1. Introduction
High and low clouds make an important contribution to the uncertainties in predicting climate change (e.g., Hartmann et al. 1992); the uncertainty of cloud simulations in general circulation models (GCMs) affects the energy balance of the atmosphere and therefore the general circulation (Soden and Held 2006; Dufresne and Bony 2008; Bony et al. 2016). Furthermore, their complex interaction with atmospheric radiation processes, such as absorption and emission in the atmosphere, is strongly related to the altitude and optical depths of clouds (Koren et al. 2010), so the necessity of simulating cloud process appropriately in GCMs is widely recognized. All the models in phase 3 of the Coupled Model Intercomparison Project (CMIP3) and most models of phase 5 (CMIP5) considered only prognostic cloud ice (floating ice), but some GCMs are exceptions. For example, CESM1-CAM5 considers falling ice, such as snow diagnostically, whereas CESM2-CAM6, NorESM2, and MIROC6-SPRINTARS consider falling ice (i.e., snow) prognostically (Waliser et al. 2009; Gettelman and Morrison 2015; Gettelman et al. 2010; Li et al. 2013, 2020; Michibata et al. 2019). This work focuses on the variability of high clouds, especially the high thin or cirrus type clouds, because the outgoing longwave radiation (OLR) at the top of the atmosphere (TOA) seen from space is strongly affected by these clouds, as they have large radiative effects on the OLR and wide coverage in the tropics (Waliser et al. 2009; Li et al. 2013, 2016).
Several observational studies on high cloud variations have revealed aspects of their relationship with sea surface temperature (SST). Zelinka and Hartmann (2011) investigated the link between high cloud fraction obtained from satellites [such as cloud fraction collected by MODIS (the Moderate Resolution Imaging Spectroradiometer; Platnick et al. 2003), AIRS (the Atmospheric Infrared Sounder; Aumann et al. 2003), and ISCCP (the International Satellite Cloud Climatology Project; Rossow and Schiffer 1999)], cloud-top frequency collected by CloudSat (Stephens et al. 2002), and SST observations (HadCRUT3v; Brohan et al. 2006). They found that the fraction or frequency of the top level of high cloud (above 200 hPa) shows a positive relation with SST variation, whereas the fractions or frequencies of other levels of high cloud (below 200 hPa) show a negative relation with SST variation. Liu et al. (2017) investigated high cloud variability and its relation to surface temperature from 2002 to 2015 using MODIS Level 3 Collection 6 monthly cloud products from Aqua and surface temperature from the HadCRUT4 data. Although the total amount of high cloud decreases with increasing SST, the cloud fraction of thin cirrus clouds increases as the surface temperature becomes warmer, whereas cirrostratus and clouds associated with deep convection decrease. Su and Jiang (2013) focused on high cloud variations and their links to the cloud water content (CWC) obtained from CloudSat level 2B-CWC-RO data and 3B-GEOPROF-lidar cloud fraction data obtained from combined CloudSat radar and CALIPSO lidar to evaluate tropical cloud change during the 2006/07 and 2009/10 El Niño events. Similar to the results of Liu et al. (2017), an increase of both CWC and cloud fraction of high clouds in the layers above 14-km altitude was confirmed in the two events, although the amount of high cloud below 14 km decreased.
Under the framework of the model intercomparison projects IPCC AR5 and AR6, the average cloud fraction change under global warming in CMIP3 models showed that the total amount of high cloud decreases under the warming climate. In detail, the top layer of high cloud (at pressure levels 180–50 hPa) increases, whereas other high cloud (at pressure levels 440–50 hPa) decreases (Zelinka et al. 2012). Similar results are also found in Su et al. (2014), who showed that GCMs simulate a cloud fraction decrease in the troposphere but a cloud fraction increase in the tropopause under the RCP4.5 scenario. Using the cloud radiative kernel proposed by Zelinka et al. (2012), they indicated that the large uncertainty in cloud feedback estimation mainly comes from the uncertainty in the response of simulated high and thin cloud to the warming conditions in GCMs. Chen et al. (2016) furthermore compared the enhanced cloud feedback obtained from NICAM (Nonhydrostatic Icosahedral Atmospheric Model; Satoh et al. 2008, 2014) with CFMIP2 (Cloud Feedback Model Intercomparison Project 2; https://www.cfmip.org). Their results showed that the longwave (LW) cloud feedbacks in NICAM are stronger than those in CFMIP2 models, and this difference comes mainly from the effect of high thin clouds that are formed mainly by the ice hydrometeor, cloud ice.
As mentioned above, the representation of clouds and their effects on radiation remain a great challenge in climate simulations (Y.-W. Chen et al. 2018). In GCMs, although computing skill has progressed, the grid spacing (both horizontal and vertical) is generally still too coarse to simulate clouds. For instance, GCMs participating in CMIP3 and CMIP5 use horizontal grid spacing on the order of 100 km, which is far too coarse to simulate clouds. Noda and Satoh (2014) showed that the advantage of using cloud system–resolving models (or storm-resolving models) such as NICAM is that the model-resolvable high cloud sizes are more appropriately simulated by GCRMs (global cloud-resolving models) than conventional GCMs. They statistically analyzed cloud size distributions of high clouds simulated in the 14- and 7-km mesh NICAM and compared them with those of the observations. They discussed in more detail the positive LW cloud feedbacks in NICAM and reported that they are larger than that simulated by conventional GCMs because of the response of small high clouds to global warming in NICAM. In addition, they indicated that the increase in high clouds in NICAM is attributed mainly to these small high clouds and the contribution of these small clouds to the change of cloud radiative forcing (CRF) under global warming is important because the change in the number of these small clouds is much greater than that of large clouds. Therefore, a model with finer resolution is considered better (or more realistic) than GCMs with coarse resolution for simulating small clouds.
Recently, the importance of the accuracy of simulated ice hydrometeors has been recognized. Several conventional models have started to improve their cloud physical scheme by implementing diagnostic or prognostic schemes in snow for a better evaluation of the ice process and results (Gettelman and Morrison 2015; Gettelman et al. 2010; Li et al. 2013, 2020; Michibata et al. 2019). Meanwhile, a new model intercomparison project has been conducted, DYAMOND (Dynamics of the Atmospheric General Circulation Modeled on Nonhydrostatic Domains; Stevens et al. 2019; Satoh et al. 2019), which includes storm-resolving models (also called cloud-system resolving models) focusing on high-resolution nonhydrostatic models. The present paper analyzes the long-term behavior of high clouds simulated by the nonhydrostatic icosahedral atmospheric model NICAM (Satoh et al. 2008, 2014), one of the global convection-permitting models participating in the DYAMOND project, which uses cloud microphysics schemes instead of cumulus parameterization to simulate cloud behaviors. Before the DYAMOND project, a number of studies have investigated the high clouds related to ice hydrometeors using short-term NICAM simulations (e.g., Kodama et al. 2012; Seiki et al. 2015a; Chen et al. 2016; Y.-W. Chen et al. 2018). Chen et al. (2016) analyzed how the emissivity of ice clouds varies with the type of ice hydrometeors in the high clouds in NICAM simulations. For the same ice water path, clouds containing both cloud ice and snow have larger emissivity than clouds containing only snow. Therefore, clouds containing both ice and snow produce stronger longwave cloud radiative forcing than clouds containing snow only. Further analysis in Y.-W. Chen et al. (2018) revealed that the model-simulated OLR is sensitive to the treatment of ice hydrometeors in terms of both type and altitude. In other models, Gasparini et al. (2021) analyzed the life cycle of anvil clouds and showed that the cloud ice increases while the cloud fraction decreases under a warmer climate. In addition, Sullivan and Voigt (2021) indicated that the energy balance in the tropics is strongly modified by the hydrometeor distributions, which are dependent on the simulated microphysical processes.
Recently, it has been realized that the treatment of the ice hydrometeors that form the high/ice clouds in the atmosphere in GCM/GCRM simulations has a large impact not only on OLR but also on the dynamic processes. Waliser et al. (2011) and Y.-W. Chen et al. (2018) showed how the ice hydrometeor treatment may affect OLR simulations, using CloudSat observations and NICAM simulations, respectively. Li et al. (2012, 2015, 2021) focused on how the floating (cloud ice) and falling (snow) ice simulated in GCMs affect the OLR simulation. The ice hydrometeors within high clouds have large impacts on both the OLR balance at the top of the atmosphere and the heating profile of the atmosphere (e.g., Waliser et al. 2011; Y.-W. Chen et al. 2018) and their long-term variations related to SST variation are not known in detail. Hence, this work aims to reveal how the long-term high cloud distribution varies with SST forcing under the present climate and how it changes under a warming climate forced by prescribed Niño-like SST warming in multidecadal-scale simulations with NICAM (Kodama et al. 2015). Unlike most of the conventional GCMs with horizontal resolution of ∼100 km, the NICAM 14-km simulation may help us to analyze more precisely how the ice hydrometeors such as cloud ice, snow, and graupel that are directly associated with the formation of high clouds behave under warming conditions. This study is the first result of the analysis of clouds using the 20-yr 14-km simulations obtained from NICAM, the cloud system–resolving GCM, which allows us to investigate interannual variations of clouds. In the following, the experimental configuration and the model used together with the observational data used as reference are described in section 2. The results for the annual variation of high cloud for the present climate and for warming conditions and how ice hydrometeors behave in the present and warming conditions are presented in section 3. The roles that ice hydrometeors play and a discussion are given in section 4, and our conclusions and suggestions for future work are presented in section 5.
2. Data and method
a. Model data
This study analyzes a set of 20-yr climate simulations for the present (January 1989–December 2008) condition follows the AMIP protocol with 14-km horizontal resolution obtained from runs of NICAM [NICAM-AMIP; see Kodama et al. (2015) for detailed configuration of the present climate]. The cloud processes in this simulation are directly calculated by the single-moment bulk cloud microphysics scheme (Tomita 2008). This cloud microphysics scheme determines six categories of mixing ratio: water vapor, rain, cloud water, cloud ice, snow, and graupel. The general performance for present climate shows that ENSO-related, seasonal, and diurnal variations are reasonably well reproduced, and the meridional circulation, clouds, and TOA radiation are simulated qualitatively, although significant biases remain, such as underestimation of low cloud fraction and shortwave reflection (Kodama et al. 2015).
For the warming climate, the SSTs produced by Mizuta et al. (2008) from the WCRP (World Climate Research Program) CMIP3 dataset were used. Changes in the climatological pattern and the trend (but not variability) of SST and sea ice concentration derived from 18 models were taken as boundary conditions. The CO2 concentrations follow the Special Report on Emission Scenarios (SRES) A1B (IPCC 2000). The A1 scenario describes a future world of very rapid economic growth, with global population that peaks in the mid-twenty-first century and declines thereafter. Under the A1 scenario there are three subscenarios: fossil intensive (A1FI), non-fossil energy sources (A1T), and a balance between A1FI and A1T (A1B). The global emissions of CO2 range from 15 to 18 GtC yr−1 in 2075–98. Figure 1a illustrates the mean state of the SST distribution under the present climate and how this mean state changes with the warming climate; Fig. 1b illustrates the mean SST distribution obtained from HadISST (January 2007–December 2015). Figures 1a and 1b show that the simulated and observed SST distributions are similar in the present climate, whereas the SST in the eastern Pacific (the east coast of South America) is higher in HadISST than in NICAM-AMIP. This difference may result from the analysis record length and duration difference in HadISST and NICAM-AMIP. Meanwhile, Fig. 1a shows a prescribed El Niño–type SST warming pattern with the magnitude of the tropical SST temperature increase ranging from 1 to 2.5 K. The SST increases by more than 2 K over the eastern and central Pacific, Indian Ocean, and tropical Atlantic. Detailed settings for NICAM–AMIP under the warming conditions may be found in Satoh et al. (2015).
SST distributions (K) obtained from (a) NICAM-AMIP and (b) HadISST. Contours and shadings in (a) denote the mean SST distribution for the present climate (January 1989–December 2007) and the difference between the warming (January 2078–December 2097) and present climate (January 1989–December 2007), respectively. Contours in (b) denote the observed SST (January 2007–December 2015).
Citation: Journal of Climate 35, 11; 10.1175/JCLI-D-21-0150.1
SST distributions (K) obtained from (a) NICAM-AMIP and (b) HadISST. Contours and shadings in (a) denote the mean SST distribution for the present climate (January 1989–December 2007) and the difference between the warming (January 2078–December 2097) and present climate (January 1989–December 2007), respectively. Contours in (b) denote the observed SST (January 2007–December 2015).
Citation: Journal of Climate 35, 11; 10.1175/JCLI-D-21-0150.1
SST distributions (K) obtained from (a) NICAM-AMIP and (b) HadISST. Contours and shadings in (a) denote the mean SST distribution for the present climate (January 1989–December 2007) and the difference between the warming (January 2078–December 2097) and present climate (January 1989–December 2007), respectively. Contours in (b) denote the observed SST (January 2007–December 2015).
Citation: Journal of Climate 35, 11; 10.1175/JCLI-D-21-0150.1
b. Observational data
The cloud fraction derived from CALIPSO-GOCCP (GCM Oriented Cloud CALIPSO Product, Chepfer et al. 2010) is used to evaluate the cloud fraction in NICAM–AMIP under the present climate. The GOCCP dataset is designed for diagnosing cloud properties from CALIPSO observations in exactly the same way as the Cloud Feedback Model Intercomparison Program (CFMIP; Webb et al. 2017) Observation Simulator Package (COSP) under the CFMIP framework to ensure the discrepancies between GCM and observations appear from biases in GCMs. It has been reported that high-level (ice) clouds are well retrieved, whereas midlevel (mixed-phase) and low-level (water-phase) clouds are overestimated (Cesana et al. 2016). The cloud fraction is provided as averages over 480-m layers, with 40 layers from the surface to 19.2 km. Data with horizontal resolution of 2° × 2° are used for evaluating cloud behavior in NICAM-AMIP, as this resolution is recommended for model evaluations. The analyzed duration is 9 years, from January 2007 to December 2015. The SST data for the same period are obtained from the HadISST (Rayner et al. 2003) monthly mean with horizontal resolution of 1° × 1°.
c. Analysis method
The definition of the cloud fraction in NICAM is explained here. The cloud fraction is flagged as 0 or 1 at each grid box based on the simulated properties of solid hydrometeors. The cloud fraction is flagged as 1 in a grid box if the sum of the mass concentration of cloud water, rain, cloud ice, snow, and graupel in the grid box is greater than 0.5 mg m−3. Note that a column accumulated cloud optical depth of less than 0.3 is the cutoff in the following analysis because the GOCCP data provide cloud information with optical depth larger than 0.3. This cutoff is applied as follows. The column accumulated cloud optical depth is retrieved from the column volume of ice water path and liquid path. If the optical depth in a column is larger than 0.3, the cloud fraction flags of the vertical grid boxes in that column are retained; otherwise, the cloud flags are set to 0. Previous studies on cloud properties in NICAM simulations have shown that NICAM can reproduce the global-scale statistical properties of clouds (e.g., Satoh et al. 2010; Kodama et al. 2012; Noda et al. 2012) although individual cloud processes cannot be simulated at 14-km resolution. Furthermore, Hashino et al. (2013) especially indicated the similarity of the statistical cloud properties between the 14- and 3.5-km resolution simulations. Therefore, the 14-km simulation is considered as having a suitable resolution to evaluate the relation between ice cloud and SST in this study from a climate point of view.
The reason why the cloud fraction definition in NICAM is chosen in a different way from GOCCP is that we have compared NICAM-AMIP cloud fraction derived from ice hydrometeor, derived from the COSP-GOCCP simulator and GOCCP. Kodama et al. (2015) showed the NICAM-AMIP cloud fraction derived from the ISCCP simulator under the present climate, which is consistent with other NICAM experiments (Satoh et al. 2012; Kodama et al. 2012; Chen et al. 2016). The bias (overestimating on high clouds) is thought to come from the wider horizontal spread in the upper troposphere of cloud ice simulated by NICAM, which is detected as optically thin high clouds in the ISCCP categorization. Similar to the above studies, the NICAM cloud fraction derived from the COSP-CALIPSO simulator faces a similar issue. Note also that another possible reason is that CALIPSO is missing larger particles under high-level clouds due to attenuation.
Similar to the method used in Zelinka and Hartmann (2011), the anomalies shown in this study are calculated as follows. The tropical mean monthly means over the data period are calculated. Then, the anomalies of each monthly mean are taken as the difference between the tropical mean (30°N–30°S) monthly data and the tropical mean monthly means over the data period. The regression coefficients, which are treated as the sensitivities to tropical mean surface temperature, are calculated based on the anomalies of the monthly means of each variable and SST. Note that we have regridded all the data to 2° × 2° to show the analysis results at the same horizontal resolution.
3. Results
a. Climatology and variability of high clouds in present climate
Figure 2 shows the tropically averaged (30°N–30°S) longitude–height section for altitudes above 6.0 km for the visible cloud fraction for GOCCP (Fig. 2a) and NICAM-AMIP (Fig. 2b) under the present climate. Qualitatively, the cloud fraction distribution in NICAM-AMIP shows a similar behavior to that in GOCCP. The maximum of the cloud fraction appears at altitudes between 12 and 17 km located at longitudes of 80°–160°E and 100°–60°W. The large difference of the cloud fraction values between GOCCP and NICAM-AMIP is thought to result from the difference in cloud fraction definition. The GOCCP cloud fraction is retrieved from the satellite simulator, whereas the NICAM cloud fraction is defined using the solid hydrometeors amount. Cesana et al. (2016) noted that the cloud fraction retrieved from CALIPSO is largely dependent on the retrieval method. With different cloud detection thresholds, the maxima of the ice cloud fraction may range from 0.1 to 0.3 in the tropics. Meanwhile, it is found that NICAM tends to simulate the most cloud ice at the top of the atmosphere of the GCRMs participating in the intermodel comparison project DYAMOND (Roh et al. 2021; Turbeville et al. 2021). When comparing the simulated IWC (ice water content) in NICAM with IWC retrieved by the CloudSat 2C-ICE product (Deng et al. 2015), results showed that NICAM is generally unable to simulate the convective ice hydrometeor (graupel) in the cloud bottom (details are shown in appendix B). Note that similar results were found in Turbeville et al. (2021) using the DYAMOND dataset, showing that NICAM has too much cloud cover in the tropopause but has too few low clouds. Despite the fact that the cloud fraction obtained from CALIPSO is greatly affected by the retrieval method and the large spread of the amount of cloud ice simulated in high-resolution models, this study focuses on the pattern similarity of simulated cloud ice in NICAM and GOCCP for the present climate. The vertical motion in NICAM–AMIP is also shown in Fig. 2b (shading, upward positive) to show the relation between these two variables in NICAM-AMIP. The locations of the two peaks of high cloud fraction peaks correspond well to where upward motion of more than 0.0025 m s−1 reaches about 14 km in altitude; the cloud fraction decreases rapidly with height above 14 km. Focusing on the peak of cloud fraction, in GOCCP it lies over the Indian Ocean, whereas in NICAM it lies over the western Pacific. This implies that the deep convection properties may be different in observations and models. Note that the issue that NICAM faces the overestimation on high clouds so far. As shown in Kodama et al. (2015), NICAM overestimates the ISCCP high cloud fraction compared with ISCCP daytime cloud fraction using the same dataset. Kodama et al. (2012) showed that the NICAM cloud fraction retrieved by the COSP-CALIPSO simulator is overestimated because the small cloud ice was widely spread in the horizontal in the upper troposphere. Once the ice cloud is detected at the upper troposphere, the ice cloud below is underestimated. Because we focus on how the cloud variation is linked to the ice hydrometeors, we consider the cloud fraction derived from the ice hydrometeors in the following analysis. Also note that our main purpose is not to compare NICAM with GOCCP but to use the GOCCP data as a reference. Further information on the difference between the cloud fraction derived from ice hydrometeors and that derived from the GOCCP simulator is shown in appendix A.
Tropically (30°N–30°S) averaged vertical cloud fraction distribution from 6.5- to 20-km height for (a) GOCCP (January 2007–December 2015) and (b) NICAM–AMIP (contours; January 1989–December 2008). Vertical motion is shown in (b) (shading; upward positive).
Citation: Journal of Climate 35, 11; 10.1175/JCLI-D-21-0150.1
Tropically (30°N–30°S) averaged vertical cloud fraction distribution from 6.5- to 20-km height for (a) GOCCP (January 2007–December 2015) and (b) NICAM–AMIP (contours; January 1989–December 2008). Vertical motion is shown in (b) (shading; upward positive).
Citation: Journal of Climate 35, 11; 10.1175/JCLI-D-21-0150.1
Tropically (30°N–30°S) averaged vertical cloud fraction distribution from 6.5- to 20-km height for (a) GOCCP (January 2007–December 2015) and (b) NICAM–AMIP (contours; January 1989–December 2008). Vertical motion is shown in (b) (shading; upward positive).
Citation: Journal of Climate 35, 11; 10.1175/JCLI-D-21-0150.1
The vertical profile of the tropical mean cloud fraction variation against tropical mean SST (regression) is shown in Fig. 3a, for both NICAM-AMIP and GOCCP. Note that the GOCCP cloud fraction is regressed onto HadISST. Also note that only results above 7 km are shown because the clouds at these altitudes are defined as high clouds. This figure shows the cloud fraction and SST have a positive (negative) relationship at altitudes above (below) 14–16 km in both NICAM simulations and observations. This figure also shows that the responses of cloud fraction to the SST simulated in NICAM are stronger than results in Zelinka and Hartmann (2011), which revealed that the high cloud response to SST variation obtained from satellite observations ranges from 0% to 1% K−1 above 200 hPa. Figure 3b furthermore shows the vertical distribution of IWC from cloud ice, snow, and graupel regressed onto SST, as the high clouds are formed mainly by ice hydrometeors. The patterns of cloud ice and snow are consistent with the cloud fraction shown in Fig. 3a, although the peak of the positive regression is lower than that for cloud fraction. The response of graupel mainly dominates in the lower part of the high cloud region and does not have a large influence on the distribution of cloud fraction. Note that the cloud variation in GOCCP may lack cloud information in the altitudes below 11 km, where graupel may predominate, because CALIPSO misses larger particles under high-level clouds due to attenuation. The high cloud variation is strongly related to the vertical velocity variation in this region in NICAM (figures not shown). Although there are no global observational data for vertical velocity, we consider that the stronger high cloud fraction variation in NICAM than in the observations may occur because NICAM simulates stronger vertical velocity variation than the real atmosphere. The longitude–height distributions of the regression results are also shown in appendix A. A comparison of Figs. A1a and A2 shows that the difference of the vertical profiles of cloud fraction in GOCCP and NICAM results from the differing amplitude of cloud responses to the tropical mean SST in the Indian Ocean and the western Pacific Ocean. In GOCCP, the negative regressions in the Indian Ocean cancel the positive regressions in the western Pacific Ocean, whereas the negative regressions in the Indian Ocean are weaker than the positive regressions in the western Pacific Ocean in NICAM.
(a) Vertical distribution of cloud fraction (January 1989–December 2008) regressed onto the SST in NICAM-AMIP (aqua) and the GOCCP cloud fraction (January 2007–December 2015) regressed onto HadISST (orange). (b) Vertical distribution of ice hydrometeors regressed onto the SST in NICAM–AMIP, where yellow, light green, and blue denote ice water content from cloud ice, snow, and graupel, respectively. Note that the results here are for the tropics (30°N–30°S).
Citation: Journal of Climate 35, 11; 10.1175/JCLI-D-21-0150.1
(a) Vertical distribution of cloud fraction (January 1989–December 2008) regressed onto the SST in NICAM-AMIP (aqua) and the GOCCP cloud fraction (January 2007–December 2015) regressed onto HadISST (orange). (b) Vertical distribution of ice hydrometeors regressed onto the SST in NICAM–AMIP, where yellow, light green, and blue denote ice water content from cloud ice, snow, and graupel, respectively. Note that the results here are for the tropics (30°N–30°S).
Citation: Journal of Climate 35, 11; 10.1175/JCLI-D-21-0150.1
(a) Vertical distribution of cloud fraction (January 1989–December 2008) regressed onto the SST in NICAM-AMIP (aqua) and the GOCCP cloud fraction (January 2007–December 2015) regressed onto HadISST (orange). (b) Vertical distribution of ice hydrometeors regressed onto the SST in NICAM–AMIP, where yellow, light green, and blue denote ice water content from cloud ice, snow, and graupel, respectively. Note that the results here are for the tropics (30°N–30°S).
Citation: Journal of Climate 35, 11; 10.1175/JCLI-D-21-0150.1
Figure 4 shows composites of cloud fraction anomaly patterns for positive SST deviation with respect to the monthly climatological values for GOCCP and NICAM cloud fraction under the present climate. The left panel shows the composite patterns for positive SST anomaly based on GOCCP and HadISST (January 2007–December 2015), and the right panel shows the results obtained from NICAM (January 1989– December 2008). Note that the composite results for individual ice hydrometeors, such as cloud ice, snow, and graupel, are also shown in the vertical section in Fig. 4d. Figures 4c and 4f show the distributions of HadISST and NICAM SST anomalies, respectively. The positive SST patterns in both HadISST and NICAM show El Niño patterns, with the maxima of warming located off Peru and extending to the west Pacific. HadISST shows larger SST deviations ranging from Δ0.5 to Δ1.0 K, whereas those in NICAM only reach Δ0.5 K in the same region. Figures 4a and 4d show that the high cloud fraction responses to positive SST patterns in GOCCP and NICAM-AMIP show similar behavior, whereas the positive NICAM cloud fraction increase over 180°–60°W is stronger than that in GOCCP. In addition, the maximum cloud fraction increase in GOCCP appears at 15 km, whereas that in NICAM appears at 16.8 km. The relationship between cloud fraction and ice hydrometeors as SST increases can also be found in Fig. 4d, showing that the cloud ice, snow, and graupel mainly dominate the cloud fraction increase in the altitudes above 14 km, centered at 12 km, and below 11 km, respectively. Figures 4b and 4e show the horizontal distribution of cloud fraction responses to the SST deviation at altitude 16 km for GOCCP and NICAM-AMIP. The horizontal distribution of cloud fraction in NICAM is more extensive than that in GOCCP. Note that the pattern composited for negative SST deviation has similar results to the positive SST deviation case but with opposite sign in same order (figures not shown). Note that Saint-Lu et al. (2020) recently showed that the total high cloud area decreases on an interannual time scale by combining GOCCP data, HadCRUT4 SST data, and ERA5 reanalysis. In contrast, Avery et al. (2017) showed that the cloud ice at the tropopause increased during the 2015/16 El Niño event. Our results here are consistent with the result in Avery et al. (2017) because we composited the cloud fraction based on the positive deviation of SST, which is an El Niño–type response.
Composite results based on the positive SST deviation for the present climate, showing the (left) observational (GOCCP and HadISST; January 2007–December 2015) and (right) NICAM results (January 1989–December 2008). (a) Tropically (30°N–30°S) averaged GOCCP vertical cloud fraction structure, (b) horizontal GOCCP distribution at 15-km height, and (c) HadISST SST deviation. (d) Tropically (30°N–30°S) averaged NICAM vertical cloud fraction (shading), cloud ice (yellow contours), snow (green contours), and graupel (blue contours) structures, (e) horizontal NICAM cloud distribution at 16-km height, and (f) NICAM SST deviation.
Citation: Journal of Climate 35, 11; 10.1175/JCLI-D-21-0150.1
Composite results based on the positive SST deviation for the present climate, showing the (left) observational (GOCCP and HadISST; January 2007–December 2015) and (right) NICAM results (January 1989–December 2008). (a) Tropically (30°N–30°S) averaged GOCCP vertical cloud fraction structure, (b) horizontal GOCCP distribution at 15-km height, and (c) HadISST SST deviation. (d) Tropically (30°N–30°S) averaged NICAM vertical cloud fraction (shading), cloud ice (yellow contours), snow (green contours), and graupel (blue contours) structures, (e) horizontal NICAM cloud distribution at 16-km height, and (f) NICAM SST deviation.
Citation: Journal of Climate 35, 11; 10.1175/JCLI-D-21-0150.1
Composite results based on the positive SST deviation for the present climate, showing the (left) observational (GOCCP and HadISST; January 2007–December 2015) and (right) NICAM results (January 1989–December 2008). (a) Tropically (30°N–30°S) averaged GOCCP vertical cloud fraction structure, (b) horizontal GOCCP distribution at 15-km height, and (c) HadISST SST deviation. (d) Tropically (30°N–30°S) averaged NICAM vertical cloud fraction (shading), cloud ice (yellow contours), snow (green contours), and graupel (blue contours) structures, (e) horizontal NICAM cloud distribution at 16-km height, and (f) NICAM SST deviation.
Citation: Journal of Climate 35, 11; 10.1175/JCLI-D-21-0150.1
b. Response of high cloud to warming conditions
Figure 5a shows cloud fraction and vertical motion as tropical longitude–height sections under the warming climate (January 2078–December 2097) and Fig. 5b shows the difference between warming (January 2078–December 2097) and present (January 1989–December 2008) climate of the two variables. The relation between cloud fraction and vertical motion remains similar to that in the present climate (Fig. 2b). The location of the maxima of the high cloud fraction corresponds to the distribution of vertical motion, and the position of peak cloud fraction shifts upward under warming conditions. In detail, the cloud fraction (Fig. 5b, shading) generally increases (decreases) at altitudes above (below) 15 km, corresponding to an upward shift of clouds under warming conditions. The peaks of cloud fraction appear at 120°–160°W and 80°–60°E, the same longitudinal region as under the present climate, whereas the peaks are located at higher altitudes (16 km) than those (15 km) under the present climate. This pattern corresponds well to the change in vertical motion (shading) under the warming condition shown in Fig. 5b. The upward vertical motion extends to higher altitudes, shifts eastward, and the downward motion centered at 120°W weakens under the warming condition. The cell of upward motion located at longitude 120°W–160°E becomes stronger than under present climate (Fig. 2b). The edge of the upward motion in this longitudinal band reaches above ∼18 km under the warming conditions compared with ∼17 km under present climate. The width of this upward motion cell is broader than that under the present climate. The downward motion weakens under warming conditions, and in particular the width of the downward motion located at 120°–80°W narrows under warming conditions. The peak of the downward motion under warming conditions is not as prominent as that under the present climate. This change is thought to correspond to the vertical motion becoming stronger and shifting eastward in response to the El Niño–type SST warming pattern. The depth of cloud fraction becomes thicker than that under present climate at 130°W, and this region is associated with the weaker downward motions under warming conditions.
(a) Tropically (30°N–30°S) averaged cloud fraction (contours) and vertical motion (shading) from 6.5 to 20 km under the warming condition. (b) As in (a), but for the differences between warming and present climate. The analyzed duration for present and warming climate is January 1989–December 2008 and January 2078–December 2097, respectively.
Citation: Journal of Climate 35, 11; 10.1175/JCLI-D-21-0150.1
(a) Tropically (30°N–30°S) averaged cloud fraction (contours) and vertical motion (shading) from 6.5 to 20 km under the warming condition. (b) As in (a), but for the differences between warming and present climate. The analyzed duration for present and warming climate is January 1989–December 2008 and January 2078–December 2097, respectively.
Citation: Journal of Climate 35, 11; 10.1175/JCLI-D-21-0150.1
(a) Tropically (30°N–30°S) averaged cloud fraction (contours) and vertical motion (shading) from 6.5 to 20 km under the warming condition. (b) As in (a), but for the differences between warming and present climate. The analyzed duration for present and warming climate is January 1989–December 2008 and January 2078–December 2097, respectively.
Citation: Journal of Climate 35, 11; 10.1175/JCLI-D-21-0150.1
Using the same method as in Fig. 3 and Zelinka and Hartmann (2011), Fig. 6 shows the variation in cloud fraction with SST (Fig. 6a) and the variation of each ice hydrometeor (cloud ice, snow, and graupel) against SST (Fig. 6b) for the present and warming conditions. In Fig. 6a, the shape of the variation of cloud fraction against SST remains the same while the profile shifts upward, and the peak of the positive relation appears at an altitude around 18.5 km under warming conditions. This upward shifting profile corresponds to the position of upward shifts of high cloud under the warming condition shown in Fig. 5b. The magnitude of the peak of the regression between SST and cloud fraction becomes larger under the warming condition than that under the present climate. In other words, in addition to the increase in the basic state of cloud fraction under the warming condition (Fig. 5), the response of cloud fraction to SST is also amplified under the warming. Similar responses are also found for the relation between SST and each ice hydrometeor (cloud ice, snow, and graupel) shown in Fig. 6b. The profiles of cloud ice, snow, and graupel shift upward under the warming condition. Note that the amplitudes of SST annual variation in the present and warming climates have similar scale while convective activity becomes stronger in the central/eastern Pacific and cloud fraction doubles under the warming climate (figures not shown). We suggest that the doubling of cloud fraction variation in the warming climate may be strongly related to the difference in convective activity. More information on the longitude–height distributions of the regression results is also given in appendix A. Additional analysis results using a 5-yr SST+4K experiment are shown in appendix A to indicate how the prescribed SST pattern affects the pattern of variation of cloud fraction and ice hydrometeors.
As in Fig. 3a, but for NICAM-AMIP present (January 1989–December 2008) and warming (January 2078–December 2097) conditions. In (a), aqua and pink colored lines denote present and warming conditions, respectively. In (b), solid and dashed lines denote present and warming conditions, respectively. Error bars shows the standard deviation in the tropical region (30°S–30°N) for each variable.
Citation: Journal of Climate 35, 11; 10.1175/JCLI-D-21-0150.1
As in Fig. 3a, but for NICAM-AMIP present (January 1989–December 2008) and warming (January 2078–December 2097) conditions. In (a), aqua and pink colored lines denote present and warming conditions, respectively. In (b), solid and dashed lines denote present and warming conditions, respectively. Error bars shows the standard deviation in the tropical region (30°S–30°N) for each variable.
Citation: Journal of Climate 35, 11; 10.1175/JCLI-D-21-0150.1
As in Fig. 3a, but for NICAM-AMIP present (January 1989–December 2008) and warming (January 2078–December 2097) conditions. In (a), aqua and pink colored lines denote present and warming conditions, respectively. In (b), solid and dashed lines denote present and warming conditions, respectively. Error bars shows the standard deviation in the tropical region (30°S–30°N) for each variable.
Citation: Journal of Climate 35, 11; 10.1175/JCLI-D-21-0150.1
In Fig. 7, similar to Fig. 4, the left panel shows the composite cloud fraction pattern based on the positive SST monthly anomaly under the warming climate (January 2078–December 2097) and the right panel shows the difference between warming (January 2078–December 2097) and present (January 1989–December 2008) climates. Figures 7c and 7f show the distribution of positive SST anomalies under the warming climate and the difference between the warming and present climate, respectively. Similar to the results in the present climate, the warming pattern corresponds to the El Niño SST pattern, with the maximum of warming reaching 0.5 K, which is stronger than under the present climate (0.3 K; Fig. 4f). Figures 7a and 7d show the composite longitude–height sections under the warming climate and the difference between warming and present climate, respectively. Note that the composite result for each ice hydrometeor is also plotted in Fig. 7a. The position of the core of maximum increase in cloud fraction shifts eastward and upward under the warming condition, and a similar pattern shift is also found in ice hydrometeors. This pattern is similar to the main changes in vertical motion under the warming condition shown in Fig. 5. Figures 7b and 7e show the horizontal distribution of the composite results at 18 km. Similar to Figs. 7a and 7b, the cloud fraction coverage extends to the eastern Pacific and extratropics under the warming condition. This pattern corresponds closely to the SST deviation pattern. Note that the pattern composited by the negative SST deviation is similar to the pattern composited by the positive SST deviation but with opposite sign in same order (figures not shown).
As in Fig. 4, but under the NICAM (left) for warming condition (January 2079–December 2098) and (right) the difference between warming (January 2079–December 2098) and present (January 1989–December 2008) climate. Note that only the difference in cloud fraction is plotted in (d).
Citation: Journal of Climate 35, 11; 10.1175/JCLI-D-21-0150.1
As in Fig. 4, but under the NICAM (left) for warming condition (January 2079–December 2098) and (right) the difference between warming (January 2079–December 2098) and present (January 1989–December 2008) climate. Note that only the difference in cloud fraction is plotted in (d).
Citation: Journal of Climate 35, 11; 10.1175/JCLI-D-21-0150.1
As in Fig. 4, but under the NICAM (left) for warming condition (January 2079–December 2098) and (right) the difference between warming (January 2079–December 2098) and present (January 1989–December 2008) climate. Note that only the difference in cloud fraction is plotted in (d).
Citation: Journal of Climate 35, 11; 10.1175/JCLI-D-21-0150.1
The above analysis shows the high cloud variation behavior associated with SST variation in NICAM-AMIP under the warming condition. Generally, the behavior of high cloud fraction in NICAM-AMIP under the warming condition is similar to that under the present climate while the vertical distribution of high cloud is shifted upward and the responses to the SST are amplified under the warming condition. The pattern of the upward shift of high cloud is similar to the pattern of the response of vertical motion to the warming condition in the tropics (Fig. 5). In addition, the high cloud response to the SST becomes stronger under the warming condition (Figs. 6 and 7), whereas the SST monthly anomalies do not change dramatically under the warming condition.
c. Role of ice hydrometeors in the model
Figures 8a and 8b show the meridional distribution of temperature and cloud fraction, respectively. The contours show the conditions under the present climate and the shading shows the response to the warming climate. Consistent with the results of various GCM studies (e.g., Su et al. 2014), the temperature increase occurs in the troposphere and reaches 6 K at altitudes between 10 and 14 km in the tropics under the warming condition. Cloud fraction responses to the warming condition (Fig. 8b) in the tropics show that the cloud fraction mainly decreases at altitudes below 14 km in the tropics and increases at altitudes above 14 km in the tropics. This vertical shift is also consistent with results obtained from GCMs that simulate a decrease in cloud fraction in the troposphere but an increase in the tropopause under the RCP4.5 scenario (Su et al. 2014). The NICAM-AMIP experiment shows consistent results with GCMs in basic fields.
Meridional distribution of (a) temperature (K) and (b) cloud fraction. Contours denote the distributions for the present climate (January 1989–December 2008) and shading shows the responses to the warming condition (January 2079–December 2098).
Citation: Journal of Climate 35, 11; 10.1175/JCLI-D-21-0150.1
Meridional distribution of (a) temperature (K) and (b) cloud fraction. Contours denote the distributions for the present climate (January 1989–December 2008) and shading shows the responses to the warming condition (January 2079–December 2098).
Citation: Journal of Climate 35, 11; 10.1175/JCLI-D-21-0150.1
Meridional distribution of (a) temperature (K) and (b) cloud fraction. Contours denote the distributions for the present climate (January 1989–December 2008) and shading shows the responses to the warming condition (January 2079–December 2098).
Citation: Journal of Climate 35, 11; 10.1175/JCLI-D-21-0150.1
Figure 9 shows meridional sections of total IWC (the sum of IWC attributed to cloud ice, snow, and graupel) and vertical motion under the present (January 1989–December 2008; Fig. 9a) and warming (January 2078–December 2097; Fig. 9b) climate. In general, the distribution of IWC is consistent with that of vertical motion, similar to the results shown in Figs. 2b and 5a. Under the present climate (Fig. 9a), the IWC reaches an altitude ∼17 km in the latitude band 15°N–15°S, where upward motion dominates. Outside this latitude band, the altitude where IWC exists descends to 14 km while descending motion is predominant. Figure 9b shows that the core of upward motion in the latitude band 15°N–15°S becomes stronger and the IWC distribution reaches higher altitudes (∼18 km) under the warming condition. Similarly, the pattern of change in the vertical motion corresponds to the position where the IWC is shifted upward (Fig. 9c). In addition, the upward motion becomes stronger while its width decreases below 14 km while the region of stronger vertical motion extends to 30°N and 30°S above 14 km under the warming condition. The IWC increase in this region contributes the cloud fraction increase in the tropical tropopause shown in Fig. 8b. The similarity in vertical motion, IWC, and cloud fraction in the responses to the warming condition is seen in Figs. 9d, 9e, and 9f; they show vertical profiles of these fields in the upward motion region between 10°S and 10°N for the present and warming conditions. All of the profiles of vertical motion, IWC and cloud fraction shift upward under the warming condition. The vertical motion basically becomes stronger under warming. As a result, the profiles of IWC and cloud fraction are shifted upward under the warming condition and the IWC and cloud fraction increase above 14 and 16 km height, respectively.
Meridional distribution of vertical motion (shading; m s−1; upward positive) and ice water content (IWC; contours; mg m−3) for the (a) present (January 1989–December 2008) and (b) warming climate (January 2079–December 2098), and (c) their difference. (d)–(f) Vertical profiles in the central tropics (10°N–10°S) for the present and warming climate for vertical motion, IWC, and cloud fraction, respectively.
Citation: Journal of Climate 35, 11; 10.1175/JCLI-D-21-0150.1
Meridional distribution of vertical motion (shading; m s−1; upward positive) and ice water content (IWC; contours; mg m−3) for the (a) present (January 1989–December 2008) and (b) warming climate (January 2079–December 2098), and (c) their difference. (d)–(f) Vertical profiles in the central tropics (10°N–10°S) for the present and warming climate for vertical motion, IWC, and cloud fraction, respectively.
Citation: Journal of Climate 35, 11; 10.1175/JCLI-D-21-0150.1
Meridional distribution of vertical motion (shading; m s−1; upward positive) and ice water content (IWC; contours; mg m−3) for the (a) present (January 1989–December 2008) and (b) warming climate (January 2079–December 2098), and (c) their difference. (d)–(f) Vertical profiles in the central tropics (10°N–10°S) for the present and warming climate for vertical motion, IWC, and cloud fraction, respectively.
Citation: Journal of Climate 35, 11; 10.1175/JCLI-D-21-0150.1
Figure 10 illustrates how ice hydrometeors behave under the warming condition in terms of the IWC for cloud ice (Fig. 10a), snow (Fig. 10b), and graupel (Fig. 10c) in NICAM for the present climate (January 1989–December 2008; contours) and their responses to the warming climate (January 2078–December 2098; shading). First, the height structure of ice hydrometeors is revealed: the cloud ice, snow, and graupel dominate the altitudes centered at ∼14, ∼10, and 6 km, respectively. This layered structure of ice hydrometeors may be found in other studies using NICAM when the cloud microphysics scheme is replaced by a double-moment scheme (Y.-W. Chen et al. 2018). Under the warming climate, the altitudes for each type of IWC shift upward and the pattern of this upward shift is similar to the upward shift in the cloud fraction under the warming condition (Fig. 8b). Meanwhile, the prominent cloud fraction increases at the tropopause in the tropics shown in Fig. 8b are mainly due to the upward shift of IWC attributed to the cloud ice in this region.
Meridional distribution of ice water content (IWC) from (a) cloud ice, (b) snow, and (c) graupel. Contours denote the IWC distributions for the present climate and shading denotes the IWC responses to the warming climate. (d) Tropical vertical profiles of IWC from cloud ice (light blue), snow (red), and graupel (blue); crosses and circles denote the profiles for the present (January 1989–December 2008) and warming (January 2078–December 2097) climate, respectively.
Citation: Journal of Climate 35, 11; 10.1175/JCLI-D-21-0150.1
Meridional distribution of ice water content (IWC) from (a) cloud ice, (b) snow, and (c) graupel. Contours denote the IWC distributions for the present climate and shading denotes the IWC responses to the warming climate. (d) Tropical vertical profiles of IWC from cloud ice (light blue), snow (red), and graupel (blue); crosses and circles denote the profiles for the present (January 1989–December 2008) and warming (January 2078–December 2097) climate, respectively.
Citation: Journal of Climate 35, 11; 10.1175/JCLI-D-21-0150.1
Meridional distribution of ice water content (IWC) from (a) cloud ice, (b) snow, and (c) graupel. Contours denote the IWC distributions for the present climate and shading denotes the IWC responses to the warming climate. (d) Tropical vertical profiles of IWC from cloud ice (light blue), snow (red), and graupel (blue); crosses and circles denote the profiles for the present (January 1989–December 2008) and warming (January 2078–December 2097) climate, respectively.
Citation: Journal of Climate 35, 11; 10.1175/JCLI-D-21-0150.1
Figure 11 illustrates the tropical distributions of ice water path (IWP). The left panel shows the IWP attributed to all ice hydrometeors (Fig. 11a), cloud ice (Fig. 11b), snow (Fig. 11c), and graupel (Fig. 11d). The right panel shows the IWP difference between the warming (January 2078–December 2097) and the present climate attributed to all ice hydrometeors (Fig. 11e), cloud ice (Fig. 11f), snow (Fig. 11g), and graupel (Fig. 11h). Under the present climate, the horizontal distribution of IWP from all ice hydrometeors in NICAM–AMIP shows good agreement with satellite observations such as MODIS, CloudSat, and CERES (Waliser et al. 2009; Li et al. 2012). The maximum of IWP appears at the equator, is concentrated over the Maritime Continent, and a branch extends from the equator to 10°S along the ITCZ. According to the observational results in Waliser et al. (2009), the maximum IWP ranges from 100 to 300 g m−2, whereas the IWP in NICAM-AMIP reaches a maximum of about 200 g m−2 (in the region 80°–75°W, 3°–10°N) within the observational range. The horizontal distributions of the IWP for cloud ice (Fig. 11b), snow (Fig. 11c), and graupel (Fig. 11d) show similar patterns although the contribution of each component is different. Cloud ice, snow, and graupel all contribute to the IWP in the region from the Indian Ocean to the Maritime Continent, whereas the IWP from cloud ice and graupel dominates in the eastern Pacific region. Under the warming condition, the increase of IWP mainly occurs in the tropics, especially in the eastern Pacific, similar to the SST warming pattern shown in Fig. 1. The IWP decreases over South America and the Indian Ocean. In response to the warming condition, the IWP from cloud ice increases over the Pacific Ocean region and decreases in the Indian Ocean region. In the Pacific Ocean region, the IWP from cloud ice not only increases in the tropics but also extends to higher latitudes, particularly over the eastern Pacific Ocean region. This pattern is similar to the increase in high cloud coverage under the warming condition shown in Fig. 7b. The responses of snow and graupel (Figs. 11g,e) to warming have similar distribution to that of cloud ice, although the increase in extent of ice hydrometeors is most prominent in cloud ice. Note that IWC changes attributed to cloud ice increases under the warming climate shown in Figs. 10a and 11b are consistent with results in Dessler et al. (2016), who showed that the IWC at the tropopause increases under the warming climate.
Horizontal distribution of ice water path (IWP; g m−2) for (a) all ice hydrometeors, (b) cloud ice, (c) snow, and (d) graupel under the present climate (January 1989–December 2008). The IWP difference (g m−2) between the warming (January 2079–December 2098) and present climate for (e) all ice hydrometeors, (f) cloud ice, (g) snow, and (h) graupel.
Citation: Journal of Climate 35, 11; 10.1175/JCLI-D-21-0150.1
Horizontal distribution of ice water path (IWP; g m−2) for (a) all ice hydrometeors, (b) cloud ice, (c) snow, and (d) graupel under the present climate (January 1989–December 2008). The IWP difference (g m−2) between the warming (January 2079–December 2098) and present climate for (e) all ice hydrometeors, (f) cloud ice, (g) snow, and (h) graupel.
Citation: Journal of Climate 35, 11; 10.1175/JCLI-D-21-0150.1
Horizontal distribution of ice water path (IWP; g m−2) for (a) all ice hydrometeors, (b) cloud ice, (c) snow, and (d) graupel under the present climate (January 1989–December 2008). The IWP difference (g m−2) between the warming (January 2079–December 2098) and present climate for (e) all ice hydrometeors, (f) cloud ice, (g) snow, and (h) graupel.
Citation: Journal of Climate 35, 11; 10.1175/JCLI-D-21-0150.1
The global mean amounts of IWP are listed in Table 1 for both the present and warming conditions. The main contributor to the total IWP (31.69 g m−2) is graupel (14.17 g m−2), while cloud ice (8.81 g m−2) and snow (8.71 g m−2) make almost equal contributions. Under the warming condition, the total IWP decreases slightly (−0.26 g m−2) while the graupel increases and cloud ice and snow decrease. Note that the decrease of IWP attributed to snow is much greater than that due to cloud ice. The IWP distribution becomes more concentrated over the central/eastern Pacific under the warming climate. This pattern is similar to that for the vertical velocity distribution under the warming climate (figures not shown). It is considered that the convective activity becomes stronger and concentrated over the central/eastern Pacific, especially the eastern Pacific. Because the convection region becomes more concentrated over the eastern Pacific, the vertical motion in the tropical region shrinks under the warming climate (Fig. 9c). The collision process in clouds is the main mechanism for graupel formation, so shrinking of the region of vertical motion in the tropics under the warming climate may be linked to the reduced formation of graupel in convective activity in the tropics. The graupel increase in the global mean (figures not shown) is contributed by an increase at high latitudes, which implies that other synoptic process changes are linked to the increase of graupel. Because the active convection region shrinks under the warming climate, the supply of snow and cloud ice from the lower atmosphere decreases and this may explain why the IWP attributed to cloud ice and snow decreases under the warming climate.
Global mean of total IWP and IWP from cloud ice, snow, and graupel in NICAM–AMIP under the present and warming climate and the differences (units are g m−2).
4. Discussion and concluding remarks
In section 3, we have shown how the cloud fraction variation is related to the SST variation under the present climate, how the cloud fraction variation is related to SST changes under the prescribed El Niño–like SST warming pattern, and how these changes may be related to the ice hydrometeors by using NICAM AMIP-type simulation pairs for the present and warming climate. Although the link between high cloud variation and ice hydrometeors has been revealed in the current analysis, the physical process behind these behaviors has not. Previous studies using NICAM with various experimental designs, which focused on the ice cloud responses to the warming climate, may provide some hints. Noda et al. (2010) conducted a sensitivity study on how the turbulence scheme modulated the low- and high-cloud behavior and found that the turbulent transport process modulated by the subgrid scale clouds affects low-, middle-, and high-level clouds as well. They reported that, although the high-cloud fraction in NICAM is related to the boundary scheme, the large high cloud fraction remains in NICAM when the mixing effect in the boundary layer scheme is changed. Analysis of the CMIP3 and CMIP5 models has revealed that the bias in low clouds in the present climate may be related to the climate sensitivity simulated in the models (Fasullo and Trenberth 2012; Sherwood et al. 2014). Iga et al. (2011) and Chen et al. (2016), using 3.5- and 14-km horizontal resolution NICAM experiments, respectively, with short simulation periods (within several months), showed that the high cloud increase under the warming climate is strongly tied to the change of cloud ice under the warming climate. Furthermore, Noda et al. (2019) found the importance of the changes of large-scale atmospheric circulation for high cloud change under the warming climate. In contrast, Chen et al. (2016) showed that although the IWP decreases under the warming climate, the PDF of occurrence of IWP less than 10 g m−2 increases under the warming climate, which implies that the thin cloud is linked to the cloud ice under the warming climate. Furthermore, Ohno and Satoh (2018), using RCE (radiative–convective equilibrium) settings, showed that the increase of high clouds in response to the increased SST in NICAM may be linked to the condition of the tropopause. They showed that the increase of the high clouds to the increase SST in NICAM can be linked to the condition of the tropopause. When the tropopause is wet, the radiative driven upward moisture transport below the wet tropopause layer increases and the upward transportation (supply) of the ice condensate to the lower layer through the sedimentation process, which cannot be observed by observations. With a wet tropopause, the radiatively driven upward moisture transport below the wet tropopause layer increases, causing more ice condensation that then supplies ice condensate to the lower layer through the sedimentation process, which cannot be seen in the observations. Although we do not conduct the same analysis as Ohno and Satoh (2018), the process they proposed may occur in the NICAM–AMIP experiment used in this study because Noda et al. (2019) showed that the relative humidity increases in the tropopause under the warming climate in the NICAM–AMIP experiment. So far, with short-period and long-term simulations of the real atmosphere, NICAM results have shown that the high cloud increases under the warming climate, which is different from conventional GCMs. Previous studies noted here showed that the high cloud change under the warming climate is strongly related to the interaction between dynamical processes and cloud microphysics, and needs to be studied from various aspects. The ice condensate transportation process proposed by Ohno and Satoh (2018) may provide a hint to help understand why the high cloud responses to the SST in NICAM are different from those in GCMs. The difference between NICAM and GCM simulations will be addressed in future work.
This study reveals how tropical high cloud responds to SST changes under the present and warming conditions in a global cloud system–resolving model, NICAM, and how these behaviors are related to ice hydrometeors, such as cloud ice, snow, and graupel. The main results are highlighted as follows.
-
Under the present climate: The variation in tropical high cloud with SST shows similar behaviors in the model and observational data, although the magnitude of the high cloud variation appears stronger in the model than in observational results. Furthermore, the longitudinal distribution of high cloud is strongly related to the vertical motion in models.
-
Under the warming climate: The variation in tropical high cloud with SST shows similar behavior under the warming condition, while the vertical profile shifts upward because the altitudes at which high clouds exist shift upward under the warming climate. The amplitude of the variation of high cloud under the warming condition is larger than that under the present climate, about double. The composite analyses showed that the magnitude of the cloud fraction increase when the positive SST deviation is larger than that of the cloud fraction decrease with the negative SST deviation. The different responses to the positive and negative SST patterns lead to the high cloud fraction increase under the warming condition.
-
Properties of ice hydrometeors: The total IWP amount decreases under the warming climate. This result is consistent with other short-term simulation results obtained from NICAM (e.g., Satoh et al. 2012; Chen et al. 2016). This study further indicated that the different ice hydrometeors behave differently under the warming climate: The amount of cloud ice and snow decrease under the warming condition while the amount of graupel increases. The cloud fraction at the top of the tropopause is very sensitive to the distribution of cloud ice. The coverage of cloud ice under the warming condition extends in the east of the Pacific Ocean and the increase is strongly linked to the greater cloud fraction increase in ENSO events under the warming condition. Some of the details of the SST forcing pattern used in the current warming experiment are as follows. The CMIP3 multimodel-mean (MMM) projected SST is used to drive the NICAM-AMIP warming climate (mentioned in section 2a and shown in Fig. 2a). It is well known that the ENSO signal in the CMIP3 MMM projected SST is stronger than that used in CMIP5 simulations. The warm bias in CMIP3 MMM SST causes a radiative bias that reaches 30 W m−2 in the tropics, especially over the ITCZ (intertropical convergence zone), SPCZ (South Pacific convergence zone), and warm pool regions (e.g., Li et al. 2015; C.-A. Chen et al. 2018). Similar issues are thought to occur in the current NICAM–AMIP simulation result and may affect the interactions in physical and dynamic processes through the hydrometeor–radiation–circulation coupling over the trade winds, ITCZ/SPCZ, and warm pool in NICAM–AMIP. Under the present climate, NICAM–AMIP results show that the IWC near the tropopause increases corresponding to the El Niño pattern warming, which is consistent with observational results for El Niño events (Su and Jiang 2013; Avery et al. 2017). The IWC increase at the tropopause under the warming climate is considered to be due to the strong prescribed SST pattern. Because the ice hydrometeor interannual variation is very sensitive to the prescribed SST pattern (Fig. A1), the way in which the prescribed SST pattern affects the physical and dynamical processes should be studied in detail in the future.
Through this work, we have shown how the temperature, cloud fraction, and ice hydrometeors respond under the warming climate in NICAM–AMIP and that the high cloud behavior is closely linked to the behavior of ice hydrometeors, especially cloud ice. The wider horizontal spread of ice hydrometeors, especially cloud ice, in the eastern Pacific is also considered as one reason for the high cloud fraction increase under the warming condition (Fig. 11) and this pattern seems closely related to the increase in high cloud coverage under the warming climate (Fig. 7b). The behavior of ice hydrometeors under the warming condition and their link to the high cloud fraction responses to the global warming addressed by a long-term cloud system–resolving model are shown for the first time in this study. Here is one notation for the relation between cloud fraction and IWC. The high cloud fraction shows an increase and an upward shift under the warming climate (Figs. 6a and 9f), and this change is closely related to the vertical profile of IWC attributed to vertically upward shifts of cloud ice under the warming climate (Figs. 6b, 9e, and 11f). In contrast, the vertically integrated IWC attributed to cloud ice (IWP attributed to cloud ice; Fig. 10f and Table 1) decreases under the warming climate. This inconsistency implies that the one-to-one relation between high cloud and cloud ice may be modified by some unresolved process and should be addressed in future work. Furthermore, the result that the simulated high cloud becomes thinner and higher under the warming climate may constitute a climate feedback that is different from the CMIP mean (e.g., Chen et al. 2016; Zelinka et al. 2016). Note also that Gasparini et al. (2021) showed that the optical depths of anvil clouds become thicker under the warming climate, using high-resolution simulation data (0.25° resolution) from the Exascale Earth System Model (E3SM), which is the opposite result from this study. The causes of these inconsistencies should be studied in the future.
Cloud-resolving models are expected to reduce the uncertainty of cloud responses to global warming (Zelinka et al. 2017) although the discrepancy in different model simulations remains. This work revealed the projected variations of high clouds related to warming conditions in a high-resolution global cloud system–resolving model. This study has shown that the basic state of the ice clouds under the present and warming condition is strongly related to the distribution of vertical motion. In response to the warming, both the vertical motion and ice clouds shift upward and eastward, which is thought to be linked to changes in the location of vertical motion associated with the El Niño–type SST warming pattern under the warming condition. With increases in computational power, climate simulations are shifting toward higher resolution and new model intercomparisons such as HighResMIP (Haarsma et al. 2016) under the framework of CMIP6 and the DYAMOND project (Stevens et al. 2019; Satoh et al. 2019) have been launched. Ahead of these projects, this study indicates that the high cloud responses to the warming condition may change as the resolution becomes finer and the microphysics of cloud processes are considered in a more realistic way in model simulations. This study also suggests the necessity for simulating the ice hydrometeors separately because the high clouds, especially those at the tropopause, seem strongly linked to cloud ice. Chen et al. (2016) and Y.-W. Chen et al. (2018) using NICAM’s short-term simulations showed that the ice hydrometeors in models directly affect how the longwave cloud radiative forcing (LWCRF) and LW radiation are simulated in the model. Thus, more detailed investigation of how the LWCRF and LW radiation are associated with ice cloud and ice hydrometeors, and how the radiative circulation behaviors change in long-term simulation is required in the future.
Acknowledgments.
We gratefully thank two reviewers and Dr. Blaž Gasparini for their constructive comments. We also thank the NICAM development team for developing NICAM. This work was supported by the Integrated Research Program for Advancing Climate Models (TOUGOU), and the Program for Risk Information on Climate Change (SOUSEI) supported by the Ministry of Education, Culture, Sports, Science, and Technology, Japan (MEXT), and JSPS KAKENHI Grant-in-Aid for Scientific Research on Innovative Areas (B) Grants 20H05728 and 20H05730. The NICAM-AMIP simulation is performed on the K computer at the RIKEN Advanced Institute for Computational Science (Proposals hp120279, hp130010, and hp140219). All the figures are plotted with the Grid Analysis and Display System (GrADS) software.
Data availability statement.
The GOCCP cloud fraction data are available at http://climserv.ipsl.polytechnique.fr/cfmip-obs/Calipso_goccp.html, and the HadISST data are available at https://www.metoffice.gov.uk/hadobs/hadisst/. The CloudSat 2C-ICE product is available at https://www.cloudsat.cira.colostate.edu/data-products/2c-ice. The whole set of NICAM–AMIP data is archived on JAMSTEC’s private server and available upon request by contacting the corresponding author.
APPENDIX A
High Cloud Response to SST
The tropical-mean longitude–height distributions are shown in Fig. A1 for high cloud fraction (shading), and IWC attributed to cloud ice (black contours), snow (blue contours), and graupel (yellow contours) regressed onto tropical SST for the present climate, warming climate, and a 5-yr SST+4K experiment. A comparison of the present climate and Niño-like warming pattern shows that the cloud fraction variation pattern remains similar under the Niño-like warming pattern while the relation strengthens in a warming climate. Results obtained from the SST+4K experiment show that the pattern responses to the SST change. The positive regression region for each variable is more concentrated over the eastern Pacific than for the Niño warming pattern, while the negative regression region shifts from the Indian Ocean to the eastern Pacific. Although the high cloud variation in the upper troposphere remains closely related to the cloud ice, the contribution of snow becomes more important than in the Niño-like warming pattern. These results showed that the detailed cloud fraction variation is very sensitive to the SST pattern. Although we did not directly analyze how the convective activity changes under the different SST warming patterns, the behaviors of ice hydrometeors changed dramatically, which indicates that the convective activity in the tropics also changes. For reference, Fig. A2 shows the tropical-mean longitude–height distributions for GOCCP cloud fraction regressed onto HadISST.
Longitude–height sections from NICAM-AMIP experiments for tropical averages of variables regressed onto the tropical mean SST for (a) the present climate (1989–2008), (b) the warming climate (2078–97), and (c) the 5-yr SST+4K experiment (1979–83). Shading denotes the regression results for cloud fraction, and black, blue, and yellow contours (solid lines: positive; dashed lines: negative) denote the results for cloud ice, snow, and graupel, respectively.
Citation: Journal of Climate 35, 11; 10.1175/JCLI-D-21-0150.1
Longitude–height sections from NICAM-AMIP experiments for tropical averages of variables regressed onto the tropical mean SST for (a) the present climate (1989–2008), (b) the warming climate (2078–97), and (c) the 5-yr SST+4K experiment (1979–83). Shading denotes the regression results for cloud fraction, and black, blue, and yellow contours (solid lines: positive; dashed lines: negative) denote the results for cloud ice, snow, and graupel, respectively.
Citation: Journal of Climate 35, 11; 10.1175/JCLI-D-21-0150.1
Longitude–height sections from NICAM-AMIP experiments for tropical averages of variables regressed onto the tropical mean SST for (a) the present climate (1989–2008), (b) the warming climate (2078–97), and (c) the 5-yr SST+4K experiment (1979–83). Shading denotes the regression results for cloud fraction, and black, blue, and yellow contours (solid lines: positive; dashed lines: negative) denote the results for cloud ice, snow, and graupel, respectively.
Citation: Journal of Climate 35, 11; 10.1175/JCLI-D-21-0150.1
Longitude–height section as in Fig. A1, but for GOCCP cloud fraction (2007–15) regressed onto HadISST.
Citation: Journal of Climate 35, 11; 10.1175/JCLI-D-21-0150.1
Longitude–height section as in Fig. A1, but for GOCCP cloud fraction (2007–15) regressed onto HadISST.
Citation: Journal of Climate 35, 11; 10.1175/JCLI-D-21-0150.1
Longitude–height section as in Fig. A1, but for GOCCP cloud fraction (2007–15) regressed onto HadISST.
Citation: Journal of Climate 35, 11; 10.1175/JCLI-D-21-0150.1
APPENDIX B
Ice Hydrometer Properties in NICAM-AMIP and CloudSat
We compare the IWP and IWC distribution in observations (CloudSat 2C-ICE product; Deng et al. 2015) and in the 20-yr NICAM–AMIP under the present climate. The IWP distributions between 60°N and 60°S for CloudSat 2C-ICE and NICAM–AMIP are plotted in Figs B1a and B1b. The IWP in NICAM–AMIP has a similar horizontal distribution to IWP in CloudSat 2C-ICE, although the total amount of IWP in NICAM is much less than in CloudSat 2C-ICE, especially in South America, the central and western Pacific, and the region along the storm track. This bias can also be found in other model products such as the ECMWF reanalysis data ERA-Interim (Li et al. 2021). A comparison of the meridional section of IWC for CloudSat 2C-ICE and NICAM-AMIP is plotted in Figs B1c and B1d. In the top layer of IWC, NICAM–AMIP shows good agreement with CloudSat 2C-ICE, whereas NICAM simulates less IWC in the middle and bottom layer at all latitudes. Focusing on the tropical region, the vertical profiles averaged over 30°N–30°S for CloudSat 2C-ICE and NICAM–AMIP are plotted in Figs B1e and B1f, respectively. Note that in Fig. B1f, the profiles are also divided into IWC attributed to cloud ice, snow, and graupel. In Figs B1c and B1d, it is seen that NICAM does not simulate enough large ice particles such as snow and graupel. In general, NICAM produces less IWP than the CloudSat 2C-ICE product (Figs. B1a,b). This may result from NICAM producing less convective ice hydrometeors such as graupel and snow in the cloud bottom (Figs. B1c–f). In contrast, NICAM may produce more cloud ice at the top of the cloud (Figs. B1c,d).
General comparison of ice water content (IWC; mg m−3) retrieved in CloudSat 2C-ICE product and IWC in the NICAM–AMIP present climate. Horizontal distributions for the (a) CloudSat 2C-ICE and (b) NICAM–AMIP present climate. Latitude–height sections for the (c) CloudSat 2C-ICE and (d) NICAM–AMIP present climate. Vertical profiles in the tropical belt (30°S–30°N) for the (e) CloudSat 2C-ICE and (f) NICAM–AMIP present climate. Note that the data period for CloudSat 2C-ICE is 2007–10 and that for NICAM–AMIP is 1989–2008.
Citation: Journal of Climate 35, 11; 10.1175/JCLI-D-21-0150.1
General comparison of ice water content (IWC; mg m−3) retrieved in CloudSat 2C-ICE product and IWC in the NICAM–AMIP present climate. Horizontal distributions for the (a) CloudSat 2C-ICE and (b) NICAM–AMIP present climate. Latitude–height sections for the (c) CloudSat 2C-ICE and (d) NICAM–AMIP present climate. Vertical profiles in the tropical belt (30°S–30°N) for the (e) CloudSat 2C-ICE and (f) NICAM–AMIP present climate. Note that the data period for CloudSat 2C-ICE is 2007–10 and that for NICAM–AMIP is 1989–2008.
Citation: Journal of Climate 35, 11; 10.1175/JCLI-D-21-0150.1
General comparison of ice water content (IWC; mg m−3) retrieved in CloudSat 2C-ICE product and IWC in the NICAM–AMIP present climate. Horizontal distributions for the (a) CloudSat 2C-ICE and (b) NICAM–AMIP present climate. Latitude–height sections for the (c) CloudSat 2C-ICE and (d) NICAM–AMIP present climate. Vertical profiles in the tropical belt (30°S–30°N) for the (e) CloudSat 2C-ICE and (f) NICAM–AMIP present climate. Note that the data period for CloudSat 2C-ICE is 2007–10 and that for NICAM–AMIP is 1989–2008.
Citation: Journal of Climate 35, 11; 10.1175/JCLI-D-21-0150.1
REFERENCES
Aumann, H. H., and Coauthors, 2003: AIRS/AMSU/HSB on the Aqua mission: Design, science objectives, data products and procession systems. IEEE Trans. Geosci. Remote Sens., 41, 253–264, https://doi.org/10.1109/TGRS.2002.808356.
Avery, M. A., S. M. Davis, K. H. Rosenlof, Y. Hao, and A. E. Dessler, 2017: Large anomalies in lower stratospheric water vapour and ice during the 2015–2016 El Niño. Nat. Geosci., 10, 405–409, https://doi.org/10.1038/ngeo2961.
Bony, S., and Coauthors, 2016: Clouds, circulation and climate sensitivity. Nat. Geosci., 8, 261–268, https://doi.org/10.1038/ngeo2398.
Brohan, P., J. J. Kennedy, I. Harris, S. F. B. Tett, and P. D. Jones, 2006: Uncertainty estimates in regional and global observed temperature changes: A new data set from 1850. J. Geophys. Res., 111, D12106, https://doi.org/10.1029/2005JD006548.
Cesana, G., and Coauthors, 2016: Using in situ airborne measurements to evaluate three cloud phase products derived from CALIPSO. J. Geophys. Res. Atmos., 121, 5788–5808, https://doi.org/10.1002/2015JD024334.
Chen, C.-A., and Coauthors, 2018: Falling snow radiative effects enhance the global warming response of the tropical Pacific atmosphere. J. Geophys. Res. Atmos., 123, 10 109-10 124, https://doi.org/10.1029/2018JD028655.
Chen, Y.-W., T. Seiki, C. Kodama, M. Satoh, A. T. Noda, and Y. Yamada, 2016: High cloud responses to global warming simulated by two different cloud microphysics schemes implemented in the Nonhydrostatic Icosahedral Atmospheric Model (NICAM). J. Climate, 29, 5949–5964, https://doi.org/10.1175/JCLI-D-15-0668.1.
Chen, Y.-W., T. Seiki, C. Kodama, M. Satoh, and A. T. Noda, 2018: Impact of precipitating ice hydrometeors on longwave radiative effect estimated by a global cloud-system resolving model. J. Adv. Model. Earth Syst., 10, 284–296, https://doi.org/10.1002/2017MS001180.
Chepfer, H., S. Bony, D. Winker, G. Cesana, J. L. Dufresne, P. Minnis, C. J. Stubenrauch, and S. Zeng, 2010: The GCM-oriented CALIPSO cloud product (CALIPSO-GOCCP). J. Geophys. Res., 115, D00H16, https://doi.org/10.1029/2009JD012251.
Deng, M. G., G. G. Mace, Z. Wang, and E. Berry, 2015: CloudSat 2C-ICE product update with a new Ze parameterization in lidar-only region. J. Geophys. Res. Atmos., 120, 12 198–12 208, https://doi.org/10.1002/2015JD023600.
Dessler, A. E., and Coauthors, 2016: Transport of ice into the stratosphere and the humidification of the stratosphere over the 21st century. Geophys. Res. Lett., 43, 2323–2329, https://doi.org/10.1002/2016GL067991.
Dufresne, J.-L., and S. Bony, 2008: An assessment of the primary sources of spread of global warming estimates from coupled atmosphere–ocean models. J. Climate, 21, 5135–5144, https://doi.org/10.1175/2008JCLI2239.1.
Fasullo, J. T., and K. E. Trenberth, 2012: A less cloudy future: The role of subtropical subsidence in climate sensitivity. Science, 338, 792–794, https://doi.org/10.1126/science.1227465.
Gasparini, B., P. J. Rasch, D. L. Hartmann, C. J. Wall, and M. Dütsch, 2021: A Lagrangian perspective on tropical anvil cloud lifecycle in present and future climate. J. Geophys. Res. Atmos., 126, e2020JD033487, https://doi.org/10.1029/2020JD033487.
Gettelman, A., and H. Morrison, 2015: Advanced two-moment bulk microphysics for global models. Part I: Off-line tests and comparison with other schemes. J. Climate, 28, 1268–1287, https://doi.org/10.1175/JCLI-D-14-00102.1.
Gettelman, A., and Coauthors, 2010: Global simulations of ice nucleation and ice supersaturation with an improved cloud scheme in the Community Atmosphere Model. J. Geophys. Res., 115, D18216, https://doi.org/10.1029/2009JD013797.
Haarsma, R. J., and Coauthors, 2016: High Resolution Model Intercomparison Project (HighResMIP v1.0) for CMIP6. Geosci. Model Dev., 9, 4185–4208, https://doi.org/10.5194/gmd-9-4185-2016.
Hartmann, D. L., M. E. Ockert-Bell, and M. L. Michelsen, 1992: The effect of cloud type on Earth’s energy balance: Global analysis. J. Climate, 5, 1281–1304, https://doi.org/10.1175/1520-0442(1992)005<1281:TEOCTO>2.0.CO;2.
Hashino, T., M. Satoh, Y. Hagihara, T. Kubota, T. Matsui, T. Nasuno, and H. Okamoto, 2013: Evaluating cloud microphysics from NICAM against CloudSat and CALIPSO. J. Geophys. Res. Atmos., 118, 7273–7292, https://doi.org/10.1002/jgrd.50564.
Iga, S., H. Tomita, Y. Tsushima, and M. Satoh, 2011: Sensitivity of Hadley circulation of physical parameters and resolution through changing upper-tropospheric ice clouds using a global cloud-system resolving model. J. Climate, 24, 2666–2679, https://doi.org/10.1175/2010JCLI3472.1.
IPCC, 2000: Special Report on Emissions Scenarios. N. Nakicenovic et al., Eds., Cambridge University Press, 570 pp.
Kodama, C., A. T. Noda, and M. Satoh, 2012: An assessment of the cloud signals simulated by NICAM using ISCCP, CALIPSO, and CloudSat satellite simulators. J. Geophys. Res., 117, D12210, https://doi.org/10.1029/2011JD017317.
Kodama, C., and Coauthors, 2015: A 20-year climatology of a NICAM AMIP-type simulation. J. Meteor. Soc. Japan, 93, 393–424, https://doi.org/10.2151/jmsj.2015-024.
Koren, I., L. A. Remer, O. Altaratz, J. V. Martins, and A. Davidi, 2010: Aerosol-induced changes of convective cloud anvils produce strong climate warming. Atmos. Chem. Phys., 10, 5001–5010, https://doi.org/10.5194/acp-10-5001-2010.
Li, J.-L. F., and Coauthors, 2012: An observationally based evaluation of cloud ice water in CMIP3 and CMIP5 GCMs and contemporary reanalyses using contemporary satellite data. J. Geophys. Res., 117, D16105, https://doi.org/10.1029/2012JD017640.
Li, J.-L. F., D. E. Waliser, G. Stephens, S. Lee, T. L’Ecuyer, S. Kato, N. Loeb, and H.-Y. Ma, 2013: Characterizing and understanding radiation budget biases in CMIP3/CMIP5 GCMs, contemporary GCM, and reanalysis. J. Geophys. Res. Atmos., 118, 8166–8184, https://doi.org/10.1002/jgrd.50378.
Li, J.-L. F., W.-L. Lee, T. Lee, E. Fetzer, J.-Y. Yu, T. L. Kubar, and C. Boening, 2015: The impacts of cloud snow radiative effects on Pacific Ocean surface heat fluxes, surface wind stress, and ocean temperatures in coupled GCM simulations. J. Geophys. Res. Atmos., 120, 2242–2260, https://doi.org/10.1002/2014JD022538.
Li, J.-L. F., and Coauthors, 2016: Considering the radiative effects of snow on tropical Pacific Ocean radiative heating profiles in contemporary GCMs using A‐Train observations. J. Geophys. Res. Atmos., 121, 1621–1636, https://doi.org/10.1002/2015JD023587.
Li, J.-L. F., and Coauthors, 2020: An overview of CMIP5 and CMIP6 simulated cloud ice, radiation fields, surface wind stress, sea surface temperatures and precipitation over tropical and subtropical oceans. J. Geophys. Res. Atmos., 125, e2020JD032848, https://doi.org/10.1029/2020JD032848.
Li, J.-L. F., and Coauthors, 2021: Improved ice content, radiation, precipitation and low-level circulation over the tropical Pacific from ECMWF ERA-Interim to ERA5. Environ. Res. Commun., 3, 081006, https://doi.org/10.1088/2515-7620/ac1bfe.
Liu, R., K.-N. Liou, H. Su, Y. Gu, B. Zhao, J. H. Jiang, and S. C. Liu, 2017: High cloud variations with surface temperature from 2002 to 2015: Contributions to atmospheric radiative cooling rate and precipitation changes. J. Geophys. Res. Atmos., 122, 5457–5471, https://doi.org/10.1002/2016JD026303.
Michibata, T., K. Suzuki, M. Sekiguchi, and T. Takemura, 2019: Prognostic precipitation in the MIROC6‐SPRINTARS GCM: Description and evaluation against satellite observations. J. Adv. Model. Earth Syst., 11, 839–860, https://doi.org/10.1029/2018MS001596.
Mizuta, R., Y. Adachi, S. Yukimoto, and S. Kusunoki, 2008: Estimation of the future distribution of sea surface temperature and sea ice using the CMIP3 multi-model ensemble mean. MRI Tech. Rep. 56, 28 pp., https://doi.org/10.1143/mritechrepo.56.
Noda, A. T., and M. Satoh, 2014: Responses of tropical and subtropical high-cloud statistics to global warming. J. Climate, 27, 7753–7768, https://doi.org/10.1175/JCLI-D-14-00179.1.
Noda, A. T., K. Oouchi, M. Satoh, H. Tomita, S. Iga, and Y. Tsushima, 2010: Importance of the subgrid-scale turbulent moist process: Cloud distribution in global cloud-resolving simulations. Atmos. Res., 96, 208–217, https://doi.org/10.1016/j.atmosres.2009.05.007.
Noda, A. T., K. Oouchi, M. Satoh, and H. Tomita, 2012: Quantitative assessment of diurnal variation of tropical convection simulated by a global nonhydrostatic model without cumulus parameterization. J. Climate, 25, 5119–5134, https://doi.org/10.1175/JCLI-D-11-00295.1.
Noda, A. T., C. Kodama, Y. Yamada, M. Satoh, T. Ogura, and T. Ohno, 2019: Responses of clouds and large-scale circulation to global warming evaluated from multidecadal simulations using a global nonhydrostatic model. J. Adv. Model. Earth Syst., 11, 2980–2995, https://doi.org/10.1029/2019MS001658.
Ohno, T., and M. Satoh, 2018: Doles of cloud microphysics on cloud responses to sea surface temperatures in radiative-convective equilibrium experiments using a high-resolution global nonhydrostatic model. J. Adv. Model. Earth Syst., 10, 1970–1989, https://doi.org/10.1029/2018MS001386.
Platnick, S., M. D. King, S. A. Ackerman, W. P. Menzel, B. A. Baum, J. C. Riedi, and R. A. Frey, 2003: The MODIS cloud products: Algorithms and examples from Terra. IEEE Trans. Geosci. Remote Sens., 41, 459–473, https://doi.org/10.1109/TGRS.2002.808301.
Rayner, N. A., D.E. Parker, E. B. Horton, C. K. Folland, L. V. Alexander, D. K. Rowell, E. C. Kent, A. Kaplan, 2003: Global analyses of sea surface temperature, sea ice, and night marine air temperature since the late nineteenth century. J. Geophys. Res., 108, 4407, https://doi.org/10.1029/2002JD002670.
Roh, W., M. Satoh, and C. Hohengger, 2021: Intercomparison of cloud properties in DYAMOND simulations over the Atlantic Ocean. J. Meteor. Soc. Japan, 99, 1439–1451, https://doi.org/10.2151/jmsj.2021-070.
Rossow, W. B., and R. A. Schiffer, 1999: Advances in understanding clouds from ISCCP. Bull. Amer. Meteor. Soc., 80, 2261–2287, https://doi.org/10.1175/1520-0477(1999)080<2261:AIUCFI>2.0.CO;2.
Saint-Lu, M., S. Bony, and J.-L. Dufresne, 2020: Observational evidence for a stability iris effect in the tropics. Geophys. Res. Lett., 47, e2020GL089059, https://doi.org/10.1029/2020GL089059.
Satoh, M., T. Matsuno, H. Tomita, H. Miura, T. Nasuno, and S. Iga, 2008: Nonhydrostatic Icosahedral Atmospheric Model (NICAM) for global cloud resolving simulations. J. Comput. Phys., 227, 3486–3514, https://doi.org/10.1016/j.jcp.2007.02.006.
Satoh, M., T. Inoue, and H. Miura, 2010: Evaluations of cloud properties of global and local cloud system resolving models using CALIPSO and CloudSat simulators. J. Geophys. Res., 115, D00H14, https://doi.org/10.1029/2009JD012247.
Satoh, M., S. Iga, H. Tomita, Y. Tsushima, and A. T. Noda, 2012: Response of upper clouds in global warming experiments obtained using a global nonhydrostatic model with explicit cloud processes. J. Climate, 25, 2178–2191, https://doi.org/10.1175/JCLI-D-11-00152.1.
Satoh, M., and Coauthors, 2014: The Non-Hydrostatic Icosahedral Atmospheric Model: Description and development. Prog. Earth Planet. Sci., 1, 18, https://doi.org/10.1186/s40645-014-0018-1.
Satoh, M., Y. Yamada, M. Sugi, C. Kodama, and A. T. Noda, 2015: Constraint on future change in global frequency of tropical cyclones due to global warming. J. Meteor. Soc. Japan, 93, 489–500, https://doi.org/10.2151/jmsj.2015-025.
Satoh, M., B. Stevens, F. Judt, M. Khairoutdinov, S.-J. Lin, W. M. Putman, and P. Düben, 2019: Global cloud-resolving models. Curr. Climate Change Rep., 5, 172–184, https://doi.org/10.1007/s40641-019-00131-0.
Seiki, T., C. Kodama, A. T. Noda, and M. Satoh, 2015a: Improvements in global cloud-system-resolving simulations by using a double-moment bulk cloud microphysics scheme. J. Climate, 28, 2405–2419, https://doi.org/10.1175/JCLI-D-14-00241.1.
Sherwood, S. C., S. Bony, and J.-L. Dufresne, 2014: Spread in model climate sensitivity traced to atmospheric convective mixing. Nature, 505, 37–42, https://doi.org/10.1038/nature12829.
Soden, B. J., and I. M. Held, 2006: An assessment of climate feedbacks in coupled ocean–atmosphere models. J. Climate, 19, 3354–3360, https://doi.org/10.1175/JCLI3799.1.
Stephens, G. L., and Coauthors, 2002: The CloudSat mission and the A-Train: A new dimension of space-based observations of clouds and precipitation. Bull. Amer. Meteor. Soc., 83, 1771–1790, https://doi.org/10.1175/BAMS-83-12-1771.
Stevens, B., and Coauthors, 2019: DYAMOND: The Dynamics of the Atmospheric General Circulation Modeled on Non-Hydrostatic Domains. Prog. Earth Planet. Sci., 6, 61, https://doi.org/10.1186/s40645-019-0304-z.
Su, H., and J. H. Jiang, 2013: Tropical clouds and circulation changes during the 2006/07 and 2009/10 El Niños. J. Climate, 26, 399–413, https://doi.org/10.1175/JCLI-D-12-00152.1.
Su, H., J. H. Jiang, C. Zhai, T. J. Shen, J. D. Meelin, G. L. Stephens, and Y. L. Yung, 2014: Weakening and strengthening structures in the Hadley circulation change under global warming and implications for cloud responses and climate sensitivity. J. Geophys. Res. Atmos., 119, 5787–5805, https://doi.org/10.1002/2014JD021642.
Sullivan, S. C., and A. Voigt, 2021: Ice microphysical processes exert a strong control on the simulated radiative energy budget in the tropics. Commun. Earth Environ., 2, 137, https://doi.org/10.1038/s43247-021-00206-7.
Tomita, H., 2008: New microphysical schemes with five and six categories by diagnostic generation of cloud ice. J. Meteor. Soc. Japan, 86A, 121–142, https://doi.org/10.2151/jmsj.86A.121.
Turbeville, S. M., J. M. Nugent, T. P. Ackerman, C. S. Bretherton, and P. N. Blossey, 2021: Tropical cirrus in global storm-resolving models: 2. Cirrus life cycle and top-of-atmosphere radiative fluxes. Earth Space Sci., 9, e2021EA001978, https://doi.org/10.1029/2021EA001978.
Waliser, D. E., and Coauthors, 2009: Cloud ice: A climate model challenge with signs and expectations of progress. J. Geophys. Res., 114, D00A21, https://doi.org/10.1029/2008JD010015.
Waliser, D. E., J.-L. F. Li, T. S. L’Ecuyer, and W.-T. Chen, 2011: The impact of precipitating ice and snow on the radiation balance in global climate models. Geophys. Res. Lett., 38, L06802, https://doi.org/10.1029/2010GL046478.
Webb, M. J., and Coauthors, 2017: The Cloud Feedback Model Intercomparison Project (CFMIP) contribution to CMIP6. Geosci. Model Dev., 10, 359–384, https://doi.org/10.5194/gmd-10-359-2017.
Zelinka, M. D., and D. L. Hartmann, 2011: The observed sensitivity of high clouds to mean surface temperature and anomalies in the tropics. J. Geophys. Res., 116, D23103, https://doi.org/10.1029/2011JD016459.
Zelinka, M. D., S. A. Klein, and D. L. Hartmann, 2012: Computing and partitioning cloud feedbacks using cloud property histograms. Part I: Cloud radiative kernels. J. Climate, 25, 3715–3735, https://doi.org/10.1175/JCLI-D-11-00248.1.
Zelinka, M. D., C. Zhou, and S. A. Klein, 2016: Insights from a refined decomposition of cloud feedbacks. Geophys. Res. Lett., 43, 9259–9269, https://doi.org/10.1002/2016GL069917.
Zelinka, M. D., D. A. Randall, M. J. Webb, and S. A. Klein, 2017: Clearing clouds of uncertainty. Nat. Climate Change, 7, 674–678, https://doi.org/10.1038/nclimate3402.