Diversity of Lagged Relationships in Global Means of Surface Temperatures and Radiative Budgets for CMIP6 piControl Simulations

Ko Tsuchida aDepartment of Earth and Planetary Sciences, Kyushu University, Fukuoka, Japan

Search for other papers by Ko Tsuchida in
Current site
Google Scholar
PubMed
Close
https://orcid.org/0000-0001-8605-0233
,
Takashi Mochizuki aDepartment of Earth and Planetary Sciences, Kyushu University, Fukuoka, Japan

Search for other papers by Takashi Mochizuki in
Current site
Google Scholar
PubMed
Close
,
Ryuichi Kawamura aDepartment of Earth and Planetary Sciences, Kyushu University, Fukuoka, Japan

Search for other papers by Ryuichi Kawamura in
Current site
Google Scholar
PubMed
Close
,
Tetsuya Kawano aDepartment of Earth and Planetary Sciences, Kyushu University, Fukuoka, Japan

Search for other papers by Tetsuya Kawano in
Current site
Google Scholar
PubMed
Close
, and
Youichi Kamae bFaculty of Life and Environmental Sciences, University of Tsukuba, Tsukuba, Japan

Search for other papers by Youichi Kamae in
Current site
Google Scholar
PubMed
Close
Open access

Abstract

Radiative feedbacks over interannual time scales can be potentially useful for global warming estimation. However, the diversity of the lead–lag relationships in global mean surface temperature (GMST) and net radiation flux at the top of the atmosphere (GMTOA) create uncertainty during the estimation of radiative feedbacks. In this study, key physical processes controlling lead–lag relationships were elucidated by categorizing preindustrial control simulations of CMIP6 into three groups based on cross correlation values of GMTOA against GMST at lag 0 and lag +1 year. The diversity in the lead–lag relationships was primarily caused by the climatological state difference of the atmosphere over the equatorial Pacific, which modulated the strength of convective activity and sensitivity of low-level clouds. Diminished atmospheric stability caused enhanced convective activity, more efficient energy release, and smaller lags. In addition, enhanced stability in the lower atmosphere rendered the low-level clouds more sensitive to sea surface temperature changes and considerably delayed the radiative response. The climatological state difference of the atmosphere resulted from model-inherent atmospheric conditions. These findings suggest that the diversity of lead–lag relationships of GMST and GMTOA over interannual time scales could represent the characteristics of general atmospheric circulation models and possible solutions of the actual atmosphere, which could also affect long-term feedback features.

© 2023 American Meteorological Society. This published article is licensed under the terms of the default AMS reuse license. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Corresponding author: Ko Tsuchida, tsuchida.kou.070@s.kyushu-u.ac.jp

Abstract

Radiative feedbacks over interannual time scales can be potentially useful for global warming estimation. However, the diversity of the lead–lag relationships in global mean surface temperature (GMST) and net radiation flux at the top of the atmosphere (GMTOA) create uncertainty during the estimation of radiative feedbacks. In this study, key physical processes controlling lead–lag relationships were elucidated by categorizing preindustrial control simulations of CMIP6 into three groups based on cross correlation values of GMTOA against GMST at lag 0 and lag +1 year. The diversity in the lead–lag relationships was primarily caused by the climatological state difference of the atmosphere over the equatorial Pacific, which modulated the strength of convective activity and sensitivity of low-level clouds. Diminished atmospheric stability caused enhanced convective activity, more efficient energy release, and smaller lags. In addition, enhanced stability in the lower atmosphere rendered the low-level clouds more sensitive to sea surface temperature changes and considerably delayed the radiative response. The climatological state difference of the atmosphere resulted from model-inherent atmospheric conditions. These findings suggest that the diversity of lead–lag relationships of GMST and GMTOA over interannual time scales could represent the characteristics of general atmospheric circulation models and possible solutions of the actual atmosphere, which could also affect long-term feedback features.

© 2023 American Meteorological Society. This published article is licensed under the terms of the default AMS reuse license. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Corresponding author: Ko Tsuchida, tsuchida.kou.070@s.kyushu-u.ac.jp

1. Introduction

The rise of global mean surface temperature (GMST; note that abbreviations are also listed in the appendix) is a fundamental and useful indicator in global warming evaluations. However, constructing precise estimates of climate system response and projections of GMST increases remains an elusive challenge. From a global perspective, the strength of climate system responses is depicted as a climate feedback parameter, as defined according to Eq. (1):
N=F+R=F+αT,
where N is the net downward radiation flux change (positive downward) at the top of the atmosphere (TOA) (W m−2; hereafter expressed as GMTOA), F is the external forcing (W m−2) affecting the climate system (positive downward), R is the climate radiative response (W m−2) associated with the change in surface temperature T (K; positive upward), and α is the climate feedback parameter (W m−2 K−1; Gregory et al. 2004; Gregory and Webb 2008). The value of −F/α is the climate sensitivity, which is conventionally estimated via the GMST increase resulting from a doubling of atmospheric CO2 concentrations (IPCC 2023).
Climate sensitivity is subdivided into two categories according to this calculation method: 1) equilibrium climate sensitivity (ECS), reflecting the GMST rise in an equilibrium climate state due to the doubling of atmospheric CO2, and 2) effective climate sensitivity (EffCS), reflecting the transient warming rate at a nonequilibrium climate state. From a cost perspective, the application of EffCS is preferred because ECS estimation is computationally expensive, while EffCS can approximate ECS and be calculated more easily. In addition, EffCS enables the estimation of climate sensitivity from either observational data or general circulation models (GCMs) based on data collected over a specific period (Gregory et al. 2002; Forster and Gregory 2006; Otto et al. 2013). Both parameters were derived with the aid of the climate feedback parameter, as shown by Eqs. (2)(4):
ECS=F2×αeq,αeq=FT,
EffCSeff=F2×αeff,αeff=NFT,
EffCSdiff=F2×α˜,α˜=δNδT,
where F is the radiative forcing change resulting from doubled CO2 concentrations (W m−2); αeq, αeff, and α˜ are the equilibrium, effective, and differential feedback parameters (W m−2 K−1), respectively; EffCSeff is the effective climate sensitivity calculated by the effective feedback parameter; and EffCSdiff is the effective climate sensitivity based on the differential feedback parameter (K; see also Rugenstein and Armour 2021). When estimating radiative feedbacks from a short data period, the differential feedback parameter α˜ is commonly used in calculations (Gregory et al. 2004; Tett et al. 2007; Andrews 2014; Gregory and Andrews 2016; Zhou et al. 2016; Andrews et al. 2018; Gregory et al. 2020). For ECS, only the contribution from external forcing should be considered in the climate feedback estimate, while for EffCS, both the anthropogenic forcing and internal variability that controls radiative feedback strength must be considered, making EffCS more difficult to evaluate and interpret.

Proistosescu et al. (2018) showed that radiative feedback as estimated by regressing fluctuations in GMTOA against GMST are mainly associated with forcing sources that are characterized by internal variability. Therefore, they suggested that radiative feedback based on short-term observations that primarily represent interannual fluctuations would not be identical to the feedback governing a climate response to greenhouse warming. However, the short-term radiative feedback correlates well with long-term ones among models (Zhou et al. 2015; Colman and Hanson 2017; Lutsko and Takahashi 2018; Gregory et al. 2020). Therefore, it can be beneficial to investigate the variability in radiative feedback strength on an interannual time scale and across models, thereby deepening our understanding of long-term radiative feedback and the corresponding ECS.

The model ensemble spread of internally generated radiative feedback as calculated via regression can be affected not only by radiative processes but also by model-dependent characteristics with interannual variability, such as that of El Niño–Southern Oscillation (ENSO; Chen et al. 2017). One of the forcing elements related to ENSO produces a time lag between GMST and GMTOA (e.g., Spencer and Braswell 2011; Trenberth et al. 2015; Xie et al. 2016; Xie and Kosaka 2017; Proistosescu et al. 2018; Wills et al. 2021; Ceppi and Fueglistaler 2021), and the diversity of lead–lag relationships among models may influence the estimated radiative feedback values.

However, the intermodal spread of time lag—as a potential explanation for the ECS/EffCS diversity observed within Coupled Model Intercomparison Project phase 6 (CMIP6) models—has not yet been clarified. In this study, we elucidate the lead–lag relationship between GMST and GMTOA associated with interannual variability of CMIP6 preindustrial control (piControl) simulations based on multiple influential factors. We classify the model results into groups according to time scales of lag relationships, and we clarify the governing physical atmospheric and oceanic processes. Notably, determining how internal variability may modulate the radiative feedback in time-lag relationships is imperative for more precise global climate projections, as well as furthering our collective understanding of the climate system response.

The remainder of this paper is organized as follows. Section 2 describes our data sources and methodology, including the calculations used to determine the lead–lag relationship and estimated inversion strength. Section 3 describes the properties of the lead–lag relationship between GMST and GMTOA in the annual (section 3a) and 12-month running (section 3b) mean data, as well as the properties of regression analysis against GMTOA (positive upward), together with group classifications; section 3c explores the comparison of GMST and GMTOA relationships in the Atmospheric Model Intercomparison Project (AMIP) historical simulations and piControl simulations via correlations with low-level cloud changes. Section 4 provides context for the results and provides an overall summary of the study findings.

2. Materials and methods

a. CMIP6 piControl, abrupt-4xCO2, and historical AMIP simulation outputs

A total of 36, 36, and 29 model output datasets from piControl, abrupt-4xCO2, and historical AMIP simulations, respectively, were analyzed as derived via CMIP6 (Eyring et al. 2016). Each piControl and abrupt-4xCO2 output set was based on an atmosphere–ocean general circulation model. For piControl simulations, external forcing from greenhouse gases, aerosols, ozone, and solar variability was fixed at preindustrial levels throughout the simulation, while, for abrupt-4xCO2 simulations, greenhouse gases were abruptly quadrupled at the beginning of each simulation. Meanwhile, each AMIP simulation output was calculated via an atmospheric general circulation model (AGCM) constrained by realistic sea surface temperatures (SSTs) and sea ice values (Gates 1992; Gates et al. 1999; Taylor et al. 2000). The data spanned 500 years (except CESM2-WACCM: 499 years) in piControl simulations, 130 years in abrupt-4xCO2 simulations (of which the first 20 years are regarded as the spinup period), and the period 1979–2014 in historical AMIP simulations. One member of each model was selected as the representative of historical AMIP simulation analysis. The output was linearly interpolated to 1.25° × 1.25° grid intervals. Annual means and 12-month running mean data from the piControl output were analyzed. Each averaged data point was defined by an anomaly relative to the 30-yr running mean centered on each year. The anomalies were analyzed with respect to the monthly climatology defined by 1981–2010 averages for each historical AMIP output. Due to database limitations, the physical variables used in this study were not made publicly available for all models; therefore, only 17 and 11 model datasets from piControl and historical AMIP outputs were analyzed for cloud water and ice content in the present study.

b. International Satellite Cloud Climatology Project data

Data from the International Satellite Cloud Climatology Project (ISCCP), established as part of the World Climate Research Program, served as observational data for model comparisons. The ISCCP data products primarily consist of visible and infrared radiance images from operational weather satellites, with the imagery having been converted to a common format and normalized to a standard reference for subsequent calibration (Schiffer and Rossow 1985; Rossow and Schiffer 1991). Various cloud forms are present in the atmosphere, and the ISCCP defines nine cloud types according to optical thickness and cloud-top pressure (Kurihana et al. 2022). In addition to the cloud data, we also analyzed the surface skin temperature, downward net radiation, longwave (LW) radiation, and net shortwave (SW) radiation at TOA. The monthly values for each variable were obtained and interpolated to 1.25° × 1.25° grid intervals. The timeframe of the data comprised 1984–2014, and anomalies during this period regarding mean monthly climatological values were calculated.

c. Japanese 55-year Reanalysis data

Monthly data from the Japanese 55-year Reanalysis (JRA-55; Kobayashi et al. 2015) database of the Japan Meteorological Agency were used for comparing the estimated inversion strength to discuss the atmospheric stability (section 2e) of historical events with the historical AMIP simulations. The horizontal grid intervals were 1.25° × 1.25°. The data period used in this analysis was from 1981 to 2011, and anomalies regarding mean monthly climatological values were calculated for 1981–2010.

d. Lead–lag correlation and regression analyses

To clarify the lead–lag relationship between GMST and GMTOA, the lead–lag correlation/regression of each physical variable against GMST or GMTOA (positive downward and upward for correlation and regression analyses, respectively) was performed. Lag correlation analysis against GMST enabled a demonstration of the range of responses of GMTOA to GMST across model groups.

e. Estimated inversion strength

To estimate the static stability in the lower atmosphere layers, together with low cloud change, the estimated inversion strength (EIS; Wood and Bretherton 2006) was calculated as follows:
EIS=LTSΓm850(z700LCL),
where LTS, Γm850, LCL, and z700, respectively, represent the lower tropospheric stability (Slingo 1980; Klein and Hartmann 1993; Klein 1997; Lawrence 2005; Wood and Hartmann 2006), moist adiabatic potential temperature gradient at 850 hPa, lifting condensation level, and geopotential height at 700 hPa. Moreover, LTS, Γm850, and LCL values were defined as follows:
LTS=θ700θ0,
Γm(T,p)=gcp[11+Lυqs(T,p)RaT1+Lυ2qs(T,p)cpRυT2],
LCL=z+(20+T273.15K5K)(100m)(1RHl),
where θ is the potential temperature, g is the gravitational acceleration, cp is the specific heat of air at a constant pressure, Lυ is the latent heat of vaporization, qs is the saturation mixing ratio dependent on pressure p and temperature T, z is the parcel’s height, K and m denote units of kelvins and meters, respectively, RHl is the parcel’s relative humidity with respect to liquid (ranging from 0 to 1), and Ra and Rυ are the gas constants for dry and water vapor, respectively.

3. Results

a. Annual mean data

1) Lead–lag relationship and classification

Figure 1a shows a scatterplot of the lagged correlation values of annual mean GMTOA (positive downward) against annual mean GMST at lag 0 (x axis) and lag +1 (y axis). A substantial difference is evident in the lag relationships between the CMIP6 models. According to these differences, all models can be roughly classified into three groups based on the relative magnitude of the lag correlations in lags 0 and +1: group A, 0.75 ≤ |lag +1|/|lag 0| ≤ 1.33, with correlation peaks at both lag 0 and lag +1 (depicted by red letters); group B, |lag +1|/|lag 0| > 1.33, with a correlation peak at lag +1 (blue letters); and group C, |lag +1|/|lag 0| < 0.75, with a correlation peak at lag 0 (gray letters). Groups A, B, and C consisted of 14, 15, and 7 models, respectively.

Fig. 1.
Fig. 1.

Relationship between global mean surface temperature (GMST; K) and positive downward net radiation flux at the top of the atmosphere (GMTOA; W m−2) on interannual variability in Coupled Model Intercomparison Project phase 6 (CMIP6) preindustrial control simulations. (a) Scatterplot of positive downward GMTOA cross-correlated to GMST at lag 0 and +1, as denoted by the x and y axes, respectively. Each model is classified into three groups: groups A (|lag 0| = |lag +1|), B (|lag 0| < |lag +1|), and C (|lag 0| > |lag +1|), colored red, blue, and gray, respectively. Black and gray lines denote |lag 0|:|lag +1| = 0.75:1, 1:1, and 1:1.33, respectively. (b) Lead–lag correlation of positive downward GMTOA against GMST, with the same colors as in (a). Positive lags indicate that GMSTs lead. Autocorrelation of GMST is also shown. Shadings represent the 95% confidence interval on an assumption of the Student’s t distribution in each group.

Citation: Journal of Climate 36, 24; 10.1175/JCLI-D-23-0045.1

Figure 1b illustrates the lead–lag correlation of GMTOA with GMST. The correlation values (±95% confidence interval on the assumption of a Student’s t distribution) of groups A and C at lag 0 were −0.46 (±0.07) and −0.51 (±0.07), respectively, which were significantly larger than that of group B, at −0.21 (±0.07). Furthermore, the correlation values (±95% confidence interval on the assumption of a Student’s t distribution) of groups A and B at lag +1 were −0.43 (±0.07) and −0.46 (±0.07), respectively, which were significantly larger than that of group C, for which the corresponding value was −0.18 (±0.12). Therefore, the multimodel averages of the three groups corresponded to the differences in their lag relationships, as recognized via grouping.

2) Intermodel relationship of radiative feedbacks

Figure 2 shows the intermodel relationships of radiative feedbacks between interannual (based on piControl simulations) and long-term time scales (based on 21–150 years in abrupt-4xCO2 simulations). The radiative feedbacks on interannual time scales were calculated as the ratio of GMST to GMTOA at the same time point, as based on α˜ in Eq. (4).

Fig. 2.
Fig. 2.

Intermodel relationship of radiative feedbacks between interannual and long-term time scales: scatterplot of radiative feedback values for each model based on preindustrial control (piControl) and abrupt-4xCO2 simulations, as denoted by the x and y axes, respectively. Each plot is colored based on the categorized groups A, B, and C (shown in Fig. 1).

Citation: Journal of Climate 36, 24; 10.1175/JCLI-D-23-0045.1

First, we compared the mean strength of radiative feedbacks based on the categorized groups. The averaged values (±95% confidence interval on the assumption of a Student’s t distribution) of groups A and C for interannual time scales, calculated using the simultaneous regression method of Gregory and Andrews (2016), were −1.40 (±0.28) and −1.57 (±0.46) W m−2 K−1, respectively, indicating stronger feedbacks than in group B, with an average value of −0.73 (±0.15) W m−2 K−1. For the long-term radiative feedbacks based on abrupt-4xCO2 simulations, the averaged feedback values were −0.97 (±0.24), −0.76 (±0.16), and −0.96 (±0.32) W m−2 K−1 in groups A, B, and C, respectively. Even though group B showed a weaker long-term feedback than groups A and C, the difference among the groups was not statistically significant (the Student’s t test). Some high ECS models such as CESM2, CanESM5, and E3SM-1-0 were categorized into group B (Li and Huang 2022), suggesting that group B may have some common features causing high ECS values, such as weak radiative feedbacks.

Second, upon comparing the radiative feedback relationships for all the models, we found that the radiative feedback values estimated from the piControl and abrupt-4xCO2 simulations were significantly correlated among the models (r = 0.56), similar to findings described for CMIP5 (Zhou et al. 2015; Colman and Hanson 2017; Gregory et al. 2020). However, when considering the categorized groups based on the lead–lag relationships in piControl simulations, we found smaller correlation values as the lag increased [group C (r = 0.89) > group A (r = 0.50) > group B (r = 0.26)]. Group B, in particular, did not display a statistically significant correlation. Therefore, a further understanding of lead–lag relationships between GMST and GMTOA on interannual time scales can contribute to clarifying the intermodal relationships/differences of long-term radiative feedbacks and thus ECS.

3) Fundamental properties of ENSO

The lead–lag relationship between GMST and GMTOA is presumably attributed to ENSO (Proistosescu et al. 2018), which is a dominant contributor to interannual GMST changes (Pan and Oort 1983; Wigley 2000; Trenberth et al. 2002; Kumar and Hoerling 2003; Chiang and Lintner 2005). We demonstrated some of the fundamental features in ENSO based on the categorized group.

The standard deviations of averaged SST values over the Niño-3 (5°S–5°N, 150°–90°W) and Niño-4 areas (5°S–5°N, 160°E–150°W) were investigated to assess the ENSO amplitude and their differences among the categorized groups. The standard deviations (±95% confidence interval on the assumption of a Student’s t distribution) of Niño-3 SST values were 0.67 (±0.10), 0.65 (±0.11), and 0.65 (±0.13) for groups A, B, and C, respectively, with no statistically significant differences among them. For Niño-4 SST values, the standard deviations were 0.63 (±0.10), 0.58 (±0.10), and 0.50 (±0.08) for groups A, B, and C, respectively. Small differences between groups were attributed to outlying models and were not statistically significant (Fig. S1 in the online supplemental material). Therefore, the ENSO amplitude difference did not affect the lead–lag relationship between GMST and GMTOA.

Figure 3a shows the power spectral density (PSD) function of the averaged SST over the Niño-3.4 area (5°S–5°N, 170°–120°W). Notably, this study focused on interannual time-scale characteristics. Group C displayed a peak in the high-frequency period (approximately 4 years) and smaller PSD value than the other groups in the lower-frequency period (approximately <7 years), whereas groups A and B possessed relatively broad PSD peaks on the interannual time scales (4–10 years). These PSD features were also simulated in the squared coherence between Niño-3.4 SST and GMTOA (Fig. 3b) and between GMST and GMTOA (Fig. 3c). This suggests that characteristics related to the ENSO frequency could contribute to controlling the time scales in which GMST and GMTOA are well correlated. The time lag between decadal GMST and GMTOA variations was larger than that between annual GMST and GMTOA variations (annual; Spencer and Braswell 2011; Xie et al. 2016). The difference of time lag between the decadal and interannual variations in CMIP6 (Fig. S2) suggests that the difference of time lag could be related to the difference in time scales of predominant variability controlling the GMST and GMTOA relationship. Group C had a relatively stronger coherence peak at a higher frequency band and smaller coherence values in a lower-frequency band (approximately < 7 years) compared with group A or B.

Fig. 3.
Fig. 3.

(a) Power spectral density (PSD) function of the averaged sea surface temperatures (SSTs) in Niño-3.4 (5°S–5°N, 170°–120°W) and (b),(c) squared coherence values between GMTOA (W m−2) and Niño-3.4 SST (K) and GMST (K), respectively, based on CMIP6 piControl simulations. Group averages were calculated from the squared coherence values of each model. Shadings represent 95% confidence intervals for PSD/squared coherence values in each group based on the bootstrap test. PSD and squared coherence values were obtained by performing a smoothing processing (50-point running means) for the frequency.

Citation: Journal of Climate 36, 24; 10.1175/JCLI-D-23-0045.1

4) Atmospheric field difference in a climatological state

In terms of ENSO properties, the lead–lag relationship between GMST and GMTOA can correspond to the efficiency of the upward atmospheric transport of the anomalous heat released by the ENSO-related SST anomaly, thereby suggesting contributions of the atmospheric field difference in a climatological state of the models.

Figures 4a and 4b show the climatological differences among the groups in terms of specific humidity and air temperatures of the equatorial Pacific as averaged over 10°S–10°N (see also Figs. S3a,b for the differences from the ensemble mean of all models). In the central and eastern Pacific, group C had larger specific humidity in the middle level (300–700 hPa) and near the sea surface, as well as lower air temperatures in the middle to upper levels (>700 hPa), compared with those of groups A and B (upper and middle panels of Figs. 4a,b). Groups A and B show similar differences, particularly in the middle and low levels, such as the reduced moisture in the middle level (500–700 hPa) and enhanced temperature lapse rate < 600 hPa (lower panels of Figs. 4a,b). These features suggest that the atmospheric field is more unstable in the following order: group C, group A, and group B, leading to enhanced convective activity and a strong upward transport of moisture.

Fig. 4.
Fig. 4.

(a),(b) Group differences in the climatological states of specific humidity (g kg−1) and air temperature (K). (c) Group differences in the vertical flow anomaly as regressed onto the Niño-3.4 SST, based on CMIP6 piControl simulations. Each value is averaged for the 10°S–10°N area. Climatological values were defined across the simulation period (500 years). Crosshatches indicate significantly different values at the 95% confidence level (t test). Gray boxes denote missing values.

Citation: Journal of Climate 36, 24; 10.1175/JCLI-D-23-0045.1

To clarify the relationship between convective activity strength and atmospheric climatological conditions, we investigated the association between vertical flow and SST anomalies. Figure 4c demonstrates the group differences in vertical flow as regressed onto the averaged SST in Niño-3.4 regions (see also Fig. S3c for the differences from the ensemble mean of all models). Compared with groups A and B, group C showed a negative value (stronger ascending flow) over the equatorial central Pacific (approximately 180°–150°W) and a positive value (stronger descending flow) over the equatorial western Pacific (approximately 120°–150°E) (Fig. 4c). Corresponding to the climatological atmospheric field differences shown in Figs. 4a and 4b, group C displayed stronger convective activity than groups A and B in response to SST increases due to the unstable atmosphere in its middle and upper levels (Figs. 4a–c). Similar features were also observed for the differences between groups A and B despite the negative value (stronger ascending flow) slightly shifting to the western region (approximately 180°; Fig. 4c). Historical AMIP simulations also revealed that the convective activity strength was mainly contingent upon the climatological features of atmospheric models, regardless of air–sea interactions (Fig. S4). Across groups, the order of the strength in convective activity over the equatorial Pacific (group C > group A > group B) corresponded to the ascending order of lead–lag relationship between GMST and GMTOA (group C < group A < group B). This suggests that enhanced convective activity leads to a rapid radiative energy release in a global average and thus smaller lags between GMST and GMTOA. The efficient energy release can also affect the dominant ENSO frequency, as shown in section 3a(2); group C has a faster radiative energy release that can potentially result in a high-frequency peak of ENSO (Fig. 3a).

5) Lead–lag regression analysis against positive upward GMTOA

Differences in atmospheric stability affect the strength of convective activity over the equatorial Pacific. In addition, they can also contribute to the spatial variation in GMTOA patterns for this region as well as the off-equatorial area and eastern tropical Pacific. The spatial characteristics that regulate the peak of GMTOA were explored through lag regression analyses. Because we focused on the net radiative response to the GMST rise (and hence the equatorial Pacific SST rise), each regression value was calculated against the positive upward GMTOA, leading to outward radiation and energy release. Figure 5 shows the surface temperature lead–lag regression onto a positive upward GMTOA anomaly at lags −1, 0, and +1, wherein lag −1 represents positive downward GMTOA lag surface temperatures for 1 year and lag +1 represents positive downward GMTOA lead surface temperatures for 1 year. Groups A and C displayed an El Niño–like surface temperature pattern at lags −1 and 0, shifting to a La Niña–like pattern at lag +1. Group B exhibited similar features at lag −1 and lag +1; however, a negative surface temperature anomaly in the equatorial east Pacific was simulated at lag 0.

Fig. 5.
Fig. 5.

Maps of lead–lag regressions for surface temperatures (K) against positive upward GMTOA (W m−2) based on CMIP6 piControl simulations: (a)–(c) Regression values at lag −1, lag 0, and lag +1, respectively. Only statistically significant regression values (95% confidence level; t test) are depicted.

Citation: Journal of Climate 36, 24; 10.1175/JCLI-D-23-0045.1

Similarly, the net total, LW, and net SW radiations regressed onto GMTOA showed significant differences among the groups at lag 0 (Fig. 6). Notably, all radiation values depicted in the figures were positive downward. The net total radiation indicated a negative value in the subtropical Pacific in groups A and C (Fig. 6a), together with negative LW and positive net SW regional radiation levels (Figs. 6b,c), as simulated in previous research (Wills et al. 2021). In group B, the net radiation was largely negative in the eastern equatorial Pacific (Fig. 6a), which was associated with the negative surface temperature anomaly in the same region, with LW and net SW radiations both being negative (Figs. 6b,c).

Fig. 6.
Fig. 6.

Maps of downward (a) net total radiation, (b) longwave (LW) radiation, and (c) net shortwave (SW) radiation at the top of the atmosphere, regressed to positive upward GMTOA values as based on CMIP6 piControl simulations (all units in W m−2). Only statistically significant regression values (95% confidence level; t test) are depicted.

Citation: Journal of Climate 36, 24; 10.1175/JCLI-D-23-0045.1

Because these differences in radiative anomalies were likely partially due to convective activity and its resultant structural changes, the meridional and zonal circulation changes were investigated (Figs. 7 and 8). Figure 7 shows the mass streamfunction (Oort and Yienger 1996) regressed onto the GMTOA at lag 0. Positive and negative values represent clockwise and anticlockwise circulation, respectively. Groups A and C produced strong dipole-like regression patterns of the mass streamfunction across the equator but were more confined to the equator than in climatological peaks (Figs. 7a,c). These mass streamfunction patterns showed stronger active convection, with a narrower convective region due to the warm equatorial pool, causing a decrease in LW radiation (i.e., enhanced outgoing LW radiation) and thus a larger GMTOA. These findings agree with the projected decrease in liquid and ice water paths under a warming global climate (Bui et al. 2019). The meridional extent of local Hadley circulation had decreased, not the intertropical convergence zone (ITCZ), as the equatorial Pacific warmed (see section 4a); however, group B did not display such a dipole structure, possessing a positive core on the equator (Fig. 7b).

Fig. 7.
Fig. 7.

Mass streamfunction (kg s−1) regressed to positive upward GMTOA (W m−2) based on CMIP6 piControl simulations. Only statistically significant regression values (95% confidence level; t test) are depicted. Contours show the climatological mass streamfunction values averaged in (a) group A, (b) group B, and (c) group C.

Citation: Journal of Climate 36, 24; 10.1175/JCLI-D-23-0045.1

Fig. 8.
Fig. 8.

Longitudinal-height cross section maps of vertical wind (shading; Pa s−1), mass fraction of cloud water and ice (contours: 5.0 × 10−4 interval, depicted values: absolute value > 1.0 × 10−3; g kg−1), and zonal circulation winds (vectors; zonal: m s−1, vertical: Pa s−1) along the equator (with values averaged for the 10°S–10°N region) regressed to positive upward GMTOA values, based on CMIP6 piControl simulations. Vertical vector components were multiplied by a factor of 100. Only statistically significant regression values (95% confidence level; t test) are depicted.

Citation: Journal of Climate 36, 24; 10.1175/JCLI-D-23-0045.1

Figure 8 shows a longitudinal–height cross section map of vertical flow, averaged over 10°S–10°N and regressed onto GMTOA at lag 0. The negative (ascending flow) anomaly around 180° in groups A and C corresponded well to the El Niño–like surface temperature patterns and resultant weakened Walker circulation. The positive anomaly of a mass fraction of cloud water and ice was simulated in the lower layer of the eastern equatorial Pacific and middle upper layer of the whole equatorial Pacific in groups A and C (Figs. 8a,c). Meanwhile, a small negative (weak descending flow) anomaly in the eastern equatorial Pacific was encountered in group B, together with an increasing mass fraction of cloud water and ice in the lower layer (Fig. 8b).

Overall, the patterns of net radiation (Fig. 6a) in groups A and C were attributed to the decreased LW radiation in the subtropics due to an equatorward shift of deep convection following its width change, in conjunction with the enhanced convective activity in response to El Niño–like surface temperature patterns. However, the pattern of net radiation change in group B was likely caused by decreasing LW and net SW radiation in the eastern equatorial Pacific, together with negative surface temperatures and increasing low-level clouds in that region.

b. 12-month running mean data

To further understand the distinct low-level cloud increases in group B and changes in deep convective activity, 12-month running mean data were analyzed (as opposed to the annual means discussed in section 3a) to determine the detailed temporal evolution of these phenomena. Figure 9 shows the lead–lag correlations of GMTOA against GMST (similar to Fig. 1b) for the 12-month running mean data. GMTOA displayed negative peaks at +4, +5, and +8 months in groups C, A, and B, respectively, corresponding to nonzero cross-correlation values at lag 0, due to the superposition of radiative and nonradiative forcing (Spencer and Braswell 2011). The orders of lag magnitude (group B > group A > group C) that were simulated using annual data (Fig. 1b) were also displayed with 12-month running mean data.

Fig. 9.
Fig. 9.

Lead–lag correlation of positive downward GMTOA against GMST (similar to Fig. 1b but derived from 12-month running mean data) based on CMIP6 piControl simulations. Circles demarcate the months when GMTOA peaks were negative.

Citation: Journal of Climate 36, 24; 10.1175/JCLI-D-23-0045.1

Figure 10 shows surface temperature, EIS, positive downward net total, LW, and net SW radiations at TOA. Plotted values are the lead–lag regressions against positive upward GMTOA averaged over 10°S–10°N. In all three groups, surface temperature shifted from El Niño–like to La Niña–like patterns (Fig. 10a), consistent with the patterns evident for annual mean data (Fig. 5). The LW radiation anomaly was positive (negative) before (after) lag 0 (Fig. 10c), corresponding to the surface temperature patterns; the inverse was observed in net SW radiation (Fig. 10d). The surface temperature anomaly at lag 0 was nearly neutral in group B, whereas positive surface temperature and weak negative net radiation flux anomalies were recorded in groups A and C (Figs. 10a,b), corresponding to the differences in time lag between GMTOA and GMST. Group B also showed a strong negative net radiation flux anomaly around 120°W (similar to the data in Fig. 6a). Furthermore, the negative net SW radiation anomaly in group B was strongest after lag 0 (near +4 months on the y axis) in the eastern Pacific (near 120°W on the x axis), while no clear peaks were simulated in groups A and C (Fig. 10d). The negative net SW radiation in group B was accompanied by positive EIS (Fig. 10d). Thus, stronger static stability contributed to a low-cloud increase after lag 0 in the equatorial east Pacific through cloud feedbacks (Norris 1998; Norris and Iacobellis 2005). Net SW radiation decreased (i.e., enhanced net outward SW radiation occurred) (Fig. 10d) and induced regional peaks of net radiation (Fig. 10b). Radiation properties prior to lag 0 in group B should be also noted. The net SW radiation showed a positive peak in the eastern Pacific (near 120°W on the x axis) near lag −1 (near −12 months on the y axis) and was associated with a positive surface temperature anomaly in group B (Fig. 10), suggesting that group B exhibited low-level clouds with a higher sensitivity to surface temperature than those in groups A and C.

Fig. 10.
Fig. 10.

Maps of longitude–time cross sections for (a) surface temperature (K), (b) positive downward net radiation (W m−2), (c) LW radiation (W m−2), and (d) net SW radiation (W m−2) at the top of the atmosphere as regressed to positive upward GMTOA (W m−2) spanning 10°S–10°N, based on CMIP6 piControl simulations. Vectors in all panels show surface wind regressed to positive upward GMTOA. The contours in (d) show EIS regressed to positive upward GMTOA at 0.2-K intervals. Light brown and blue-green contour lines correspond to positive and negative anomalies, respectively.

Citation: Journal of Climate 36, 24; 10.1175/JCLI-D-23-0045.1

c. Temporal variation in net radiation at TOA and low clouds in historical AMIP simulations

The observed El Niño– to La Niña–like surface temperature shift in the piControl simulations (group B) may have caused low-level cloud increases in the eastern equatorial Pacific, possibly influencing the global mean energy budget. To clarify the underlying driving factors, specifically how low cloud coverage and net downward radiation changes at TOA were associated with the actual El Niño to La Niña shifts, historical AMIP simulations were investigated. The surface temperatures for groups A, B, and C in historical AMIP simulations were identical because all simulations used the prescribed SST; groups A, B, and C consisted of 9, 13, and 7 models, respectively.

Three specific events were selected for further analysis, namely, 1987–89, 1997–99, and 2009–11, because they exhibited clear shifts from El Niño to La Niña. Figure 11 presents the 1987–89 case. In response to surface temperature changes due to the shift from El Niño to La Niña, low-level cloud cover decreased around July 1987, when the equatorial surface temperature showed a positive anomaly in all three groups. Thereafter, cloud cover increased again as the equatorial surface temperature anomaly declined around July 1988. This finding agrees well with the enhanced low-level cloud cover in the ISCCP observation (Fig. 11e), which caused a negative net SW radiation anomaly (Fig. 11d) and thus a negative net radiation flux anomaly on the equatorial eastern Pacific (Fig. 11b), while the surface temperature presented a negative anomaly (Fig. 11a). Hereafter, we mainly discuss the radiation properties that were related to a negative surface temperature anomaly, but the same mechanism can be inversely applied for a positive surface temperature anomaly. Although a negative net radiation flux was also found in the western and central equatorial Pacific (Fig. 11b), the negative LW and positive net SW radiation anomalies suggested a convective cloud decrease, rather than low-level cloud cover (Figs. 11c,d). Group B showed a much larger increase in low-level cloud cover and a stronger EIS than the other groups, creating a stronger reflection of solar radiation. The 2009–11 case showed similar low-level cloud cover and EIS features that were associated with comparable surface temperature transition patterns (Fig. S5).

Fig. 11.
Fig. 11.

Longitudinal–time cross sections for (a) surface temperature (K), (b) positive downward net radiation (W m−2), (c) LW radiation (W m−2), (d) net SW radiation (W m−2) at the top of the atmosphere, and (e) low cloud coverage (%) anomalies from climatological values averaged for the 10°S–10°N region during 1987–89. Data were obtained from the International Satellite Cloud Climatology Project (ISCCP), Japanese 55-year Reanalysis (JRA-55), and CMIP6 historical Atmospheric Model Intercomparison Project (AMIP) simulations. Vectors and contours in all panels show surface wind anomalies (m s−1) and EIS (0.1-K intervals), respectively. Light brown and blue-green contour lines correspond to positive and negative anomalies, respectively. Gray shading denotes missing values.

Citation: Journal of Climate 36, 24; 10.1175/JCLI-D-23-0045.1

Figure 12 illustrates the 1997–99 case. Because an eastern Pacific (EP) El Niño–like surface temperature pattern was evident (Kao and Yu 2009; Takahashi et al. 2011; Capotondi et al. 2015), the region where the radiation peaks in this case was not identical to that of the other two events. Nevertheless, the overall properties found in the 1987–89 and 2009–11 cases (low-level cloud change and a radiative response related to the equatorial surface temperature change) can be also applied for 1997–99. EP El Niño–like surface temperatures led to positive LW and net SW radiations on the eastern equatorial Pacific, while leading to stronger positive LW than negative net SW radiations on the central equatorial Pacific (Figs. 12c,d). Thus, a positive net radiation flux was present at TOA as long as the positive surface temperature anomaly was maintained in the region. Following the El Niño event, negative surface temperatures resulting from the La Niña event (i.e., non-EP-like patterns) induced an increase in low-level cloud coverage around the central equatorial Pacific for the models and ISCCP. Regardless of the surface temperature pattern differences with the other two cases, group B showed a more positive anomaly of low-level cloud cover associated with negative SW radiation, corresponding to a more negative value of net TOA radiation flux (Figs. 12b,d,e).

Fig. 12.
Fig. 12.

Longitudinal–time cross sections for (a) surface temperature (K), (b) positive downward net radiation (W m−2), (c) LW radiation (W m−2), (d) net SW radiation (W m−2) at the top of the atmosphere, and (e) low cloud coverage (%) anomalies from the climatological value averaged for the 10°S–10°N region during 1997–99. Data were obtained from the ISCCP, JRA-55, and CMIP6 historical AMIP simulations. Vectors and contours in all panels show surface wind anomalies (m s−1) and EIS (0.1-K intervals), respectively. Light brown and blue-green contour lines correspond to positive and negative anomalies, respectively. Gray shading denotes missing values.

Citation: Journal of Climate 36, 24; 10.1175/JCLI-D-23-0045.1

These findings indicate that an El Niño to La Niña event can lead to a change in low-level cloud cover, especially in the eastern equatorial Pacific. An El Niño–like pattern of a surface temperature anomaly leads to a decrease in low-level cloud, thereby enhancing net downward radiation flux. Conversely, when surface temperatures shift to a La Niña–like pattern, low-level cloud cover increases, resulting in a reduction of net downward radiation flux. These changes in low-level cloud cover and associated radiation flux peak in the eastern or central Pacific according to the surface temperature pattern (as well as the ISCCP) and peaks with the largest magnitudes are found in group B. The substantial change in low cloud coverage due to such El Niño to La Niña shifts in group B was also simulated as a common response to the equatorial surface temperature change and the resultant changes in low-level atmospheric stability (Fig. S6). Therefore, substantial changes in the extent of low-level cloud cover over the eastern equatorial Pacific led to a stronger net radiation change, dominating control of the total energy budget in group B.

4. Summary and discussion

a. Summary

The lead–lag relationships between GMST and GMTOA were analyzed via CMIP6 piControl and historical AMIP simulations. The time lag between GMST and GMTOA peaks were scattered among the models, which could be classified into three groups according to the relative magnitude of lag correlations in lag 0 and +1: group A (0.75 ≤ |lag +1|/|lag +0| ≤ 1.33, correlation peaks at lag 0 and +1), group B (|lag +1|/|lag +0| > 1.33, correlation peak at lag +1), and group C (|lag +1|/|lag +0| < 0.75, correlation peak at lag 0; see section 3a). The drivers of the simulated differences in the ratio of correlation values at lag 0 and lag +1 could be attributed to the climatological state differences of atmospheric models across the groups, resulting in the two key factors described below.

One key factor was the efficiency of an upward transport of anomalous heat via convective activity in the equatorial Pacific. A more unstable atmosphere was characterized by a higher specific humidity and lower air temperature in the middle layer, which induced a stronger convection over the equatorial Pacific for a unit increase of surface temperature, thereby accelerating energy release with corresponding smaller lags. The efficient energy release can potentially affect the highly frequent ENSO in group C (section 3a).

A second key factor is the difference in low-level cloud sensitivity. Groups A and C produced high values for net TOA downward radiation in the subtropical Pacific, whereas group B displayed higher values over the eastern equatorial Pacific. GMST peaked in an El Niño–like surface temperature state and shifted to a La Niña–like pattern across all the models and groups in the CMIP6 piControl simulations. Therefore, groups A and C displayed smaller lags than group B as they aligned more closely with an El Niño–like surface temperature pattern, which led to maximum GMTOA through enhanced convective activity and local Hadley circulation. In contrast, an El Niño–like surface temperature pattern led to less low-level cloud cover over the eastern equatorial Pacific in group B than in groups A and C, and the resultant enhanced positive cloud feedback inhibited the upward GMTOA peak at this stage. As the surface temperature pattern shifted to a La Niña–like phase, low-level cloud cover increased substantially over the eastern equatorial Pacific. Consequently, the negative cloud feedback and associated radiation output were enhanced.

Historical AMIP experimental analyses revealed that both the substantial increase in low-level clouds as well as the different strengths of convective activity were primarily due to atmospheric model characteristics—including higher humidity in the lower layers and strengthened tropospheric stability in the equatorial Pacific—rather than being effects of complicated atmospheric–ocean interactions.

b. Discussion

1) Convective activity change according to the equatorial Pacific warming

El Niño–like surface temperature patterns led to a marked increase in LW (and corresponding decrease in net SW) radiations in the equatorial Pacific and an LW decrease (with corresponding net SW increase) in the subtropics (Figs. 6b,c), suggesting that convective clouds decreased in the subtropics; however, changes in the precipitation rate commonly represent a wider and narrower ITCZ during El Niño and La Niña events, respectively (Dias and Pauluis 2011; Wodzicki and Rapp 2016, 2020). Bui et al. (2019) stated that the enhancement of both deep and shallow convection (with the latter occurring as a result of diminished deep convection) increases precipitation, potentially complicating the relationship between radiation and precipitation changes. Notably, in the present study, the active convection region was defined from a radiation perspective, not one of precipitation, potentially leading to a different view of ITCZ changes compared with previous research.

2) Climatological state of atmosphere and its application for ECS

The climatological state inherent to atmospheric models controls the strength of convective activity and sensitivity of low-level clouds, thereby affecting the lead–lag relationship between GMST and GMTOA. Concerning the sensitivity of the low-level clouds, Ceppi and Fueglistaler (2021) demonstrated that the low-level cloud increase during an El Niño decay phase was due to an EIS increase. Compared with the ISCCP observations (section 3c), groups A and C appeared better at reproducing observed radiation properties. However, the enhanced atmospheric warming over the middle to high layers and oceanic evaporation resulting from warming global temperatures could lead to a more stable and moist environment in the lower layers similar to that of group B, suggesting that the lead–lag relationship between GMST and GMTOA may vary in the future. In this respect, group B may provide an indication of the relationship between GMST and GMTOA under future climate change. Low-level cloud sensitivity as well as convective activity strength have great impacts on the cloud feedback components of ECS estimates based on CMIP3, CMIP5, and CMIP6 (Ceppi et al. 2017; Schlund et al. 2020; Zelinka et al. 2020). The feedback strength between abrupt-4xCO2 experiments and piControl simulations based on CMIP5 may have had modest but significant correlation values (Gregory et al. 2020), which was confirmed for CMIP6 in this study (Fig. 2), possibly originating from cloud feedback differences (Zhou et al. 2015; Colman and Hanson 2017; Lutsko and Takahashi 2018). Previous research suggests that air–sea interactions play a significant role in linking SST and radiation budget changes with ENSO variability (Ceppi and Fueglistaler 2021). Meanwhile, our results reveal that the atmospheric conditions in a climatological state that is inherent to a model can also significantly influence the intermodal relationships between those two components. The radiative feedback difference from the lag relationships between GMST and GMTOA implies that a deeper understanding of the atmospheric field characteristics in a climatological state could improve our comprehension and the applicability of radiative feedback mechanisms, in addition to ECS levels derived from climate feedback cycles.

Acknowledgments.

This work was supported by the Japan Society for the Promotion of Science KAKENHI (JP19H05703), Japan Science and Technology Agency SPRING (JPMJSP2136), and the Integrated Research Program for Advancing Climate Models (Tougou) (JPMXD0717935457, JPMXD0717935715). Author contributions are as follows: Ko Tsuchida: conceptualization, data analysis and investigation, and writing (original draft); Takashi Mochizuki: funding acquisition, resources, supervision, writing, review, and editing; Youichi Kamae, Ryuichi Kawamura, and Tetsuya Kawano: discussion and advice.

Data availability statement.

The CMIP6 and ISCCP datasets supporting the conclusions of this study are available at https://esgf-node.llnl.gov/search/cmip6/ and https://isccp.giss.nasa.gov/products/onlineData.html, respectively.

APPENDIX

Abbreviations

AGCM

Atmospheric general circulation model

AMIP

Atmospheric Model Intercomparison Project

ECS

Equilibrium climate sensitivity

EffCS

Effective climate sensitivity

ENSO

El Niño–Southern Oscillation

GCM

General circulation model

GMTOA

Net radiation flux at the top of the atmosphere

GMST

Global mean surface temperature

LW

Longwave radiation

SST

Sea surface temperature

SW

Shortwave radiation

TOA

Top of the atmosphere

REFERENCES

  • Andrews, T., 2014: Using an AGCM to diagnose historical effective radiative forcing and mechanisms of recent decadal climate change. J. Climate, 27, 11931209, https://doi.org/10.1175/JCLI-D-13-00336.1.

    • Search Google Scholar
    • Export Citation
  • Andrews, T., and Coauthors, 2018: Accounting for changing temperature patterns increases historical estimates of climate sensitivity. Geophys. Res. Lett., 45, 84908499, https://doi.org/10.1029/2018GL078887.

    • Search Google Scholar
    • Export Citation
  • Bui, H. X., J.-Y. Yu, H.-W. Liu, C.-Y. Tu, P.-G. Chiu, and H.-H. Hsu, 2019: Convective structure changes over the equatorial Pacific with highly increased precipitation under global warming simulated in the HiRAM. SOLA, 15, 119124, https://doi.org/10.2151/sola.2019-022.

    • Search Google Scholar
    • Export Citation
  • Capotondi, A., and Coauthors, 2015: Understanding ENSO diversity. Bull. Amer. Meteor. Soc., 96, 921938, https://doi.org/10.1175/BAMS-D-13-00117.1.

    • Search Google Scholar
    • Export Citation
  • Ceppi, P., and S. Fueglistaler, 2021: The El Niño–Southern Oscillation pattern effect. Geophys. Res. Lett., 48, e2021GL095261, https://doi.org/10.1029/2021GL095261.

    • Search Google Scholar
    • Export Citation
  • Ceppi, P., F. Brient, M. D. Zelinka, and D. L. Hartmann, 2017: Cloud feedback mechanisms and their representation in global climate models. Wiley Interdiscip. Rev.: Climate Change, 8, e465, https://doi.org/10.1002/wcc.465.

    • Search Google Scholar
    • Export Citation
  • Chen, L., T. Li, Y. Yu, and S. K. Behera, 2017: A possible explanation for the divergent projection of ENSO amplitude change under global warming. Climate Dyn., 49, 37993811, https://doi.org/10.1007/s00382-017-3544-x.

    • Search Google Scholar
    • Export Citation
  • Chiang, J. C. H., and B. R. Lintner, 2005: Mechanisms of remote tropical surface warming during El Niño. J. Climate, 18, 41304149, https://doi.org/10.1175/JCLI3529.1.

    • Search Google Scholar
    • Export Citation
  • Colman, R., and L. Hanson, 2017: On the relative strength of radiative feedbacks under climate variability and change. Climate Dyn., 49, 21152129, https://doi.org/10.1007/s00382-016-3441-8.

    • Search Google Scholar
    • Export Citation
  • Dias, J., and O. Pauluis, 2011: Modulations of the phase speed of convectively coupled Kelvin waves by the ITCZ. J. Atmos. Sci., 68, 14461459, https://doi.org/10.1175/2011JAS3630.1.

    • Search Google Scholar
    • Export Citation
  • Eyring, V., S. Bony, G. A. Meehl, C. A. Senior, B. Stevens, R. J. Stouffer, and K. E. Taylor, 2016: Overview of the Coupled Model Intercomparison Project Phase 6 (CMIP6) experimental design and organization. Geosci. Model Dev., 9, 19371958, https://doi.org/10.5194/gmd-9-1937-2016.

    • Search Google Scholar
    • Export Citation
  • Forster, P. M. F., and J. M. Gregory, 2006: The climate sensitivity and its components diagnosed from Earth radiation budget data. J. Climate, 19, 3952, https://doi.org/10.1175/JCLI3611.1.

    • Search Google Scholar
    • Export Citation
  • Gates, W. L., 1992: AMIP: The Atmospheric Model Intercomparison Project. Bull. Amer. Meteor. Soc., 73, 19621970, https://doi.org/10.1175/1520-0477(1992)073<1962:ATAMIP>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Gates, W. L., and Coauthors, 1999: An overview of the results of the Atmospheric Model Intercomparison Project (AMIP I). Bull. Amer. Meteor. Soc., 80, 2956, https://doi.org/10.1175/1520-0477(1999)080<0029:AOOTRO>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Gregory, J. M., and M. Webb, 2008: Tropospheric adjustment induces a cloud component in CO2 forcing. J. Climate, 21, 5871, https://doi.org/10.1175/2007JCLI1834.1.

    • Search Google Scholar
    • Export Citation
  • Gregory, J. M., and T. Andrews, 2016: Variation in climate sensitivity and feedback parameters during the historical period. Geophys. Res. Lett., 43, 39113920, https://doi.org/10.1002/2016GL068406.

    • Search Google Scholar
    • Export Citation
  • Gregory, J. M., R. J. Stouffer, S. C. B. Raper, P. A. Stott, and N. A. Rayner, 2002: An observationally based estimate of the climate sensitivity. J. Climate, 15, 31173121, https://doi.org/10.1175/1520-0442(2002)015<3117:AOBEOT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Gregory, J. M., and Coauthors, 2004: A new method for diagnosing radiative forcing and climate sensitivity. Geophys. Res. Lett., 31, L03205, https://doi.org/10.1029/2003GL018747.

    • Search Google Scholar
    • Export Citation
  • Gregory, J. M., T. Andrews, P. Ceppi, T. Mauritsen, and M. J. Webb, 2020: How accurately can the climate sensitivity to CO2 be estimated from historical climate change? Climate Dyn., 54, 129157, https://doi.org/10.1007/s00382-019-04991-y.

    • Search Google Scholar
    • Export Citation
  • IPCC, 2023: Climate Change 2022: Impacts, Adaptation, and Vulnerability. Cambridge University Press, 3056 pp., https://doi.org/10.1017/9781009325844.

  • Kao, H.-Y., and J.-Y. Yu, 2009: Contrasting eastern-Pacific and central-Pacific types of ENSO. J. Climate, 22, 615632, https://doi.org/10.1175/2008JCLI2309.1.

    • Search Google Scholar
    • Export Citation
  • Klein, S. A., 1997: Synoptic variability of low-cloud properties and meteorological parameters in the subtropical trade wind boundary layer. J. Climate, 10, 20182039, https://doi.org/10.1175/1520-0442(1997)010<2018:SVOLCP>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Klein, S. A., and D. L. Hartmann, 1993: The seasonal cycle of low stratiform clouds. J. Climate, 6, 15871606, https://doi.org/10.1175/1520-0442(1993)006%3C1587:TSCOLS%3E2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Kobayashi, S., and Coauthors, 2015: The JRA-55 reanalysis: General specifications and basic characteristics. J. Meteor. Soc. Japan, 93, 548, https://doi.org/10.2151/jmsj.2015-001.

    • Search Google Scholar
    • Export Citation
  • Kumar, A., and M. P. Hoerling, 2003: The nature and causes for the delayed atmospheric response to El Niño. J. Climate, 16, 13911403, https://doi.org/10.1175/1520-0442(2003)16%3C1391:TNACFT%3E2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Kurihana, T., E. Moyer, R. Willett, D. Gilton, and I. Foster, 2022: Data-driven cloud clustering via a rotationally invariant autoencoder. IEEE Trans. Geosci. Remote Sens., 60, 125, https://doi.org/10.1109/TGRS.2021.3098008.

    • Search Google Scholar
    • Export Citation
  • Lawrence, M. G., 2005: The relationship between relative humidity and the dewpoint temperature in moist air: A simple conversion and applications. Bull. Amer. Meteor. Soc., 86, 225234, https://doi.org/10.1175/BAMS-86-2-225.

    • Search Google Scholar
    • Export Citation
  • Li, S., and P. Huang, 2022: An exponential-interval sampling method for evaluating equilibrium climate sensitivity via reducing internal variability noise. Geosci. Lett., 9, 34, https://doi.org/10.1186/s40562-022-00244-9.

    • Search Google Scholar
    • Export Citation
  • Lutsko, N. J., and K. Takahashi, 2018: What can the internal variability of CMIP5 models tell us about their climate sensitivity? J. Climate, 31, 50515069, https://doi.org/10.1175/JCLI-D-17-0736.1.

    • Search Google Scholar
    • Export Citation
  • Norris, J. R., 1998: Low cloud type over the ocean from surface observations. Part I: Relationship to surface meteorology and the vertical distribution of temperature and moisture. J. Climate, 11, 369382, https://doi.org/10.1175/1520-0442(1998)011<0369:LCTOTO>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Norris, J. R., and S. F. Iacobellis, 2005: North Pacific cloud feedbacks inferred from synoptic-scale dynamic and thermodynamic relationships. J. Climate, 18, 48624878, https://doi.org/10.1175/JCLI3558.1.

    • Search Google Scholar
    • Export Citation
  • Oort, A. H., and J. J. Yienger, 1996: Observed interannual variability in the Hadley circulation and its connection to ENSO. J. Climate, 9, 27512767, https://doi.org/10.1175/1520-0442(1996)009<2751:OIVITH>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Otto, A., and Coauthors, 2013: Energy budget constraints on climate response. Nat. Geosci., 6, 415416, https://doi.org/10.1038/ngeo1836.

    • Search Google Scholar
    • Export Citation
  • Pan, Y. H., and A. H. Oort, 1983: Global climate variations connected with sea surface temperature anomalies in the eastern equatorial Pacific Ocean for the 1958–73 period. Mon. Wea. Rev., 111, 12441258, https://doi.org/10.1175/1520-0493(1983)111<1244:GCVCWS>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Proistosescu, C., A. Donohoe, K. C. Armour, G. H. Roe, M. F. Stuecker, and C. M. Bitz, 2018: Radiative feedbacks from stochastic variability in surface temperature and radiative imbalance. Geophys. Res. Lett., 45, 50825094, https://doi.org/10.1029/2018GL077678.

    • Search Google Scholar
    • Export Citation
  • Rossow, W. B., and R. A. Schiffer, 1991: ISCCP cloud data products. Bull. Amer. Meteor. Soc., 72, 220, https://doi.org/10.1175/1520-0477(1991)072<0002:ICDP>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Rugenstein, M. A. A., and K. C. Armour, 2021: Three flavors of radiative feedbacks and their implications for estimating equilibrium climate sensitivity. Geophys. Res. Lett., 48, e2021GL092983, https://doi.org/10.1029/2021GL092983.

    • Search Google Scholar
    • Export Citation
  • Schiffer, R. A., and W. B. Rossow, 1985: ISCCP global radiance data set: A new resource for climate research. Bull. Amer. Meteor. Soc., 66, 14981505, https://doi.org/10.1175/1520-0477(1985)066<1498:IGRDSA>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Schlund, M., A. Lauer, P. Gentine, S. C. Sherwood, and V. Eyring, 2020: Emergent constraints on equilibrium climate sensitivity in CMIP5: Do they hold for CMIP6? Earth Syst. Dyn., 11, 12331258, https://doi.org/10.5194/esd-11-1233-2020.

    • Search Google Scholar
    • Export Citation
  • Slingo, J. M., 1980: A cloud parametrization scheme derived from GATE data for use with a numerical model. Quart. J. Roy. Meteor. Soc., 106, 747770, https://doi.org/10.1002/qj.49710645008.

    • Search Google Scholar
    • Export Citation
  • Spencer, R. W., and W. D. Braswell, 2011: On the misdiagnosis of surface temperature feedbacks from variations in Earth’s radiant energy balance. Remote Sens., 3, 16031613, https://doi.org/10.3390/rs3081603.

    • Search Google Scholar
    • Export Citation
  • Takahashi, K., A. Montecinos, K. Goubanova, and B. Dewitte, 2011: ENSO regimes: Reinterpreting the canonical and Modoki El Niño. Geophys. Res. Lett., 38, L10704, https://doi.org/10.1029/2011GL047364.

    • Search Google Scholar
    • Export Citation
  • Taylor, K. E., D. Williamson, and F. Zwiers, 2000: The sea surface temperature and sea–ice concentration boundary conditions for AMIP II simulations. PCMDI Rep. 60, 28 pp., https://www.researchgate.net/publication/247997830.

  • Tett, S. F. B., and Coauthors, 2007: The impact of natural and anthropogenic forcings on climate and hydrology since 1550. Climate Dyn., 28, 334, https://doi.org/10.1007/s00382-006-0165-1.

    • Search Google Scholar
    • Export Citation
  • Trenberth, K. E., J. M. Caron, D. P. Stepaniak, and S. Worley, 2002: Evolution of El Niño-Southern Oscillation and global atmospheric surface temperatures. J. Geophys. Res., 107, 4065, https://doi.org/10.1029/2000JD000298.

    • Search Google Scholar
    • Export Citation
  • Trenberth, K. E., Y. Zhang, J. T. Fasullo, and S. Taguchi, 2015: Climate variability and relationships between top-of-atmosphere radiation and temperatures on Earth. J. Geophys. Res. Atmos., 120, 36423659, https://doi.org/10.1002/2014JD022887.

    • Search Google Scholar
    • Export Citation
  • Wigley, T. M. L., 2000: ENSO, volcanoes and record-breaking temperatures. Geophys. Res. Lett., 27, 41014104, https://doi.org/10.1029/2000GL012159.

    • Search Google Scholar
    • Export Citation
  • Wills, R. C. J., K. C. Armour, D. S. Battisti, C. Proistosescu, and L. A. Parsons, 2021: Slow modes of global temperature variability and their impact on climate sensitivity estimates. J. Climate, 34, 87178738, https://doi.org/10.1175/JCLI-D-20-1013.1.

    • Search Google Scholar
    • Export Citation
  • Wodzicki, K. R., and A. D. Rapp, 2016: Long-term characterization of the Pacific ITCZ using TRMM, GPCP, and ERA-Interim. J. Geophys. Res. Atmos., 121, 31533170, https://doi.org/10.1002/2015JD024458.

    • Search Google Scholar
    • Export Citation
  • Wodzicki, K. R., and A. D. Rapp, 2020: Variations in precipitating convective feature populations with ITCZ width in the Pacific Ocean. J. Climate, 33, 43914401, https://doi.org/10.1175/JCLI-D-19-0689.1.

    • Search Google Scholar
    • Export Citation
  • Wood, R., and C. S. Bretherton, 2006: On the relationship between stratiform low cloud cover and lower-tropospheric stability. J. Climate, 19, 64256432, https://doi.org/10.1175/JCLI3988.1.

    • Search Google Scholar
    • Export Citation
  • Wood, R., and D. L. Hartmann, 2006: Spatial variability of liquid water path in marine low cloud: The importance of mesoscale cellular convection. J. Climate, 19, 17481764, https://doi.org/10.1175/JCLI3702.1.

    • Search Google Scholar
    • Export Citation
  • Xie, S.-P., and Y. Kosaka, 2017: What caused the global surface warming hiatus of 1998–2013? Curr. Climate Change Rep., 3, 128140, https://doi.org/10.1007/s40641-017-0063-0.

    • Search Google Scholar
    • Export Citation
  • Xie, S.-P., Y. Kosaka, and Y. Okumura, 2016: Distinct energy budgets for anthropogenic and natural change during global warming hiatus. Nat. Geosci., 9, 2933, https://doi.org/10.1038/ngeo2581.

    • Search Google Scholar
    • Export Citation
  • Zelinka, M. D., T. A. Myers, D. T. McCoy, S. Po-Chedley, P. M. Caldwell, P. Ceppi, S. A. Klein, and K. E. Taylor, 2020: Causes of higher climate sensitivity in CMIP6 models. Geophy. Res. Lett., 47, e2019GL085782, https://doi.org/10.1029/2019GL085782.

    • Search Google Scholar
    • Export Citation
  • Zhou, C., M. D. Zelinka, A. E. Dessler, and S. A. Klein, 2015: The relationship between interannual and long-term cloud feedbacks. Geophys. Res. Lett., 42, 10 46310 469, https://doi.org/10.1002/2015GL066698.

    • Search Google Scholar
    • Export Citation
  • Zhou, C., M. D. Zelinka, and S. A. Klein, 2016: Impact of decadal cloud variations on the Earth’s energy budget. Nat. Geosci., 9, 871874, https://doi.org/10.1038/ngeo2828.

    • Search Google Scholar
    • Export Citation

Supplementary Materials

Save
  • Andrews, T., 2014: Using an AGCM to diagnose historical effective radiative forcing and mechanisms of recent decadal climate change. J. Climate, 27, 11931209, https://doi.org/10.1175/JCLI-D-13-00336.1.

    • Search Google Scholar
    • Export Citation
  • Andrews, T., and Coauthors, 2018: Accounting for changing temperature patterns increases historical estimates of climate sensitivity. Geophys. Res. Lett., 45, 84908499, https://doi.org/10.1029/2018GL078887.

    • Search Google Scholar
    • Export Citation
  • Bui, H. X., J.-Y. Yu, H.-W. Liu, C.-Y. Tu, P.-G. Chiu, and H.-H. Hsu, 2019: Convective structure changes over the equatorial Pacific with highly increased precipitation under global warming simulated in the HiRAM. SOLA, 15, 119124, https://doi.org/10.2151/sola.2019-022.

    • Search Google Scholar
    • Export Citation
  • Capotondi, A., and Coauthors, 2015: Understanding ENSO diversity. Bull. Amer. Meteor. Soc., 96, 921938, https://doi.org/10.1175/BAMS-D-13-00117.1.

    • Search Google Scholar
    • Export Citation
  • Ceppi, P., and S. Fueglistaler, 2021: The El Niño–Southern Oscillation pattern effect. Geophys. Res. Lett., 48, e2021GL095261, https://doi.org/10.1029/2021GL095261.

    • Search Google Scholar
    • Export Citation
  • Ceppi, P., F. Brient, M. D. Zelinka, and D. L. Hartmann, 2017: Cloud feedback mechanisms and their representation in global climate models. Wiley Interdiscip. Rev.: Climate Change, 8, e465, https://doi.org/10.1002/wcc.465.

    • Search Google Scholar
    • Export Citation
  • Chen, L., T. Li, Y. Yu, and S. K. Behera, 2017: A possible explanation for the divergent projection of ENSO amplitude change under global warming. Climate Dyn., 49, 37993811, https://doi.org/10.1007/s00382-017-3544-x.

    • Search Google Scholar
    • Export Citation
  • Chiang, J. C. H., and B. R. Lintner, 2005: Mechanisms of remote tropical surface warming during El Niño. J. Climate, 18, 41304149, https://doi.org/10.1175/JCLI3529.1.

    • Search Google Scholar
    • Export Citation
  • Colman, R., and L. Hanson, 2017: On the relative strength of radiative feedbacks under climate variability and change. Climate Dyn., 49, 21152129, https://doi.org/10.1007/s00382-016-3441-8.

    • Search Google Scholar
    • Export Citation
  • Dias, J., and O. Pauluis, 2011: Modulations of the phase speed of convectively coupled Kelvin waves by the ITCZ. J. Atmos. Sci., 68, 14461459, https://doi.org/10.1175/2011JAS3630.1.

    • Search Google Scholar
    • Export Citation
  • Eyring, V., S. Bony, G. A. Meehl, C. A. Senior, B. Stevens, R. J. Stouffer, and K. E. Taylor, 2016: Overview of the Coupled Model Intercomparison Project Phase 6 (CMIP6) experimental design and organization. Geosci. Model Dev., 9, 19371958, https://doi.org/10.5194/gmd-9-1937-2016.

    • Search Google Scholar
    • Export Citation
  • Forster, P. M. F., and J. M. Gregory, 2006: The climate sensitivity and its components diagnosed from Earth radiation budget data. J. Climate, 19, 3952, https://doi.org/10.1175/JCLI3611.1.

    • Search Google Scholar
    • Export Citation
  • Gates, W. L., 1992: AMIP: The Atmospheric Model Intercomparison Project. Bull. Amer. Meteor. Soc., 73, 19621970, https://doi.org/10.1175/1520-0477(1992)073<1962:ATAMIP>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Gates, W. L., and Coauthors, 1999: An overview of the results of the Atmospheric Model Intercomparison Project (AMIP I). Bull. Amer. Meteor. Soc., 80, 2956, https://doi.org/10.1175/1520-0477(1999)080<0029:AOOTRO>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Gregory, J. M., and M. Webb, 2008: Tropospheric adjustment induces a cloud component in CO2 forcing. J. Climate, 21, 5871, https://doi.org/10.1175/2007JCLI1834.1.

    • Search Google Scholar
    • Export Citation
  • Gregory, J. M., and T. Andrews, 2016: Variation in climate sensitivity and feedback parameters during the historical period. Geophys. Res. Lett., 43, 39113920, https://doi.org/10.1002/2016GL068406.

    • Search Google Scholar
    • Export Citation
  • Gregory, J. M., R. J. Stouffer, S. C. B. Raper, P. A. Stott, and N. A. Rayner, 2002: An observationally based estimate of the climate sensitivity. J. Climate, 15, 31173121, https://doi.org/10.1175/1520-0442(2002)015<3117:AOBEOT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Gregory, J. M., and Coauthors, 2004: A new method for diagnosing radiative forcing and climate sensitivity. Geophys. Res. Lett., 31, L03205, https://doi.org/10.1029/2003GL018747.

    • Search Google Scholar
    • Export Citation
  • Gregory, J. M., T. Andrews, P. Ceppi, T. Mauritsen, and M. J. Webb, 2020: How accurately can the climate sensitivity to CO2 be estimated from historical climate change? Climate Dyn., 54, 129157, https://doi.org/10.1007/s00382-019-04991-y.

    • Search Google Scholar
    • Export Citation
  • IPCC, 2023: Climate Change 2022: Impacts, Adaptation, and Vulnerability. Cambridge University Press, 3056 pp., https://doi.org/10.1017/9781009325844.

  • Kao, H.-Y., and J.-Y. Yu, 2009: Contrasting eastern-Pacific and central-Pacific types of ENSO. J. Climate, 22, 615632, https://doi.org/10.1175/2008JCLI2309.1.

    • Search Google Scholar
    • Export Citation
  • Klein, S. A., 1997: Synoptic variability of low-cloud properties and meteorological parameters in the subtropical trade wind boundary layer. J. Climate, 10, 20182039, https://doi.org/10.1175/1520-0442(1997)010<2018:SVOLCP>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Klein, S. A., and D. L. Hartmann, 1993: The seasonal cycle of low stratiform clouds. J. Climate, 6, 15871606, https://doi.org/10.1175/1520-0442(1993)006%3C1587:TSCOLS%3E2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Kobayashi, S., and Coauthors, 2015: The JRA-55 reanalysis: General specifications and basic characteristics. J. Meteor. Soc. Japan, 93, 548, https://doi.org/10.2151/jmsj.2015-001.

    • Search Google Scholar
    • Export Citation
  • Kumar, A., and M. P. Hoerling, 2003: The nature and causes for the delayed atmospheric response to El Niño. J. Climate, 16, 13911403, https://doi.org/10.1175/1520-0442(2003)16%3C1391:TNACFT%3E2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Kurihana, T., E. Moyer, R. Willett, D. Gilton, and I. Foster, 2022: Data-driven cloud clustering via a rotationally invariant autoencoder. IEEE Trans. Geosci. Remote Sens., 60, 125, https://doi.org/10.1109/TGRS.2021.3098008.

    • Search Google Scholar
    • Export Citation
  • Lawrence, M. G., 2005: The relationship between relative humidity and the dewpoint temperature in moist air: A simple conversion and applications. Bull. Amer. Meteor. Soc., 86, 225234, https://doi.org/10.1175/BAMS-86-2-225.

    • Search Google Scholar
    • Export Citation
  • Li, S., and P. Huang, 2022: An exponential-interval sampling method for evaluating equilibrium climate sensitivity via reducing internal variability noise. Geosci. Lett., 9, 34, https://doi.org/10.1186/s40562-022-00244-9.

    • Search Google Scholar
    • Export Citation
  • Lutsko, N. J., and K. Takahashi, 2018: What can the internal variability of CMIP5 models tell us about their climate sensitivity? J. Climate, 31, 50515069, https://doi.org/10.1175/JCLI-D-17-0736.1.

    • Search Google Scholar
    • Export Citation
  • Norris, J. R., 1998: Low cloud type over the ocean from surface observations. Part I: Relationship to surface meteorology and the vertical distribution of temperature and moisture. J. Climate, 11, 369382, https://doi.org/10.1175/1520-0442(1998)011<0369:LCTOTO>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Norris, J. R., and S. F. Iacobellis, 2005: North Pacific cloud feedbacks inferred from synoptic-scale dynamic and thermodynamic relationships. J. Climate, 18, 48624878, https://doi.org/10.1175/JCLI3558.1.

    • Search Google Scholar
    • Export Citation
  • Oort, A. H., and J. J. Yienger, 1996: Observed interannual variability in the Hadley circulation and its connection to ENSO. J. Climate, 9, 27512767, https://doi.org/10.1175/1520-0442(1996)009<2751:OIVITH>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Otto, A., and Coauthors, 2013: Energy budget constraints on climate response. Nat. Geosci., 6, 415416, https://doi.org/10.1038/ngeo1836.

    • Search Google Scholar
    • Export Citation
  • Pan, Y. H., and A. H. Oort, 1983: Global climate variations connected with sea surface temperature anomalies in the eastern equatorial Pacific Ocean for the 1958–73 period. Mon. Wea. Rev., 111, 12441258, https://doi.org/10.1175/1520-0493(1983)111<1244:GCVCWS>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Proistosescu, C., A. Donohoe, K. C. Armour, G. H. Roe, M. F. Stuecker, and C. M. Bitz, 2018: Radiative feedbacks from stochastic variability in surface temperature and radiative imbalance. Geophys. Res. Lett., 45, 50825094, https://doi.org/10.1029/2018GL077678.

    • Search Google Scholar
    • Export Citation
  • Rossow, W. B., and R. A. Schiffer, 1991: ISCCP cloud data products. Bull. Amer. Meteor. Soc., 72, 220, https://doi.org/10.1175/1520-0477(1991)072<0002:ICDP>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Rugenstein, M. A. A., and K. C. Armour, 2021: Three flavors of radiative feedbacks and their implications for estimating equilibrium climate sensitivity. Geophys. Res. Lett., 48, e2021GL092983, https://doi.org/10.1029/2021GL092983.

    • Search Google Scholar
    • Export Citation
  • Schiffer, R. A., and W. B. Rossow, 1985: ISCCP global radiance data set: A new resource for climate research. Bull. Amer. Meteor. Soc., 66, 14981505, https://doi.org/10.1175/1520-0477(1985)066<1498:IGRDSA>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Schlund, M., A. Lauer, P. Gentine, S. C. Sherwood, and V. Eyring, 2020: Emergent constraints on equilibrium climate sensitivity in CMIP5: Do they hold for CMIP6? Earth Syst. Dyn., 11, 12331258, https://doi.org/10.5194/esd-11-1233-2020.

    • Search Google Scholar
    • Export Citation
  • Slingo, J. M., 1980: A cloud parametrization scheme derived from GATE data for use with a numerical model. Quart. J. Roy. Meteor. Soc., 106, 747770, https://doi.org/10.1002/qj.49710645008.

    • Search Google Scholar
    • Export Citation
  • Spencer, R. W., and W. D. Braswell, 2011: On the misdiagnosis of surface temperature feedbacks from variations in Earth’s radiant energy balance. Remote Sens., 3, 16031613, https://doi.org/10.3390/rs3081603.

    • Search Google Scholar
    • Export Citation
  • Takahashi, K., A. Montecinos, K. Goubanova, and B. Dewitte, 2011: ENSO regimes: Reinterpreting the canonical and Modoki El Niño. Geophys. Res. Lett., 38, L10704, https://doi.org/10.1029/2011GL047364.

    • Search Google Scholar
    • Export Citation
  • Taylor, K. E., D. Williamson, and F. Zwiers, 2000: The sea surface temperature and sea–ice concentration boundary conditions for AMIP II simulations. PCMDI Rep. 60, 28 pp., https://www.researchgate.net/publication/247997830.

  • Tett, S. F. B., and Coauthors, 2007: The impact of natural and anthropogenic forcings on climate and hydrology since 1550. Climate Dyn., 28, 334, https://doi.org/10.1007/s00382-006-0165-1.

    • Search Google Scholar
    • Export Citation
  • Trenberth, K. E., J. M. Caron, D. P. Stepaniak, and S. Worley, 2002: Evolution of El Niño-Southern Oscillation and global atmospheric surface temperatures. J. Geophys. Res., 107, 4065, https://doi.org/10.1029/2000JD000298.

    • Search Google Scholar
    • Export Citation
  • Trenberth, K. E., Y. Zhang, J. T. Fasullo, and S. Taguchi, 2015: Climate variability and relationships between top-of-atmosphere radiation and temperatures on Earth. J. Geophys. Res. Atmos., 120, 36423659, https://doi.org/10.1002/2014JD022887.

    • Search Google Scholar
    • Export Citation
  • Wigley, T. M. L., 2000: ENSO, volcanoes and record-breaking temperatures. Geophys. Res. Lett., 27, 41014104, https://doi.org/10.1029/2000GL012159.

    • Search Google Scholar
    • Export Citation
  • Wills, R. C. J., K. C. Armour, D. S. Battisti, C. Proistosescu, and L. A. Parsons, 2021: Slow modes of global temperature variability and their impact on climate sensitivity estimates. J. Climate, 34, 87178738, https://doi.org/10.1175/JCLI-D-20-1013.1.

    • Search Google Scholar
    • Export Citation
  • Wodzicki, K. R., and A. D. Rapp, 2016: Long-term characterization of the Pacific ITCZ using TRMM, GPCP, and ERA-Interim. J. Geophys. Res. Atmos., 121, 31533170, https://doi.org/10.1002/2015JD024458.

    • Search Google Scholar
    • Export Citation
  • Wodzicki, K. R., and A. D. Rapp, 2020: Variations in precipitating convective feature populations with ITCZ width in the Pacific Ocean. J. Climate, 33, 43914401, https://doi.org/10.1175/JCLI-D-19-0689.1.

    • Search Google Scholar
    • Export Citation
  • Wood, R., and C. S. Bretherton, 2006: On the relationship between stratiform low cloud cover and lower-tropospheric stability. J. Climate, 19, 64256432, https://doi.org/10.1175/JCLI3988.1.

    • Search Google Scholar
    • Export Citation
  • Wood, R., and D. L. Hartmann, 2006: Spatial variability of liquid water path in marine low cloud: The importance of mesoscale cellular convection. J. Climate, 19, 17481764, https://doi.org/10.1175/JCLI3702.1.

    • Search Google Scholar
    • Export Citation
  • Xie, S.-P., and Y. Kosaka, 2017: What caused the global surface warming hiatus of 1998–2013? Curr. Climate Change Rep., 3, 128140, https://doi.org/10.1007/s40641-017-0063-0.

    • Search Google Scholar
    • Export Citation
  • Xie, S.-P., Y. Kosaka, and Y. Okumura, 2016: Distinct energy budgets for anthropogenic and natural change during global warming hiatus. Nat. Geosci., 9, 2933, https://doi.org/10.1038/ngeo2581.

    • Search Google Scholar
    • Export Citation
  • Zelinka, M. D., T. A. Myers, D. T. McCoy, S. Po-Chedley, P. M. Caldwell, P. Ceppi, S. A. Klein, and K. E. Taylor, 2020: Causes of higher climate sensitivity in CMIP6 models. Geophy. Res. Lett., 47, e2019GL085782, https://doi.org/10.1029/2019GL085782.

    • Search Google Scholar
    • Export Citation
  • Zhou, C., M. D. Zelinka, A. E. Dessler, and S. A. Klein, 2015: The relationship between interannual and long-term cloud feedbacks. Geophys. Res. Lett., 42, 10 46310 469, https://doi.org/10.1002/2015GL066698.

    • Search Google Scholar
    • Export Citation
  • Zhou, C., M. D. Zelinka, and S. A. Klein, 2016: Impact of decadal cloud variations on the Earth’s energy budget. Nat. Geosci., 9, 871874, https://doi.org/10.1038/ngeo2828.

    • Search Google Scholar
    • Export Citation
  • Fig. 1.

    Relationship between global mean surface temperature (GMST; K) and positive downward net radiation flux at the top of the atmosphere (GMTOA; W m−2) on interannual variability in Coupled Model Intercomparison Project phase 6 (CMIP6) preindustrial control simulations. (a) Scatterplot of positive downward GMTOA cross-correlated to GMST at lag 0 and +1, as denoted by the x and y axes, respectively. Each model is classified into three groups: groups A (|lag 0| = |lag +1|), B (|lag 0| < |lag +1|), and C (|lag 0| > |lag +1|), colored red, blue, and gray, respectively. Black and gray lines denote |lag 0|:|lag +1| = 0.75:1, 1:1, and 1:1.33, respectively. (b) Lead–lag correlation of positive downward GMTOA against GMST, with the same colors as in (a). Positive lags indicate that GMSTs lead. Autocorrelation of GMST is also shown. Shadings represent the 95% confidence interval on an assumption of the Student’s t distribution in each group.