1. Introduction
Recent studies have revealed pronounced interactions between the climate and ocean biology in the tropical Pacific. ENSO is a major driver that strongly affects the upwelling and mixing processes that induce changes in ocean biology in the region (Philander 1983; Zhang and Levitus 1997). For example, ocean observational data indicate large anomalies of ocean physical and biological fields during El Niño and La Niña events, including chlorophyll (Chl) and sea surface temperature (SST). Indeed, coherent relationships exist between variations in ocean biological and physical fields. On interannual time scales, observed changes in ocean biology (e.g., Chl) closely follow those in physical fields (e.g., SST) in association with ENSO, which indicates that changes in ocean physical fields are a dominant driving force and that changes in ocean biological fields represent a response.
On the other hand, changes in ocean biology can produce feedback to the ocean physics. Solar radiation is the main energy source for Earth’s climate system. Ocean biological components [such as phytoplankton, detritus, and colored dissolved organic matter (CDOM)] can absorb this radiation in the visible spectrum (380–700 nm) and further modify the vertical penetration of the incoming solar radiation in the upper ocean (Lewis et al. 1990; Morel and Antoine 1994; Siegel et al. 2002; Strutton and Chavez 2004). A penetration depth (Hp) can be introduced to represent the ocean biology–induced heating (OBH) effects (Murtugudde et al. 2002; Zhang et al. 2011), serving as a linkage between ocean biology and physics. For example, large Chl anomalies are observed during ENSO cycles, which has been demonstrated to have significant effects on the redistribution of solar radiation and stratification in the upper ocean. These effects can modulate SST in the tropical Pacific, a field that affects the atmosphere, which in turn feeds back to the ocean. Thus, interannual variations in Chl not only indicate a response to ENSO-induced physical changes, but also simultaneously serve as a feedback to ENSO. To quantify these effects, several heating terms can be defined that are directly related to Chl (Zhang 2015a). Through modulating these heating terms, the biological conditions can affect the mean climate and its variability over the tropical Pacific. The interactions between ocean biological and physical processes have thus received considerable attention in recent years.
Various approximations and simplifications are often made in parameterizing bioheating processes in climate models; there thus exist large and strongly model-dependent uncertainties in representing their effects. For example, previous studies have investigated the impacts of phytoplankton (i.e., chlorophyll) on the mean ocean state in the tropical Pacific. Even in ocean-only modeling studies with prescribed chlorophyll concentration taken from satellite ocean color data, chlorophyll-induced effects on SSTs in the eastern equatorial Pacific indicate a cooling effect (e.g., Nakamoto et al. 2001; Anderson et al. 2007, 2009; Lin et al. 2008; Gnanadesikan and Anderson 2009), whereas other models indicate a warming effect (e.g., Murtugudde et al. 2002). In terms of the effects on interannual variations, phytoplankton has been demonstrated to exert a significant influence on ENSO. Using an intermediate coupled model, Timmermann and Jin (2002) found that phytoplankton blooms tend to damp La Niña conditions. Using fully represented biogeochemical models, recent studies have investigated bioclimate interactions and biological impacts on ENSO, which can be significantly different across models (Manizza et al. 2005, 2008; Marzeion et al. 2005; Wetzel et al. 2006; Lengaigne et al. 2007; Löptien et al. 2009; Jochum et al. 2010; Patara et al. 2012; Park et al. 2011, 2014a,b). For example, Marzeion et al. (2005) and Lengaigne et al. (2007) suggested that the ENSO amplitude is increased when biofeedback is included, which can be attributed to weaker seasonal variability because the reduced amplitude of seasonal cycles can give more freedom for ENSO to develop. By contrast, Jochum et al. (2010) suggested that interannual chlorophyll variation acts to reduce the ENSO amplitude when using a coupled atmosphere–ocean physics and ocean biology models. Park et al. (2014b) further demonstrated that the presence of chlorophyll tends to increase the ENSO amplitude resulting from the shoaling of the mean thermocline in the eastern equatorial Pacific.
We investigated bioeffects using a simplified modeling configuration in our previous studies (Zhang et al. 2009, 2015; Zhang 2015a,b). The statistical relationship between interannual variations in SST and Hp is used to derive an empirical model for interannual variability of Hp from historical data. The derived empirical Hp model was then used to represent the bioeffects on physics in a hybrid coupled model (HCM), in which anomaly coupling was considered between ocean biological and physical processes (i.e., interannual Hp anomalies were represented as a linear response to SST). To understand the processes involved, the Hp field has been used to quantify the effect on Chl-induced heating terms (Ballabrera-Poy et al. 2007; Zhang 2014). The relationships among related variables were analyzed in the HCM simulations to illustrate where and how Hp could most affect heating terms in the tropical Pacific. It was demonstrated that interannual Hp anomalies acted as a negative feedback on ENSO in the HCM-based simulations, a result that is different from that of a previous similar study by Marzeion et al. (2005), who identified a positive feedback in their model. In terms of the physical processes involved, pathways through which the bioinduced heating effects are realized were further investigated. The dominant processes that determine the biofeedback in our HCM-based simulations were attributed to an indirect heating effect on the subsurface layer and stratification in the upper ocean (Zhang 2015a,b), which is apparently different from the direct heating effect on SST within the mixed layer proposed by Timmermann and Jin (2002), who used a simplified coupled ocean–atmosphere model. These different and even contradictory modeling results indicate that biofeedback effects and their underlying mechanisms are sensitively dependent on the way Hp (which can be estimated from Chl) is represented in the model simulations. Indeed, as is evident in the previous modeling studies, a subtle change in Hp in model representations can have completely different effects and can even lead to sign changes in SST simulations.
In this work, we continued to investigate the bioeffects on the climate in the tropical Pacific using an improved ocean biology model (Wang et al. 2015), which is used to depict Chl and Hp fields. This biomodel consisted of 12 components, including six nutrients (nitrate, silicon, dissolved inorganic carbon, ammonium, dissolved oxygen, and dissolved iron) and six biological components (large and small sizes of phytoplankton, large and small sizes of zooplankton, and large and small detritus) (Wang et al. 2006; Christian et al. 2001; Wang et al. 2008). The ocean biological model was coupled to an ocean general circulation model (OGCM) in which Chl concentration was interactively and dynamically calculated. The model can simulate not only physical fields but also ocean biological fields. Because Chl was a major component, in this paper we focused on its effect using ocean-only modeling experiments forced by prescribed atmospheric forcing fields; the use of ocean-only simulations enabled us to exclude ocean–atmosphere coupling effects. Two experiments were performed. In a control run (referred to as Chlinter), the Chl concentration varied interannually and was interactively calculated from the ocean biology model. In a climatological run (referred to as Chlclim), the seasonally varying climatological Chl field derived from the Chlinter run was prescribed. Even though the wind forcing was prescribed to be the same in the two runs, SST simulations were found to have systematic differences. Because ENSO is the strongest interannual signal with large climate effects in the tropical Pacific, our analyses in this paper will focus on the modulating effects on ENSO. Some specific questions will be addressed: How are the Chl-induced feedback effects realized? What are the relationships between the differences in SST simulations and bioeffects in the ocean-only experiments? Can the differences in the way Chl represented cause systematic effects on SST in these ocean-only model simulations? Are the relationships between Chl-induced effects and SST differences consistent with those derived from our previous HCM-based simulations?
This paper is organized as follows. Section 2 describes the ocean biology model, the data used for validation, and the experimental design. The heating terms affected by Hp and Chl are described in section 3. Sections 4 and 5 present the results of the two simulations, highlighting the biological impacts on the mean state and ENSO. Section 6 provides a summary and discussion.
2. Model and data
a. A coupled ocean physics–biology model
The ocean physics model used in this study was a reduced-gravity, primitive equation, sigma-coordinate ocean general circulation model specifically developed for the upper equatorial ocean (Gent and Cane 1989). Some notable features included a bulk mixed layer model embedded into the OGCM (Chen et al. 1994) and an advective atmosphere mixed layer (AML) model coupled into the ocean model to calculate sea surface heat fluxes (Siegel et al. 1995; Murtugudde et al. 1996). The configuration of the ocean model used in this study has been described in detail by Wang et al. (2009). The model had 20 vertical layers, including an explicitly represented mixed layer at the top with its depth determined by a bulk mixed layer model (Chen et al. 1994). The model domain is in the tropical Pacific from 120°E to 76°W and from 30°S to 30°N. The horizontal resolutions varied: zonal resolutions were 1° in the central basin and gradually decreased to 0.4° in the western and eastern boundaries, and meridional resolutions were from 0.3° to 0.6° between 15°S and 15°N and increased to 2° at the northern and southern boundaries. The temperature, salinity, nitrate, and other attributes were gradually relaxed to their climatological fields obtained from the World Ocean Atlas 1998 (WOA98) (http://www.nodc.noaa.gov/OC5/indprod.html) in the sponge layer within the 10° domain near the northern and southern boundaries.
A schematic diagram illustrating the ocean biology model with its 12 biological components.
Citation: Journal of Climate 31, 5; 10.1175/JCLI-D-17-0505.1
Additionally, some diagnostic variables were introduced. For example, phytoplankton carbon biomass (mol N m−3) was converted to mol C m−3 according to the Redfield ratio (C:N = 6.625), and chlorophyll concentration was then calculated using a ratio between phytoplankton biomass and chlorophyll, called the C:Chl ratio (Wang et al. 2009). The biological model can also produce the detritus field in association with phytoplankton and zooplankton mortality and zooplankton excretion; for example, detritus can be grazed on by small zooplankton and decomposed by bacteria.
b. An optical model in the upper ocean
c. Experimental designs
The ocean model was initiated from temperature and salinity fields obtained from the World Ocean Atlas 2001 data (WOA01) and then integrated for 30 years using atmospheric climatological forcing fields. Then, the model was integrated from 1980 to 2007, forced by 6-day means of surface wind stress from the National Centers for Environmental Prediction (NCEP)–National Center for Atmospheric Research (NCAR) reanalysis (Kalnay et al. 1996). Note that solar radiation and cloudiness were prescribed to be their climatological monthly fields, so their interannual variations were not considered in the model simulations.
To investigate the effect of chlorophyll-induced heating, two experiments were performed by using the coupled ocean physics–biology model with Chl being represented interannually or climatologically. In an interannual Chl run (called Chlinter), the interannually varying Chl concentration was interactively calculated from the ocean biology model with an embedded bio-optical scheme to represent direct ocean biology–induced heating effects on the ocean thermodynamics. In a climatological run (called Chlclim), the seasonally varying climatological Chl field was prescribed that was derived from the Chlinter run. The model outputs from 1985 to 2007 were used for analyses.
d. Observational data used for model validations
Some observational data were used to validate the model simulations. Surface chlorophyll datasets were obtained from the GlobColour project from 1998 to 2007, which supplied continuous datasets for merged Level-3 Ocean Color products (including SeaWIFS, MODIS, MERIS, and VIIRS sensors; see details at http://hermes.acri.fr/index.php; Maritorena et al. 2010). Then, monthly CHL-1 data [chlorophyll concentration (mg m−3) for case 1 waters] were interpolated from 0.25° × 0.25° grids to the model grids. The surface wind stress fields used to force the ocean model were obtained from the NCEP–NCAR reanalysis (Kalnay et al. 1996). The SST fields were obtained from Reynolds et al. (2002).
3. The heating terms affected (Qpen, Qabs, and Rsr)
The ocean biology–induced heating effect is represented by Hp because these terms are explicitly related to Hp. As a function of both Hm and Hp, for example, Qabs increases exponentially with Hm but decreases exponentially with Hp; that is, the deeper the ML is, the more solar radiation is directly absorbed within the ML. Similarly, these expressions indicate that the larger Hp is, the less the solar radiation directly absorbed within the ML. Therefore, changes in Hm and Hp tend to have an opposite effect on Qabs. That is, a negative (positive) perturbation of Hm produces a reduction (an increase) in Qabs, whereas a negative (positive) perturbation of Hp produces an increase (a reduction) in Qabs. The absorbed part (Qabs) within the ML directly produces a perturbation to Rsr and thus a change to SST. However, the extent to which the Qabs and Rsr fields are affected by Hp can be different as indicated by these mathematical expressions. In addition to being proportional to Qabs (which has an exponential relation with Hm and Hp), Rsr is also inversely proportional to Hm. As such, changes in both Qabs and Hm can give rise to the effects on Rsr, which needs to be accounted for when analyzing the ocean biology–induced heating effects on Rsr (as indicated by Hp). More specifically, Hm can have twofold effects on Rsr that tend to be opposite. On the one hand, because Hm is exponentially related to Qabs, a deepening (shoaling) of the ML, which causes an increase (a decrease) in Qabs, leads to an increase (a decrease) in Rsr. On the other hand, because Hm is inversely proportional to Rsr as well, the deepening (shoaling) of the ML leads to a decrease (an increase) in Rsr. With regards to the Qpen field, its exponential relationships with Hm and Hp are opposite to those of Qabs.
The effects of ocean biology–induced heating on the penetrating solar radiation are realized through these terms that are explicitly related to Hp. Therefore, the relationships among these related fields can be analyzed to understand the effects of the OBH-related feedback and the processes involved. Two influence pathways are possible by which variations in Hp can affect SST in the equatorial Pacific, depending on the relative dominance of which term is modulated by Hp most significantly. On one hand, if a change in Hp induces a significant change to Rsr, the biofeedback can be realized through a direct heating effect on SST [i.e., the change to Qabs/ (ρ0CpHm) induced by Hp can directly heat the ML, which causes changes to SST]; this is referred to as a direct effect on SST. If a change in Hp does not significantly affect Rsr, then direct heating effects within the ML [Qabs/ (ρ0CpHm)] are not a major process that causes SSTs to change. On the other hand, if a change in Hp causes a significant modulation of Qpen, a differential heating can then be induced vertically between the ML (Qabs, the absorbed part in the ML) and subsurface layers (Qpen, the penetrated part through the bottom of the ML and into the subsurface layers). These processes can further modify the stratification and vertical mixing in the upper ocean and thus affect SST; this is referred to as an indirect effect on SST. Therefore, the extent to which these heating terms are affected by Hp implies different dominant processes that are involved in the bioeffects. How the Rsr and/or Qpen fields are affected by changes in Hp can be a good indicator of which influence pathways are taken and the underlying processes are operating in association with the ocean biology–induced heating effects. These will be analyzed using ocean-only modeling experiments.
Some issues will be addressed in detail: Which term is affected dominantly by Hp? In which regions can these heating terms be mostly affected by Hp? How are these heating terms collectively modulated by the interannual variability of Hm (a major physical factor) and/or of Hp (a biological factor)? Model outputs were used to reveal the structure of these affected fields and their relationships to understand the bioeffects.
4. A simulation with biofeedback on the physics included
A reference run was performed in which Chl field is interannually varying, which was interactively calculated in the ocean biology model (denoted as Chlinter). The model with biofeedback explicitly included can very well simulate the mean state and interannual variations of not only physical but also biological fields, including Chl. Then, Chl was used to calculate Hp, which was explicitly related to the heating terms. The relationships among these related physical and biological fields (e.g., including heating terms affected by Hp) were analyzed to understand the biofeedback effects and the underlying processes involved.
a. Model validations of the simulated Chl field
The chlorophyll concentration simulated from Chlinter was compared with that observed from satellites. Figure 2 displays the annual-mean chlorophyll concentration simulated in the ML. Compared with the observations, the model reproduced the basic pattern of mean chlorophyll concentration in most regions of the tropical Pacific. However, there were notable differences between simulations and observations in the eastern coastal region and the equatorial region (Figs. 2a,c). For example, the mean concentration in the model was slightly higher than the observations (≈0.3 mg m−3) in the equatorial region. In terms of interannual variations, a large variability in chlorophyll concentration was observed in the equatorial and coastal regions. The spatial pattern of the standard deviation (STDV) of chlorophyll in the model (Fig. 2d) closely resembled the satellite observations (Fig. 2b). However, the STDV was weaker than observations in the equatorial region, especially in the 5°–10°N region. Overall, the model captured the main characteristics of the mean chlorophyll distribution and its variability in the tropical Pacific. As described above, Chl can be used to estimate Hp, which appears explicitly in the three heating terms.
Horizontal distributions of annual-mean fields (top) observed from the Global Color Project and (bottom) simulated from the model, (a),(c) for Chl and (b),(d) for the corresponding standard deviation of interannual anomalies. The observed fields are calculated during 1997–2015 and the simulated ones are calculated during 1980–2007. The contour interval is 0.05 mg m−3 in (a) and (c), and 0.02 mg m−3 in (b) and (d).
Citation: Journal of Climate 31, 5; 10.1175/JCLI-D-17-0505.1
b. Mean climatological fields
Figure 3 shows the horizontal distributions of annual-mean fields for Hm and Hp from the Chlinter run. The values of Hp were small compared with those of Hm in most of the tropical Pacific. The incoming solar radiation tended to decay with depth in the upper ocean; some portion can penetrate the bottom of the ML. As expressed above, the three heating terms are dominantly affected by both Hm and Hp.
Horizontal distributions of annual-mean fields simulated from the model for (a) Hm and (b) Hp. The contour interval is 10 m in (a) and 1 m in (b).
Citation: Journal of Climate 31, 5; 10.1175/JCLI-D-17-0505.1
Figure 4 shows the horizontal distributions of annual-mean fields for Qpen, Rsr, and Qabs. As indicated by the Qpen field. the structure of the Qpen field was similar to that of Hm, and so Hm was a dominant factor in dictating the spatial pattern of these heating terms. As indicated by Qabs (Fig. 4b), for example, most of the solar radiation was absorbed within the ML; only small portion can penetrate the bottom of the ML into the subsurface layer, which is indicated by Qpen (Fig. 4a). As defined above, the effects of a change in Hm on the Qabs and Qpen fields tended to be opposite. For example, a deep ML led to the incoming solar radiation that was absorbed more within the ML (a large Qabs portion) and less penetrated into the subsurface layers (a smaller Qpen portion). In contrast, a shallow ML resulted in the solar radiation that can penetrate more through the bottom of the ML (a larger Qpen portion), and a reduction of the heat uptake within the ML (a smaller Qabs portion). Climatologically, the Hp effects on these heating terms are not evident.
Horizontal distributions of annual-mean fields simulated from the model for (a) Qpen, (b) Qabs, and (c) Rsr. The contour interval is 4 W m−2 in (a), 10 W m−2 in (b), and 1°C month−1 in (c).
Citation: Journal of Climate 31, 5; 10.1175/JCLI-D-17-0505.1
c. Interannual variability
Figure 5 displays the time–longitude sections along the equator for total fields of SST and Chl simulated from the model. Pronounced anomalies are associated with ENSO. The corresponding interannual anomalies along the equator for SST and Chl are shown in Fig. 6. Large anomalies of physical and biological fields in the tropical Pacific are induced by ENSO events because ENSO is a major driver of interannual variations in the region. For example, changes in ocean physics produce changes to ocean biology as represented by Chl. During ENSO cycles, pronounced interannual Chl anomalies were observed in the tropical Pacific, with regions of large variability occurring in the western-central regions. For example, low Chl concentrations were observed during an El Niño event in the western-central Pacific. In contrast, a La Niña event was accompanied by high Chl concentrations in the western-central and eastern equatorial Pacific, and low Chl concentrations in the far western region. Therefore, the interannual variations in Chl correlated well with those in SSTs over the tropical Pacific, with the former closely following the latter during ENSO evolution.
Time–longitude sections along the equator for total fields of (a) SST and (b) Chl simulated from the model. For better visibility, only the model results for 1981–92 are displayed. The contour interval is 1.0°C in (a) and 0.05 mg m−3 in (b).
Citation: Journal of Climate 31, 5; 10.1175/JCLI-D-17-0505.1
Time–longitude sections along the equator for interannual anomalies of (a) SST and (b) Chl simulated from the model. The contour interval is 0.5°C in (a) and 0.05 mg m−3 in (b).
Citation: Journal of Climate 31, 5; 10.1175/JCLI-D-17-0505.1
Figure 7 displays interannual variations in Hm and Hp along the equator from the Chlinter run. Pronounced interannual Hm anomalies were seen in the tropical Pacific during ENSO evolution (Fig. 7a). For instance, La Niña events are accompanied with an anomalously deep ML in the western-central basin (Fig. 7a) and anomalously shallow ML in the east. El Niño events are accompanied with an opposite situation (Fig. 7).
Time–longitude sections along the equator for interannual anomalies of (a) Hm and (b) Hp simulated from the model. The contour interval is 4.0 m in (a) and 1 m in (b).
Citation: Journal of Climate 31, 5; 10.1175/JCLI-D-17-0505.1
Large variability occurred in the western-central regions. As analyzed by Zhang et al. (2011), large interannual Hp anomalies were induced by ENSO cycles. During La Niña events, for example, a negative Hp anomaly was observed in the western-central and eastern equatorial Pacific, accompanied with a positive Hp anomaly in the far western region. During El Niño events, an opposite anomaly pattern was observed. Also evident, interannual variations in Hp and SST are correlated well over the tropical Pacific, with the former closely following the latter during ENSO cycles. Note that the interannual Hp variability simulated from the model (Fig. 7b) closely resembled that estimated from satellite data (Zhang et al. 2011). For example, the amplitudes of the interannual Hp variations simulated in the Chlinter run were comparable to the satellite data.
Comparisons between interannual variations in these anomaly fields indicated their well-defined structure and relationship during ENSO cycles. During El Niño events, the ML was shallow in the western-central regions, which was accompanied by an increase in Hp; during La Niña events, the ML was deep in the western-central equatorial Pacific, accompanied by a decrease in Hp. Moreover, in the western-central basin, the interannual variations of Hp tended to have amplitudes that were comparable to those of Hm during ENSO events. The relatively large variability of Hp in the western-central equatorial Pacific is expected to exert a strong influence on the OBH terms in the region. However, the amplitude of interannual variations of Hp (Fig. 7b) was relatively small in the eastern equatorial region compared with that of Hm (Fig. 7a); so, the modulating effects of the Hp variability on these OBH terms were small in the east. Quantitatively, the standard deviations (STDV) of Hp and Hm anomalies were 1.21 and 3.85 m in the Niño-4 region, and 1.12 and 4.72 m in the Niño-3 region, respectively (Table 1).
The standard deviation of anomaly fields simulated in the Chlinter run from 1981 to 1992.
d. Modulating effects of interannual Hp variability on the OBH terms
The interannual variability of simulated Hp exhibited a well-defined relationship with the interannual variability of the heating terms. Although Hm is a major factor determining the vertical partitioning of solar radiation between the ML and subsurface layers, Hp can have modulating effects. Model simulations were used to illustrate the combined effects of Hm and Hp on the partitioning of solar radiation in the ML and subsurface layers. The effects of the interannual Hp variability on these terms were seen to be term dependent and geographically sensitive. Note that the signs and variabilities of the Qpen and Qabs fields were out of phase as they were defined above. The analyses below will show results only for the Qpen and Rsr fields, but not for Qabs because changes in Qabs are the same as those for Qpen (but with the opposite sign).
Figure 8 shows the interannual anomalies of Qpen and the effects induced by those of Hp from the Chlinter run. Large interannual variability of Qpen was found in the equatorial regions in association with ENSO events. As noted above, the interannual variability of Hm is a dominant contributor to that of Qpen. However, the amplitude and structure of the interannual variations in Hm, Hp, and Qpen clearly indicated that the Qpen field can also be significantly altered by Hp in the western-central regions (Figs. 8b,c) where interannual variations in Hp (Fig. 7b) tended to have the amplitude that was comparable with that of Hm (Fig. 7a). In the western-central regions, for example, Qpen had high values during El Niño events but low values during La Niña events.
Time–longitude sections along the equator for interannual anomalies simulated from the Chlinter run: (a) Qpen estimated with the interannual Hp effect explicitly included 
Citation: Journal of Climate 31, 5; 10.1175/JCLI-D-17-0505.1
Figures 9 and 10 illustrate the horizontal structure of several related interannual anomaly fields for La Niña and El Niño conditions. Coherent relationships existed between these interannual fields during ENSO evolution. In the western-central basin, for example, the magnitudes of the interannual variabilities of Hm and Hp were comparable. This indicates that their anomalies can equally contribute to the interannual variability of Qpen, with their effects on Qpen being comparable. During ENSO cycles, the interannual variations in Hp and Hm tended to be out of phase in the western-central region, and their effects on Qpen were in phase. For example, La Niña events (Fig. 9) are characterized by a cold SST anomaly in the central and eastern equatorial Pacific, and a positive Hm anomaly and a negative Hp anomaly in the western-central basin (Fig. 9b). The deep ML led to less solar radiation reaching the bottom of the ML (a negative Qpen anomaly; Fig. 9d), and the negative Hp anomaly also allowed less solar radiation penetration to the bottom of the ML (a negative Qpen anomaly). Therefore, the negative Hp anomaly tended to have the effect on Qpen that is in the same direction as that of the positive Hm anomaly (Fig. 9e); the combined effects of the positive Hm anomaly and negative Hp anomaly on Qpen produced a more negative Qpen anomaly during La Niña events (Fig. 9f).
Horizontal patterns of interannual anomaly fields for a La Niña condition in January 1989: (a) SST, (b) Hm, (c) Hp, (d) Qpen estimated with the interannual Hp effect included, (e) Qpen estimated without the interannual Hp effect, and (f) the differences between (d) and (e). The contour interval is 0.5°C in (a), 5 m in (b), 1 m in (c), 2 W m−2 in (d) and (e), and 1 W m−2 in (f).
Citation: Journal of Climate 31, 5; 10.1175/JCLI-D-17-0505.1
As in Fig. 9, but for an El Niño condition in January 1987.
Citation: Journal of Climate 31, 5; 10.1175/JCLI-D-17-0505.1
El Niño event (Fig. 10) are characterized by a warm SST anomaly in the central and eastern equatorial Pacific, and a negative Hm anomaly and a positive Hp anomaly in the western-central basin (Fig. 10b). The negative Hm anomaly allowed solar radiation more to penetrate to the bottom of the ML (a positive Qpen anomaly; Fig. 10d), and the positive Hp anomaly also allowed solar radiation more to penetrate to the bottom of the ML (Fig. 10f). Thus, the combined effects of the negative Hm anomaly and positive Hp anomaly on Qpen caused the more positive Qpen anomaly, as compared with the case in which the positive Hp anomaly is not considered (Fig. 10e). That is, the effects of interannual Hp anomalies lead to the Qpen field that is more negative during La Niña and more positive during El Niño, respectively. So, the contribution of interannual Hp anomalies led to an enhanced interannual Qpen variability in the western-central equatorial Pacific.
Next, let us move to the effects on Rsr. Although large interannual Hp variability in the western-central regions was observed to have a notable modulating effect on Qpen and Qabs, this was not the case for Rsr. Figure 11 exhibits the time–longitude sections along the equator for interannual anomalies of Rsr and the effects induced by those of Hp, which are simulated from the Chlinter run. As was previously expressed, Rsr is proportional to Qabs (which decreases exponentially with Hm but increases exponentially with Hp) and inversely proportional to Hm. Therefore, the net effect of Hp on Rsr was additionally modulated by Hm, which has twofold effects on Rsr. During an El Niño event, for example, Hp exhibited a positive anomaly in the western-central regions, which led to a decrease in Qabs; at the same time, the ML tended to be shallow (a negative Hm anomaly). Because the decreased Qabs field (due to the effect of the positive Hp anomaly) acted on the shallow ML, the change to Rsr [= Qabs/ (ρ0CpHm)] induced by the reduced Qabs was compensated by the shoaling effect of the ML. Therefore, Rsr was not as significantly modulated by the positive Hp anomaly as Qabs and Qpen. Similarly, during a La Niña event, the negative Hp anomaly led to an increase in Qabs in the western-central regions; at the same time, the ML deepened (a positive Hm anomaly). The change to Rsr induced by the increased Qabs was compensated by the deepening effect of the ML. The value of Rsr was not seen to be affected by Hp as significantly as Qabs and Qpen were. Figure 11c clearly indicates that Rsr was not as strongly affected by a change in Hp, whereas Qpen and Qabs were significantly affected (Fig. 8c). Quantitatively, the amplitude of the interannual variability of Hp can be more than 30% of that of Hm in the western-central basin, and the amplitude of the interannual variability for Qpen can be modulated by more than 60% in the region, whereas that for Rsr is less than 5%. Therefore, interannual variations in Hp are demonstrated to have significant modulating effects on Qpen in the western-central equatorial Pacific.
Time–longitude sections along the equator for interannual anomalies simulated from the Chlinter run: (a) Rsr estimated with the interannual Hp effect explicitly included 
Citation: Journal of Climate 31, 5; 10.1175/JCLI-D-17-0505.1
The situation in the eastern equatorial basin is very different. In this region, Hm was a predominant factor that determined the structure and variability of Qpen, and the modulating effect of Hp on these heating terms was not significant in the eastern equatorial Pacific compared with that in the western-central region. This can be attributed to a few factors. First, the range of the interannual variability of Hp in the east was approximately one order of magnitude smaller than that of Hm; so, the effects of interannual variations of Hp on these heating terms are negligible, compared with those of Hm. Second, the interannual variations in Hp and Hm were in phase in the eastern basin (east of 150°W), and their effects on Qpen were out of phase. So, the effects of the interannual Hp anomalies only led to a small reduction of Qpen (a small offset).
These results have implications for revealing the underlying processes that can be responsible for the bioeffects that are strongly model dependent. When Rsr is not dominantly modulated by Hp, the direct heating effect on the SST may not be a major process in association with the bioeffects. When Qpen is significantly affected by Hp in the western-central equatorial Pacific, the indirect heating effect on SST can be realized via the modulation of Qpen, which can induce a differential heating vertically in the upper ocean that further affects the stratification and vertical mixing. Therefore, there are complicated interplays among these processes that lead to a modulation of ENSO.
e. A negative feedback involved in the OBH
As demonstrated above, interannual variations in Hm and Hp can both play a role in modulating the heating terms in the western-central equatorial Pacific. For example, a negative Hm anomaly was observed in the western-central equatorial Pacific during an El Niño event, accompanied by a positive Hp anomaly. The positive Hp anomaly induced direct warming at subsurface depths (a positive Qpen anomaly) and cooling in the ML (a negative Qabs anomaly), which reduced the vertical contrast of the thermal field and thus destabilized the stratification and enhanced the mixing in the upper ocean, which decreased the SST. The positive Hp anomaly-induced effects during the El Niño event thus weakened the warm SST anomalies. Similar processes operated during La Niña events but with the opposite sense. Therefore, the oceanic processes induced by anomalous Hp perturbations acted in such a way that the interannual SST variability was weakened. The OBH effects thus act as a negative feedback on ENSO.
These analyses illustrated an ocean biological factor that modulates ENSO in the tropical Pacific. The biofeedback is realized by modulating several heating terms, which provides a biological source for heating or cooling in the upper ocean. Although the heating terms are predominately determined by Hm, they can also be modulated by Hp significantly in the western-central equatorial Pacific. Therefore, the heating terms are affected collectively by interannual anomalies of Hm and Hp. Relationships between changes in physical and biological fields indicate a negative feedback on ENSO. If these bioheating effects are not included in the model, then the related negative feedback is not operating, the effects of which on simulations will be examined below.
5. A simulation with a prescribed climatological chlorophyll field
A climatological run (Chlclim) was performed in which Chl was taken as seasonally varying. When the effects of interannual Chl variations were excluded, there was no interactive feedback from the ocean biology to physics. Note that this approximation has been commonly taken in many climate models in which seasonally varying chlorophyll fields are prescribed in long-term climate modeling. Results in the Chlclim and Chlinter runs were compared with each other. Because these are two ocean-only simulations that were forced by the same prescribed wind forcing with all other model settings being exactly the same, the differences can be solely attributed to the effects of interannual Chl variations, and the related effects can thus be isolated in a clear way.
Figure 12a shows the time–longitude sections along the equator for interannual SST anomalies simulated from the Chlclim run. The space–time evolution is similar to that in the Chlinter run but the amplitude is increased both during El Niño and La Niña events. The differences in SSTs from the two simulations are displayed in Fig. 12b. Evidently, excluding the OBH effects in the Chlclim run caused the simulated interannual SST variability to become stronger. The quantified effects are presented in Table 2. For example, the standard deviations of the Niño-3 and Niño-4 SST anomalies were 0.9° and 0.61°C in the Chlclim run but 0.86° and 0.59°C in the Chlinter run; these values represented increases of approximately 4.4% and 3.3% for the Niño-3 and Niño-4 SST variabilities, respectively. Therefore, the OBH effects led to an increase in the ENSO amplitude in the Chlclim run compared with the Chlinter run.
Time–longitude sections along the equator for (a) interannual SST anomalies simulated from the Chlclim run and (b) the differences in SST between the Chlinter and Chlclim runs. The contour interval is 0.5°C in (a) and 0.1°C in (b).
Citation: Journal of Climate 31, 5; 10.1175/JCLI-D-17-0505.1
The standard deviation of SST anomaly fields simulated in the two experiments (Chlinter and Chlclim) from 1981 to 1992.
Comparisons between these two runs indicated coherent differences in SST simulations during ENSO cycles (Fig. 12b). A cooling difference was observed during El Niño events, but a strong warming difference was observed during La Niña events. Moreover, the cooling difference during El Niño events was accompanied by a negative Hp anomaly, and the warming difference during La Niña events was accompanied by a positive Hp anomaly. Therefore, the SST differences between the Chlclim and Chlinter runs had a coherent relationship with interannual Hp anomalies. Additionally, interannual variations in Hp had an asymmetric effect on SST: stronger during La Niña events but relatively weaker during El Niño events.
The relationships between the differences in SST and the heating terms were further analyzed to aid in understanding the effects. When Hp was prescribed as seasonally varying in the Chlclim run, the interannual variations in the heating terms were solely determined by the variations in Hm. Interannual Qpen variability tended to be weaker in the Chlclim run when the effects of interannual Chl anomalies were not included. Figure 13 illustrates interannual Qpen variability simulated in the Chlclim run and the Qpen differences between the Chlinter and Chlclim runs. The interannual Qpen variability simulated in the Chlclim run was considerably weaker in the western-central regions compared with the Chlinter run (Fig. 8a). For example, the positive (negative) Qpen anomalies during El Niño (La Niña) events in the Chlclim run were both reduced compared with the Chlinter run. The increased SST variability in the Chlclim run (stronger warming SST differences during El Niño and much stronger cooling SST differences during La Niña relative to in the Chlinter run) can thus be explained by excluding the effects of the positive (negative) Hp anomalies during El Niño (La Niña) events.
Time–longitude sections along the equator for (a) interannual Qpen anomalies simulated from the Chlclim run and (b) the differences in Qpen between the Chlinter and Chlclim runs. The contour interval is 2 W m−2 in (a) and 1 W m−2 in (b).
Citation: Journal of Climate 31, 5; 10.1175/JCLI-D-17-0505.1
Furthermore, Fig. 14 presents the time–longitude sections for interannual Hm anomalies simulated from the Chlclim run and the Hm differences between the two runs. A modulating effect on Hm is also seen in the two simulations although they are forced by exactly the same prescribed wind, which is a major dynamical factor affecting Hm. In general, regions with large differences in Hm between the two runs (Fig. 14b) are in the western-central equatorial Pacific, with the positive (negative) Hm differences corresponding to positive (negative) interannual Hp anomalies (Fig. 7b). During El Niño (La Niña), for example, positive (negative) Hp anomalies in the western-central equatorial Pacific tend to increase (decrease) Hm as indicated by the Hm differences (Fig. 14b), a biological factor that can modulate Hm. As a result, a positive (negative) Hp anomaly in the western-central equatorial Pacific (Fig. 7b) can modulate Hm in such a way that interannual variability of Hm becomes stronger during El Niño–La Niña cycles. The physical mechanism for the Hm differences (modulations) in the two runs is the same as that for the Chl-induced effects as discussed above. For example, a positive Hp anomaly during El Niño acts to destabilize the stratification and enhance the vertical mixing intensity in the upper ocean, leading to an increase in Hm. Since Qpen can be caused by both Hm and Hp, the Qpen differences (i.e., modulations; Fig. 13b) can also be induced by the modulated effects on Hm (i.e., the Hm differences between two experiments in Fig. 14b). But, the amplitude of the Hm differences between two experiments (Fig. 14b) is relatively smaller compared with that of interannual Hp variability (Fig. 7b). Thus, the Qpen differences in the two simulations are mainly attributed to the contributions of interannual Hp variability. Note that spatial locations with large Chl-induced modulating effects (as represented by the differences in Qpen, Hm, and SST in the two runs) are not necessarily the same as those with large interannual variabilities (e.g., Qpen, Hm, and SST). For example, large variabilities of the heating terms, Hm, SST, and so on are located over the eastern equatorial Pacific, but the Chl-induced modulating effects on these fields are demonstrated to be significant and robust over the western-central equatorial Pacific. One of major findings in this paper is to perform the two model simulations to clearly illustrate these well-defined structure and relationship among the related physical and biological fields.
Time–longitude sections along the equator for (a) interannual Hm anomalies simulated from the Chlclim run and (b) the differences in Hm between the Chlinter and Chlclim runs. The contour interval is 4 m in (a) and 1 m in (b).
Citation: Journal of Climate 31, 5; 10.1175/JCLI-D-17-0505.1
To further study the possible biological impacts on ENSO, the relationships between some related fields (e.g., the differences in Qpen, Rsr, and SST between the two experiments) were examined by linear regression analyses (Fig. 15). ENSO events (i.e., SST anomalies) induce a systematic perturbation to Chl (or Hp), which acts to modulate the Qpen field (Fig. 15a), which can have a feedback on SST anomalies (Fig. 15b). For example, a good relationship was found between the differences in Qpen and SSTAs in the Niño-3 region (Fig. 15c); this indicates that interannual Chl anomalies induced more penetration of solar radiation to the bottom of the ML during El Niño events but less during La Niña events. This relationship was more obvious in the Niño-4 region (the corresponding figure is not given). Also, the SST differences between the two runs exhibited an asymmetric bioeffect: they were stronger during La Niña events but relatively weaker during El Niño events (Fig. 15b). The linear relationship between the differences in SST and Qpen demonstrated that interannual Chl anomalies (Fig. 15c) tended to weaken the ENSO amplitude. For example, the Qpen difference was positive during La Niña events; interannual variations in ocean biology led to less solar radiation penetration to the bottom of the ML but more absorption within the ML, which acted to stabilize the stratification, reduce vertical mixing and thus warm the SST. Ocean biology–induced heating thus served as a negative feedback on the ENSO. These relationships indicate that the SST differences between these two runs can be attributed dominantly to the Qpen differences (Fig. 15c), but not to Rsr (Fig. 15d).
Scatterplots of various fields to illustrate their relationships simulated in the model experiments in the Niño-3 region during 1981 to 1992: (a) the SSTAs in Chlclim vs the Qpen difference, (b) the SST difference vs the SSTAs in Chlclim, (c) the SST difference vs the Qpen difference, and (d) the SST difference vs the Rsr difference. The differences are calculated between the Chlinter and Chlclim runs (Chlinter minus Chlclim).
Citation: Journal of Climate 31, 5; 10.1175/JCLI-D-17-0505.1
Therefore, the relationships between the SST differences and the modulating effects of interannual Chl variability on the heating terms indicate a damping effect on interannual variability associated with ENSO. The SST differences in the two runs were consistent with previous modeling studies by Zhang (2015a,b), who used an empirical model for interannual Hp–SST relationships to examine the bioeffects on ENSO. Additionally, the SST differences in the two simulations with and without the bioeffects were consistent with the effects of interannual Hp variability inferred from the diagnostic analyses in section 3.
6. Conclusions and discussion
Recent modeling studies have identified clear climate feedback associated with ocean biology–induced heating (OBH) in the tropical Pacific. For example, biological components (such as phytoplankton and detritus) can absorb solar radiation within the mixed layer and further affect the redistribution of heat in the upper ocean. Large uncertainties exist in parameterizing this relationship, and various approximations are thus often made in representing the biofeedback. Previous modeling studies have indicated that there are differences not only in the simulated bioeffects in different models but also in the underlying processes involved (Sweeney et al. 2005; Lin et al. 2007). However, model simulations are sensitively dependent on the way this bioeffect is represented on physics. Currently, biofeedback effects on ENSO are strongly model dependent, and the underlying processes are not well understood. Moreover, large biases still exist in climate model simulations in which bioeffects are not adequately considered; the relationships between climate simulation biases and the way bioeffects are represented are not well known.
What is new with improvements in representing bioeffects in this study is using a comprehensive ocean biology processes-based model to examine the role of Chl in the climate variability associated with ENSO, whereas in our previous studies (Zhang 2015b) interannual anomalies of Chl concentration and Hp are estimated based on their statistical relationships with SST that are derived from historical data. When ocean biological fields (including Chl concentration) are estimated using an ocean ecosystem model, the relationship shown in section 2b is then used to determine Hp, which acts to have modulating effects on the solar radiation penetration. The use of an ocean biological processes–based model allowed us to consider the fully coupled interactions between ocean physics and biology. These biophysics linkages between the Chl concentration, Hp, and their heating effects on the thermodynamics in the upper ocean can be more adequately represented in this ocean-only modeling study.
In this paper, Chl was considered as a major component that affects the vertical penetration of solar radiation in the upper ocean and to represent the bioheating feedback and to study the underlying processes; then, the corresponding Hp field was estimated to quantify the heating effects. Two ocean-only experiments were performed to demonstrate the biofeedback on ENSO. In an interannual Chl run (Chlinter), interannually varying OBH effects were included by using an advanced ocean biology model. In a climatological Chl run (Chlclim run), Chl was prescribed as only seasonally varying, which was calculated from the corresponding Chlinter run. Both experiments were forced by the same prescribed surface winds. Model simulations were analyzed to illustrate the structure of related fields and their relationships, which were then used to explain the effects of ocean biology–induced heating. When the feedback effects were interactively represented in the model, Chl can modulate the shortwave radiation penetration in the upper ocean; three heating terms that are directly affected by interannual Chl (or Hp) anomalies were analyzed in detail. In the Chlinter run, pronounced variations of Hp were observed in the western-central equatorial Pacific, with amplitudes being comparable to those of Hm, whereas in the east, the amplitude of interannual Hp variability was small relative to that of Hm.
Although these two runs were forced by the same prescribed wind field, systematic differences in SST were observed in the two runs. The interannually represented chlorophyll anomalies tended to weaken the ENSO variability in the Chlinter run compared with the Chlclim run. Quantitatively, interannual variations in chlorophyll led to a decrease in the ENSO amplitude by 4.2% compared with the Chlclim run. The bioeffects were characterized by a negative feedback on the climate system in the tropical Pacific.
To understand the effect and processes involved in the SST differences, the Hp field was used to quantify the effect on these related heating terms induced by Chl. A relationship was found between Hp anomalies and SST differences in the two runs. For example, a positive Hp anomaly during an El Niño event induced a cooling effect in the western-central equatorial Pacific, whereas a negative Hp anomaly during a La Niña event induced a strong warming effect. Therefore, the SST differences in the two runs can be explained by the Chl-induced effects on the heating terms. The resultant SST differences in the two runs were consistent with our previous modeling studies; including the OBH feedback led to a negative feedback on ENSO and thus weaker interannual variability.
In this work, only the Chl component was considered. However, other biological components can also directly affect the vertical penetration of incoming solar radiation. For example, detritus can modulate the Hp field and thus induce an ocean biology-related feedback. In this study, Hp was estimated from Chl only, and the effects of other components were not considered (e.g., detritus). Currently, the model’s ability to simulate detritus remains limited, and the corresponding data are very sparse. The impacts of detritus on the simulations need to be investigated in the future. Additionally, CDOM has been shown to have an important influence on the marine ecosystems (Xiu and Chai 2014), and thus its feedback effects on the physical conditions also need to be examined.
In this work, we performed ocean-only experiments in which the air–sea coupled feedback is ignored. Thus, there is an apparent deficiency to investigate the role of chlorophyll variability on the climate variability as this study can either exaggerate the impact of chlorophyll by ignoring role of dynamical factors on Hm, or underestimate the impact of chlorophyll by ignoring positive air–sea coupled feedback (i.e., Bjerknes feedback). For example, in the context of coupled ocean–atmosphere systems, the induced SST changes can affect the atmosphere and lead to coupled interactions between the ocean and the atmosphere (e.g., Zhu et al. 2016), which means that bioeffects can be amplified by air–sea interactions, and thus larger effects can be expected when considering the Chl-induced climate feedback. Also, the ocean biology–induced heating feedbacks can interact with other forcing and feedback processes in the tropical Pacific, including the freshwater flux forcing (Zhang and Busalacchi 2009), tropical instability waves (Zhang and Busalacchi 2008), and stochastic wind forcing (Zhang et al. 2008); these complicated interplays are not well understood (Zhang et al. 2015). Further modeling studies along these lines are underway.
Acknowledgments
The authors thank Drs. Tony Busalacchi and Jieshun Zhu for their comments. The authors also wish to thank three anonymous reviewers for their insightful comments and constructive suggestions. This research was supported by the National Natural Science Foundation of China (Grants 41690122, 41690120, 41475101, 41490644, 41490640, and 41421005), the Chinese Academy of Sciences Strategic Priority Project, the Western Pacific Ocean System (Grants XDA11010105 and XDA11020306), and the AoShan Talents, the Taishan Scholarship and the Qingdao Innovative Program (Grants 2015ASTP and 2014GJJS0101).
REFERENCES
Anderson, W. G., A. Gnanadesikan, R. Hallberg, J. Dunne, and B. L. Samuels, 2007: Impact of ocean color on the maintenance of the Pacific cold tongue. Geophys. Res. Lett., 34, L11609, https://doi.org/10.1029/2007GL030100.
Anderson, W. G., A. Gnanadesikan, and A. Wittenberg, 2009: Regional impacts of ocean color on tropical Pacific variability. Ocean Sci., 5, 313–327, https://doi.org/10.5194/os-5-313-2009.
Ballabrera-Poy, J., R. Murtugudde, R.-H. Zhang, and A. J. Busalacchi, 2007: Coupled ocean–atmosphere response to seasonal modulation of ocean color: Impact on interannual climate simulations in the tropical Pacific. J. Climate, 20, 353–374, https://doi.org/10.1175/JCLI3958.1.
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, https://doi.org/10.1175/1520-0485(1994)024<2156:AHVMSA>2.0.CO;2.
Christian, J. R., M. A. Verschell, R. G. Murtugudde, A. J. Busalacchi, and C. R. McClain, 2001: Biogeochemical modelling of the tropical Pacific Ocean. II: Iron biogeochemistry. Deep-Sea Res. II, 49, 509–543, https://doi.org/10.1016/S0967-0645(01)00111-4.
Gent, P. R., and M. A. Cane, 1989: A reduced gravity, primitive equation model of the upper equatorial ocean. J. Comput. Phys., 81, 444–480, https://doi.org/10.1016/0021-9991(89)90216-7.
Gnanadesikan, A., and W. G. Anderson, 2009: Ocean water clarity and the ocean general circulation in a coupled climate model. J. Phys. Oceanogr., 39, 314–332, https://doi.org/10.1175/2008JPO3935.1.
Jochum, M., S. Yeager, K. Lindsay, K. Moore, and R. Murtugudde, 2010: Quantification of the feedback between phytoplankton and ENSO in the Community Climate System Model. J. Climate, 23, 2916–2925, https://doi.org/10.1175/2010JCLI3254.1.
Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77, 437–471, https://doi.org/10.1175/1520-0477(1996)077<0437:TNYRP>2.0.CO;2.
Lengaigne, M., C. Menkes, O. Aumont, T. Gorgues, L. Bopp, J.-M. André, and G. Madec, 2007: Influence of the oceanic biology on the tropical Pacific climate in a coupled general circulation model. Climate Dyn., 28, 503–516, https://doi.org/10.1007/s00382-006-0200-2.
Lewis, M. R., M.-E. Carr, G. C. Feldman, W. Esaias, and C. R. McClain, 1990: Influence of penetrating solar radiation on the heat budget of the equatorial Pacific Ocean. Nature, 347, 543–545, https://doi.org/10.1038/347543a0.
Lin, P., H. Liu, and X. Zhang, 2007: Sensitivity of the upper ocean temperature and circulation in the equatorial Pacific to solar radiation penetration due to phytoplankton. Adv. Atmos. Sci., 24, 765–780, https://doi.org/10.1007/s00376-007-0765-7.
Lin, P., H. Liu, and X. Zhang, 2008: Effect of chlorophyll-a spatial distribution on upper ocean temperature in the central and eastern equatorial Pacific. Adv. Atmos. Sci., 25, 585–596, https://doi.org/10.1007/s00376-008-0585-4.
Löptien, U., C. Eden, A. Timmermann, and H. Dietze, 2009: Effects of biologically induced differential heating in an eddy-permitting coupled ocean–ecosystem model. J. Geophys. Res., 114, C06011, https://doi.org/10.1029/2008JC004936.
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, https://doi.org/10.1029/2004GL020778.
Manizza, M., C. Le Quéré, A. J. Watson, and E. T. Buitenhuis, 2008: Ocean biogeochemical response to phytoplankton-light feedback in a global model. J. Geophys. Res., 113, C10010, https://doi.org/10.1029/2007JC004478.
Maritorena, S., O. H. F. d’Andon, A. Mangin, and D. A. Siegel, 2010: Merged satellite ocean color data products using a bio-optical model: Characteristics, benefits and issues. Remote Sens. Environ., 114, 1791–1804, https://doi.org/10.1016/j.rse.2010.04.002.
Marzeion, B., A. Timmermann, R. Murtugudde, and F.-F. Jin, 2005: Biophysical feedbacks in the tropical Pacific. J. Climate, 18, 58–70, https://doi.org/10.1175/JCLI3261.1.
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, https://doi.org/10.1175/1520-0485(1994)024<1652:HRWTUO>2.0.CO;2.
Murtugudde, R., R. Seager, and A. Busalacchi, 1996: Simulation of the tropical oceans with an ocean GCM coupled to an atmospheric mixed-layer model. J. Climate, 9, 1795–1815, https://doi.org/10.1175/1520-0442(1996)009<1795:SOTTOW>2.0.CO;2.
Murtugudde, R., J. Beauchamp, C. R. McClain, M. Lewis, and A. J. Busalacchi, 2002: Effects of penetrative radiation of the upper tropical ocean circulation. J. Climate, 15, 470–486, https://doi.org/10.1175/1520-0442(2002)015<0470:EOPROT>2.0.CO;2.
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, https://doi.org/10.1029/2000GL012494.
Park, J.-Y., J.-S. Kug, J. Park, S.-W. Yeh, and C. J. Jang, 2011: Variability of chlorophyll associated with El Niño–Southern Oscillation and its possible biological feedback in the equatorial Pacific. J. Geophys. Res., 116, C10001, https://doi.org/10.1029/2011JC007056.
Park, J.-Y., J.-S. Kug, H. Seo, and J. Bader, 2014a: Impact of bio-physical feedbacks on the tropical climate in coupled and uncoupled GCMs. Climate Dyn., 43, 1811–1827, https://doi.org/10.1007/s00382-013-2009-0.
Park, J.-Y., J.-S. Kug, and Y.-G. Park, 2014b: An exploratory modeling study on bio-physical processes associated with ENSO. Prog. Oceanogr., 124, 28–41, https://doi.org/10.1016/j.pocean.2014.03.013.
Patara, L., M. Vichi, S. Masina, P. G. Fogli, and E. Manzini, 2012: Global response to solar radiation absorbed by phytoplankton in a coupled climate model. Climate Dyn., 39, 1951–1968, https://doi.org/10.1007/s00382-012-1300-9.
Paulson, C. A., and J. J. Simpson, 1977: Irradiance measurements in the upper ocean. J. Phys. Oceanogr., 7, 952–956, https://doi.org/10.1175/1520-0485(1977)007<0952:IMITUO>2.0.CO;2.
Philander, S. G. H., 1983: El Niño Southern Oscillation phenomena. Nature, 302, 295–301, https://doi.org/10.1038/302295a0.
Reynolds, R. W., N. A. Rayner, T. M. Smith, D. C. Stokes, and W. Wang, 2002: An improved in situ and satellite SST analysis for climate. J. Climate, 15, 1609–1625, https://doi.org/10.1175/1520-0442(2002)015<1609:AIISAS>2.0.CO;2.
Siegel, D. A., J. C. Ohlmann, L. Washburn, R. R. Bidigare, C. T. Nosse, E. Fields, and Y. Zhou, 1995: Solar radiation, phytoplankton pigments and the radiant heating of the equatorial Pacific warm pool. J. Geophys. Res., 100, 4885–4891, https://doi.org/10.1029/94JC03128.
Siegel, D. A., S. Maritorena, N. B. Nelson, D. A. Hansell, and M. Lorenzi-Kayser, 2002: Global distribution and dynamics of colored dissolved and detrital organic materials. J. Geophys. Res., 107, 3228, https://doi.org/10.1029/2001JC000965.
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, https://doi.org/10.1175/1520-0442(2004)017<1097:BHITEP>2.0.CO;2.
Sweeney, C., A. Gnanadesikan, S. M. Griffies, M. J. Harrison, A. J. Rosati, and B. L. Samuels, 2005: Impacts of shortwave penetration depth on large-scale ocean circulation and heat transport. J. Phys. Oceanogr., 35, 1103–1119, https://doi.org/10.1175/JPO2740.1.
Timmermann, A., and F.-F. Jin, 2002: Phytoplankton influences on tropical climate. Geophys. Res. Lett., 29, 2104, https://doi.org/10.1029/2002GL015434.
Wang, X., J. R. Christian, R. Murtugudde, and A. J. Busalacchi, 2006: Spatial and temporal variability in new production in the equatorial Pacific during 1980–2003: Physical and biogeochemical controls. Deep-Sea Res. II, 53, 677–697, https://doi.org/10.1016/j.dsr2.2006.01.023.
Wang, X., R. Le Borgne, R. Murtugudde, A. J. Busalacchi, and M. Behrenfeld, 2008: Spatial and temporal variations in dissolved and particulate organic nitrogen in the equatorial Pacific: Biological regulations and physical influences. Biogeosciences, 5, 1705–1721, https://doi.org/10.5194/bg-5-1705-2008.
Wang, X., M. Behrenfeld, R. Le Borgne, R. Murtugudde, and E. Boss, 2009: Regulation of phytoplankton carbon to chlorophyll ratio by light, nutrients and temperature in the equatorial Pacific Ocean: A basin-scale model. Biogeosciences, 6, 391–404, https://doi.org/10.5194/bg-6-391-2009.
Wang, X., R. Murtugudde, E. Hackert, J. Wang, and J. Beauchamp, 2015: Seasonal to decadal variations of sea surface pCO2 and sea-air CO2 flux in the equatorial oceans over 1984–2013: A basin-scale comparison of the Pacific and Atlantic Oceans. Global Biogeochem. Cycles, 29, 597–609, https://doi.org/10.1002/2014GB005031.
Wetzel, P., E. Maier-Reimer, M. Botzet, J. H. Jungclaus, N. Keenlyside, and M. Latif, 2006: Effects of ocean biology on the penetrative radiation in a coupled climate model. J. Climate, 19, 3973–3987, https://doi.org/10.1175/JCLI3828.1.
Xiu, P., and F. Chai, 2014: Connections between physical, optical and biogeochemical processes in the Pacific Ocean. Prog. Oceanogr., 122, 30–53, https://doi.org/10.1016/j.pocean.2013.11.008.
Zhang, R.-H., 2015a: Structure and effect of ocean biology–induced heating (OBH) in the tropical Pacific, diagnosed from a hybrid coupled model simulation. Climate Dyn., 44, 695–715, https://doi.org/10.1007/s00382-014-2231-4.
Zhang, R.-H., 2015b: An ocean-biology-induced negative feedback on ENSO as derived from a hybrid coupled model of the tropical Pacific. J. Geophys. Res. Oceans, 120, 8052–8076, https://doi.org/10.1002/2015JC011305.
Zhang, R.-H., and S. Levitus, 1997: Interannual variability of the coupled tropical Pacific ocean–atmosphere system associated with the El Niño–Southern Oscillation. J. Climate, 10, 1312–1330, https://doi.org/10.1175/1520-0442(1997)010<1312:IVOTCT>2.0.CO;2.
Zhang, R.-H., and A. J. Busalacchi, 2008: Rectified effects of tropical instability wave (TIW)-induced atmospheric wind feedback in the tropical Pacific. Geophys. Res. Lett., 35, L05608, https://doi.org/10.1029/2007GL033028.
Zhang, R.-H., and A. J. Busalacchi, 2009: Freshwater flux (FWF)-induced oceanic feedback in a hybrid coupled model of the tropical Pacific. J. Climate, 22, 853–879, https://doi.org/10.1175/2008JCLI2543.1.
Zhang, R.-H., A. J. Busalacchi, and D. G. DeWitt, 2008: The roles of atmospheric stochastic forcing (SF) and oceanic entrainment temperature (Te) in decadal modulation of ENSO. J. Climate, 21, 674–704, https://doi.org/10.1175/2007JCLI1665.1.
Zhang, R.-H., A. J. Busalacchi, X. Wang, J. Ballabrera-Poy, R. G. Murtugudde, E. C. Hackert, and D. Chen, 2009: Role of ocean biology-induced climate feedback in the modulation of El Niño–Southern Oscillation. Geophys. Res. Lett., 36, L03608, https://doi.org/10.1029/2008GL036568.
Zhang, R.-H., D. Chen, and G. Wang, 2011: Using satellite ocean color data to derive an empirical model for the penetration depth of solar radiation (Hp) in the tropical Pacific ocean. J. Atmos. Oceanic Technol., 28, 944–965, https://doi.org/10.1175/2011JTECHO797.1.
Zhang, R.-H., C. Gao, X. Kang, H. Zhi, Z. Wang, and L. Feng, 2015: ENSO modulations due to interannual variability of freshwater forcing and ocean biology-induced heating in the tropical Pacific. Sci. Rep., 5, 18506, https://doi.org/10.1038/srep18506.
Zhu, J., A. Kumar, B. Huang, M. A. Balmaseda, Z.-Z. Hu, L. Marx, and J. L. Kinter III, 2016: The role of off-equatorial surface temperature anomalies in the 2014 El Niño prediction. Sci. Rep., 6, 19677, https://doi.org/10.1038/srep19677.