1. Introduction
The impact of turbidity and biological production on the optical properties of the oceans has been of great interest for some time (e.g., Jerlov 1968) especially since the advent of remote sensing (McClain et al. 1998). It was not long before the in-water spectral radiances and the radiant heating associated with the absorptive properties of the suspended sediments and marine phytoplankton started to be investigated in models and from observations (Denman 1973; Paulson and Simpson 1977; Simpson and Dickey 1981b; Siegel et al. 1995; Ohlmann et al. 1996). The role of proper representation of the penetrative solar radiation for accurate SST simulation was first proposed by Denman (1973) followed by Simpson and Dickey (1981a) and Dickey and Simpson (1983). This was later confirmed from observations by Siegel et al. (1995) and Ohlmann et al. (1996, 1998. Direct estimates of the biological heating at large scales were the natural next step following the availability of basin-scale chlorophyll concentrations provided by remote sensing (Lewis et al. 1990; Sathyendranath et al. 1991; Strutton and Chavez 2004). A potential positive feedback between the mixed layer depth, due to water column stabilization, and phytoplankton growth in the Arabian Sea was proposed by Sathyendranath et al. (1991), while the ∼3°C overestimate of the western Pacific warm pool sea surface temperature (SST) in model simulations was blamed on the misrepresentation of the penetrative loss of surface solar radiation by Lewis et al. (1990). The appropriate representation of solar transmission for large-scale models continues to be discussed (e.g., Morel 1988; Mobley 1989; Ohlmann et al. 2000; Ohlmann 2003) even though the first-order effect seems to be largely captured by the single exponential formulation, the Beer–Lambert law, of Morel (1988) or the first exponential term of Ohlmann (2003) for the vertical resolutions of state-of-the-art ocean and climate models.
Most ocean general circulation models (OGCMs) and coupled climate models neglected the penetrative part of the solar radiation until studies such as those by Chen et al. (1994) and Schneider et al. (1996) began to report the significant impact of solar transmission on SSTs, the variable of most interest for seasonal to interannual prediction. Schneider and Zhu (1998) were the first to hint, with a coupled climate model, that some chronic coupled model problems (e.g., the simulation of the annual cycle in the crucial east Pacific cold tongue region) may be related to the inaccurate representation of solar transmission in the ocean. In the Tropics, the mixed layer–thermocline interaction or the Bjerknes feedback (Bjerknes 1969) is what is expected to be affected by the subsurface radiative heating. However, different ocean models result in different mixed layer–thermocline interactions due to differences in physical parameterizations, model configurations, and surface forcings, especially in regions where oceanic mixed layers tend to be relatively shallow, such as the deep Tropics. These intermodel differences are noticeable. For example, the results of Nakamoto et al. (2000, 2001), where enhanced radiative heat attenuation resulted in a cooling in their isopycnal model, are not consistent with the results of Murtugudde et al. (2002, hereafter M02), where the enhanced radiative heat attenuation resulted in a warming in the east equatorial Pacific in their sigma coordinate model. The intermodel difference was also noted by Kara et al. (2004) who used the same model used by Rochford et al. (2001) and contrasted their results to other OGCM and coupled model studies. Also interesting are the differences reported by Manizza et al. (2005), where different bio-optical models provide opposing changes on the tropical region, while maintaining similar significant response at mid- and high latitudes. The differences of turbidity effects in different models should not demote the interest of the scientific community as they all indicate the importance of penetrative radiation and the need for exploring various models and the available parameterizations to understand the nature of potential feedbacks. As will be shown in section 5, changes in the heat trapped in the surface mixed layer arise as a combination of the direct thermal term due to the change of the attenuation depth plus an indirect dynamical term related to the ocean dynamical response. The relative amplitude of each term determines if a warming or cooling will occur. Thus, additional sensitivity studies are required to fully understand how the usual physical parameterizations of ocean models affect the system response to biogenic heating.
It is thus encouraging that the complexity of models varies from one-dimensional, Secchi disk–based turbidities (Kantha and Clayson 1994) to OGCMs with satellite-derived chlorophyll concentrations (Kara et al. 2004) and from coupled ocean–atmosphere general circulation models (Schneider and Zhu 1998) and atmospheric GCMs (Shell et al. 2003) to simple process-based models (Timmermann and Jin 2002) and statistical atmospheric coupling (Marzeion et al. 2005). The processes of interest vary from intraseasonal variability (Gildor et al. 2003) to fronts (Edwards et al. 2001), from the Tropics (Nakamoto et al. 2001, 2002) to midlatitudes (Oschlies 2004) and from mean pigment concentrations (M02) to interannual variations in chlorophyll concentrations (Gildor and Naik 2005). Some heuristic arguments for bioclimate feedbacks at interdecadal time scales in the Pacific Ocean are presented in Miller et al. (2003).
The ultimate goal of all these and other related studies is to improve SST simulations in models with the expectation that coupled climate predictions (with special emphasis on ENSO prediction) can be improved in turn. The simple coupled model experiments of Timmermann and Jin (2002) used a temperature–chlorophyll relation to argue that the biophysical feedbacks will be intrinsically asymmetric with a cold ENSO event involving a negative feedback associated with anomalously high near-surface chlorophyll concentrations driven by the anomalously strong upwelling. A warm ENSO event on the other hand would not be strongly affected by biological feedbacks due to reduced surface chlorophyll concentrations (Barber and Chavez 1983; Murtugudde et al. 1999). The intuitive biological feedbacks were essentially prescribed in this simple model to affirm the processes involved. Shell et al. (2003) forced an AGCM with SST anomalies produced by the biological feedbacks in the forced OGCM experiments of Nakamoto et al. (2000, 2001) to show that the spatiotemporal scales of the SST anomalies were able to produce local and remote atmospheric circulation anomalies throughout the troposphere. The potential damping/amplification of the feedbacks due to interactive ocean–atmosphere coupling was not addressed in this study. Schneider and Zhu (1998) compared coupled model experiments with and without penetrative radiation to compare the annual cycle and did not address the impacts of turbidity on interannual variability. The first step toward quantifying the impacts of biological feedbacks on interannual variability was reported by Marzeion et al. (2005) using a statistical atmospheric coupling to the same OGCM used by M02 albeit with much coarser horizontal and vertical resolutions and with an embedded ocean ecosystem model (Christian et al. 2002; Christian and Murtugudde 2003). This study generated the chlorophyll variability and the associated attenuation depth changes internally to confirm the results of M02 that the dynamic feedbacks associated with the subsurface heating due to chlorophyll absorption of solar radiation lead to deeper mixed layers and warmer SSTs. The interannual variability was consistent with the asymmetric feedback suggested by Timmermann and Jin (2002). However, the ENSO variability in this model was not robust and not very realistic (their Fig. 12).
The motivation of the present study is to examine the impact of seasonal variability of the satellite-derived attenuation depths on the annual cycle and the interannual variability in the tropical Pacific with a statistical atmosphere coupled to an OGCM that produces not only a realistic Niño-3 index, but also the spatial structure that is consistent with nearly all state-of-the-art coupled climate models. The OGCM is the same one used by M02 with a 1/3° meridional resolution in the deep Tropics and vertical resolution of ∼5 m below the mixed layer in the top 100 m. For a direct comparison to our earlier study, we carried out simulations with a constant attenuation depth and spatially varying attenuation depths. Additional sensitivity studies should involve seasonal variability of attenuation depths. This research is essential for understanding the processes involved in realistic seasonal to interannual variability of the coupled bioclimate feedbacks in the crucial tropical Pacific region. As such it is a necessary prerequisite before a statistical relationship between observed attenuation depths and physical parameters such as the mixed layer and the thermocline depths can be constructed to be used in full coupled climate models. The OGCM resolves the mixed layer, the thermocline, and the interactions between them as realistically as any state-of-the-art model. The processes diagnosed here will provide the basis for the feedback from time-varying chlorophyll distributions to ocean circulation and SSTs.
The description of the OGCM, the statistical atmosphere, the coupling, and the data used in this work are provided in section 2. The experimental setup is described in section 3. The results are presented in section 4 and discussed in section 5. The summary and conclusions are contained in section 6.
2. Model and data
a. Hybrid coupled model
The hybrid coupled model (HCM) used in this work couples a reduced-gravity, primitive equation OGCM to a statistical wind stress anomaly model. Details of the OGCM can be found in M02. The main changes in the physical configuration are that the model used here has 31 layers and the domain extends from 25°S to 25°N. The statistical wind stress anomaly (τ) model is derived from a SVD analysis of the covariance matrix from time series of monthly mean SST and wind stress fields (e.g., Syu et al. 1995; Chang et al. 2001). The τ model is constructed from the ECHAM4.5 (Roeckner et al. 1996) ensemble simulations and observed SST during the period 1963–96. The coupling between the atmospheric τ model and the OGCM is as follows: the OGCM calculates SST anomalies (relative to its uncoupled mean climatology) that are used to calculate wind anomalies, which in turn are added to the observed mean seasonal climatology of wind stress to drive the OGCM. Previous work with the statistical atmosphere model (Zhang et al. 2003; Zhang and Busalacchi 2005; Zhang et al. 2005) provided a retrospective prediction of ENSO with a skill comparable to the most advanced coupled systems.
The HCM is initiated with an imposed SST anomaly for eight months. Evolution of anomalous conditions thereafter is determined solely by coupled interaction in the system. As examined previously by numerous studies (e.g., Barnett et al. 1993; Syu et al. 1995), coupled behaviors depend on the so-called relative coupling coefficient, α, multiplying the wind anomalies by a scalar parameter before being added to the climatological wind stress fields driving the OGCM. Several tuning experiments have been performed with different values of coefficient α to examine the coupled interannual variability. In general, α > 1 is required to produce sustainable interannual variability. Values larger than 1.4 have been found to produce a too strong upwelling in the equatorial region, which eventually erodes the western warm pool. A value of 1.2 was shown to produce ENSO oscillations with a reasonable amplitude and frequency. All coupled simulations begin with this value. However, both simulations with realistic attenuation depth do not sustain interannual variability shortly after the coupled simulations begin, indicating an increase on the inertia of the ocean component compared with the shallow attenuation depth of 17 m. Thus, coupled simulations have been restarted with a coupling coefficient of 1.3.
b. Ocean color data
The spectrum of solar insolation reaching the surface of the ocean contains energy in a wide range of frequencies. Below the ocean surface, the penetration of solar radiation follows a Beer–Lambert law with a wavelength-dependent absorption coefficient that, in general, is expressed in terms of a solar transmission function, TR(z), accounting for the fraction surface solar radiation left at depth z [i.e., En(z) = En(0−)TR(z)]. In this expression, En(0−) represents the solar flux just below the sea surface and En(z) is the solar flux at depth z. As in M02, a single absorption coefficient is used to account for the average attenuation over the visible band (380–700 nm) that is, TR(z) = γ exp(−z/Hp), where γ is the fraction of the radiation available to penetrate to depths beyond the very first centimeters, and Hp is the inverse of the diffuse attenuation coefficient Kp. Following M02, maps of attenuation depth are calculated from chlorophyll concentrations using the following empirical relationship: Kp(x, y) = Kw + aChl(x, y)b. The values used here are γ = 0.33, Kw = 0.027 m−1, a = 0.0518 m−1/(mg m−3)b, and b = 0.428. Maps of monthly chlorophyll fields come from level-3 monthly composites of the Sea-viewing Wide Field-of-view Sensor (SeaWiFS) chlorophyll from September 1997 to September 2005. The 9-km resolution maps are binned to the grid of the model, and the monthly median is used as the monthly climatology in order to reduce the sensitivity to the extreme El Niño and La Niña events of 1997 and 1998, respectively. Figure 1a shows the spatial distribution of the mean of the 12-month climatology. The Beer–Lambert law implies that regions with the smallest attenuation depth correspond to the regions where downwelling solar irradiance is absorbed the fastest. Thus, Fig. 1a features regions of elevated biological activity (coastal and equatorial upwelling) as the regions of small attenuation depth (<19 m). Values larger than 25 m are found in the oligotrophic subtropical gyres. The standard deviation in Fig. 1b shows that the ITCZ and the South Pacific convergence zone (SPCZ) are the regions with the largest seasonal changes of the attenuation depth, probably reflecting the seasonal changes of upwelling and solar radiation. Low seasonal variability is found in both the subtropical gyres and the equatorial cold tongue. Figure 2 shows the seasonal modulation (deviations from the annual mean) of the constructed climatology of the attenuation depth at the equator. The amplitude of the seasonal cycle is largest in the western part of the equatorial region, which is not likely to be due to equatorial dynamical processes but rather to the southwestern incline of the ITCZ (see Fig. 1b). In the eastern equatorial Pacific, attenuation depth is correlated with SST (Fig. 8a). During the first part of the year, seasonal anomalies of both SST and attenuation depth are positive (i.e., surface water is seasonally warmer with reduced biological activity). During boreal spring, weak wind speeds inhibit vertical upwelling and entrainment into the surface mixed layer leading to warm SST and low biological activity. As winds strengthen after May, the thermocline shoals, the SST begins to cool, and the attenuation depth rapidly decreases, indicating an increase of the biological activity. In the western part of the basin, where the equatorial thermocline is deeper, the seasonal modulation of both SST and attenuation depth is mostly unrelated to the seasonal changes of the deep thermocline. The largest negative anomaly of HP490 occurs during boreal spring, which is a period of negative SST seasonal anomalies, slightly lagging the period of strongest wind speed west of the international date line. The weakness of the seasonal modulation of SST is related to the presence of a climatological barrier layer (Lukas and Lindstrom 1991). The eolian supply of iron in the western equatorial Pacific sustains a low-nutrient surface regime, whose modulation is set by the entrainment of subsurface nutrients (Barber and Chavez 1991). In such a context, strong winds deepen the mixed layer, leading to increased surface nutrients, enhancing biological activity, and reducing attenuation depths.
c. Validation data
Validation of the simulations is done by comparing the SST, the depth of the 20°C isotherm (D20), and ocean surface current against available analyzed fields. SST fields are obtained from the optimal interpolation analysis of Reynolds and Smith (1994). Although a mixed layer depth (MLD), climatology have been estimated for the World Ocean (e.g., Monterey and Levitus 1997), validation of this physical parameter is not carried out here as MLD is not a precisely defined quantity and it is often very difficult to match model MLDs to observed MLDs since the model computes the entrainment required to balance the available surface turbulent kinetic energy (TKE), whereas observed estimates use hydrographic data to estimate the depth of vertically homogeneous layer. Keeping in mind that the mixed layer–thermocline interactions is the key process in ENSO cycles, this issue has to be addressed in more detail in both observations and models [which coincidentally is one of the goals of the Pacific Upwelling and Mixing Physics (PUMP) experiment organized under CLIVAR; more information available online at http://www.pmel.noaa.gov/kessler/clivar/pump.html].
The estimates of D20 come from the Tropical Atmosphere Ocean (TAO) array (McPhaden et al. 1998) products distributed by the TAO project (more information available online at http://www.pmel.noaa.gov/tao/). Tropical currents fields come from the Ocean Surface Current Analyses—Real Time (OSCAR) project (Bonjean and Lagerloef 2002), where velocities are derived from satellite altimeter and vector wind data. Because of the intermittency of some sensors, the disparity of coverage between subsurface temperature and salinity, and the length of the different products, the seasonal cycle calculated from these products, and the model/data discrepancies must be considered with caution.
3. Experimental setup
To investigate the ability of using remotely sensed ocean color to parameterize the effects of biological heat entrapment in the HCM, we will examine three different simulations differing in the value of the attenuation depth. In the control simulation (named HP17), a constant attenuation depth is specified everywhere. As in M02, a value of 17 m, the nearest value to the global spatial average of the annual mean has been selected [note that M02 used Coastal Zone Color Scanner (CZCS)-derived attenuation depths]. Next, two additional simulations are carried out. In the first additional simulation (HPAM), the spatially varying annual mean of SeaWiFS attenuation depth is used to account for the penetrating radiation. The last simulation (HPSC) considers the seasonal cycle of the attenuation depth.
All three ocean models (differing on the parameterization of the attenuation of solar radiation) are spun up for 20 yr. During this spinup, each run is forced by monthly climatological winds from the European Centre for Medium-Range Weather Forecasts (ECMWF) wind stress (ECMWF 1994). At the end of the 20th year, the ocean model is coupled to the statistical atmosphere model. Inside the statistical atmosphere model, both the wind stress climatology and the SST-τ covariance defining the statistical atmosphere model is the same in all three simulations. However, the SST climatology used to calculate SST anomalies differs between simulations. In each coupled simulation, the SST climatology is calculated from the respective ocean spinup. Failure to adapt the climatological SST to each ocean configuration will introduce spurious biases on the SST anomalies that can critically affect the performance of the coupled system.
In each case, the HCM is initialized as described in section 2a. Initializing from the spinups, each coupled model is spun up with a coupling factor of 1.2 for an additional 20 yr. After this time, (year 2000 of the simulations) and, as pointed out in section 2a, the simulations HPAM and HPSC are continued with a coupled coefficient of 1.3 for a period of 35 yr (α kept to 1.2 for HP17). The last 30 yr are used to study the differences between the simulations. Note that, similar to the case of Marzeion et al. (2005), no effort has been made to tune the ocean model in the context of the HCM, nor after introducing the effects of seasonal cycle of the attenuation depth.
4. Results
Figure 3 displays the longitude–time plot of the equatorial SST variability during the last 30 yr of each simulation. The control simulation (HP17) exhibits seven warm events that roughly correspond to a 4.3 yr period. Note that the longitudinal excursions of the 26°C are restricted to the central Pacific, which translates to weak warm events in the eastern equatorial Pacific. When satellite data are used to account for the spatial structure of the attenuation depth (simulations HPAM and HPSC), a significant impact on the frequency, regularity, and amplitude of the warm events is obtained. Changes in the variability in both the warm pool and the cold tongue are substantial, especially when the seasonal cycle of the attenuation depth is accounted for (HPSC). In the following sections we present the differences between these simulations in terms of the annual mean, the seasonal cycle, and the interannual variability.
a. Annual mean
The last 30 yr of each coupled simulation are used to calculate the annual mean of SST, D20, ocean currents, and surface wind stress. Figure 4 shows the annual mean of Reynolds SST (Fig. 4a) and the differences between the annual mean of each simulation minus the observed values (Figs. 4b–d). Compared to observations, the SST of the control simulation (Fig. 4b) has a cold bias in the whole basin. As noted in M02, the forced OGCM has a cold (warm) bias in the cold tongue (warm pool). Here, the largest bias of the coupled model, in absolute value, is found in the cold tongue (>3°C). Realistic attenuation depths warms most of the eastern tropical Pacific, reducing the cold bias in that region (Figs. 4c,d). Note that this is, in fact, the region where the differences are the more accentuated. Note however that in a fully coupled model, the warming of the cold tongue region and the resultant flattening of the east–west gradient of SST would lead to weaker winds, reduced upwelling, and farther warming of the cold tongue.
Positive and neutral impacts are found for the depth of the annual mean thermocline depth (Fig. 5) and also the zonal currents (Fig. 6). The annual mean of the depth of the D20 from the TAO array is shown in Fig. 5a. Figure 5b shows the annual mean of the control simulation minus the TAO field. The model thermocline is deeper than the observed in the central equatorial region and shallower everywhere else. SeaWiFS attenuation depth sharpens the east–west thermocline gradient. Some dynamical effects of these changes are seen in Fig. 6, which shows the latitudinal velocity structure at 165°E (Fig. 6a), 155°W (Fig. 6b), and 110°W (Fig. 6c). Each plot in Fig. 6 compares the OSCAR analysis and the three simulations. Each simulation is found to correctly reproduce the main structure of the surface current except in the western part of the basin, where none of the simulations are able to capture the latitudinal structure of the South Equatorial Current (SEC). In the central and eastern part of the basin, the coupled model shows good agreement with the observed structure. In general, the attenuation depth has very little impact on the annual mean of the ocean currents. In a fully coupled model, equatorial SST changes would result in off-equatorial wind changes and the impacts on the surface currents would be larger (Schneider and Zhu 1998).
Changes in the vertical structure of the ocean due to a realistic attenuation depth are significant. Figure 7a shows the vertical structure of the annual mean temperature at the equator for the control simulation. Impacts on the annual mean at the equator due to SeaWiFS attenuation depths are shown in Figs. 7b,c. In the eastern part of the basin, where the depth of the 20°C isotherm has increased (see Fig. 5), there is a clear subsurface warming below the mixed layer. This warming in the subsurface eastern equatorial Pacific, noticeable at depths of 200 m, agrees with the results of M02 and Marzeion et al. (2005). In the west there is a significantly reduced subsurface heating. In M02, where winds were independent of the state of the system, both east and west had similar heating below the MLD. Figure 7 also shows that realistic attenuation depth increases the depth of the mixed layer almost everywhere in the domain (e.g., Figs. 7b,c) as was already pointed out by M02 and Marzeion et al. (2005).
b. Seasonal cycle
As shown by Xie (1994) the amplitude and phase of the seasonal westward propagation of the SST is a function of the mean thermocline. However, the slight change of the thermocline shown in Figs. 5c,d is too weak to modify the seasonal cycle of the SST (Fig. 8). Figure 8 shows the monthly climatology minus the annual mean. That is, model bias does not contribute to the difference fields shown in Fig. 8. The results indicate that there are no significant differences between the errors of the seasonal cycle between the three simulations. The largest error in these simulations are the amplitude of the spring warming in the eastern Pacific (−1.5°C). Although not represented in Fig. 8, all simulations correctly reproduce the eastern cooling at the beginning of the boreal summer and the subsequent cooling during fall. Although realistic attenuation depth does not provide significant changes, the root-mean-square (RMS) of the fields shown in Figs. 8b–d indicates that the amplitude of the error is slightly lower in simulations HPAM and HPSC.
A similar lack of significant changes is obtained with respect to the seasonal cycle of D20 and surface ocean currents (not shown). Just note that the seasonal cycle of the D20 of the three simulations overestimate the seasonal anomaly in the eastern basin during spring and underestimates the fall deepening of the thermocline. The zonal current of the three simulations, averaged between 1° and 3°N [i.e., the northern branch (NB) of the SEC] has also been investigated because it plays a significant role on the lateral temperature advection (i.e., affects the strength of the SST anomalies in the eastern equatorial Pacific). The largest seasonal changes observed in the OSCAR data (not shown) are located approximately between 100° and 140°W, where the annual mean velocity is about −0.4 m s−1. At these longitudes, the westward component is strongly reduced in spring, when the SST is warm, and it is increased at the end of summer, when the SST is cold. In the warm pool the annual mean of the SEC is weak and the significant westward component only occurs at the beginning of the year. Neither simulation properly reproduces the seasonal modulation of the NB SEC (i.e., too strong SEC on spring and too weak at the end of summer). Note however, that in each case, realistic attenuation depths slightly reduce the amplitude of the error (calculated as the RMS of the difference to observations) for every parameter under consideration (i.e., SST, D20, and ocean currents). These results should be considered only as an example of the tendency of the coupled system to veer toward observations when realistic attenuation depths are being used. The shortness of the climate satellite record still introduces uncertainties on the seasonal cycle and annual means of surface chlorophyll. Better estimates of these quantities during the following years [combining the SeaWiFS and the Moderate Resolution Imaging Spectroradiometer (MODIS) records] might provide sensible improvements.
c. Interannual variability
The longitude–time plots of the SST anomalies with respect to each climatological field are shown in Figs. 9a–c. In the control simulation, temperature anomalies in the western Pacific both originate and mature in the band 160°E–160°W with limited eastward extension. Realistic attenuation depths allow for an increase of the amplitude of the SST anomalies, reduction of the the frequency of the warm events, and separation of the region of origin (western of 160°E) and maturity (central and eastern Pacific) of ENSO events. The amplitude of the eastern SST anomalies dramatically increases in the case of seasonal attenuation depth, which is consistent with Schneider and Zhu (1998). Improvements in the amplitude of eastern equatorial anomalies can be explicitly shown by the SST average on the Niño-3 region (5°S–5°N, 150°–90°W) as shown in Fig. 10. Note that the changes shown in Fig. 9c are not limited only to the eastern Pacific, but also affect the genesis of the events, which now originate farthest to the west and propagate eastward before reaching their maturity. Figures 9d–f shows similar effects on the amplitude of the zonal wind stress anomalies: the HPSC simulation yields stronger anomalies reaching eastward of the 160°W longitude. The westward propagation during the major warm events of 2009, 2016, 2022, and 2029 points out the local coupling between both SST and zonal wind stress anomalies present in the HPSC simulation. The wind stress anomalies have been multiplied by 1.3/1.2 = 1.08 to account for the difference in the coupling coefficient between HP17 and the other two simulations. Also note that while the exact nature of the ENSO evolution in a statistical atmospheric model is clearly linked to the period over which the regression relations between SST and winds are constructed, our interest here is simply in contrasting different simulations to quantify the impact of biological feedbacks.
As shown in Fig. 7, the impact of realistic attenuation depths is not limited to the surface temperature, but also on the vertical structure of the equatorial region (Figs. 9g–i and Figs. 9j–l show the thermocline depth and the MLD, respectively). Longitude–depth composites of warm events (not shown) illustrate the increase of the discharge process in the warm pool mentioned in section 4a. Figures 9g–i show the enhancement of the thermocline seesaw during ENSO events that project onto the annual mean differences shown in Fig. 7. As shown in Fig. 3, realistic attenuation depths intensifies the amplitude of the warm events, increasing the subsurface warming in the eastern equatorial Pacific (on the order of 1°C), and enhances the discharge mechanism in the western Pacific, which is characterized by an equatorial upwelling and eastward advection of temperature at the surface. The boost in the western discharge results in a colder temperature in the west during ENSO events, which has feedbacks to the ecosystem response to ENSO (Turk et al. 2001; Murtugudde et al. 1999). The reduced subsurface warming on the western part of the equator, observed in Figs. 7b,c, illustrates the signature of the ENSO amplification on the annual mean of the ocean. The role of the increased amplitude and persistence of the MLD anomalies (Fig. 9l) in the eastern equator will be discussed in the next section, as its interaction with the seasonal modulation of the attenuation depth helps maintaining long-lasting SST anomalies in the HPSC simulation.
The spectrum of the Niño-3 index (Fig. 10b) shows that realistic attenuation depths increase the variance of the signal at longer periods. Interestingly, the 2-yr peak of the control simulation (not observed in the spectrum of the time series of Niño-3 index from the Reynolds SST) disappears from the HPSC spectrum (shifts to a shorter period). Note however, that the larger energy to longer periods, and a more realistic quasi-biannual mode does not necessarily mean a uniform improvement in the variability of the SST across the basin. For example, Fig. 11 shows the standard deviation of the monthly anomalies of the Reynolds SST and the three simulations. It is evident HPAM and HPSC simulations (Figs. 11c,d) do not show improved variance patterns of SST (excess of variability around the date line and very weak variability in the eastern equatorial Pacific). The sponge regions of the ocean model (poleward of 20° latitude), artificially damps the variance of the SST near the sponge regions. The seasonal attenuation depth shows only a scant increase of the variance in the central and eastern regions.
A striking result is shown in Fig. 12, where the frequency of warm events is shown as function of the month of the temperature peak. In the control simulation, warm events peak in June (two), July (four), and November (one). Warm events of the constant attenuation depth (HPAM) occur in May (one), June (two), July (three), and December (one). In HPSC, warm events peak in July (two), November (one), and December (three). In this simulation, the stronger warm events peak either in November or December. To the first order, the improved annual cycle leads to better phase locking of model ENSO events. It is usually accepted (see Tziperman et al. 1998) that the observed phase locking indicates the importance of the seasonal cycle on the dynamics of ENSO. As HPAM does not show any improvement on the phase locking, it can be conjectured that if the realistic attenuation affects the phase locking of ENSO, it has to do so via its seasonal variability. The deepening of the thermocline and the associated reduction in the biological activity and deeper attenuation depths during warm events are not represented in any of these simulations since the interannual variability of the attenuation depths is not included. A stationary attenuation depth traps more radiation within the mixed layer during warm events and leads to maximum warming during summer and fall months when the climatological upwelling is at its peak. As explained in the next section, the interaction between the ENSO-related mixed layer depth variability and the seasonal cycle of turbidity may explain the longer-lasting warm anomalies in the eastern Pacific, leading to a better rendition of the mature phase during the end of the calendar year. This indicates that a statistical representation of the attenuation depth variability at interannual time scales can be expected to significantly improve ENSO simulations and subsequently, ENSO prediction. The interactions of the mixed layer and thermocline with the background variations in mixed layer and attenuation depths will be further discussed in the following section.
5. Discussion
The mechanisms by which biogenic absorption of solar radiation lead to an increase of the equatorial mixed layer depth and the warming of the equatorial cold tongue have been already reported by several authors as, for example, M02, Timmermann and Jin (2002), and Marzeion et al. (2005). As the OGCM used in this study is the same as in M02 and Marzeion et al. (2005), the same mechanisms (direct and indirect) described by these authors apply here. The direct mechanism, reducing the amount of absorbed radiative heat in the mixed layer, seems to be the key factor in the cooling of the western Pacific and subtropical gyres (Figs. 4b–d), as already pointed out by Lewis et al. (1990). The indirect mechanism, associated with changes in the subsurface stratification and mixed layer depths leading to the weakening of the northern branch of the SEC in the eastern and central Pacific (Figs. 6b,c), actually warms the SSTs in the eastern equatorial Pacific. Together, the warming of the eastern cold tongue and the cooling of the western warm pool weaken the zonal SST gradient, abate the equatorial winds of the coupled model between 180° and 130°W (Figs. 9a –f), and enhance the eastern warming and western cooling via Bjerknes feedbacks (Bjerknes 1969).
As pointed out by Schneider and Zhu (1998), the amplitude of the SST seasonal cycle is affected by the absorption of radiative heating. The eigenvalue factorization of the total SST signal (not shown) indicates a slight decrease of the amplitude of the annual cycle (the variance accounted by the first two modes, strongly associated to the seasonal cycle, drops from 86% to 81% when realistic attenuation depth is used). Changes in MLD and SST agree with the linear model of Xie (1994), as the deepening of the mixed layer reduces the amplitude of the SST seasonal cycle.
The impact of a weaker seasonal cycle on the amplitude of ENSO has been previously reported (Jin et al. 1994; Tziperman et al. 1994; Liu 2002). Figures 9, 10 and 11 indicate that realistic attenuation depths, which deepen the mixed layer in the eastern equatorial Pacific and reduce the amplitude of the seasonal cycle of SST, amplify the eastern SST anomalies, especially in the HPSC simulation. Moreover, realistic attenuation depth increases the period between events, illustrated as a shift to longer periods on the power spectrum of the Niño-3 index (Fig. 10b). Both HPAM and HPSC simulations provide similar changes in the annual mean and the seasonal cycle. However, the amplitude of ENSO events is different in these simulations. This result illustrates that, by itself, a better seasonal cycle of SST or thermocline depth in the eastern equatorial Pacific does not automatically translate into an improvement in the simulation of ENSO (see Meehl et al. 2004). Causes for the different interannual variability between HPAM and HPSC cannot be retrieved from departures of the annual mean and seasonal cycle (as shown by Figs. 4 –10), because the difference in ENSO amplitude strongly project on these fields. For this reason, the discussion is restricted to the impact of the seasonal modulation of attenuation depth in the eastern equatorial Pacific.
To further hone these ideas, consider the 110°W longitude at the equator. The annual mean of the attenuation depth is 18.6 m (just 1.6 m deeper than in the control simulation). The range of the annual cycle of the attenuation depth is barely 1.0 m (Fig. 2; i.e., smaller than the difference respect to the control simulation). At this location, the attenuation depth is below the annual mean during May through mid-October, and above the annual mean the rest of the year. From June to October, a reduction in the attenuation depth increases the heat trapped in the mixed layer unless ocean dynamics dwindles the mixed layer below some critical value. When this effect is superimposed on the low frequency of ENSO dynamics (Fig. 9), in which both the thermocline depth and the MLD deepen several months before the maturity of a warm event, it always contributes to a enhanced warming of the eastern Pacific, which is completely absent in HPAM. The larger warming in the eastern Pacific further reduces the zonal SST gradient and reinforces the positive feedback leading to the maturity of the warm event in the eastern Pacific. During October–December, when the attenuation depth increases, the mixed layer is deep enough (anomalies as large as 30 m, see Fig. 9l) to make the absorbed radiative heat independent of changes in the attenuation depth. This combination of processes allows warm SST anomalies to persist for longer periods of time, increase their intensity, allowing for an enhanced eastward displacement of the wind stress anomalies (not shown), which positively feeds back to maintain and amplify these anomalies. The process reverses during a cold event. During a cold event, the thermocline rises, the mixed layer depth shrinks, and the temperature is reduced in the eastern Pacific. From June to September, the attenuation depth is seasonally reduced by about 0.5 m. This ought to warm the surface water; however, the decrease of the MLD (about −10 m) is larger than the critical value (h1 = 28 m, hp = 19 m, and Δh1 = −0.7 m), so the seasonal reduction of the attenuation depth does not contribute to the warming of the cold anomaly. After September, when the attenuation depth increases, more radiative heating leaves the mixed layer (not present in HPAM) and further cools the SST. The enhanced cooling of the eastern Pacific enhances the zonal SST gradient, strengthens the equatorial winds, increases equatorial upwelling that further feeds back to maintain the negative SST anomaly. Note that at seasonal scales, the role of the attenuation depth is rather symmetrical (Figs. 9 and 10) as both warm and cold anomalies are enhanced.
Finally, the composite of the heat leaving the mixed layer during warm events at this location is shown in Fig. 14a. The composite is constructed by averaging the months of the calendar for the years when the warm events peak. Note that in HPAM, warm events peak on different seasons, weakening the amplitude of the composite. The composite of the time series illustrate the reduction of heat leaving the mixed layer during the months of increased turbidity (the direct thermal effect during May–September), and the larger impact of the indirect dynamical mechanism at the end of the year, helping to maintain the warm anomalies in the eastern Pacific.
6. Summary and conclusions
The ability of using data from remotely sensed ocean color to parameterize biogenic heating in a hybrid coupled ocean–atmosphere model is investigated. Previous studies (see Marzeion et al. 2005) have presented evidence that ocean biology has an effect on the annual mean, the seasonal cycle, and the interannual variability in the tropical Pacific Ocean. In their study, Marzeion et al. (2005) used a three-component, ocean dynamics–ecosystem–statistical atmosphere model, to explore the influence of simulated chlorophyll on the coupled system. Here, a two-component HCM system is used with the parameterizations of biogeochemical feedbacks from remotely sensed ocean color.
The impact of the seasonal cycle of water turbidity on the annual mean, seasonal cycle, and interannual variability of the coupled model is investigated using three simulations differing in the parameterization of the attenuation depth: a control simulation (HP17) using a constant 17-m attenuation depth, and two additional simulations, HPAM and HPSC, which use a realistic attenuation depth spatial distribution. The first additional simulation, HPAM, accounts only for the annual mean of the attenuation depths. The last simulation, HPSC, accounts for the seasonal cycle of the attenuation depths. Our results indicate that accounting for a more realistic attenuation (independently of being constant in time or resolving the seasonal cycle) has a similar impact on both the annual mean and the seasonal cycle. However, the interannual variability of the coupled model is significantly affected when the seasonal cycle of the attenuation depth is used (HPSC). In that case, the amplitude of the warm events increase, the time period between events increases, and a more realistic phase locking of the events to the calendar year is simulated.
The strong impact of the seasonal cycle of the attenuation depths on the amplitude, frequency, and phase locking of ENSO agrees with the postulate that ENSO depends on the annual cycle of the ocean (see, e.g., Gu and Philander 1995). Simulation HPSC modifies the behavior of the Niño-3 index in the right direction. This indicates that the effects of biogenic heating are nonnegligible and affect the model biases, the seasonal cycle, and the interannual variability in the tropical Pacific.
Longitude–depth maps across the equator (Fig. 9) and the time series of the Niño-3 index (Fig. 10) indicate that seasonal attenuation depth has a rather symmetric impact on the amplitude of warm and cold events. This symmetry does not contradict the asymmetry proposed by Timmermann and Jin (2002), which applies to the interannual variability of the attenuation depth: they propose that the increase of biogenic activity should have a larger impact during cold events (warming the surface by increasing the amount of heat absorbed in the mixed layer) than during warm events. The parameterization of the interannual variability of the attenuation depth based on the remote sensing ocean color has not been investigated here.
Model simulations were also carried out with the diffuse attenuation at 490 nm provided by SeaWiFS for the period 1997–2003. The impact of the seasonal cycle of the attenuation depths on the ENSO behavior was nearly identical to the results presented here despite the significant differences in the amplitude of the seasonal cycle compared to the attenuation depths used here. This highlights the robustness of the proposed feedbacks between the seasonal cycle of the mixed layer and the radiative attenuation depths in both cases since the attenuation depths are deeper than the mixed layer during the boreal spring months and shallower than the mixed layer during the upwelling season of boreal summer and fall months. The potential role of the boreal spring subsurface chlorophyll maximum proposed by Murtugudde et al. (2002) is thus reaffirmed further underscoring the need for sufficient vertical resolution below the mixed layer to capture these feedbacks.
Acknowledgments
The authors want to thank the journal editor and the anonymous reviewers for their comments and contributions that improved the quality of this manuscript. We also acknowledge the TAO Project Office Director, Dr. Michael J. McPhaden, for providing the TAO observations and the OSCAR Project Office for providing OSCAR data. The ECHAM4.5 ensemble mean wind stress data were provided by Dr. M. Tippett at the International Research Institute for Climate Prediction (IRI). We acknowledge the NASA MODIS and NOAA/NESDIS grants for partial support to JB and RM. Partial support from NASA-Carbon grants are also acknowledged gratefully.
REFERENCES
Barber, R. T., and F. P. Chavez, 1983: Biological consequences of El Niño. Science, 222 , 1203–1210.
Barber, R. T., and F. P. Chavez, 1991: Regulation of primary productivity rate in the equatorial Pacific. Limnol. Oceanogr., 36 , 1803–1815.
Barnett, T. P., M. Latif, N. Graham, M. Flugel, S. Pazan, and W. White, 1993: ENSO and ENSO-related predictability. Part I: Prediction of equatorial Pacific sea surface temperature with a hybrid coupled ocean–atmosphere model. J. Climate, 6 , 1545–1566.
Bjerknes, J., 1969: Atmospheric teleconnections from equatorial Pacific. Mon. Wea. Rev., 97 , 163–172.
Bonjean, F., and G. S. E. Lagerloef, 2002: Diagnostic model and analysis of the surface currents in the tropical Pacific Ocean. J. Phys. Oceanogr., 32 , 2938–2954.
Chang, P., L. Ji, and R. Saravanan, 2001: A hybrid coupled model study of tropical Atlantic variability. J. Climate, 14 , 361–390.
Chen, D., L. M. Rothstein, and A. J. Busalacchi, 1994: A hybrid vertical mixing scheme and its application to tropical ocean models. J. Phys. Oceanogr., 24 , 2156–2179.
Christian, J. R., and R. Murtugudde, 2003: Tropical Atlantic variability in a coupled physical-biogeochemical ocean model. Deep-Sea Res. II, 50 , 2947–2969.
Christian, J. R., M. A. Verschell, R. Murtugudde, A. J. Busalacchi, and C. R. McClain, 2002: Biogeochemical modelling of the tropical Pacific Ocean. I: Seasonal and interannual variability. Deep-Sea Res. II, 49 , 509–543.
Denman, K. L., 1973: A time-dependent model of the upper ocean. J. Phys. Oceanogr., 3 , 173–184.
Dickey, T. D., and J. J. Simpson, 1983: The influence of optical water type on the diurnal response of the upper ocean. Tellus, 35B , 142–154.
ECMWF, 1994: Description of the ECMWF/WCRP Level-III. A global atmospheric data archive, 72 pp.
Edwards, A. M., T. Platt, and D. G. Wright, 2001: Biologically induced circulation at fronts. J. Geophys. Res., 106 , 7081–7095.
Gildor, H., and N. H. Naik, 2005: Evaluating the effect of interannual variations of surface chlorophyll on upper ocean temperature. J. Geophys. Res., 110 .C07012, doi:10.1029/2004JC002779.
Gildor, H., A. H. Sobel, M. A. Cane, and R. N. Sambrotto, 2003: A role for ocean biota in tropical intraseasonal atmospheric variability. Geophys. Res. Lett., 30 .1460, doi:10.1029/2002GL016759.
Gu, D., and S. G. H. Philander, 1995: Secular changes of annual and interannual variability in the Tropics during the past century. J. Climate, 8 , 864–876.
Jerlov, N. G., 1968: Optical Oceanography. Elsevier Oceanography Series, Vol. 5, Elsevier, 154 pp.
Jin, F. F., J. D. Neelin, and M. Ghil, 1994: El Niño on the devil’s staircase—Annual subharmonic steps to chaos. Science, 264 , 70–72.
Kantha, L. H., and C. A. Clayson, 1994: An improved mixed-layer model for geophysical applications. J. Geophys. Res., 99 , 25235–25266.
Kara, A. B., H. E. Hurlburt, P. A. Rochford, and J. J. O’Brien, 2004: The impact of water turbidity on interannual sea surface temperature simulations in a layered global ocean model. J. Phys. Oceanogr., 34 , 345–359.
Lewis, M. R., M. E. Carr, G. C. Feldman, W. Esaias, and C. McClain, 1990: Influence of penetrating solar-radiation on the heat-budget of the equatorial Pacific Ocean. Nature, 347 , 543–545.
Liu, Z. G., 2002: A simple model study of ENSO suppression by external periodic forcing. J. Climate, 15 , 1088–1098.
Lukas, R., and E. Lindstrom, 1991: The mixed layer of the western equatorial Pacific Ocean. J. Geophys. Res., 96 , 3343–3357.
Manizza, M., C. Le Quéré, A. J. Watson, and E. T. Buitenhuis, 2005: Bio-optical feedbacks among phytoplankton, upper ocean physics and sea-ice in a global model. Geophys. Res. Lett., 32 .L05603, doi:10.1029/2004GL020778.
Marzeion, B., A. Timmermann, R. Murtugudde, and F-F. Jin, 2005: Biophysical feedbacks in the tropical Pacific. J. Climate, 18 , 58–70.
McClain, C. R., M. L. Cleave, G. C. Feldman, W. W. Gregg, S. B. Hooker, and N. Kuring, 1998: Science quality SeaWiFS data for global biosphere research. Sea Technol., 39 , 10–16.
McPhaden, M. J., and Coauthors, 1998: The Tropical Ocean Global Atmosphere observing system: A decade of progress. J. Geophys. Res., 103 , 14169–14240.
Meehl, G. A., W. M. Washington, J. M. Arblaster, and A. X. Hu, 2004: Factors affecting climate sensitivity in global coupled models. J. Climate, 17 , 1584–1596.
Miller, A. J., and Coauthors, 2003: Potential feedbacks between Pacific Ocean ecosystems and interdecadal climate variations. Bull. Amer. Meteor. Soc., 84 , 617–633.
Mobley, C. D., 1989: A numerical-model for the computation of radiance distributions in natural-waters with wind-roughened surfaces. Limnol. Oceanogr., 34 , 1473–1483.
Monterey, G. I., and S. Levitus, 1997: Climatological cycle of mixed layer depth in the World Ocean. NOAA/NESDIS, 5 pp.
Morel, A., 1988: Optical modeling of the upper ocean in relation to its biogenous matter content (Case I waters). J. Geophys. Res., 93 , 10749–10768.
Murtugudde, R. G., S. R. Signorini, J. R. Christian, A. J. Busalacchi, C. R. McClain, and J. Picaut, 1999: Ocean color variability of the tropical Indo-Pacific basin observed by SeaWiFS during 1997–1998. J. Geophys. Res., 104 , 18351–18366.
Murtugudde, R., J. Beauchamp, C. R. McClain, M. Lewis, and A. J. Busalacchi, 2002: Effects of penetrative radiation on the upper tropical ocean circulation. J. Climate, 15 , 470–486.
Nakamoto, S., S. P. Kumar, J. M. Oberhuber, K. Muneyama, and 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. P. Kumar, J. M. Oberhuber, J. Ishizaka, K. Muneyama, and R. Frouin, 2001: Response of the equatorial Pacific to chlorophyll pigment in a mixed layer isopycnal ocean general circulation model. Geophys. Res. Lett., 28 , 2021–2024.
Nakamoto, S., S. P. Kumar, J. Oberhuber, H. Saito, K. Muneyama, and R. Frouin, 2002: Chlorophyll modulation of mixed layer thermodynamics in a mixed-layer isopycnal General Circulation Model—An example from Arabian Sea and equatorial Pacific. Proc. Ind. Acad. Sci.-Earth Planet. Sci., 111 , 339–349.
Ohlmann, J. C., 2003: Ocean radiant heating in climate models. J. Climate, 16 , 1337–1351.
Ohlmann, J. C., D. A. Siegel, and C. Gautier, 1996: Ocean mixed layer radiant heating and 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 Pacific during TOGA-COARE. J. Geophys. Res., 103 , 5379–5395.
Ohlmann, J. C., D. A. Siegel, and C. D. Mobley, 2000: Ocean radiant heating. Part I: Optical influences. J. Phys. Oceanogr., 30 , 1833–1848.
Oschlies, A., 2004: Feedbacks of biotically induced radiative heating on upper-ocean heat budget, circulation, and biological production in a coupled ecosystem-circulation model. J. Geophys. Res., 109 .C12031, doi:10.1029/2004JC002430.
Paulson, C. A., and J. J. Simpson, 1977: Irradiance measurements in the upper ocean. J. Phys. Oceanogr., 7 , 952–956.
Reynolds, R. W., and T. M. Smith, 1994: Improved global sea surface temperature analyses using optimum interpolation. J. Climate, 7 , 929–948.
Rochford, P. A., A. B. Kara, A. J. Wallcraft, and R. A. Arnone, 2001: Importance of solar subsurface heating in ocean general circulation models. J. Geophys. Res., 106 , 30923–30938.
Roeckner, E., and Coauthors, 1996: The atmospheric general circulation model ECHAM-4: Model description and simulation of present-day climate. Tech. Rep. 218, Max Planck Institute for Meteorology, Hamburg, Germany, 90 pp.
Sathyendranath, S., A. D. Gouveia, S. R. Shetye, P. Ravindran, and T. Platt, 1991: Biological-control of surface-temperature in the Arabian Sea. Nature, 349 , 54–56.
Schneider, E. K., and Z. X. Zhu, 1998: Sensitivity of the simulated annual cycle of sea surface temperature in the equatorial Pacific 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.
Shell, K. M., R. Frouin, S. Nakamoto, and R. C. J. Somerville, 2003: Atmospheric response to solar radiation absorbed by phytoplankton. J. Geophys. Res., 108 .4445, doi:10.1029/2003JD003440.
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 Pacific warm pool. J. Geophys. Res., 100 , 4885–4891.
Simpson, J. J., and T. D. Dickey, 1981a: The relationship between downward irradiance and upper ocean structure. J. Phys. Oceanogr., 11 , 309–323.
Simpson, J. J., and T. D. Dickey, 1981b: Alternative parameterizations of downward irradiance and their dynamical significance. J. Phys. Oceanogr., 11 , 876–882.
Strutton, P. G., and F. P. Chavez, 2004: Biological heating in the equatorial Pacific: Observed variability and potential for real-time calculation. J. Climate, 17 , 1097–1109.
Syu, H. H., J. D. Neelin, and D. Gutzler, 1995: Seasonal and interannual variability in a hybrid coupled GCM. J. Climate, 8 , 2121–2143.
Timmermann, A., and F. F. Jin, 2002: Phytoplankton influences on tropical climate. Geophys. Res. Lett., 29 .2104, doi:10.1029/2002GL015434.
Turk, D., M. R. Lewis, G. W. Harrison, T. Kawano, and I. Asanuma, 2001: Geographical distribution of new production in the western/central equatorial Pacific during El Niño and non-El Niño conditions. J. Geophys. Res., 106 , 4501–4515.
Tziperman, E., L. Stone, M. A. Cane, and H. Jarosh, 1994: El-Niño chaos—Overlapping of resonances between the seasonal cycle and the Pacific Ocean–atmosphere oscillator. Science, 264 , 72–74.
Tziperman, E., M. A. Cane, S. E. Zebiak, Y. Xue, and B. Blumenthal, 1998: Locking of El Niño peak time to the end of the calendar year in the delayed oscillator picture of ENSO. J. Climate, 11 , 2191–2199.
Xie, S-P., 1994: On the genesis of the equatorial annual cycle. J. Climate, 7 , 2008–2013.
Zhang, R-H., and A. J. Busalacchi, 2005: Interdecadal changes in properties of El Niño in an intermediate coupled model. J. Climate, 18 , 1369–1380.
Zhang, R-H., S. E. Zebiak, R. Kleeman, and N. Keenlyside, 2003: A new intermediate coupled model for El Niño simulation and prediction. Geophys. Res. Lett., 30 .2012, doi:10.1029/2003GL018010.
Zhang, R-H., S. E. Zebiak, R. Kleeman, and N. Keenlyside, 2005: Retrospective El Niño forecasts using an improved intermediate coupled model. Mon. Wea. Rev., 133 , 2777–2802.
(a) Average distribution of the SeaWiFS-based attenuation depth in meters. The 9-km, level-3 monthly anomalies have been binned to the ocean model grid. (b) Standard deviation of the attenuation depth. Shaded areas indicate regions where the standard deviation is larger than 2 m.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3958.1
Seasonal modulation around the annual mean at the equator (2°S–2°N) for SeaWiFS-derived attenuation depth (m). Shaded region indicates negative values. Contour interval is 0.2 m. First positive and negative intervals are +0.1 and −0.1 m, respectively.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3958.1
Temporal evolution of the equatorial (2°S–2°N) SST for simulation (a) HP17, (b) HPAM, and (c) HPSC. The contour interval is 1°C. Highlighted are temperatures lower than 24°C in the east and larger than 26°C in the west.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3958.1
(a) Annual mean of Reynolds SST in the OGCM spatial domain. The sponge layer regions (between 20° and 25° latitudes) have been cropped. Differences between the (b) HP17 SST annual mean and the Reynolds SST annual mean, (c) HPAM and Reynolds SST, and (d) HPSC and Reynolds SST.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3958.1
Same as in Fig. 4 but for the depth of the 20°C isotherm. Units are in m.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3958.1
Annual mean of the latitudinal structure of the zonal surface current at (a) 165°E, (b) 155°W, and (c) 110°W. Units are in m s−1. Each plot compares the estimated value from the OSCAR estimates and for each simulation.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3958.1
Longitude–depth distribution of temperature at the equator. (a) Annual mean of HP17. (b) Difference between the annual mean of HPAM and the annual mean of HP17. (c) Same as in (b), but for HPSC. Units are in °C. Superimposed to (b) and (c) are the annual mean of the MLD. The dashed MLD corresponds to HP17. The solid MLD corresponds to (b) HPAM and (c) HPSC.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3958.1
Longitude–time maps accounting for the seasonal cycle of equatorial SST in °C: (a) Reynolds data (contour interval 0.5; shaded indicates negative values); (b) difference between HP17 and Reynolds seasonal cycles; differences between (c) HPAM and (d) HPSC, respectively, to Reynolds. Contour interval for (b), (c), and (d) is 0.4; first positive and negative contours are +0.2 and −0.2, respectively.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3958.1
Longitude–time plots of the time evolution of the equatorial (2°S–2°N) anomalies for the three simulations. (a), (b), (c) SST. Contour interval is 0.5°C. The first solid contour corresponds to 0.25°C. (d), (e), (f) Zonal wind stress. Contour interval is 0.010 dyn cm−2. The first solid contour corresponds to 0.005 dyn cm−2. (g), (h), (i) Depth of the 20°C isotherm. Contour interval is 10 m. The first solid contour corresponds to 5 m. (j), (k), (l) Depth of the mixed layer. Contour interval is 10 m. The first solid contour corresponds to 5 m.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3958.1
(a) Time series of the Niño-3 index for the control, HPAM, and HPSC simulations. (b) Power spectrum of the normalized Niño-3 indices (all three time series have variance equal to 1).
Citation: Journal of Climate 20, 2; 10.1175/JCLI3958.1
Standard deviation of monthly anomalies for the (a) Reynolds SST, (b) the control simulation, (c) HPAM, and (d) HPSC. Units are in °C.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3958.1
Histogram with the frequency distribution as a function of the month of the year in which the warm events peak.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3958.1
Map of exp[−MLD(HPSC)/HP] − exp[-MLD(HP17)/17], i.e., the difference on radiative flux across the bottom of the mixed layer. The annual mean of the MLD and attenuation depth are considered. This map indicates where the flow of radiant heat across the lower limit of the mixed layer increases (positive values) or decreases (negative values). The contour interval is 0.05.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3958.1
(a) Time series of the fraction of heat leaving the mixed layer for a 25-month warm event composite. The higher the value, the more heat leaves the mixed layer. Horizontal lines indicate the 10%, 30%, and 50% of heat leaving the surface layer. (b) Seasonal cycle of the SeaWiFS attenuation depth. Both series correspond to 110°W at the equator.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3958.1
