in solar attenuation for open ocean waters that are largest in the deep penetrating wavebands between 400 and 600 nm. Chlorophyll concentration thus in uences how visible energy ( 350 to 750 nm) is attenuated in the water column, and ultimately de nes the vertical distribution of radiant heating within the upper ocean. An exact calculation of in-water solar uxes requires solving the in-water radiative transfer equation with the spectral and angular distribution of the incident solar energy as a boundary condition. This is rarely carried out as the surface boundary condition is not easily determined and solving the radiative transfer equation is computationally intensive. Use of a diuse attenuation coecient spectrum allows for a very accurate approximation of the solar ux at depth, performed in the following way. First, diuse attenuation coecient spectra for pure seawater and its chlorophyll concentration are summed at each wavelength to determine a total diuse attenuation coecient spectrum (Kd(z;); e.g. Morel 1988; Morel and Antoine 1994) as a function of depth. Next, the Beer-Lambert relation is used along with the diuse attenuation coecient to calculate the solar ux at depth from the surface 11
ux for each wavelength after accounting for surface albedo. The downward irradiance (Ed) in wavelength , at depth z, is then Z
Ed(z;) = Ed(0;)()exp(? Kd(z0 ;)dz0
where () is a spectral albedo value, and Ed(0;) s the downward spectral irradiance just above the air-sea interface. Finally, integration over the solar spectrum gives the solar ux at depth. Such "full spectral" calculations have been pointed out in numerous studies including Lewis et al. (1983), Lewis (1987), and Ohlmann et al. (1996). However, the formulation is computationally intense, requires spectral resolution of the incident solar radiation and the sea-surface albedo, and requires vertical resolution of the diuse attenuation coecient spectrum. In-water solar uxes are rarely calculated in this manner, and never in models interested in climate scales (but see Woods 1980; Woods et al. 1984). Solar transmission parameterizations used in climate models are generally based on (1) and some simplifying assumptions. Albedo is considered spectrally uniform. Energy in the UV and near-IR wavelengths is assumed completely attenuated in the top well-mixed model layer (upper few meters). Energy in the deep-penetrating visible wavebands is divided among a small number of bins each with a dierent e-folding scale that is uniform with depth. Mathematically the parameterizations give the solar ux at depth as
Ed(z) = Ed(0)iAi exp(?Biz)
where the number of empirically determined Ai and Bi parameters ranges from one or two (Denman 1973; Paulson and Simpson 1977; Ohlmann et al. 1998) to nine or more (Simpson and Dickey 1981; Zaneveld et al. 1981). The Ai and Bi parameters are mostly a function of Jerlov water type, a discrete integer index (I, IA, IB, II, or III) developed in the 1960's as a proxy for chlorophyll concentration (Jerlov 1976). Recent advances in measurement techniques yield near real-time, global observations of chlorophyll concentration (SeaWiFS, MODIS) rendering Jerlov indices obsolete. Radiant heating parameterization that relate chlorophyll concentration and solar transmission are based on empirical ts of observed diuse attenuation coecient spectra and chlorophyll (Morel and Antoine 1994). Alternatively, parameters of equation (2) were determined by regressing chlorophyll concentration, solar zenith angle and cloud amount to solar radiation pro les obtained from coupled atmospheric and oceanic radiative transfer models (Ohlmann and Siegel 2000). However, clouds and solar zenith angle aect solar uxes only in the top 10 m of the ocean, and such detailed dependencies are not appropriate for coarse resolution climate models of today. Parameterizations of equation (2) with a single exponen12
tial de ned in terms of chlorophyll concentration and appropriate for climatic integrations are currently being developed. The sensitivity of upper ocean evolution to variations in solar transmission has long been investigated in a number of 1-D mixed layer modeling studies performed on diurnal to seasonal time scales. Denman (1973) ran simulations with the Kraus-Turner mixed layer model and showed that a proper solar transmission parameterization allowing solar penetration beyond the mixed layer base can give an increase in mixed layer depth of up to 70%. In addition, he found that he needed to double the light attenuation coecient, from Jerlov type II values, due to summertime increase in phytoplankton biomass and particulate organic matter, in order to obtain reasonable summer mixed layer depths during periods of low winds. Charlock (1982) changed solar transmission parameters from those for Jerlov I to Jerlov II and noted a corresponding change in SST that exceeded 1 C. Lewis et al. (1990) examined data from the central equatorial Paci c and found that up to 40 W m?2 penetrates the upper ocean mixed layer. These (and many other) studies show the importance of correctly parameterizing solar transmission in one dimensional mixed layer models but do not explicitly address the in uence of solar transmission on three dimensional climate scale variations (decadal and basin scale changes circulation patterns, for example). The vertical scale of ocean general circulation models (OGCMs) is suciently coarse that the deposition of solar energy can be parameterized in a simpler manner than for ner scale mixed layer models. OGCM's have occasionally handled radiant heating by depositing all the available solar energy uniformly within the topmost (mixed) layer. This scheme eliminates the need for computations associated with a solar transmission parameterization and is thought to introduce little error. Data for clear open-ocean waters indicate that more than 15% of the incident ux can pass beyond 20m, a typical OGCM layer thickness (Lewis et al. 1990; Siegel et al. 1995). Error due to neglect of solar uxes beyond the topmost layer can thus exceed 30 W m?2 in equatorial regions (assuming a 200 W m?2 surface ux). Employing a parameterization to calculate the solar ux at the base of the topmost model layer introduces only a small amount of computational overhead and can potentially give much more realistic OGCM model results. The solar transmission in uence on upper ocean evolution for the Indo-Paci c warm pool region was rst investigated in a three dimensional OGCM by Schneider at al. (1996). They suggest that subsurface radiant heating and surface cooling promotes the convective mixing that contributes to the maintenance of the vertical structure of the western Paci c warm pool. In-water solar ux data collected in the western equatorial Paci c during TOGACOARE indicate that 15 W m?2 penetrates beyond the mixed layer which supports the accuracy of the solar transmission formulation used in their model. 13
There are only a few studies that address the sensitivity of OGCM results to variations in solar transmission. These illustrate that biologically mediated solar transmission can signi cantly aect ocean temperature and currents, both locally and remotely. Schneider and Zhu (1998) investigated the eects of solar transmission beyond 30 m. Inclusion of solar penetration causes an increase in mixed layer depth by up to 30 m. The associated increased heat capacity decreases the annual SST cycle and annual mean SSTs both by as much as 1 C. Solar penetration increased the annual zonally averaged temperature near 50 m by as much as 5 C. The deeper mixed layer in the western equatorial Paci c with penetration causes a decrease in the sensitivity of SST to upwelling. Reduced amplitude in annual SST cycle o the equator leads to lighter easterly winds (by up to 0.2 dyn cm?2 ) in the equatorial region, ultimately reducing zonal currents (by more than 4 cm s?1), and eastern Paci c upwelling. Compared to observations the mean and annual cycle of SST are more realistic when solar penetration is considered. Ohlmann et al. (1996) use a chlorophyll-based solar transmission parameterization and show that transmission at 30 m is similar to Schneider and Zhu (1998). This implies a chlorophyll concentration of 0.2 mg m?3 which is more realistic for low-tomid latitude open ocean areas than is the chlorophyll concentration implied by complete absorption in the top 30 m. Nakamoto et al. (2001a) compared OGCM simulations with either xed or spatially and seasonally changing global ocean chlorophyll concentration. In the xed concentration run, absorption of solar radiation is parameterized according to Paulson and Simpson (1977) and corresponds to clear, Jerlov type I waters. In the changing concentration run the parameterization of Morel and Antoine (1994) is used instead, with observed Coastal Zone Color Scanner chlorophyll data (Feldman et al., 1989). This study addressed the effects of a space-time varying chlorophyll distribution on SST and mixed layer dynamics (see also Nakamoto et al., 2001b, for eects of living versus dead phytoplankton). The chlorophyll-dependent simulation yielded shallower mixed layer depths throughout most of the equatorial region (more solar energy trapped near the surface). A shallower mixed layer with decreased penetration of solar radiation is consistent with previous results (Schneider et al., 1996; Schneider and Zhu, 1998, Nakamoto et al., 2000). The simulation with explicit chlorophyll dependence produced SST values that are up to 2 C smaller in the eastern equatorial Paci c when compared with the results for Jerlov clear waters. This cooling is due to three-dimensional dynamical eects where mixed layer shoaling along the equator results in anomalous westward currents. The consequent increased westward surface ow indirectly increases upwelling and horizontal temperature advection that cools SST. This mechanism suggests that anomalous upwelling in the eastern equatorial Paci c is associated with the horizontal thermal gradient induced by chlorophyll pigments. The Nakamoto et al. (2001a) 14
modeling work shows that the circulation of the equatorial Paci c is quite sensitive to light penetration and more generally to processes that maintain the thermocline. Murtugudde et al. (2001) compare OGCM simulations with constant solar attenuation depths to simulations where the annual mean spatially variable attenuation depths are computed from the CZCS pigment data (Morel 1988). The study focuses on the chronic problem of a colder than observed cold tongue that plagues many state-of-the art coupled oceanatmosphere models and forced OGCMs. Murtugudde et al. (2001) argue that using an appropriate attenuation depth provides a heat trapping below the mixed layer that slightly weakens the strati cation, providing natural restrati cation of the water column. This leads to deeper mixed layers for the same wind generated TKE , to weaker surface currents, reduced divergence, and hence warmer SSTs. The problem of the colder than observed cold tongue can thus be remedied by using an accurate solar heating parameterization. The amplitude and patterns of the sensitivities will depend to some extent upon other feedbacks such as the o-equatorial SST in uence on equatorial zonal winds as pointed out by Schneider and Zhu (1998). The ocean's feedbacks to the atmosphere following chlorophyll induced SST changes have only recently been investigated. Absorption of solar radiation by phytoplankton can act to amplify the positive feedback between boundary layer stratiform cloudiness and SST. Norris and Leovy (1994) found summertime anomalies of boundary layer stratiform cloud amount and SST show a strong inverse correlation over much of the midlatitude and subtropical ocean. Decreased (increased) SST favors increased (deceased) cloud amount through modi cation of boundary layer processes and increased (decreased) cloud amount favors decreased (increased) SST by changing surface radiative uxes. Increased SST due to phytoplankton absorption would act to decrease cloud amount, increase transmitted radiation, and thus produce an additional SST increase. The phytoplankton SST anomaly patterns found by Nakamoto et al. (2001a) correspond to regions where cloud-SST anomaly correlations are greatest and likely most sensitive to changes in SST. K. Shell and colleagues have work in progress that uses SST (Figure 3) from the Nakamoto et al. (2001a) study to force an atmospheric GCM (CCM3). The primary in uence of decreased solar penetration and elevated SST on the atmosphere is a 0.3 C ampli cation of the seasonal cycle in the lowest atmospheric layer (Figure 4). Additionally, the mean atmospheric temperature is 0.05 C warmer when chlorophyll concentration is properly considered in determining SST. Atmospheric temperatures over land can change by 1 C. Additional studies of the atmospheric sensitivity to SST anomalies induced by ocean biology are clearly of interest. Even though surface albedo is generally small, as noted above it varies depending on water composition. In the open ocean, a change in chlorophyll concentration from 0.03 to 30 15
mg m?3 decreases surface albedo by about 0.005. The resulting change in re ected solar ux is about 1 W m?2 at the top of the atmosphere (Frouin and Iacobellis 2001). This value is not negligible compared with the radiative forcing since pre-industrial times due to greenhouse gases, which over the ocean is fairly homogeneous spatially with values in the range 1-3 W m?2 (e.g., Kiel and Briegleb 1993). Therefore, the distribution of phytoplankton, by modifying the radiation balance regionally, can aect predictions of climate change. If the phytoplankton abundance of the oceans should increase due to global warming, it would contribute to a warmer climate, not only by trapping more heat near the surface, but also by decreasing surface albedo. However, the relationship between phytoplankton abundance and surface albedo is non-linear, and a large increase in abundance from current levels would be required for the negative climate feed-back to be eective (Frouin and Iacobellis 2001). ii. Phytoplankton eects on DMS uxes Dimethylsul de is the most abundant form of volatile sulfur (S) in the ocean and is the main source of biogenic reduced S to the global atmosphere (Andreae and Crutzen 1997). The sea-to-air ux of S due to DMS is estimated to be in the range of 15-33 Tg S/yr which constitutes about 40% of the total atmospheric sulfate burden (Erickson et al. 1990; Chin and Jacob 1996). Once ventilated to the atmosphere, DMS is rapidly oxidized to form nonsea-salt sulfate (nss-SO42- ) and methanesulfonate (MSA) aerosols. A signi cant fraction of these aerosol particles may act as cloud condensation nuclei and may in uence the radiation budget of the atmosphere via both the direct and indirect eects. The alterations to the atmospheric radiation budget by DMS may play a role in climate variability at many time scales. Various species of phytoplankton produce diering amounts of dimethylsulfoniopropionate (DMSP), the precursor to DMS. In general, coccolithophorids and small agellates have higher intracellular concentrations of DMSP, which is thought to act as an osmolyte in the algal cell. Shaw (1983) and then Charlson et al. (1987) postulated links between DMS, atmospheric sulfate aerosols and global climate. It was hypothesized that an increase in biogenicly produced sulfate aerosols would lead to formation of more cloud condensation nuclei (CCN), and brighter clouds. This change in cloud microphysics could cool the earth's surface and thus stabilize climate against perturbations due to greenhouse warming. While phytoplankton are protagonists in this feed-back loop, recent advances in understanding suggest that it is the entire food web that determines net DMS production and not just algal taxonomy (Simo 2001). The proposed DMS-climate link, later called the CLAW hypothesis after the authors of the Charlson et al. (1987) paper, stimulated a urry of research in the 1990's and several hundred scienti c publications, but is still to be veri ed. Attempts to assess the direction 16
and magnitude of the DMS-climate feedback (Foley et al. 1991; Lawrence 1993; Gabric et al. 1998) in the context of global warming due to increased greenhouse gasses suggest the likelihood of a small, negative feedback (stabilizing), with magnitude of order 10%, and considerable uncertainty. These studies have all concluded that a feedback would occur over multi-decadal time-scales. They did not, however, attempt to link the spatial structures of interdecadal climate variations to regional alterations of DMS production by the ecosystem and hence to possible feedback loops. Moreover, seawater DMS time series long enough to enable an evaluation of the CLAW hypothesis on interdecadal time scales are non-existent. Bates and Quinn (1997) collated data from 11 cruises in the equatorial Paci c undertaken from 1982 to 1996 and reported that mean DMS levels during El Ni~no periods were not signi cantly dierent from those in normal years. Despite the major physical changes that occurred during the well-documented 1992 El Ni~no, the chemical and biological variability was small (Murray et al. 1994). Even though primary production decreased during the ENSO event, this appeared to be due to a reduction in the numbers of larger diatoms, which are not major DMS producers. In contrast to the Bates and Quinn (1997) study, Legrand and FenietSaigne (1991) found a good correlation between El Ni~no events and high MSA concentrations in south polar snow layers deposited over the 1922-1984 time period presumably due to enhanced DMS concentrations at high southern latitudes during El Ni~no years. Legrand and Feniet-Saigne (1991) suggest this could have been due to higher sea surface wind speed (implying increased sea-to-air exchange), or variations in sea-ice cover, which can aect ocean salinity and hence the osmotic balance in the algal cell for which DMSP is thought to have a regulating role. Analysis of an 8-year time series of atmospheric measurements at Cape Grim, Tasmania (41S, 145E), illustrates the strong seasonality in DMS, and has con rmed the connection between atmospheric DMS and aerosol sulfur species in this region (Ayers et al. 1991; Boers et al. 1994). A multi-decadal times series of MSA observations at Cape Grim (Figure 5) shows considerable interannual variability in the magnitude of the MSA peak, the strong seasonality and early January timing of the MSA maximum is remarkably consistent. In the absence of long-term oceanic time series, modeling can provide some insight into the potential for an interdecadal feedback. Gabric et al (2002) forced a regional DMS production model in the Subantarctic Southern Ocean with data on temperature, cloud, wind speed and mixed layer depth under enhanced greenhouse conditions derived from a coupled general circulation model. The GCM and DMS models were run in transient mode over the time period 1961-2080. The results showed considerable interdecadal variability in the annual integrated DMS ux (Figure 6), suggesting the potential for a signi cant DMS response to changes in the physical forcings. 17
As the 1976 climate shift occurred, it is plausible that the species assemblage shift that occurred during this time may have also resulted in changes in the distributions of the DMS producing species. This may have resulted in a dierent distribution of DMS related atmosphere particles interaction with atmospheric radiation after 1976 as compared to before. This is an example of unexplored biological consequences on physics after interdecadal climate changes. Note that another sulfur-containing biogenic gas, carbonyl sul de (OCS), is produced in the ocean from dissolved organic sulfur compounds. It is the most abundant sulfur gas in the background atmosphere. Ocean emission contributes to about 20% of the OCS sources (Andreae and Crutzen 1997). Like DMS, OCS can be oxidized to sulfur oxide gas and then to sulfate aerosols, but only in the mid- to lower stratosphere where wavelengths are suciently energetic to break the C-S bond (e.g., Chin and Davis 1995). This process has produced a veil of aerosols that re ects incoming solar radiation and helps to cool the planet, the socalled Junge layer. Since OCS is obtained from biological reactions at the surface and is not reactive in the troposphere, it may reside in the atmosphere for a long time. Overall chemical lifetime is 29 years. On a decadal time scale, changes in OCS emissions to the atmosphere due to changes in phyto- plankton biomass and marine ecosystems (e.g., in response to global warming), can aect the sulfate burden of the stratosphere, re ection of solar radiation back to space, and therefore climate. iii. Additional considerations Nutrient cycling changes may modulate the above two mechanisms because nutrients limit the growth of phytoplankton and because much of the Paci c is nutrient limited. The limiting nutrients include nitrate, phosphate, silicate and iron. Ocean physics can control the ux of these nutrients in to the regions where radiation eects and DMS uxes in uence climate variability and must therefore be accounted for on interdecadal timescales. The frequency and intensity of Asia dust storms (which carry iron to the open ocean) may also have some interdecadal signals that could alter ocean productivity on the this scale. Other, more subtle, eects may also come into play. The transfer velocity of gases at the air-sea interface is a function of sea surface turbulence (Wanninkho 1992) so there may be signi cant feedbacks with climate that are altered through this dependence on wind speed. Should the climate system change so that the surface wind speeds and air-sea interaction change, the transfer velocity would also change. An example would be that for a xed surface ocean concentration of DMS, an increase in the wind speed would increase the transfer velocity, in a non-linear way, and increase the ux of DMS from the ocean to the atmosphere. Ecosystems change the surfactants on the sea surface and hence modulate the wind stress magnitude. These eects, though, are probably much smaller than the ones already 18
discussed. It is unlikely that changes in CO2 in the atmosphere (and its consequent eect on radiation) due to changing oceanic ecosystems is important on interdecadal timescales. The reservoir of CO2 in the atmosphere is almost certainly far too large to be impacted by oceanic ecosystem CO2 ux or sequestration. Shifts in ENSO occurrence rates or deep water formation rates, however, which cause suciently large changes in the ecosystem may yet be shown to impact the atmospheric CO2 in a dynamically important way. The possible drawdown of nutrients (nitrate, N, and phosphate, P) in high-nutrient, low-chlorophyll (HNLC) regions, particularly in the Southern Ocean, and changes in the production of calcium carbonate are potentially major negative feedbacks on the accumulation of atmospheric CO2. Strati cation of the Southern Ocean, whether due to temperature or salinity eects (e.g., partial melting of Antarctic ice cap), would reduce the input of macronutrients to surface waters and give Aeolian iron inputs a chance to catch up with the input of N and P from upwelling. A complete drawdown via photosynthesis of excess N and P in the ocean could reduce atmospheric CO2 concentrations to 100-140 ppm (U.S. Global Change Research Program 1999). Since atmospheric CO2 concentrations are currently increasing at a rate of (3.3 ppm y?1, such a drawdown of excess N and P would oset the current rate of accumulation of CO2 in the atmosphere for 70-80 years. Studies with both coral reef communities and coccolithophorids have shown that a reduction in the saturation state of calcium carbonate due to the accumulation of CO2 in the surface water of the ocean will very likely decrease the rate of calcium carbonate production (Gattuso et al. 1998; Kleypas et al. 1999; Leclercq et al. 2000). Riebesell et al. (2000), for example, have shown that increasing the atmospheric CO2 concentration from 280 to 750 ppm reduces the calci cation rates of the coccolithophorids Emiliania huxleyi and Gephyrocapsa oceanica by 16-83%. However, the increase in seawater CO2 concentrations will also reduce the buer capacity of seawater, causing more CO2 to be released per CaCO3 precipitated. Furthermore, reduction in pelagic CaCO3 production may lead to less ecient ballasting and ultimately burial of exported organic carbon. While a reduction in CaCO3 production will exert a negative feedback on the accumulation of CO2 in the atmosphere, the concomitant reduction in buer capacity of surface seawater and ballasting of exported organics will reduce the impact of this feedback to an extent that is unclear at this time. A major uncertainty in quantifying the feedback between biological systems and their environment is the resiliency and adaptability of biological communities due to their genetic diversity, phenotypic plasticity and evolutionary potential. Eorts to develop a theoretical understanding of the manner in which ecosystems adapt to environmental change date from the work of Lotka (1922), who postulated that natural selection tends to maximize the energy ux through a system, at least within the constraints to which the system is subject. 19
Odum (1983) expanded on Lotka's theory. He argued that natural systems tend to maximize power, and that theories and corollaries derived from the maximum power principle could explain much about the structure and processes of these systems. In rationalizing the application of the maximum power principle to ecosystems, Odum drew analogies between ecosystem behavior and the laws of thermodynamics. A number of authors have explored the application of analogues of thermodynamic principles to the behavior of natural systems (Jorgensen 2000; Jorgensen and Straskraba 2000), and in many cases these thermodynamic approaches have met with considerable success in estimating parameters to describe real ecosystems. In a recent paper Cropp and Gabric (2001) employed a genetic algorithm to simulate the evolutionary response of the biota of a model ecosystem. The model ecosystem consisted of a simple autotroph-herbivore-nutrient oceanic mixed-layer. One of the interesting results of the simulations was that the optimal parameter values proved to be very insensitive to the choice of selection pressure. In particular, the simulations suggested the hypothesis that within the constraints of the external environment and the genetic potential of their constituent biota, ecosystems evolve to the state most resilient to perturbation. Laws et al. (2000) have applied the hypothesis of maximum resilience to a more complex food web model of an open-ocean pelagic ecosystem, similar to that of Cropp and Grabric (2001). Most of the parameter values were chosen from information in the literature or were otherwise constrained in a deterministic manner. Two parameters, however, were allowed to adapt so as to maximize the resiliency of the steady state solution. These two adaptive parameters were the relative growth rate (sensu Goldman 1980) of the large phytoplankton and the biomass of the lter feeders. In this case, it was possible to compare the predictions of the model with results of eld studies carried out as a part of the Joint Global Ocean Flux Study and related work (Figure 7). Because the system was assumed to be in steady state, the export ratio equals the f ratio (Eppley and Peterson 1979) and was designated the ef ratio. The predicted ef ratios based on the principle of maximum resiliency are in remarkable agreement with observed ef ratios, and there is likewise remarkable agreement between predicted and observed heterotrophic bacterial biomass. These comparisons clearly support the assumption that pelagic marine ecosystems tend to evolve toward a condition of maximum resiliency, as predicted by the results of the Cropp and Gabric (2001) simulations. Relevant to climate feedback is the fact that the model ef ratio is negatively correlated with temperature. Hence an increase in the temperature of the surface waters of the ocean due to global warming would lead to less ecient export of organic matter to the interior of the ocean. More of the organic matter that was exported would probably take the form of dissolved organic carbon, and the eciency of the biological pump would be reduced. Thus 20
the response of the ef ratio to global warming would amount to a positive feedback on the climate system. The drawdown of macronutrients (N and P) in HNLC regions, particularly the southern ocean, and changes in ballasting of exported organics (calcium carbonate versus silica) are possibly major feedbacks. The most likely mechanism leading to a switch from calcium carbonate to silica would be a reduction in pH. However, diatoms cannot make silica without silicate, so there is a limit on how far that transition can go.
4. Coupled physical-biological eects in Paci c interdecadal variability a. Stochastic excitation Even in the context of stochastic excitation of oceanic decadal variability, biological feedbacks may be important. For example, in the simplest framework of thermal forcing of SST anomalies (e.g., Hasselmann 1976; Barsugli and Battisti 1998), the strength of SST response may be modulated by the ecosystem if it exerts a feedback on the SST (even without considering SST in uencing the atmosphere). For negative (positive) feedbacks, the spectrum of oceanic SST response becomes less (more) red. If one considers an SST eld cooled by upwelling of nutrient rich waters, then an increased concentration of phytoplankton driven by the nutrient injection would warm SST through increased surface layer absorption and be a negative feedback. If one considers an SST eld driven by surface heat uxes, then warmer SST could increase the phytoplankton growth rates, increase surface layer absorption, and lead to a positive feedback. Obviously, dierent regions of the Paci c could be controlled by any of a number of local scenarios. b. The Subduction Mode The Subduction Mode could be in uenced by biological processes in several ways (Figure 8). The eect of upwelling on tropical Paci c SST might include a response to the changing ecosystem so that the amplitude of tropical SST anomalies would be changed in the presense of biology (cf. Murtugudde et al. 2001; Nakamoto et al. 2001a). This could, in turn, aect the strength of the atmospheric teleconnections to the midlatitude. But the atmospheric teleconnections could also be in uenced by DMS in uence on CCNs. The midlatitude SST anomalies could also be altered by biological feedbacks including phytoplankton radiation absorption and local DMS eects on cloud formation and atmospheric response. If water masses that are subducted from the midlatitudes or subtropics into the tropical upwelling zones contain nutrients that limit the growth of tropical ecosystems, then the strength of tropical biological response can be modulated as well. Recent studies suggest that the subduction eect is particularly signi cant at multi-decadal time scales (e.g. 50 years or longer) (Shin and Liu 2000). Therefore, it is possible that the coupled biological eect for 21
this mechanism would also play a more important role at multi-decadal time scales. These possibilities depend, however, on the ability of the biology to signi cantly aect the climate system either directly, e.g. by changing the pro le of heating in the upper ocean, or indirectly, e.g. by aecting the ux of radiation by via changes of clouds. Additionally, properties of subducted waters are believed to be set during times of deepest mixed layer formation during strong storms (Stommel daemon). It remains to be seen if variations of the nutrients at these source regions or changes in buildup of nutrients in the thermocline during the equatorward transit control the concentrations of waters upwelled at the equator. The ecosystem is likely a negative feedback in the equatorial Paci c (Murtugudde et al. 2001). When the tropical trades begin to strengthen, stronger upwelling will lead to surface blooms which will trap heat and reduce the SST gradient and consequently weaken the winds. When the trades get weaker, reduced upwelling, a relaxed thermocline and nitracline, and reduced entrainment can lead to enhanced subsurface warming by weakening the strati cation, deepening the mixed layer, reducing the divergence and upwelling further. Since the zonal winds along the equator have strong correlations to o-equatorial SSTs (Schneider and Zhu 1998), the equatorial ecosystem eects on circulation may aect subtropical subduction zones. c. The Midlatitude Gyre Mode The Midlatitude Gyre Mode may be in uenced by biological processes as well (Figure 8). The Kuroshio-Oyashio Extension (KOE) region is where the atmosphere is most sensitive to SST anomalies in uncoupled atmospheric models (Peng et al. 1997) and in full-physics coupled models (Schneider et al. 2002). Since ocean dynamics control the SST in the KOE on interdecadal timescales (e.g., Schneider and Miller 2001), changes in the ecosystem in that region could in uence a coupled feedback loop in two ways. First, the amplitude of the SST anomaly may be altered by the phytoplankton radiation absorption eect. This may be a negative feedback if upwelling (which drives the cold phase of SST) increases phytoplankton absorption eects. Secondly, the atmospheric response which completes the gyre mode feedback loop could potentially be aected by the ux of DMS into the atmosphere, its eect on clouds, and the ow of energy. Such a feedback is rather hypothetical at this point but potential eects are unknown and need to be explored in this context. The seasonality of the physical and biological processes are particularly important. The physical processes associated with the gyre mode are dominated by winter conditions, while biological activity and eect on solar radiation peaks during the spring portion of the year. d. The re-emergence mechanism There are several ways in which biological processes could interact with the reemer22
gence mechanism and the atmospheric bridge. Both physical processes in uence SST and MLD, which aect primary productivity and thus the amount of light absorbed in the water column. The latter, can in turn, feed back upon the temperature pro le of the upper ocean and can, during years with low productivity and solar absorption partially destroy temperature anomalies sequestered below the seasonal thermocline. In contrast, high solar absorption insulates the deeper layers and should increase the year to year persistence of winter temperature anomalies. In addition to its in uence on temperature, the reemergence process may alter the seasonal evolution of other quantities such as nutrients or plankton. For example, nutrient rich (or poor) water sequestered below the mixed layer at the end of one winter may return in the subsequent fall/winter. Biological processes could also aect the atmospheric bridge (a remote forcing in this context) indirectly by changing the amplitude and/or frequency of SST anomalies in the equatorial Paci c by changing the oceanic absorption or directly by changing the uxes of DMS into the tropical atmosphere. While it is likely that feedbacks between ocean biology and the reemergence-atmospheric-bridge process in uence climate variability, the strength and importance of the eect is unclear. It is unlikely that the result would lead to oscillations with a preferred decadal period. e. Indirect biology/physical feedback mechanisms Biology may introduce a redistribution of heat by aecting absorption in the ocean column, for instance, or by altering the existence or characteristics of another component of the system which itself interacts with the heat budget (e.g. as in the proposed DMS/cloud eect). Other indirect feedback eects may also occur; for instance, a mechanism whereby changes in temperature aect biological processes which in turn aect the ice albedo and/or the albedo of the ocean at the ice margin. The albedo eect of the biology would be enhanced by the strong physical ice/albedo feedback itself so that the biological eect would be magni ed by the physical eect. Biological modi cation/enhancement of strong physical feedback mechanisms including remote eects oer a way of amplifying eects and may well provide the most likely route for important biological/physical coupling. f. Non-linear tropical modes If tropical decadal variability results from the nonlinear interaction of the seasonal cycle and ENSO (Jin et al. 1994; Tziperman et al. 1994), the documented sensitivity of the former to changes of absorption (Schneider and Zhu 1998; Murtugudde et al. 2001) might aect the decadal time scales. Assuming that without biology, ENSO and the annual cycle operate in a frequency locked regime, say the ENSO frequency is 4 years, a modi cation of the amplitude of the annual cycle by about 0.3 C might lead to a break up of the frequency locked ENSO state and to chaotic ENSO oscillations. The existence of chaotic ENSO variability implies 23
that on decadal and interdecadal timescales variability is excited. Hence, the modi cation of the annual cycle due to biology might generate changes in the amplitude of decadal variability. How such changes feed back onto the biology, to form a coupled physical-biological feedback loop, is unclear. g. Synopsis Oceanic ecosystems may in uence previously hypothesized interdecadal climate variability mechanisms in the Paci c. Biological processes do not appear capable of establishing a timescale via introducing a delayed feedback in these modes, however. Instead, biology is likely to sensitize or de-sensitize the physical ocean-atmosphere interactions, mainly through altered SST through phytoplankton-radiation eects and altered atmospheric dynamics through DMS uxes. Determining the strength and importance of these potential feedbacks requires further study. The equatorial region is a key area requiring particular attention. The large incident solar radiation and importance of solar absorption oers the possibility that biological mechanisms aect the physical environment directly. In turn, the strong atmospheric response to SST anomalies in the tropical region (and its role in the Subduction Mode) provides the possibility of closing a feedback loop. The Kuroshio-Oyashio Extension region (the heart of the Midlatitude Gyre Mode) is another hot spot where midlatitude ocean-to-atmosphere feedbacks are possible. In both regions, the eects of phytoplankton on upper-ocean absorption and changes in the atmospheric response due to DMS eects on cloudiness are the postulated mechanisms, although their strengths and eectiveness are currently unknown. The possible ampli cation of small biological perturbations by powerful physical feedback mechanisms, such as the ice/snow albedo feedback, give the possibility also of high latitude eects, although no obvious mechanisms have been proposed.
5. Research Challenges
Given the complexity of the physical-biological feedback problem, we now discuss what lines of research are best followed to better understand these eects in interdecadal Paci c climate variations. a. Modeling strategies The identi cation of biological mechanisms of potential importance using local or column atmosphere, ocean, and ecosystem models of varying complexity (both coupled and uncoupled) is clearly the rst step in the process. Some appreciable biologically mediated perturbation to the radiative balance, either directly or through local interactions with strong physical feedback mechanisms, is required and this can best be tested with such local models. Since it is undoubtedly the case that a range of climate related biological-physical interac24
tions remains unrecognized, it is particularly important to be sensitive to the possibility of such interactions and to test them in such models. Oceanic biological models now range from simple single currency nitrate-phytoplanktonzooplankton model to multiple currency models that have scores of degrees of freedom. It is dicult to validate the more complicated models due to a lack of observations. But it would be ideal to intercompare these models both in uncoupled mode and in coupled column model to assess the necessity of including a large number of dependent variables insofar that they aect the structure of radiative and DMS response for instance. Variations in upper ocean phytoplankton biomass can promote changes in solar transmission within the water column, which can subsequently impact climate. Climate models must adequately represent in-water solar ux divergences if accurate simulations are to be performed. While a sophisticated representation is computationally impracticable, a simple single exponential solar transmission parameterization based on chlorophyll concentration is a necessary component of climate models hoping to produce accurate results. Such a parameterization should give transmission values with less than 10% error for depths beyond 10 m (Ohlmann et al. 1998). The description of solar transmission in modeling studies is often limited to a statement of solar transmission parameter values with little regard for what the values actually mean in terms of upper ocean chlorophyll concentration. For investigating biological feedbacks to climate it is necessary to think rst of changes in phytoplankton biomass, followed by changes in the corresponding solar transmission parameters rather than bypassing any thought or mention of chlorophyll. This paradigm is necessary if the in uence of chlorophyll concentration on climate evolution and associated feedback mechanisms are to be adequately investigated. Local models necessarily lack the dynamical aspect which redistributes local perturbations throughout the system and which allows new modes of variation and new feedback mechanisms to exist. The cost and complexity of integrating dynamical models of atmosphere and ocean for the hundreds of years needed to investigate interdecadal climate variations means that they must be operated at comparatively coarse resolution with grid sizes ranging from several hundreds of km for global models to the order of tens of km for regional models. At these larger scales, subgrid scale physical and biological processes must be parameterized in terms of the grid scale variables available in the large-scale models. The second step is thus the parameterization of interacting physical-biological processes suitable for use in global and regional coupled models. Uncoupled global and regional models are used to determine the sensitivity of the individual components in the subsequent step. For instance, the sensitivity of the atmosphere to changes in the DMS uxes due to ocean biology may be investigated using observed es25
timates of the the DMS ux in an uncoupled atmospheric model. The atmospheric model results re ect the eects of this DMS ux in the context of the model's atmospheric physics and dynamics and results can be compared against a simulation made without DMS uxes. In a similar vein, an ocean-ecosystem model could be forced with observed atmospheric forcing over decadal timescales and the results qualitatively and quantitatively veri ed against available observations with special attention to the key regions of interdecadal interaction (the tropical Paci c and the KOE region). The nal step is to carry out and analyze long term multi-century simulations with fully coupled atmosphere-ocean-ecosystem models. The global climate system participates in interdecadal variability and it may not be suitable to isolate parts of the Paci c using regional models, although this remains a possibility depending on the investigation. Clearly the signal-to-noise ratio of the bio-ocean-atm eects must be assessed by simulations with and without ocean biology. Follow-up runs should include successively more sophisticated biology to determine the importance of re nements and the conditions for importance. An important goal of the coupled biological/physical models will be to assess the potential of HNLC regions to draw down excess N and P in response to changes in physical circulation and to quantify the impact on the biological and solubility pumps of the combined eect of reduced calci cation, reduced buer capacity, and reduced ballasting of exported organics caused by increases in upper ocean CO2 concentrations. Biological models targeting changes that may occur on an interdecadal time scale must allow for the evolutionary potential of biological communities, and particularly of organisms such as plankton with short generation times. As noted by Hutchinson (1967, p. 376), "The annual cycle [of a plankton community] is thus, in terms of generation time, the equivalent of perhaps up to ten thousand years in the successional history of....some forest trees." Resilience of biological communities appears to be a common consequence of evolutionary pressures. It is quite possible that in a decade or so, the advent of computational power will allow integration of fully coupled ocean-atmosphere-biogeochemical models for routine climate forecasts. The models and the technology exists currently to quantify the potential feedback of the ecosystem to the coupled climate variability on ENSO time-scales. Is there a amplitude, frequency, or phase modulation during the preconditioning, onset, evolution, and the decay phase of the ENSO events? What are the feedbacks of the equatorial/tropical ecosystems to extratropical climate variability and vice versa? b. Observational strategies Sustained long-term observations of biological and physical variables in key regions of the Paci c Ocean will be necessary to determine the importance of biological feedbacks on interdecadal climate variations. These will will also help clarify whether there are truly 26
preferred timescales of climate variability in the Paci c. This eort will require commitments by the international community. For example, CalCOFI has been observing biology and physics for 50 years, but this region turns out not be be a key region of ocean-atmosphere interaction. Our knowledge now is advanced to the point that we can make a good guess where the hot spots are (while still allowing for as yet unrecognized possibilities). At least two potential hot spots of biological response and physical sensitivity have been identi ed. The primary region is the tropical Paci c, because of the large ow of energy that can be perturbed by the biology and its strong dynamical response to change. The second region is Kuroshio-Oyashio Extension, because of its central importance in midlatitude ocean-atmosphere feedbacks. These two key regions require long-term measurements of the crucial biological variables, such as chlorophyll and DMS uxes, as well as basic physical oceanographic variables (temperature, salinity, and horizontal currents) and basic atmospheric variables (pressure, winds, clouds and surface heat uxes). Additional resources would also be useful in measuring supplementary biological variables such as nutrient concentrations (NO3 , Fe, SiO4), biomass, productivity, vertical distributions, zooplankton, sedimentation and CO2 in order to better elucidate the fundamental biological processes. Existing remote sensing strategies that measure ocean color, surface temperature, winds, humidity, sea level, etc., from satellites and subsurface temperature, salinity and velocity from pro ling oats will prove invaluable to the long-term study of these coupled biologicalphysical climate processes. New satellite strategies also need to be devised to measure large scale phytoplankton species distributions and aerosol characteristics, including their relation to cloud microphysical properties and other atmospheric sensitivities. Oceanic biology is rich in fascinating detail, and marine ecosystems can exhibit extremely complex dynamical behavior. Naturally, there is no end to the list of observations that biologists would consider essential to document biotic responses and feedbacks to the climate system. Thus a key element in strategies for the future must be the use of the coupled models in identifying gaps in our knowledge of how the coupled biological-physical systems function, and in identifying the variables and parameters to which the models are most sensitive. Only in this way can we hope to optimize our observing strategy in what will still be an undersampled system.
Biological processes are largely absent in the collection of global coupled climate models used to study climate and climate change in the recent IPCC assessment (e.g. Table 9.1, Chapter 9, IPCC, 2001). In these models, biology does not interact with other climate variables to determine the variability of climate nor its change under speci ed forcing changes. This situation is rapidly changing and the representation of basic biological processes on 27
land and in the oceans will be part of future climate variability and change investigations. Here we have attempted to assess the potential role of marine ecosystems in interdecadal climate variation. In particular, we have asked if biological processes are important for in uencing climate variability (i.e. how biological processes engender, enhance, and/or suppress internally generated climate variability or externally forced change). We concentrated our attention on the Paci c Ocean ecosystem and on interdecadal climate variation and change, but the problem can be extended globally. In summary, Figure 9 displays many of the key processes that were discussed in this paper and that need to be better understood. This gure illustrates the relationships of varying certainty between boundary layer cloud cover, SST, ocean strati cation, and plankton concentration. Arrows indicate the direction of forcing. The sign indicates whether an increase in the forcing parameter causes an increase (+) or decrease (-) in the forced parameter. (?) indicates those processes which especially need more investigation. Increased plankton concentration could cause increased sea surface temperature and ocean strati cation by absorbing solar radiation closer to the surface. Increased sea surface temperature would cause boundary-layer stratiform cloud cover to decrease and would thus allow more solar radiation to reach the ocean. This set of interactions would act as a local positive feedback on the climate system. Plankton concentration could aect cloud cover directly through DMS production but this needs to be better quanti ed. The sensitivity of plankton concentration to light level changes due to clouds is uncertain. The sensitivity of plankton concentration to changes in nutrient mixing due to ocean strati cation is also uncertain. We have outlined the modeling and observational strategy that will be necessary to guide future investigations of the fascinating, but as yet unclear, possibility that biological-physical interactions play a role in establishing interdecadal climate variability. Acknowledgments. This synopsis was conceived during the Surfside Climate Workshop on \Climate Forcing of Oceanic Ecosystems: Are Signi cant Biological Feedbacks Possible on Interdecadal Timescales?" held April 18-20, 2001, in La Jolla, California. We are grateful to the Director of the Scripps Institution of Oceanography for generous funding, as well as to NSF (OCE00-82543) and NASA (NAG5-9788) for additional funding.
Alexander, M. A., and C. Deser, 1995: A mechanism for the recurrence of wintertime midlatitude SST anomalies. J. Phys. Oceanogr., 25, 122-137. Alexander, M. A., C. Deser, and M. S. Timlin, 1999: The reemergence of SST anomalies in the North Paci c Ocean. J. Climate, 12, 2419-2433. Alexander, M. A., J. D. Scott, and C. Deser, 2000: Processes that in uence sea surface temperature and ocean mixed layer depth variability in a coupled model J. Geophys. 28
Res., 105, 16823-16842. Alexander, M. A., M. S. Timlin, and J. D. Scott, 2001: Winter-to-winter recurrence of sea surface temperature, salinity and mixed layer depth anomalies. Prog. Oceanogr., 49, 41-61. Andreae, M. O., and P. J. Crutzen, 1997: Atmospheric aerosols: Biogeochemical sources and role in atmospheric chemistry. Science, 276, 1052-1058. Ayers, G. P., J. P. Ivey, and R. W. Gillett, 1991: Coherence between seasonal cycles of dimethylsul de, methanesulphonate and sulphate in marine air. Nature, 349, 404-406. Barsugli, J. J., and D. S. Battisti, 1998: The basic eects of atmosphere-ocean thermal coupling on midlatitude variability. Journal of the Atmospheric Sciences, 55, 477-493. Bates, T. S., and P. K. Quinn, 1997: Dimethylsul de (DMS) in the equatorial Paci c Ocean (1982 to 1996): Evidence of a climate feedback? Geophys. Res. Lett., 24, 861-864. Baumgartner, T. R., A. Soutar, and V. Ferreira-Bartrina, 1992: Reconstruction of the history of Paci c sardine and Northern anchovy populations over the past 2 millennia from sediments of the Santa-Barbara Basin, California. Calif. Coop. Ocean. Fish. Invest. Rep., 33, 24-40. Beamish, R. J., and D. R. Bouillon, 1993: Paci c salmon production trends in relation to climate. Can. J. Fish. Aquat. Sci., 50, 1002-1016. Boer, G. J., G. Flato, M. C. Reader, and D. Ramsden, 2000: A transient climate change simulation with greenhouse gas and aerosol forcing: experimental design and comparison with the instrumental record for the twentieth century. Clim. Dyn., 16, 405-425. Boers, R., G. P. Ayers, and J. L. Gras, 1994: Coherence between seasonal variation in satellite-derived cloud optical depth and boundary-layer CCN concentrations at a midlatitude Southern hemisphere station. Tellus B, 46, 123-131. Bograd, S. J., and R. J. Lynn, 2001: Physical-biological coupling in the California Current during the 1997-99 El Ni~no-La Ni~na cycle. Geophys. Res. Lett., 28, 275-278. Brodeur, R. D., and D. M. Ware, 1992: Interannual and interdecadal changes in zooplankton biomass in the subarctic Paci c Ocean. Fish. Oceanogr., 1, 32-38. Chai, F., R. C. Dugdale, T.-H. Peng, F. P. Wilkerson, and R. T. Barber, 2001: One dimensional ecosystem model of the equatorial Paci c upwelling system, Part I: Model development and silicon and nitrogen cycle. Deep-Sea Res., in press. Chang, P., Link, Ji, L., Hong, L. and Flugel, M., 1996: Chaotic dynamics versus stochastic Process in El Ni~no-Southern Oscillation in Coupled Ocean-Atmosphere Models. Physica D, 98, 301-320. Charlock, T. P., 1982: Mid-latitude model analysis of solar radiation, the upper layers of the sea, and seasonal climate. J. Geophys. Res., 87, 8,923-8,930. 29
Charlson, R. J., J. E. Lovelock, M. O. Andreae, and S. G. Warren, 1987: Oceanic phytoplankton, atmospheric sulphur, cloud albedo and climate. Nature, 326, 655-661. Chin, M., and D. D. Davis, 1995: A reanalysis of carbonyl sul de as a source of stratospheric background sulfur aerosols. J. Geophys. Res., 100, 8993-9005. Chin, M., and D. J. Jacob, 1996: Anthropogenic and natural contributions to tropospheric sulfate: a global model analysis. J. Geophys. Res., 101, 18691-18699. Conversi, A., and S. Hameed, 1997: Evidence for quasi-biennial oscillations in zooplankton biomass in the subarctic Paci c J. Geophys. Res., 102, 15659-15665. Cox, P. M., R. A. Betts, C. D. Jones, S. A. Spall, and I. J. Totterdell, 2000: Acceleration of global warming due to carbon-cycle feedbacks in a coupled climate model. Nature, 408, 184-187. Cropp, R., and A. Gabric, 2001: Ecosystem adaptation: Do ecosystems maximize resilience? Ecology, in press. Denman, K. L., 1973: A time-dependent model of the upper ocean. J. Phys. Oceanogr., 3, 173-184. Denman, K., E. E. Hofmann, and H. Marchant, 1996: Marine biotic responses to environmental change and feedbacks to climate. In: Climate Change 1995: The Science of Climate Change, J. T. Houghton, L. G. Meira Filho, B. A. Callander, N. Harris, A. Kattenberg and K. Maskell (Eds.), Intergovernmental Panel on Climate Change, Cambridge University Press, Cambridge, 483-516. Deser, C., and M. Blackmon, 1995: On the relationship between tropical and North Paci c sea surface temperature variations, J. Climate, 8, 1677-1680. Deser, C., M. A. Alexander, and M. S. Timlin, 1996: Upper ocean thermal variations in the North Paci c during 1970-1991. J. Climate, 9, 1840-1855. Emerson, S., S. Mecking, J. Abell, 2001: The biological pump in the subtropical North Paci c Ocean: Nutrient sources, Red eld ratios, and recent changes. Glob. Biogeochem. Cyc., 15, 535-554. En eld, D. B., and A. M. Mestas-Nun~nez, 1999: Multiscale variabilities in global sea surface temperatures and their relationships with tropospheric climate patterns. J. Climate, 12, 2719-2733. Eppley, R. W., and B. J. Peterson, 1979: Particulate organic-matter ux and planktonic new production in the deep ocean. Nature, 282, 677-680. Erickson, D. J. III, S. Ghan and J. Penner, 1990: Global ocean to atmosphere dimethyl sul de ux. J. Geophys. Res., 95, 7543-7552. Feldman, G., N. Kuring, C. McClain, J. Elrod, N. Maynard, and D. Endres, 1989: Ocean color - availability of the global data set. EOS, Trans., AGU, 70, 634-641. 30
Field, D. B., and T. R. Baumgartner, 2000: A 900-year stable isotope record of interdecadal and centennial change from the California Current. Paleooceanogr., 15, 695-708. Foley, J. A., K. E. Taylor, and S. J. Ghan, 1991: Planktonic dimethylsul de and cloud albedo - An estimate of the feedback response. Climatic Change, 18, 1-15. Francis, R. C., and S. R. Hare, 1994: Decadal scale regime shifts in the large marine ecosystems of the Northeast Paci c: a case for historical science. Fish. Oceanogr., 3, 279-291. Francis, R. C., S. R. Hare, A. B. Hollowed, and W. S. Wooster, 1998: Eects of interdecadal climate variability on the oceanic ecosystems of the NE Paci c. Fish. Oceanogr., 7, 1-21. Frankignoul, C., P. Muller, and E. Zorita, 1997: A simple model of the decadal response of the ocean to stochastic wind forcing, J. Phys. Oceanogr., 27, 1533-1546. Freeland, H., K. Denman, C. S. Wong, F. Whitney, and R. Jacques, 1997: Evidence of change in the winter mixed layer in the Northeast Paci c Ocean. Deep-Sea Res., 44, 2117-2129. Frouin, R., and S. F. Iacobellis, 2001: In uence of phytoplankton on the global radiation budget. J. Geophys. Res., sub judice. Gabric, A. J., P. H. Whetton, R. Boers, G. P. T. Ayers, 1998: The impact of simulated climate change on the air-sea ux of dimethylsulphide in the subantarctic Southern Ocean. Tellus B, 50, 388-399. Gabric, A. J., P. Whetton, and R. Cropp, 2002: Dimethylsulphide production in the subantarctic Southern Ocean under enhanced greenhouse conditions. Tellus, in press. Gattuso, J-P., M. Frankignoulle, I. Bourge, S. Romaine, and R. W. Buddemeier, 1998: Eect of calcium carbonate saturation of seawater on coral calci cation. Global Planet. Change, 18, 37-46. Goldman, J. C., 1980: Physiological processes, nutrient availability, and the concept of relative growth rate in marine phytoplankton ecology, In: Primary Productivity in the Sea, P. G. Falkowski (Ed.), Plenum, New York, 179-194. Gu, D. F. and S. G. H. Philander, 1997: Interdecadal climate uctuations that depend on exchanges between the tropics and extratropics. Science, 275, 805-807. Guilderson, T. P., and D. P. Schrag, 1998: Abrupt shift in subsurface temperatures in the Tropical Paci c associated with changes in El Ni~no. Science, 281, 240-243. Haigh, S. P., K. L. Denman, and W. W. Hsieh, 2001: Simulation of the planktonic ecosystem response to pre- and post-1976 forcing in an isopycnic model of the North Paci c. Can. J. Fish. Aquat. Sci., 58, 703-722. Hare, S. R., and N. J. Mantua, 2000: Empirical evidence for North Paci c regime shifts in 1977 and 1989 Prog. Oceanogr., 47, 103-145. 31
Hasselmann, K., 1976: Stochastic climate models. Part I. Theory. Tellus, 28, 473-485. Hutchinson, G. E. 1967. A Treatise on Limnology. Vol II. Wiley, New York. 1115 pages. IPCC, 2001: Climate Change 2001: The Scienti c Basis. Contribution of Working Group I to the Third Assessment Report of the Intergovernmental Panel on Climate Change (IPCC), J. T. Houghton, Y. Ding, D. J. Griggs, M. Noguer, P. J. van der Linden, and D. Xiaosu (Eds.), Cambridge University Press, UK, 944 pp. Jerlov, N. G., 1976: Marine Optics, Elsevier, 231 pp. Jin, F.-F., J. D. Neelin, M. and Ghil, 1994: El Ni~no on the Devil's Staircase: Annual subharmonic steps to chaos. Science, 264, 70-72. Jin, F.-F. 2001: Low frequency modes of tropical ocean dynamics. J. Climate, 14, 3874-3881. Jorgensen, S. E., 2000: A general outline of thermodynamic approaches to ecosystem theory. In: Handbook of Ecosystem Theories and Management. Jorgensen, S. E. and F. Muller (Eds.), CRC Press, Boca Raton, Florida. Jorgensen, S. E., and M. Straskraba, 2000. Ecosystems as cybernetic systems. In: Handbook of Ecosystem Theories and Management. Jorgensen, S. E. and F. Muller (Eds.), CRC Press, Boca Raton, Florida. Kiehl, J. T., and B. P. Briegleb, 1993: The relative roles of sulfate aerosols and greenhouse gases in climate forcing. Science, 260, 311-314. Kleeman R., J. P. McCreary and B. A. Klinger, 1999: A mechanism for the decadal variation of ENSO. Geophys. Res. Lett., 26, 1743-1746. Kleypas, J. A., R. W. Buddemeier, D. Archer, J-P. Gattuso, C. Langdon, and B. N. Opdyke, 1999: Geochemical consequences of increased atmospheric carbon dioxide on coral reefs. Science, 284, 118-120. Latif, M., 1998: Dynamics of interdecadal variability in coupled ocean-atmosphere models. J. Climate, 11, 602-624. Latif, M. and T. P. Barnett, 1994: Causes of decadal climate variability over the North Paci c and North America. Science, 266, 634-637. Lau, N. C., and M. J. Nath, 1996: The role of the \atmospheric bridge" in linking tropical Paci c ENSO events to extratropical SST anomalies. J. Climate, 9, 2036-2057. Lawrence, M. G., 1993: An empirical-analysis of the strength of the phytoplanktondimethylsul de-cloud-climate feedback cycle. J. Geophys. Res, 98, 20663-20673. Laws, E. A., P. G. Falkowski, W. O. Smith, H. Ducklow, and J. J. McCarthy, 2000: Temperature eects on export production in the open ocean. Global Biogeochem. Cycles, 14, 1231-1246. Leclercq, N., J. Gattuso, and J. Jaubert, 2000: CO2 partial pressure controls the calci cation rate of a coral community. Global Change Biology, 6, 329-334. 32
Legrand, M., and C. Feniet-Saigne, 1991: Methanesulfonic acid in south polar snow layers: A record of strong El Ni~no. Geophys. Res. Lett., 18, 187-190. Lewis, M. R., 1987: Phytoplankton and thermal structure in the tropical ocean. Oceanologica Acta, 6, 91-95. Lewis, M. R., J. J. Cullen, and T. Platt, 1983: Phytoplankton and thermal structure in the upper ocean: Consequences of nonuniformity in the chlorophyll pro le. J. Geophys. Res., 88, 2565-2570. Lewis, M. R., M.-E. Carr, G. C. Feldman, W. Esaias, and C. McClain, 1990: In uence of penetrating solar radiation on the heat budget of the equatorial Paci c Ocean. Nature, 347, 543-545. Limsakul, A., T. Saino, T. Midorikawa, and J. I. Goes, 2001: Temporal variations in lower trophic level biological environments in the northwestern North Paci c Subtropical Gyre from 1950 to 1997. Prog. Oceanogr., 49, 129-149. Liu, Z. Y., and S. I. Shin, 1999: On thermocline ventilation of active and passive tracers. Geophys. Res. Lett., 26, 357-360. Liu, Z., L. Wu, R. Gallimore and R. Jacob, 2001: Dual origins of Paci c decadal climate variability. Geophys. Res. Lett., sub judice. Lotka, A. J., 1922: Contribution to the energetics of evolution. Proc. Natl. Acad. Sci., 8, 147-150. Mackas, D. L., R. Goldblatt, and A. G. Lewis, 1998: Interdecadal variation in developmental timing of Neocalanus plumchrus populations at Ocean Station P in the subarctic North Paci c. Can. J. Fish. Aquat. Sci., 55, 1878-1893. Mann, M. E. and J. Park, 1996: Joint spatiotemporal modes of surface temperature and sea level pressure variability in the Northern Hemisphere during the last century. Journal of Climate, 9, 2137-2162. Mantua, N. J., S. R. Hare, Y. Zhang, J. M. Wallace, and R. C. Francis, 1997: A Paci c interdecadal climate oscillation with impacts on salmon production. Bull. Am. Meteorol. Soc., 78, 1069-1079. Matear, R.J., and A. C. Hirst, 1999: Climate change feedback on the future oceanic CO2 uptake. Tellus, B, 51, 722-733. McGowan, J. A., D. R. Cayan, and L. M. Dorman, 1998: Climate-ocean variability and ecosystem response in the northeast Paci c. Sciences, 281, 210-217. McPhaden, M. J., and D. Zhang, 2001: Decadal spin-down of the Paci c ocean shallow meridional overturning circulation. Nature, sub judice. Maier-Reimer, E., U. Mikolajewicz, and A. Winguth, 1996: Future ocean uptake of CO2: Interaction between ocean circulation and biology. Clim. Dyn., 12, 711-722. 33
Miller, A. J., D. R. Cayan, T. P. Barnett, N. E. Graham and J. M. Oberhuber, 1994: The 1976-77 climate shift of the Paci c Ocean. Oceanography, 7, 21-26. Miller, A. J., D. R. Cayan, and W. B. White, 1998: A westward intensi ed decadal in the North Paci c thermocline and gyre-scale circulation. J. Climate, 11, , 3112-3127. Miller, A. J., and N. Schneider, 2000: Interdecadal climate regime dynamics in the North Paci c Ocean: Theories, observations and ecosystem impacts. Prog. Oceanogr., 47, 355-379. Minobe, S., 1999: Resonance in bidecadal and pentadecadal climate oscillations over the North Paci c: Role in climatic regime shifts. Geophys. Res. Lett., 26, 855-858. Mitchell, J. F. B., T. C. Johns, J. M. Gregory, and S. F. B. Tett, 1995: Climate response to increasing levels of greenhouse cases and sulfate aerosols. Nature, 376, 501-504. Morel, A., 1988: Optical modeling of the upper ocean in relation to its biogenous matter content (case I waters). J. Geophys. Res., 93, 10,749-10,768. Morel, A., and D. Antoine, 1994: Heating rate within the upper ocean in relation to its bio-optical state. J. Phys. Oceanogr., 24, 1652-1665. Murray, J. W., R. T. Barber, M. R. Roman, M. P. Bacon, and R. A. Feely, 1994: Physical and biological controls on carbon cycling in the equatorial Paci c. Science, 266, 58-65. Murtugudde, R., J. Beauchamp, and A. Busalacchi, 2001: Eects of penetrative radiation on the upper tropical ocean circulation. J. Climate, sub judice. Nakamoto, S., S. Prasanna Kumar, J. M. Oberhuber, K. Muneyama, R. Frouin, 2000: Chlorophyll modulation of sea surface temperature in the Arabian Sea in a mixed layer isopycnal general circulation model. Geophys. Res. Lett., 27, 747-750. Nakamoto, S., S. Prasanna Kumar, J. M. Oberhuber, J. Ishizaka, K. Muneyama, R. Frouin, 2001a: Response of the equatorial Paci c to chlorophyll pigments in a mixed layer isopycnal ocean general circulation model. Geophys. Res. Lett., 28, 2021-2024. Nakamoto, S., S. Prasanna Kumar, J. M. Oberhuber, H. Saito, K. Muneyama, R. Frouin, 2001b: Chlorophyll modulation of mixed layer dynamics in a mixed layer isopycnal general circulation model. Proc. Indian Acad. Sci. (Earth and Planetary Sci.), in press. Nakamura, H., G. Lin, and T. Yamagata, 1997: Decadal climate variability in the North Paci c during recent decades. Bull. Am. Meteorol. Soc., 78, 2215-2225. Nonaka, M., S.-P. Xie and J. P. McCreary, 2001: Decadal variations in the subtropical cell and equatorial Paci c SST. Geophys. Res. Lett., sub judice. Norris, J. R., and C. B. Leovy, 1994: Interannual variability in stratiform cloudiness and sea-surface temperature. J. Climate, 7, 1915-1925. Odum, H. T., 1983: Systems Ecology: An Introduction. Wiley, New York. Ohlmann, J. C., D. A. Siegel, and C. Gautier, 1996: Ocean mixed layer radiant heating and 34
solar penetration: A global analysis. J. Climate, 9, 2265-2280. Ohlmann, J. C., D. A. Siegel, and L. Washburn, 1998: Radiant heating of the western equatorial Paci c during TOGA-COARE. J. Geophys. Res., 103, 5379-5395. Ohlmann, J. C., and D. A. Siegel, 2000. Ocean radiant heating. Part II: Parameterizing solar radiation transmission through the upper ocean. J. Phys. Oceanogr., 30, 1849-1865. Ono, T., T. Midorikawa, Y. W. Watanabe, K. Tadokoro, and T. Saino, 2001: Temporal increases of phosphate and apparent oxygen utilization in the subsurface waters of western subarctic Paci c from 1968 to 1998. Geophys. Res. Lett., 28, 3285-3288. Paulson, C. A., and J. J. Simpson, 1977: Irradiance measurements in the upper ocean. J. Phys. Oceanogr., 7, 953-956. Payne, R. E., 1972: Albedo of the sea surface. J. Atmos. Sci., 29, 959-970. Peng, S. L., W. A., Robinson, M. P. Hoerling, 1997: The modeled atmospheric response to midlatitude SST anomalies and its dependence on background circulation. states. J. Climate, 10, 971-987. Polovina, J., G. T. Mitchum, N. E. Graham, M. P. Craig, E. E. DeMartini, and E. N. Flint, 1994: Physical and biological consequences of a climatic event in the Central North Paci c. Fish. Oceanogr., 3, 15-21. Reader, M. C., and G. J. Boer, 1998: The modi cation of greenhouse gas warming by the direct eect of sulphate aerosols. Clim. Dyn., 14, 593-607. Riebesell, U., I. Zondervan, B. Rost, P. D. Tortell, R. E. Zeebe, and F. M. M. Morel, 2000: Reduced calci cation of marine plankton in response to increased atmospheric CO2. Nature, 407, 364-367. Roeckner, E., T. Siebert, and J. Feichter, 1995: Climatic response to anthropogenic sulfate forcing simulated with a general circulation model. In: Aerosol Forcing of Climate, R. J. Charlson and J. Heintzenberg, Eds., Wiley, 349-362. Sarmiento, J. L., T. M. C. Hughes, R. L. Stouer, and S. Manabe, 1998: Simulated response of the ocean carbon cycle to anthropogenic climate warming. Nature, 393, 245-249. Sathyendranath, S., and T. Platt, 1988: The spectral irradiance eld at the surface and in the interior of the ocean - A model for applications in oceanography and remote-sensing. J. Geophys. Res., 93, 9270-9280. Schneider, E. K., and Z. Zhu, 1998: Sensitivity of the simulated annual cycle of the sea surface temperature in the equatorial Paci c to sunlight penetration. J. Climate, 11, 1932-1950. Schneider, N., T. Barnett, M. Latif, and T. Stockdale, 1996: Warm pool physics in a coupled GCM. J. Climate, 9, 219-239. Schneider, N., A. J. Miller, M. A. Alexander, and C. Deser, 1999: Subduction of decadal 35
North Paci c temperature anomalies: Observations and dynamics. J. Phys. Oceanogr., 29, 1056-1070. Schneider, N. and A. J. Miller, 2001: Predicting western North Paci c Ocean climate. J. Climate, 14, 3997-4002. Schneider, N., A. J. Miller, and D. W. Pierce, 2002: Anatomy of North Paci c decadal variability. J. Climate, in press. Schwartzlose, R. A., J. Alheit, A. Bakun, T. R. Baumgartner, R. Cloete, R. J. M. Crawford, W. J. Fletcher, Y. Green-Ruiz, E. Hagen, T. Kawasaki, D. Lluch-Belda, S. E. LluchCota, A. D. MacCall, Y. Matsuura, M. O. Nevarez-Martinez, R. H. Parrish, C. Roy, R. Serra, K. V. Shust, M. N. Ward, and J. Z. Zuzunaga, 1999: Worldwide large-scale
uctuations of sardine and anchovy populations. S. Afr. J. Marine Sci. 21, 289-347. Shaw, G. E., 1983: Bio-controlled thermostasis involving the sulphur cycle. Climate Change, 5, 297-303. Shin S. and Z. Liu, 2000: Response of equatorial thermocline to extratropical buoyancy forcing. J. Phys. Oceanogr., 30, 2883-2905 Siegel, D. A., J. C. Ohlmann, L. Washburn, R. R. Bidigare, C. T. Nosse, E. Fields, and Y. M. Zhou, 1995: Solar-radiation, phytoplankton pigments and the radiant heating of the equatorial Paci c warm pool. J. Geophys. Res., 100, 4885-4891. Simpson, J. J. and T. D. Dickey, 1981: Alternative parameterizations of downward irradiance, and their dynamical signi cance. J. Phys. Oceanogr., 11, 876-882. Simo, R., 2001: Production of atmospheric sulfur by oceanic plankton: Biogeochemical, ecological and evolutionary links. Trends Ecol. Evol., 16, 287-294. Sugimoto, T., and K. Tadokoro, 1997: Interannual-interdecadal variations in zooplankton biomass, chlorophyll concentration and physical environment in the subarctic Paci c and Bering Sea. Fish. Oceanogr., 6, 74-93. Tanimoto, Y., N. Iwasaka, K. Hanawa, and Y. Toba, 1992: Characteristic variations of sea surface temperature with multiple time scales in the North Paci c. J. Climate, 6, 1153-1160. Timmermann, A., 2001: Decadal ENSO amplitude modulations: A nonlinear paradigm. Global Planetary Change, sub judice. Timmermann, A., and F.-F. Jin, 2001: A nonlinear mechanism for decadal El Ni~no amplitude changes. Geophys. Res. Lett., sub judice. Tourre, Y., W. B. White, Y. Kushnir, 1999: Evolution of interdecadal variability in sea level pressure, sea surface temperature, and upper ocean temperature over the Paci c Ocean. J. Phys. Oceanogr., 29, 1528-1541. Trenberth, K. E., and T. J. Hoar, 1996: The 1990-1995 El Ni~no Southern Oscillation event: 36
Longest on record. Geophys. Res. Lett., 23, 57-60. Tziperman, E., L. Stone, M. A. Cane, and H. Jarosh, 1994: El Ni~no chaos: Overlapping of resonances between the seasonal cycle and the Paci c ocean-atmosphere oscillator. Science, 264, 72-74. U.S. Global Change Research Program, 1999: A U.S. Carbon Cycle Science Plan. A Report of the Carbon and Climate Working Group, J. L. Sarmiento and S. C. Wofsy (co-chairs). Venrick, E. L., J. A. McGowan, D. R. Cayan, and T. L. Hayward, 1987: Climate and chlorophyll a: long-term trends in the central north Paci c Ocean. Science, 238, 70-72. Wanninkhof, R., 1992: Relationship between wind-speed and gas-exchange over the ocean. J. Geophys. Res., 97, 7373-7382. Ware, D. M., 1995: A century and a half of change in the climate of the NE Paci c. Fish. Oceanogr., 4, 267-277. Watanabe, M., and M. Kimoto, 2000: On the persistence of decadal SST anomalies in the North Atlantic. J. Climate, 13, 3017-3028. Whitney, F. A., C. S. Wong, and P. W. Boyd, 1998: Interannual variability in nitrate supply to surface waters of the Northeast Paci c Ocean. Mar. Ecol. Prog. Ser., 170, 15-23. Woods, J. D., 1980: Diurnal and seasonal variation of convection the wind-mixed layer of the ocean. Quart. J. Roy. Meteor. Soc., 106, 379-394. Woods, J. D., W. Barkman, and A. Horch, 1984: Solar heating of the oceans - Diurnal, seasonal and meridional variation. Q. J. R. Meteorol. Soc., 110, 633-656. Zaneveld, J. R. V., J. C. Kitchen, and H. Pak, 1981: The in uence of optical water type on the heating rate of a constant depth mixed layer. J. Geophys. Res., 86, 6426-6428. Zhang, Y., J. M. Wallace, and D. S. Battisti, 1997: ENSO-like interdecadal variability: 1900-93, J. Climate, 10, 1004-1020.
Figure 1: Sketch of basic processes involved in the Midlatitude Gyre Mode and Subduction
Mode that have been proposed to explain interdecadal variability in the Paci c. The Midlatitude Gyre Mode involves the atmosphere responding to SST which is driven by a delayed response of the midlatitude gyre current system to antecedent atmospheric forcing with opposite polarity. The Subduction Mode involves a midlatitude SST anomaly that subducts into the thermocline and upwells in the tropics after a delay to drive the atmospheric bridge which forces a midlatitude SST anomaly of opposite polarity. Figure 2: Modeled ecosystem response to the 1976-77 climate shift physical forcing (from Chai et al. 2001). The springtime (March-April) vertically integrated (0-60m) total phytoplankton biomass for the North West Paci c (35N-45N, 160E-160W) increases substantially after the shift in agreement with the observed chlorophyll changes described by Venrick et al. (1987). Figure 3: SST anomaly, averaged over the Northern Hemisphere (*), Southern Hemisphere (+) and globe (o), induced by including the annual cycle spatial distribution of observed chlorophyll from CZCS in a full-physics ocean model of Nakamoto et al. (2001a). From Shell et al. (2001, in preparation).. Figure 4: Atmospheric model response, expressed as surface air temperature averaged over the Northern Hemisphere (x), Southern Hemisphere (diamond) and globe (box), to the SST anomaly pattern of Figure 3. From Shell et al. (2001, in preparation). Figure 5: Methanesulfonate aerosols (oxidized DMS) measured at Cape Grim (Tasmania) 1978-2000 showing interdecadal variations in one of the few interdecadal-scale time series available. Figure 6: Modeled annual DMS ux in the Subantarctic Southern Ocean. The regional DMS production model, forced with temperature, cloud, wind speed and mixed-layer depth data, yields variations with interdecadal timescales. From Gabric et al. (2002). Figure 7: (a) Model ef ratios versus observed ef ratios. The model is a complex food web for an open-ocean pelagic ecosystem and the observations are from eld studies carried out as a part of the Joint Global Ocean Flux Study and related work. The straight line is the 1:1 line. (b) Total primary production versus observed ef ratios at the same locations. From Laws et al. (2000). Figure 8: Sketch of the possible eects of ocean biology on the Midlatitude Gyre Mode and Subduction Mode described in Figure 1. Changes in upper-ocean heating due to phytoplankton modify SST anomalies, while changes in DMS uxes to the atmospheric modify atmospheric clouds and dynamics. Ascertaining the strength and importance of these eects require further study. 38
Figure 9: Sketch of the various aspects of biological-physical interaction discussed in this paper. Arrows indicate the direction of forcing among boundary layer cloud cover, SST, ocean strati cation, and plankton concentration. The sign indicates whether an increase in the forcing parameter causes an increase (+) or decrease (-) in the forced parameter. (?) indicates those processes which especially need more investigation.