1. Introduction
Variability in the location of midlatitude jet streams plays a fundamental role in the climate system, impacting the tracks of extratropical cyclones and regional temperature and precipitation patterns. Variability in jet position on subseasonal-to-seasonal time scales is primarily associated with internal oscillations in the climate system (e.g., Hurrell 1995; Thompson and Wallace 2000), whereas variability in jet position on longer time scales can also be driven by anthropogenic forcing. Both observational and modeling evidence suggests that the Southern Hemisphere (SH) midlatitude jet has shifted poleward in summer months over the late twentieth century in response to stratospheric ozone depletion (Thompson and Solomon 2002; Gillett and Thompson 2003; Polvani et al. 2011), and modeling evidence suggests that increasing greenhouse gases will shift both the Northern Hemisphere (NH) and SH midlatitude jets poleward over most ocean basins in the twenty-first century (Kushner et al. 2001; Barnes and Polvani 2013; Simpson et al. 2014).
In addition to impacting weather patterns on a regional level, systematic changes in the positions of the midlatitude jets might also contribute to global-scale climate feedbacks. As the jets shift poleward with increasing greenhouse gas concentrations, the clouds associated with extratropical cyclone activity might be expected to push poleward with the jets, resulting in a positive (warming) feedback as the clouds retreat to higher latitudes where they reflect less sunlight back to space (e.g., Boucher et al. 2013). Consistent with this idea, Bender et al. (2012), Eastman and Warren (2013), and Norris et al. (2016) have noted systematic poleward shifts in observed midlatitude cloud fields over the last several decades, which they attributed to poleward shifts in the NH and SH midlatitude jets. However, contrary to this simple picture, the cloud radiative response to a poleward jet shift actually varies widely across global climate models (Grise and Polvani 2014, hereafter GP14). Furthermore, GP14 found little observational evidence supporting the notion that a poleward jet shift contributes to a net shortwave warming effect, at least based on interannual variability from the SH summer season.
In this paper, we concern ourselves with understanding why there is so much discrepancy among models and observations in the response of clouds and cloud radiative effects to variability in the position of the midlatitude jet. Our motivation here is not that a poleward jet shift is the most important process for determining future midlatitude cloud feedbacks. Kay et al. (2014), Ceppi et al. (2014), and Wall and Hartmann (2015) have all recently shown that the cloud feedbacks driven by a poleward jet shift in increasing greenhouse gas scenarios are of second-order importance to those driven by thermodynamic influences, such as the rising altitude of the melting level in a warming climate (Ceppi et al. 2016). But unlike these thermodynamic influences, studying jet variability provides us with a clear mechanism by which we can test the response of model clouds to a forcing (i.e., internal variability in the jet position) that can be directly compared to observations. As shown below, examining the response of observed and model clouds to jet variability provides insight into cloud processes poorly represented in today’s climate models, which are not immediately apparent from analysis of the mean state climatology.
Previous studies have provided some insight into the cloud and cloud radiative anomalies associated with a midlatitude jet shift. As the jet moves poleward, the large-scale vertical motion field shifts poleward with the jet (e.g., Thompson and Wallace 2000). Large-scale ascending motion at midlatitudes is associated with the increased presence of clouds with tops in the mid-to-upper troposphere (e.g., Weaver and Ramanathan 1997; Li et al. 2014b), as deep rising motion within extratropical cyclones drives nimbostratus and deep convective clouds (Lau and Crane 1995, 1997; Gordon et al. 2005). Thus, as the jet shifts poleward, the extratropical cyclone activity, upward vertical velocity anomalies, and high-topped “storm track” clouds all closely follow (Grise et al. 2013; Li et al. 2014a).
However, as the jet moves poleward, downward vertical velocity anomalies also shift poleward from the subtropics into midlatitudes. Downward vertical velocity anomalies inhibit the formation of clouds with tops in the mid-to-upper troposphere but are favorable for the formation of low-level clouds over midlatitude oceans (Booth et al. 2013; Govekar et al. 2014; Li et al. 2014b). The enhanced subsidence above cool midlatitude sea surface temperatures supports the development of a strong boundary layer temperature inversion that is conducive to the formation of highly reflective low-level stratocumulus clouds (e.g., Klein and Hartmann 1993). The strong boundary layer inversion promotes the coupling of low clouds to their surface moisture source via turbulent mixing, while at the same time, it inhibits the entrainment of dry free tropospheric air (e.g., Wood and Bretherton 2006). Consequently, as the midlatitude jet shifts poleward, storm-track clouds might decrease near 45° latitude, but low-level maritime clouds might increase there.
The competing behaviors of high and low clouds over midlatitude oceans make determining the net radiative impacts of a jet shift challenging (see also recent review by Ceppi and Hartmann 2015). Longwave cloud radiative effects are closely tied to the poleward shift in high-topped clouds (Grise et al. 2013; Li et al. 2014a), and this behavior appears to be well captured by global climate models (GP14). In contrast, shortwave cloud radiative effects are more strongly affected by the presence of highly reflective low-level liquid clouds, which are challenging to represent in global climate models (e.g., Klein et al. 2013). As a result, it is not surprising that the shortwave cloud radiative effects associated with a poleward jet shift vary widely across global climate models (GP14). However, even in observations, the shortwave cloud radiative effects associated with a jet shift vary by ocean basin and season. GP14 found little evidence for a shortwave cloud radiative warming effect associated with a SH midlatitude jet shift during summer months, whereas Tselioudis et al. (2016) have recently found evidence that a poleward shift in the North Atlantic jet during winter months is associated with a small shortwave cloud radiative warming effect. These widely varying shortwave cloud radiative effects, both in observations and models, remain poorly understood.
In this study, we build on these previous results and diagnose in detail the dynamical influences on variability in midlatitude clouds and their radiative effects. To do this, we follow several studies of tropical and subtropical clouds, which examined observed clouds within the context of midtropospheric vertical velocity and lower-tropospheric static stability anomalies (Medeiros and Stevens 2011; Myers and Norris 2013; Qu et al. 2015). Our results reveal that both midtropospheric vertical velocity and lower-tropospheric static stability are crucial in understanding observed variability in midlatitude clouds and their radiative effects. In particular, we show that the discrepancies in the cloud radiative effects associated with midlatitude jet variability between models and observations result from the overdependence of midlatitude cloud radiative effects on vertical motion in many models.
Our paper is organized as follows. Section 2 describes the data and methods. In section 3, we examine the responses of clouds and cloud radiative effects to SH midlatitude jet variability in both observations and global climate models. In section 4, we repeat our analyses for the NH midlatitudes. Section 5 concludes with a discussion and summary of our results.
2. Data and methods
a. Data
To examine the linkages among midlatitude clouds, cloud radiative effects, and jet variability in global climate models, we use monthly mean output from the global climate models that participated in CMIP5 (Taylor et al. 2012). The model data were obtained from PCMDI at Lawrence Livermore National Laboratory and from the Centre for Environmental Data Analysis. Here, we analyze output from the 20 CMIP5 models examined by GP14 (listed in Table 1). For each model, we use output from two scenarios: 1) preindustrial control (i.e., hundreds of years of unforced variability) and 2) abrupt 4×CO2 (in which atmospheric CO2 concentrations are quadrupled at the beginning of a 150-yr run). Because the control runs of each model have unequal length, we use 200 years from each model’s control run so that the analysis period for all models is of identical length. For five models (denoted by asterisks in Table 1), we also make use of the historical scenario runs for the 1979–2005 period, which provide output from the ISCCP simulator (Klein and Jakob 1999). For all scenarios, we use the first ensemble member (“r1i1p1”) from each model.
Listing of the CMIP5 models used in this study. Asterisks denote models whose historical runs are used in this study. (second column) The model types were defined by GP14 using (third column) the jet-CRE index, which is defined as the 30°–60°S average of shortwave CRE anomalies associated with a 1° poleward jet shift (as shown in Fig. 5b) over SH summer months (DJF). Positive (warming) values are type I, and negative (cooling) values are type II.
b. Methods
We calculate the time series of the position of the midlatitude jet by computing the latitude of the 850-hPa zonal-mean zonal wind maximum. At each monthly time step, a quadratic fit to the three grid points nearest the maximum value is used to determine the jet latitude to a resolution of 0.01° (see appendix of GP14 for further details). Jet latitude time series are calculated for the SH (0°–90°S), North Atlantic Ocean (0°–90°N, 60°W–0°), and North Pacific Ocean (0°–90°N, 135°E–125°W).
We calculate the radiative impacts of clouds using the cloud radiative effect (CRE), which is the difference in outgoing radiation at the top of the atmosphere between clear-sky and all-sky scenes (e.g., Ramanathan et al. 1989). The CRE approach can be biased by changes in temperature, water vapor and other greenhouse gases, and surface albedo, and in those cases, a cloud radiative kernel approach is more appropriate (Zelinka et al. 2012). However, only a small number of CMIP5 models provide the necessary output for the kernel approach (see Zelinka et al. 2013). Here, we focus on the control runs of the models, which minimizes biases due to the CRE approach.
3. Cloud and CRE anomalies associated with SH midlatitude jet shifts
In this section, we examine the cloud and CRE anomalies associated with variability in the position of the SH midlatitude jet. To do this, we regress monthly mean cloud and CRE anomalies (from the seasonally varying background climatology) on the SH jet latitude time series (as defined in section 2b). The resulting regression patterns correspond to the cloud and CRE anomalies associated with a 1° poleward shift in the SH midlatitude jet. We first examine these patterns in observations and then compare the observed results with those from global climate models.
a. Observations
Figure 1 shows the observed cloud fraction anomalies associated with a 1° poleward shift in the SH midlatitude jet, as derived from ISCCP data (see also Grise et al. 2013). When the SH jet shifts poleward, high cloud fraction increases near 30°S and between 50° and 60°S but decreases between 40° and 50°S (Fig. 1a, black-boxed region). In contrast, low cloud fraction decreases between 50° and 60°S but increases between 30° and 50°S (Fig. 1b). The magnitudes of the anomalies in Fig. 1 are small (<1%), given that the climatological high and low cloud fractions over the Southern Ocean are approximately 20% and 70%, respectively (see also Fig. 7b). The standard deviation of SH jet latitude is ~3°, so the cloud fraction anomalies associated with a typical SH jet shift (i.e., the values in Fig. 1 multiplied by 3) are roughly an order of magnitude smaller than the climatology.
Regressions of observed ISCCP (a) high cloud fraction anomalies (cloud top pressure < 440 hPa) and (b) low cloud fraction anomalies (cloud top pressure > 680 hPa) on the SH jet latitude time series. Low cloud anomalies are calculated using the random overlap assumption. Units correspond to a 1° poleward shift in the SH midlatitude jet. Anomalies are defined with respect to the seasonally varying monthly mean climatology. The regressions are taken over all months in the period July 1983–June 2008. The stippling indicates regions that are 95% statistically significant using Student’s t test, and the black-boxed region highlights the 40°–50°S latitude band.
Citation: Journal of Climate 29, 24; 10.1175/JCLI-D-16-0295.1
Figures 2a and 2b show the observed longwave and shortwave CRE anomalies associated with a 1° poleward shift in the SH midlatitude jet, as derived from CERES data. For comparison, the annual-mean climatological values of longwave and shortwave CRE over the Southern Ocean are approximately 30 and −70 W m−2, respectively. As found by GP14, the longwave CRE anomalies closely mirror the high cloud anomalies (cf. Fig. 2a with Fig. 1a), whereas the shortwave CRE anomalies have little coherent spatial structure (Fig. 2b). The increases in high cloud fraction and longwave CRE near 30°S and between 50° and 60°S are consistent with upward midtropospheric vertical velocity anomalies (i.e., negative ω at 500 hPa) in those regions, and the decreases in high cloud fraction and longwave CRE between 40° and 50°S are consistent with downward vertical velocity anomalies there (Fig. 2c).
As in Fig. 1, but for regressions of observed (a) CERES longwave (LW) CRE anomalies, (b) CERES shortwave (SW) CRE anomalies, (c) 500-hPa vertical velocity ω500 anomalies, and (d) estimated inversion strength (EIS) anomalies on the SH jet latitude time series. The regressions are taken over all months in the periods (a),(b) March 2000–February 2015 and (c),(d) January 1979–December 2015.
Citation: Journal of Climate 29, 24; 10.1175/JCLI-D-16-0295.1
As noted by Li et al. (2014b), the low cloud anomalies have an opposite relationship with vertical velocity (cf. Fig. 2c with Fig. 1b); low clouds increase in regions with downward vertical velocity anomalies (40°–50°S) and decrease in regions with upward vertical velocity anomalies (50°–60°S). Enhanced subsidence promotes warming and drying of the free troposphere, which acts to strengthen the boundary layer temperature inversion (in the absence of sea surface temperature changes). Consequently, it is not coincidental that vertical velocity and EIS anomalies closely mirror one another (cf. Figs. 2c and 2d). In fact, vertical velocity and EIS anomalies are significantly correlated with one another at most locations across the Southern Ocean (r ≈ 0.3; not shown). Note that, unlike the low cloud field, the spatial pattern of shortwave CRE anomalies (Fig. 2b) does not closely resemble that of either the vertical velocity or EIS anomalies.
To better understand the relationship of SH midlatitude cloud and CRE anomalies with vertical velocity and EIS, in Figs. 3 and 4a, we composite observed cloud fraction and CRE anomalies in a phase space defined by 500-hPa vertical velocity ω500 and EIS anomalies. The phase space is constructed as follows. First, for each oceanic grid point between 40° and 50°S, cloud, CRE, ω500, and EIS anomalies are found for all months in the data record. Next, the cloud and CRE anomalies are binned into categories according to their coinciding ω500 and EIS anomalies. Finally, the cloud and CRE anomalies in each category are averaged to yield the composites shown in Figs. 3 and 4a.
Composites of observed ISCCP (a) high cloud fraction, (b) low cloud fraction, and (c) total cloud fraction anomalies at all oceanic grid points over the Southern Ocean (40°–50°S) as a function of the coinciding vertical velocity anomalies (horizontal axis) and EIS anomalies (vertical axis). Low cloud anomalies are calculated using the random overlap assumption. Composites are taken over all months in the period July 1983–June 2008.
Citation: Journal of Climate 29, 24; 10.1175/JCLI-D-16-0295.1
As in Fig. 3, but for the composite patterns of (top row) longwave CRE anomalies and (bottom row) shortwave CRE anomalies. Composites are taken over (a) observed CERES CRE anomalies (March 2000–February 2015) and (b),(c) CRE anomalies from 200 years of preindustrial control runs of CMIP5 models. The model results are shown as the average of the composite patterns from (b) 10 type I models and (c) 10 type II models (see Table 1 for listing of model types; see also GP14). The black contours denote the anomalous percentage of grid points falling in a region of the ω500–EIS phase space for periods of a poleward jet shift ≥ 1° latitude (solid contours) and an equatorward jet shift ≥ 1° latitude (dashed contours).
Citation: Journal of Climate 29, 24; 10.1175/JCLI-D-16-0295.1
Consistent with the results in Figs. 1 and 2, at SH midlatitudes, high cloud fraction anomalies (Fig. 3a) and longwave CRE anomalies (Fig. 4a, top) are almost entirely a function of vertical velocity, with increased high clouds and longwave CRE associated with upward vertical velocity anomalies (e.g., Weaver and Ramanathan 1997; Li et al. 2014b). Middle-cloud fraction anomalies are also strongly dependent on vertical velocity (not shown). However, low cloud fraction anomalies at SH midlatitudes (Fig. 3b) are almost entirely a function of EIS, with increased low clouds associated with enhanced boundary layer inversion strength (e.g., Wood and Bretherton 2006). As a result, total cloud fraction anomalies (Fig. 3c, bottom) and shortwave CRE anomalies (Fig. 4a, bottom), which are affected by the reflection from clouds in all layers, depend on both ω500 and EIS.1
The shortwave cloud radiative cooling anomalies at SH midlatitudes are strongest for upward vertical velocity and positive EIS anomalies and weakest for downward vertical velocity and negative EIS anomalies (Fig. 4a, bottom). These relationships are consistent with those found by Myers and Norris (2013) for subtropical low cloud regimes. Because a poleward jet shift is associated with downward vertical velocity anomalies and positive EIS anomalies in the 40°–50°S latitude band (Figs. 2c,d, black-boxed region), it results in a move upward and to the right on the ω500–EIS phase space diagram (Fig. 4, black contours). Moving toward the top right on the ω500–EIS phase space diagram implies a decrease in longwave CRE with a poleward jet shift (cf. Fig. 2a) but little change in shortwave CRE (cf. Fig. 2b), as the jet shift dynamical anomalies (Fig. 4a, black contours) are nearly orthogonal to the shortwave CRE anomalies (shading in Fig. 4a, bottom). Consequently, the competing sensitivities of observed shortwave CRE anomalies to vertical velocity and EIS perturbations explain why there are only weak shortwave CRE anomalies associated with an SH midlatitude jet shift over the 40°–50°S latitude band.
b. Models
We now examine whether global climate models can capture the observed relationships among cloud anomalies, CRE anomalies, and the SH midlatitude jet. Figure 5 shows the longwave CRE, shortwave CRE, ω500, and EIS anomalies associated with a 1° poleward shift in the SH midlatitude jet, as derived from month-to-month variability in the preindustrial control runs of 20 CMIP5 models. Following GP14, we have separated the CMIP5 models into two subsets of 10 models (see Table 1): “type I” models (Fig. 5, left column) and “type II” models (Fig. 5, right column).
As in Fig. 2, but for the preindustrial control runs of CMIP5 models. Anomalies are defined with respect to the seasonally varying monthly mean climatology from 200 years of each model’s preindustrial control run. (left) The average of the regression patterns from 10 type I models and (right) the average of the regression patterns from 10 type II models (see Table 1).
Citation: Journal of Climate 29, 24; 10.1175/JCLI-D-16-0295.1
The longwave CRE, ω500, and EIS anomalies are similar between the two subsets of CMIP5 models and compare well with those found from observations (cf. Figs. 5a,c,d with Figs. 2a,c,d). But as noted by GP14, the shortwave CRE anomalies vary widely across CMIP5 models (Fig. 5b), even though the dynamical changes are nearly identical in the two subsets of models (Figs. 5c,d). Type I models produce a sizeable shortwave cloud radiative warming effect in the 40°–50°S latitude band with a poleward jet shift (Fig. 5b, left), whereas type II models indicate little net shortwave CRE there (as in observations; compare Fig. 5b, right, with Fig. 2b). The notable difference in shortwave CRE anomalies between the type I and type II models is by construction. GP14 defined “type I” models as those models that produce a positive shortwave CRE anomaly at SH midlatitudes (30°–60°S) when the SH jet shifts poleward by 1° latitude during the SH summer season, and “type II” models as those models that produce a negative shortwave CRE anomaly at SH midlatitudes when the SH jet shifts poleward (see Table 1, third column). However, GP14 did not find a conclusive physical reason to explain this varying model behavior.
The distinctive behavior of type I and type II models is more readily understood using ω500–EIS phase space diagrams (Fig. 4). The phase space diagrams reveal that 1) SH midlatitude longwave CRE anomalies are strongly a function of ω500 in observations and both subsets of CMIP5 models (Fig. 4, top) and that 2) SH midlatitude shortwave CRE anomalies are a function of both ω500 and EIS in observations and type II models but more strongly a function of ω500 in type I models (Fig. 4, bottom). As noted above, a poleward jet shift is associated with downward velocity anomalies and positive EIS anomalies between 40° and 50°S, which translates into a movement toward the top right in the ω500–EIS phase space diagram (Fig. 4, black contours). In type II models, moving toward the top right in the ω500–EIS phase space diagram implies little change in shortwave CRE with a poleward jet shift (as in observations; compare Figs. 4a,c). But in type I models, moving toward the top right in the ω500–EIS phase space diagram implies an increase in shortwave CRE (see also Fig. 5b, left), as the jet shift dynamical anomalies (Fig. 4b, bottom, black contours) are no longer orthogonal to the shortwave CRE anomalies (Fig. 4b, bottom, shading). Hence, the overdependence of SH midlatitude shortwave CRE anomalies on vertical velocity anomalies in type I models produces an erroneous warming signal at SH midlatitudes with a poleward jet shift (Fig. 4b, bottom; Fig. 5b, left), which does not occur in observations or in type II models.
The results in Fig. 4 suggest that the diverse range of shortwave CRE responses to a poleward SH jet shift across CMIP5 models (Fig. 5b; see also GP14) follows directly from the varying sensitivities of the models’ clouds to EIS perturbations. To provide more insight into these sensitivities, Fig. 6 shows spatial distributions of correlations of shortwave CRE anomalies with 500-hPa vertical velocity and EIS perturbations from both observations and CMIP5 models. Consistent with Fig. 4a, at SH midlatitudes, observed shortwave CRE anomalies are positively correlated with midtropospheric vertical velocity anomalies (i.e., anomalous rising motion is associated with increased cloud reflection) and negatively correlated with EIS anomalies (i.e., a stronger boundary layer inversion is associated with increased cloud reflection; Fig. 6, top row). Consistent with Figs. 4b,c, both subsets of CMIP5 models reproduce the positive correlations with vertical velocity anomalies (although the magnitude is overestimated; Fig. 6a), whereas only the type II models reproduce the negative correlations with EIS anomalies (Fig. 6b). Note that, in both subsets of models, correlations between vertical velocity and EIS anomalies at each grid point are also similar to those found in observations (not shown).
Correlations of monthly mean shortwave CRE anomalies at each grid point with (a) monthly mean 500-hPa vertical velocity anomalies at the same grid point and (b) monthly mean EIS anomalies at the same grid point. Results are shown for (top row) observations (March 2000–February 2015), (middle row) type I CMIP5 models, and (bottom row) type II CMIP5 models. The thin black solid line highlights where the observed value of EIS exceeds 2 K in the long-term mean climatology. The stippling in the top row indicates regions where the observed correlations are 95% statistically significant.
Citation: Journal of Climate 29, 24; 10.1175/JCLI-D-16-0295.1
Importantly, Fig. 6 also reveals that the varying EIS sensitivities of the models are not unique to the SH midlatitudes. In observations and CMIP5 models, shortwave CRE anomalies are positively correlated with ω500 anomalies across the globe (Fig. 6a) and with EIS anomalies in the deep tropics (Fig. 6b). But in observations and type II models, shortwave CRE anomalies are negatively correlated with EIS anomalies throughout the subtropics and midlatitudes (Fig. 6b). The correlations switch sign near where the observed climatological-mean value of EIS exceeds 2 K (Fig. 6, thin black line). Qu et al. (2015) have recently shown that approximately half of CMIP5 models underestimate the observed sensitivity of subtropical low cloud cover to EIS perturbations (see also Medeiros and Nuijens 2016). The results in Fig. 6 support their findings and reveal that similar biases extend into midlatitudes.
Why do type I models underestimate the observed sensitivity to EIS perturbations (Figs. 4 and 6)? To answer this question, detailed cloud output from the models is required, but only a few CMIP5 models provide diagnostic cloud output that is directly comparable to ISCCP observations (see Table 2 of Zelinka et al. 2013). To maximize the number of models that can be examined, we use output from the historical runs of five models (two type I models and three type II models, denoted by asterisks in Table 1). Note that, while these five models can provide helpful insight into the varying EIS sensitivities of CMIP5 models, their behavior is not necessarily representative of all type I and type II models.
Figure 7 explores two possible reasons for the underestimated sensitivity of shortwave CRE to EIS perturbations in type I models. First, low clouds themselves might not be sensitive enough to EIS perturbations in type I models. Second, climatological low cloud fractions might be severely underestimated in type I models, causing shortwave CRE perturbations to be governed by variability in higher-level clouds and thus more strongly linked to vertical velocity perturbations. At least for the five models examined here, the first reason appears to be more valid. In observations and the three type II models, low cloud fraction anomalies are correlated with EIS anomalies throughout the low cloud regimes of the subtropics and midlatitudes (Fig. 7a, top and bottom), but these correlations are largely absent in the two type I models (Fig. 7a, middle). The regions of underestimated correlations between low cloud and EIS perturbations in the two type I models directly coincide with those regions with underestimated correlations between shortwave CRE and EIS perturbations (Fig. 6b, middle), drawing a strong linkage between low cloud sensitivities and the shortwave CRE sensitivities that are the focus of this study.
(a) Correlations of low cloud fraction anomalies at each grid point with monthly mean EIS anomalies at the same grid point. (b) Low cloud fraction climatology. Results are shown for (top row) observations (July 1983–June 2008), (middle row) the mean of two type I CMIP5 models (CanESM2 and MPI-ESM-LR), and (bottom row) the mean of three type II CMIP5 models (HadGEM2-ES, MIROC5, and MRI-CGCM3). Model results are calculated from available data from years 1979–2005 of the historical scenario, and the model cloud fractions are output from the models’ ISCCP simulator. All low cloud fractions are calculated using the random overlap assumption. The stippling in (a, top) indicates regions where the observed correlations are 95% statistically significant.
Citation: Journal of Climate 29, 24; 10.1175/JCLI-D-16-0295.1
There is little evidence that climatological low cloud fractions are severely underestimated in type I models. In fact, for the models examined here, the climatological low cloud fractions over midlatitude oceans in type I models actually compare better with observations (Fig. 7b, top and middle), whereas the climatological low cloud fractions over midlatitude oceans in type II models are underestimated (Fig. 7b, bottom). Type I models also have more realistic magnitudes of climatological-mean shortwave CRE over midlatitude oceans (see Fig. 5 of GP14). Thus, while type II models well reproduce the observed low cloud and shortwave CRE sensitivities to EIS perturbations (Figs. 6b and 7a), it is important to note that these models have their own shortcomings, as they underestimate the magnitudes of climatological low cloud fraction and shortwave CRE over midlatitude oceans.
c. Summary
In this section, we examined the cloud and CRE anomalies associated with a poleward shift in the SH midlatitude jet. In both observations and models, high cloud and longwave CRE anomalies closely follow the poleward movement of the jet and the attendant vertical velocity field (Figs. 1, 2, and 5). Low cloud anomalies increase equatorward of the jet in observations (Fig. 1b) but are highly variable in models (Fig. 7a). The compensating effects of the high and low cloud anomalies produce few coherent changes in shortwave CRE in observations (Fig. 2b), but sizeable changes in shortwave CRE do occur in some models (Fig. 5b, left). The varying shortwave CRE responses to a jet shift among observations and models appear directly tied to varying sensitivities of low clouds to EIS perturbations (Fig. 7a). In observations, at SH midlatitudes, high clouds increase with upward vertical velocity anomalies (Fig. 3a) and low clouds increase with increasing boundary layer inversion strength (Fig. 3b), causing shortwave cloud radiative cooling to increase for both upward vertical velocity and positive EIS anomalies (Fig. 4a, bottom; Fig. 6, top row). But in a subset of CMIP5 models (type I), the sensitivity of low clouds and therefore shortwave cloud radiative effects to EIS anomalies is substantially underestimated (Fig. 6, middle row; Fig. 7a). The reduced sensitivity of low cloud and shortwave CRE anomalies to EIS perturbations in these models is not unique to the SH midlatitudes, but also occurs in the subtropics (see also Qu et al. 2015) and at NH midlatitudes (Figs. 6b and 7a). In the next section, we apply these insights to understand the cloud and CRE anomalies associated with NH midlatitude jet shifts.
4. Cloud and CRE anomalies associated with NH midlatitude jet shifts
In this section, we examine the cloud and CRE anomalies associated with variability in the position of the NH midlatitude jets. We examine the variability associated with the North Atlantic and North Pacific jets separately.
a. North Atlantic
Figures 8 and 9 show the cloud, CRE, ω500, and EIS anomalies associated with a 1° poleward shift in the North Atlantic jet. A poleward shift in the North Atlantic jet is associated with enhanced high cloud cover and longwave CRE near 60°N and in the subtropics and reduced high cloud cover and longwave CRE between 30° and 45°N (Figs. 8a and 9a; see also Li et al. 2014a). The vertical velocity anomalies have qualitatively similar structure (Fig. 9c), consistent with their driving of the high cloud and longwave CRE anomalies. The observed tripole pattern of the high cloud, longwave CRE, and vertical velocity anomalies in Figs. 8–9 is very similar to that found at SH midlatitudes (Figs. 1a, 2a, and 2c), except that the North Atlantic anomalies are tilted from southwest to northeast following the orientation of the jet. These observed anomalies are well captured by CMIP5 models (cf. left and right columns of Fig. 9a and Fig. 9c).
As in Fig. 1, but for regressions on the North Atlantic (60°W–0°) jet latitude time series. Units correspond to a 1° poleward shift in the North Atlantic jet. For reference, the climatological values of high and low cloud fraction are ~30% and ~50% in the western North Atlantic and ~15% and ~70% in the eastern North Atlantic.
Citation: Journal of Climate 29, 24; 10.1175/JCLI-D-16-0295.1
As in Fig. 2, but for regressions on the North Atlantic (60°W–0°) jet latitude time series. (left column) Results for the observations and (right column) results from the preindustrial control runs of CMIP5 models (as in Fig. 5). Units correspond to a 1° poleward shift in the North Atlantic jet. The stippling in the left column indicates regions that are 95% statistically significant using Student’s t test. For reference, the climatological values of longwave and shortwave CRE in the black-boxed regions are approximately 30 to 40 W m−2 and −55 to 65 W m−2, respectively.
Citation: Journal of Climate 29, 24; 10.1175/JCLI-D-16-0295.1
In the eastern North Atlantic, a poleward jet shift is associated with reduced low cloud cover near 60°N and enhanced low cloud cover between 30°N and 55°N centered off the coast of the Iberian Peninsula (Fig. 8b). The low cloud anomalies off of western Europe coincide with large downward vertical velocity anomalies and positive EIS anomalies there (Figs. 9c,d, right box). The dynamical anomalies in this region are thus very similar to those found in the 40°–50°S latitude band associated with an SH midlatitude jet shift (Figs. 2c,d). Additionally, as over the Southern Ocean, type I models underestimate the observed sensitivity of low clouds and shortwave CRE to EIS perturbations in this region (Figs. 6b and 7a). Thus, consistent with our SH findings, a North Atlantic jet shift is associated with positive shortwave CRE anomalies in the eastern North Atlantic near 40°N in many CMIP5 models but not in observations (Fig. 9b). Note that, while these shortwave CRE anomalies are generally larger in type I models, a subjective grouping of the models based on the Atlantic shortwave CRE anomalies does not unambiguously show two model types as it does at SH midlatitudes (not shown).
In the western North Atlantic, a poleward jet shift is associated with positive shortwave CRE anomalies off the East Coast of the United States in both observations and models (Fig. 9b, left box). These positive shortwave CRE anomalies were recently noted in observations for the first time by Tselioudis et al. (2016). The western North Atlantic is a region where low cloud coverage is less extensive in the climatology (Fig. 7b), giving higher-topped clouds a greater contribution to the total cloud fraction field. Consequently, Tselioudis et al. (2016) argue that a poleward shift in the high cloud fraction field in this region (Fig. 8a) should contribute to a poleward shift in the total cloud fraction field, creating a positive shortwave CRE anomaly equatorward of the jet. However, the shift in the North Atlantic jet latitude is actually associated with very few significant vertical velocity or EIS anomalies over the western North Atlantic (Figs. 9c,d), so the shortwave CRE anomalies there (Fig. 9b) do not appear strongly tied to large-scale midlatitude dynamics. Instead, Tselioudis et al. (2016) suggest that shortwave CRE anomalies in this region are more strongly linked to variability in the Hadley cell edge latitude. Additionally, shortwave CRE anomalies over the western North Atlantic are positively correlated with both vertical velocity and EIS anomalies (Fig. 6), suggesting that the cloud behavior in this region is characterized more by deep convective frontal clouds (cf. with deep tropical convective regimes in Fig. 6) than by the low cloud cover that dominates most midlatitude oceans (Fig. 7b).
b. North Pacific
Figures 10–12 show the cloud, CRE, ω500, and EIS anomalies associated with a 1° poleward shift in the North Pacific jet. As the anomalies associated with variability in the North Pacific jet are strongly seasonally dependent, we examine the December–February (DJF) and June–August (JJA) seasons individually.
As in Fig. 1, but for regressions on the North Pacific (135°E–125°W) jet latitude time series for (left column) the DJF season and (right column) the JJA season. Units correspond to a 1° poleward shift in the North Pacific jet. For reference, the climatological values of high and low cloud fraction in the black-boxed regions are ~25%–30% and ~70%, respectively. Note that there are virtually no high clouds in the far-eastern Pacific in the JJA climatology.
Citation: Journal of Climate 29, 24; 10.1175/JCLI-D-16-0295.1
As in Fig. 9, but for regressions on the North Pacific (135°E–125°W) jet latitude time series for the DJF season. Units correspond to a 1° poleward shift in the North Pacific jet. For reference, the climatological values of longwave and shortwave CRE in the black-boxed regions are approximately 45 and −50 W m−2, respectively.
Citation: Journal of Climate 29, 24; 10.1175/JCLI-D-16-0295.1
As in Fig. 11, but for the JJA season. For reference, the climatological values of longwave and shortwave CRE in the black-boxed regions are approximately 20–30 and −100 W m−2, respectively.
Citation: Journal of Climate 29, 24; 10.1175/JCLI-D-16-0295.1
During NH winter (DJF), the anomalies associated with a poleward jet shift over the North Pacific Ocean closely resemble those from the eastern North Atlantic and Southern Oceans: 1) a tripole pattern in high cloud fraction (Fig. 10a, left), longwave CRE (Fig. 11a), and vertical velocity anomalies (Fig. 11c) over the eastern North Pacific, 2) an increase in low cloud fraction in the central and eastern North Pacific equatorward of 45°N (particularly off the west coast of North America; Fig. 10b, left), and 3) a tripole pattern in EIS anomalies extending across the North Pacific basin (Fig. 11d). However, unlike the other two basins, weak positive shortwave CRE anomalies are also present in the region off the California coast, not only in CMIP5 models but also in observations (Fig. 11b, right box). In this region, the vertical velocity anomalies have very large magnitude (Fig. 11c), whereas the EIS anomalies are more modest (Fig. 11d). The sensitivities of shortwave CRE anomalies to vertical velocity and EIS perturbations in this region are very similar to those over the Southern Ocean (Fig. 6), but in this case, presumably owing to the overwhelming influence of the vertical velocity anomalies, the jet shift contributes to a small positive shortwave CRE anomaly for the observations, type I models, and type II models. In other words, the reduction in solar reflection by the decreases in high and midlevel clouds over the eastern North Pacific (Fig. 10a, left) appears to slightly outpace any enhancement in solar reflection from increased low cloud amount (Fig. 10b, left).
During NH summer (JJA), the high cloud, longwave CRE, and vertical velocity anomalies associated with a poleward North Pacific jet shift are similar to those in DJF but are shifted poleward and extend westward (Fig. 10a, right; Figs. 12a,c). However, the low cloud and EIS anomalies have a substantially different structure—positive anomalies in the North Pacific that extend southward along the North American coast and negative anomalies over the central Pacific (Fig. 10b, right; Fig. 12d). Consistent with Fig. 6, the combination of downward vertical velocity anomalies and negative EIS anomalies in the central North Pacific produces a strong shortwave cloud radiative warming effect in observations and all CMIP5 models (Fig. 12b, left box). This is consistent with the reduction in both high and low cloud cover in this region (Fig. 10, right).
The only major difference between observations and CMIP5 models for a North Pacific jet shift occurs in the eastern North Pacific during JJA (Fig. 12, right box). Here, as in the eastern North Atlantic and Southern Oceans, observations indicate a shortwave CRE cooling anomaly (Fig. 12b, left), whereas CMIP5 models show the shortwave CRE warming anomaly from the central Pacific extending eastward to the North American coast (Fig. 12b, right). According to Fig. 6, the downward vertical velocity anomalies in this region should contribute to shortwave CRE warming, whereas the positive EIS anomalies should contribute to shortwave CRE cooling. The net result is a shortwave CRE cooling anomaly in observations (Fig. 12b, left). As noted in section 3, the sensitivity of models’ shortwave CRE anomalies to EIS perturbations is often underestimated (Figs. 4 and 6), and consequently the model results do not closely match observations in the eastern North Pacific during JJA (Fig. 12b, right). There is large intermodel spread in the model results in the region (see also Clement et al. 2009), and we note that the type I/type II categorization is not particularly helpful in understanding the intermodel spread in this region.
c. Summary
In this section, we examined the cloud and CRE anomalies associated with a poleward shift in the North Atlantic and North Pacific jets. The anomalies associated with NH midlatitude jet shifts are more diverse than their SH counterparts (see summary in Table 2). Consistent with our findings in the SH, the high cloud and longwave CRE anomalies associated with NH jet shifts closely follow vertical motion anomalies (see also Li et al. 2014a). However, in contrast to our findings in the SH, CMIP5 models tend to reproduce the shortwave CRE signatures of NH jet shifts more accurately. The key differences between observations and models are confined to regions with similar characteristics as the Southern Ocean—that is, where 1) poleward jet shifts increase downward motion and EIS and 2) low clouds dominate the total cloud cover field. It is only in these regions (eastern North Atlantic and summertime eastern North Pacific) where the varying sensitivities of observed and model clouds to vertical velocity and EIS perturbations (Fig. 6) lead to significant model biases (Table 2).
Summary of the observed relationships documented in this study. (column 1) Black-boxed regions highlighted in figures. (columns 2 and 3) Sign of correlations between shortwave CRE anomalies and ω500 and EIS anomalies (as shown in Fig. 6). (columns 4–6) Sign of ω500, EIS, and shortwave CRE anomaly associated with 1° poleward jet shift (as shown in Figs. 2, 9, 11, and 12). The symbol “X” refers to anomalies that are of ambiguous sign. Asterisks denote where observations and models differ on sign.
5. Summary and discussion
Recent observational evidence suggests that midlatitude cloud patterns have shifted notably poleward over the last several decades (Bender et al. 2012; Eastman and Warren 2013; Norris et al. 2016). These observed cloud trends could be a signature of poleward shifts in the midlatitude jet streams and extratropical storm tracks, which are projected to occur with increasing atmospheric greenhouse gas concentrations (e.g., Kushner et al. 2001; Yin 2005; Barnes and Polvani 2013). If cloud cover at midlatitudes were governed exclusively by extratropical storm tracks, a poleward shift in the jet would reduce the total cloud cover at midlatitudes, leading to a warming feedback on the climate system as the storm-track clouds move to higher latitudes where they reflect less solar radiation (e.g., Boucher et al. 2013). However, as the jet and storm tracks shift poleward, increased subsidence and lower-tropospheric stability promote the existence of low clouds over midlatitude oceans. As a result, the net radiative effect of a jet shift is not straightforward, as it reflects the competing behaviors of high and low clouds (see also Ceppi and Hartmann 2015). Some global climate models suggest that a poleward shift in the midlatitude jet is associated with a net cloud radiative warming effect (Grise et al. 2013; GP14), but the observational support for this effect is not as conclusive (GP14; Tselioudis et al. 2016).
In this study, we reconcile the varying responses of clouds and cloud radiative effects to midlatitude jet variability among observations and global climate models. To do this, we examine the sensitivities of observed and model cloud properties to two variables: midtropospheric vertical velocity (which is known to affect high clouds; Fig. 3a) and EIS (which is known to affect marine boundary layer clouds; Fig. 3b). As the midlatitude jet shifts poleward, vertical velocity anomalies push poleward with the jet, and the high clouds and their associated longwave radiative effects closely follow the poleward shift in upward vertical velocities (see also Grise et al. 2013; Li et al. 2014a). This behavior is observed over the Southern Ocean (Figs. 1 and 2), North Atlantic Ocean (Figs. 8 and 9), and North Pacific Ocean (Figs. 10–12) and is consistent between observations and CMIP5 models.
The behavior of low clouds (and thus the total cloud fraction and shortwave cloud radiative effects) is more diverse across ocean basins, and between observations and CMIP5 models. In observations, midlatitude low cloud cover increases in regions with enhanced EIS (Fig. 3b; Wood and Bretherton 2006). Changes in total cloud fraction and shortwave cloud radiative effects thus reflect the combined effects of high and low cloud anomalies. As a result, observed midlatitude shortwave CRE anomalies have a multivariate dependence: both increases in upward vertical velocity and increases in EIS tend to increase the reflection of solar radiation by clouds (Fig. 4a, bottom; Fig. 6, top row). Thus, in regions where a poleward midlatitude jet shift is associated with increases in EIS but decreases in upward vertical velocity, the observed shortwave cloud radiative effect is relatively small: for example, over the Southern Ocean (Fig. 2; GP14) and eastern North Atlantic Ocean (Fig. 9; see also Table 2). However, in regions where the vertical velocity and EIS anomalies do not have canceling effects, a poleward jet shift can produce a sizeable shortwave cloud radiative effect: for example, over the western North Atlantic Ocean (Fig. 9; Tselioudis et al. 2016) and the summertime central Pacific Ocean (Fig. 12).
The reason that many global climate models exaggerate the shortwave cloud radiative effects associated with a jet shift is that they do not properly capture the observed multivariate dependence of cloud radiative effect anomalies on vertical velocity and EIS (Figs. 4 and 6). In particular, we find here that the shortwave CRE anomalies in a subset of CMIP5 models (the type I models of GP14) are overly dependent on vertical velocity. As a result, in regions where a poleward midlatitude jet shift is associated with increases in EIS and decreases in upward vertical velocity (e.g., Southern Ocean, eastern Atlantic and Pacific Oceans), type I models produce a shortwave cloud radiative warming effect at midlatitudes (due to the downward vertical velocity anomalies), whereas observations produce a negligible shortwave cloud radiative effect (due to the compensating effects of the downward vertical velocity and positive EIS anomalies). The underestimated sensitivity of model cloud and cloud radiative effects to EIS perturbations is not limited to midlatitude regions but also occurs in low cloud regimes of the subtropics (Fig. 6; see also Qu et al. 2015; Medeiros and Nuijens 2016).
Overall, midlatitude clouds and their radiative effects are strongly affected by changes in the large-scale circulation, and these relationships are generally well captured in global climate models. However, key discrepancies between models and observations occur in regions where observed CRE anomalies are sensitive to both vertical velocity and EIS perturbations, and model CRE anomalies have a more singular dependence on vertical velocity perturbations (Figs. 4 and 6). Consequently, we might have less confidence in future midlatitude cloud feedbacks in regions where the projected changes in ω500 and EIS are nonnegligible and of the same sign.
Figure 13 shows the CMIP5 multimodel-mean responses of shortwave CRE, ω500, and EIS to a quadrupling of atmospheric CO2 concentrations. The regions outlined in black in Fig. 13a demarcate midlatitude regions where the ω500 and EIS responses are of the same sign and would thus have competing dynamical effects on the shortwave CRE response (following from Fig. 4). Given that a poleward midlatitude jet shift is projected to occur over most ocean basins with increasing CO2 concentrations, it is not surprising that the regions highlighted in Fig. 13a closely correspond to the regions over the Southern Ocean, eastern North Atlantic Ocean, and eastern North Pacific Ocean where models have significant biases in representing the shortwave cloud radiative effects associated with natural variability in the jet position (Table 2). These results are also consistent with the findings of GP14, who found significantly larger shortwave cloud radiative responses in these regions in type I models upon an abrupt quadrupling of CO2 (see Fig. 9 in GP14). However, as stated in section 1, the midlatitude cloud feedbacks associated with changes in large-scale dynamics are likely to be of second-order importance to those associated with the thermodynamic influences of warming temperatures (Kay et al. 2014; Ceppi et al. 2014; Wall and Hartmann 2015).
CMIP5 multimodel-mean response of (a) shortwave CRE, (b) 500-hPa vertical velocity ω500, and (c) EIS to 4×CO2 forcing. The 4×CO2 response is calculated as the difference in the climatologies between the models’ preindustrial control runs and the last 50 years of the models’ abrupt 4×CO2 runs. Regions where the multimodel-mean ω500 and EIS responses are both positive (negative) at midlatitudes are outlined in solid (dashed) black lines in (a). The stippling in all three panels indicates regions where the responses are 95% statistically significant using Student’s t test.
Citation: Journal of Climate 29, 24; 10.1175/JCLI-D-16-0295.1
Finally, we note that midlatitude cloud and CRE anomalies are likely sensitive to more than just the two dynamical variables (vertical velocity and EIS) examined here (see also Bretherton 2015). In addition to EIS, Qu et al. (2014) found that subtropical low cloud anomalies were strongly related to sea surface temperature perturbations, primarily through surface latent heat fluxes and the moisture contrast between the maritime boundary layer and free troposphere (Qu et al. 2015). Additionally, Norris and Iacobellis (2005) found that low cloud fraction over the summertime North Pacific was strongly tied to the surface wind direction and low-level temperature advection. Future work might address if these or other dynamical variables are relevant in understanding the midlatitude cloud anomalies shown here. These relationships will also need to be explored on daily time scales, where a more process-level understanding can be developed.
Acknowledgments
We thank L. M. Polvani for helpful discussions during the early stages of this project and the editor (S. A. Klein) and three anonymous reviewers for helpful comments on the manuscript. We acknowledge the WCRP’s Working Group on Coupled Modelling, which is responsible for CMIP, and we thank the climate modeling groups (listed in Table 1) for producing and making available their model output. For CMIP, the U.S. Department of Energy’s PCMDI provides coordinating support and led development of software infrastructure in partnership with the Global Organization for Earth System Science Portals. This material is based upon work supported by the National Science Foundation under Division of Atmospheric and Geospace Sciences Grant AGS-1522829. B. M. acknowledges support from the Regional and Global Climate Modeling Program of the U.S. Department of Energy’s Office of Science, Biological and Environmental Research Cooperative Agreement DE-FC02-97ER62402.
REFERENCES
Barnes, E. A., , and L. M. Polvani, 2013: Response of the midlatitude jets and of their variability to increased greenhouse gases in CMIP5 models. J. Climate, 26, 7117–7135, doi:10.1175/JCLI-D-12-00536.1.
Bender, F. A.-M., , V. Ramanathan, , and G. Tselioudis, 2012: Changes in extratropical storm track cloudiness 1983–2008: Observational support for a poleward shift. Climate Dyn., 38, 2037–2053, doi:10.1007/s00382-011-1065-6.
Booth, J. F., , C. M. Naud, , and A. D. Del Genio, 2013: Diagnosing warm frontal cloud formation in a GCM: A novel approach using conditional subsetting. J. Climate, 26, 5827–5845, doi:10.1175/JCLI-D-12-00637.1.
Boucher, O., and et al. , 2013: Clouds and aerosols. Climate Change 2013: The Physical Science Basis, T. F. Stocker et al., Eds., Cambridge University Press, 571–657.
Bretherton, C. S., 2015: Insights into low-latitude cloud feedbacks from high-resolution models. Philos. Trans. Roy. Soc. London, 373, 20140415, doi:10.1098/rsta.2014.0415.
Ceppi, P., , and D. L. Hartmann, 2015: Connections between clouds, radiation, and midlatitude dynamics: A review. Curr. Climate Change Rep., 1, 94–102, doi:10.1007/s40641-015-0010-x.
Ceppi, P., , M. D. Zelinka, , and D. L. Hartmann, 2014: The response of the Southern Hemisphere eddy-driven jet to future changes in shortwave radiation in CMIP5. Geophys. Res. Lett., 41, 3244–3250, doi:10.1002/2014GL060043.
Ceppi, P., , D. L. Hartmann, , and M. J. Webb, 2016: Mechanisms of the negative shortwave cloud feedback in middle to high latitudes. J. Climate, 29, 139–157, doi:10.1175/JCLI-D-15-0327.1.
Clement, A. C., , R. Burgman, , and J. R. Norris, 2009: Observational and model evidence for positive low-level cloud feedback. Science, 325, 460–464, doi:10.1126/science.1171255.
Dee, D. P., and et al. , 2011: The ERA-Interim reanalysis: Configuration and performance of the data assimilation system. Quart. J. Roy. Meteor. Soc., 137, 553–597, doi:10.1002/qj.828.
Eastman, R., , and S. G. Warren, 2013: A 39-yr survey of cloud changes from land stations worldwide 1971–2009: Long-term trends, relation to aerosols, and expansion of the tropical belt. J. Climate, 26, 1286–1303, doi:10.1175/JCLI-D-12-00280.1.
Georgakakos, K., , and R. Bras, 1984: A hydrologically useful station precipitation model. 1. Formulation. Water Resour. Res., 20, 1585–1596, doi:10.1029/WR020i011p01585.
Gillett, N. P., , and D. W. J. Thompson, 2003: Simulation of recent Southern Hemisphere climate change. Science, 302, 273–275, doi:10.1126/science.1087440.
Gordon, N. D., , J. R. Norris, , C. P. Weaver, , and S. A. Klein, 2005: Cluster analysis of cloud regimes and characteristic dynamics of midlatitude synoptic systems in observations and a model. J. Geophys. Res., 110, D15S17, doi:10.1029/2004JD005027.
Govekar, P. D., , C. Jakob, , and J. Catto, 2014: The relationship between clouds and dynamics in Southern Hemisphere extratropical cyclones in the real world and a climate model. J. Geophys. Res. Atmos., 119, 6609–6628, doi:10.1002/2013JD020699.
Grise, K. M., , and L. M. Polvani, 2014: Southern Hemisphere cloud–dynamics biases in CMIP5 models and their implications for climate projections. J. Climate, 27, 6074–6092, doi:10.1175/JCLI-D-14-00113.1.
Grise, K. M., , L. M. Polvani, , G. Tselioudis, , Y. Wu, , and M. D. Zelinka, 2013: The ozone hole indirect effect: Cloud radiative anomalies accompanying the poleward shift of the eddy-driven jet in the Southern Hemisphere. Geophys. Res. Lett., 40, 3688–3692, doi:10.1002/grl.50675.
Hurrell, J. W., 1995: Decadal trends in the North Atlantic Oscillation: Regional temperatures and precipitation. Science, 269, 676–679, doi:10.1126/science.269.5224.676.
Kay, J. E., , B. Medeiros, , Y.-T. Hwang, , A. Gettelman, , J. Perket, , and M. G. Flanner, 2014: Processes controlling Southern Ocean shortwave climate feedbacks in CESM. Geophys. Res. Lett., 41, 616–622, doi:10.1002/2013GL058315.
Klein, S. A., , and D. L. Hartmann, 1993: The seasonal cycle of low stratiform clouds. J. Climate, 6, 1587–1606, doi:10.1175/1520-0442(1993)006<1587:TSCOLS>2.0.CO;2.
Klein, S. A., , and C. Jakob, 1999: Validation and sensitivities of frontal clouds simulated by the ECMWF model. Mon. Wea. Rev., 127, 2514–2531, doi:10.1175/1520-0493(1999)127<2514:VASOFC>2.0.CO;2.
Klein, S. A., , Y. Zhang, , M. D. Zelinka, , R. Pincus, , J. Boyle, , and P. J. Gleckler, 2013: Are climate model simulations of clouds improving? An evaluation using the ISCCP simulator. J. Geophys. Res. Atmos., 118, 1329–1342, doi:10.1002/jgrd.50141.
Kushner, P. J., , I. M. Held, , and T. L. Delworth, 2001: Southern Hemisphere atmospheric circulation response to global warming. J. Climate, 14, 2238–2249, doi:10.1175/1520-0442(2001)014<0001:SHACRT>2.0.CO;2.
Lau, N.-C., , and M. W. Crane, 1995: A satellite view of the synoptic-scale organization of cloud properties in midlatitude and tropical circulation systems. Mon. Wea. Rev., 123, 1984–2006, doi:10.1175/1520-0493(1995)123<1984:ASVOTS>2.0.CO;2.
Lau, N.-C., , and M. W. Crane, 1997: Comparing satellite and surface observations of cloud patterns in synoptic-scale circulation systems. Mon. Wea. Rev., 125, 3172–3189, doi:10.1175/1520-0493(1997)125<3172:CSASOO>2.0.CO;2.
Li, Y., , D. W. J. Thompson, , Y. Huang, , and M. Zhang, 2014a: Observed linkages between the northern annular mode/North Atlantic Oscillation, cloud incidence, and cloud radiative forcing. Geophys. Res. Lett., 41, 1681–1688, doi:10.1002/2013GL059113.
Li, Y., , D. W. J. Thompson, , G. L. Stephens, , and S. Bony, 2014b: A global survey of the instantaneous linkages between cloud vertical structure and large-scale climate. J. Geophys. Res. Atmos., 119, 3770–3792, doi:10.1002/2013JD020669.
Loeb, N. G., , S. Kato, , W. Su, , T. Wong, , F. G. Rose, , D. R. Doelling, , J. R. Norris, , and X. Huang, 2012: Advances in understanding top-of-atmosphere radiation variability from satellite observations. Surv. Geophys., 33, 359–385, doi:10.1007/s10712-012-9175-1.
Medeiros, B., , and B. Stevens, 2011: Revealing differences in GCM representations of low clouds. Climate Dyn., 36, 385–399, doi:10.1007/s00382-009-0694-5.
Medeiros, B., , and L. Nuijens, 2016: Clouds at Barbados are representative of clouds across the trade wind regions in observations and climate models. Proc. Natl. Acad. Sci. USA, 113, E3062–E3070, doi:10.1073/pnas.1521494113.
Morcrette, J.-J., , and Y. Fouquart, 1986: The overlapping of cloud layers in shortwave radiation parameterizations. J. Atmos. Sci., 43, 321–328, doi:10.1175/1520-0469(1986)043<0321:TOOCLI>2.0.CO;2.
Myers, T. A., , and J. R. Norris, 2013: Observational evidence that enhanced subsidence reduces subtropical marine boundary layer cloudiness. J. Climate, 26, 7507–7524, doi:10.1175/JCLI-D-12-00736.1.
Norris, J. R., , and S. F. Iacobellis, 2005: North Pacific cloud feedbacks inferred from synoptic-scale dynamic and thermodynamic relationships. J. Climate, 18, 4862–4878, doi:10.1175/JCLI3558.1.
Norris, J. R., , R. J. Allen, , A. T. Evan, , M. D. Zelinka, , C. W. O’Dell, , and S. A. Klein, 2016: Evidence for climate change in the satellite cloud record. Nature, 536, 72–75, doi:10.1038/nature18273.
Pincus, R., , S. Platnick, , S. A. Ackerman, , R. S. Hemler, , and R. J. P. Hoffmann, 2012: Reconciling simulated and observed views of clouds: MODIS, ISCCP, and the limits of instrument simulators. J. Climate, 25, 4699–4720, doi:10.1175/JCLI-D-11-00267.1.
Polvani, L. M., , D. W. Waugh, , G. J. P. Correa, , and S.-W. Son, 2011: Stratospheric ozone depletion: The main driver of twentieth-century atmospheric circulation changes in the Southern Hemisphere. J. Climate, 24, 795–812, doi:10.1175/2010JCLI3772.1.
Qu, X., , A. Hall, , S. A. Klein, , and P. M. Caldwell, 2014: On the spread of changes in marine low cloud cover in climate model simulations of the 21st century. Climate Dyn., 42, 2603–2626, doi:10.1007/s00382-013-1945-z.
Qu, X., , A. Hall, , S. A. Klein, , and A. M. DeAngelis, 2015: Positive tropical marine low- cloud cover feedback inferred from cloud-controlling factors. Geophys. Res. Lett., 42, 7767–7775, doi:10.1002/2015GL065627.
Ramanathan, V., , R. D. Cess, , E. F. Harrison, , P. Minnis, , B. R. Barkstrom, , E. Ahmad, , and D. Hartmann, 1989: Cloud radiative forcing and climate: Results from the Earth Radiation Budget Experiment. Science, 243, 57–63, doi:10.1126/science.243.4887.57.
Rossow, W. B., , and R. A. Schiffer, 1999: Advances in understanding clouds from ISCCP. Bull. Amer. Meteor. Soc., 80, 2261–2288, doi:10.1175/1520-0477(1999)080<2261:AIUCFI>2.0.CO;2.
Simpson, I., , T. Shaw, , and R. Seager, 2014: A diagnosis of the seasonally and longitudinally varying midlatitude circulation response to global warming. J. Atmos. Sci., 71, 2489–2515, doi:10.1175/JAS-D-13-0325.1.
Taylor, K. E., , R. J. Stouffer, , and G. A. Meehl, 2012: An overview of CMIP5 and the experiment design. Bull. Amer. Meteor. Soc., 93, 485–498, doi:10.1175/BAMS-D-11-00094.1.
Thompson, D. W. J., , and J. M. Wallace, 2000: Annular modes in the extratropical circulation. Part I: Month-to-month variability. J. Climate, 13, 1000–1016, doi:10.1175/1520-0442(2000)013<1000:AMITEC>2.0.CO;2.
Thompson, D. W. J., , and S. Solomon, 2002: Interpretation of recent Southern Hemisphere climate change. Science, 296, 895–899, doi:10.1126/science.1069270.
Tselioudis, G., , B. Lipat, , D. Konsta, , K. Grise, , and L. Polvani, 2016: Midlatitude cloud shifts, their primary link to the Hadley cell, and their diverse radiative effects. Geophys. Res. Lett., 43, 4594–4601, doi:10.1002/2016GL068242.
Wall, C. J., , and D. L. Hartmann, 2015: On the influence of poleward jet shift on shortwave cloud feedback in global climate models. J. Adv. Model. Earth Syst., 7, 2044–2059, doi:10.1002/2015MS000520.
Weare, B. C., 2000: Near-global observations of low clouds. J. Climate, 13, 1255–1268, doi:10.1175/1520-0442(2000)013<1255:NGOOLC>2.0.CO;2.
Weaver, C. P., , and V. Ramanathan, 1997: Relationships between large-scale vertical velocity, static stability, and cloud radiative forcing over Northern Hemisphere extratropical oceans. J. Climate, 10, 2871–2887, doi:10.1175/1520-0442(1997)010<2871:RBLSVV>2.0.CO;2.
Wood, R., , and C. S. Bretherton, 2006: On the relationship between stratiform low cloud cover and lower-tropospheric stability. J. Climate, 19, 6425–6432, doi:10.1175/JCLI3988.1.
Yin, J. H., 2005: A consistent poleward shift of the storm tracks in simulations of 21st century climate. Geophys. Res. Lett., 32, L18701, doi:10.1029/2005GL023684.
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, doi:10.1175/JCLI-D-11-00248.1.
Zelinka, M. D., , S. A. Klein, , K. E. Taylor, , T. Andrews, , M. J. Webb, , J. M. Gregory, , and P. M. Forster, 2013: Contributions of different cloud types to feedbacks and rapid adjustments in CMIP5. J. Climate, 26, 5007–5027, doi:10.1175/JCLI-D-12-00555.1.
Zhang, Y., and et al. , 2012: Regional assessment of the parameter-dependent performance of CAM4 in simulating tropical clouds. Geophys. Res. Lett., 39, L14708, doi:10.1029/2012GL052355.
