Radiative driving of shallow return flows from ... - Wiley Online Library

PUBLICATIONS Journal of Advances in Modeling Earth Systems RESEARCH ARTICLE

Radiative driving of shallow return flows from the ITCZ

10.1002/2015MS000606 N. Nishant1, S. C. Sherwood1, and O. Geoffroy1 Key Points: Shallow meridional circulation is strongly controlled by radiative cooling Control of the circulation by SST is mediated by radiative cooling Clouds and vertical gradient of moisture and temperature above the PBL regulates the radiative cooling

Correspondence to: N. Nishant, [email protected]

Citation: Nishant, N., S. C. Sherwood, and O. Geoffroy (2016), Radiative driving of shallow return flows from the ITCZ, J. Adv. Model. Earth Syst., 8, 831–842, doi:10.1002/2015MS000606. Received 8 DEC 2015 Accepted 2 MAY 2016 Accepted article online 6 MAY 2016 Published online 29 MAY 2016

1

Climate Change Research Centre and ARC Centre of Excellence for Climate System Science, University of New South Wales, Sydney, New South Wales, Australia

Abstract The tropical overturning circulation over the central-east Pacific not only extends through the troposphere but also includes a shallow component outflowing from the ITCZ near the planetary boundary layer (PBL) top, the Shallow Meridional Circulation (SMC). Idealized simulations with the Weather Research and Forecasting (WRF) model, which qualitatively reproduce this two-component circulation, are used to investigate the mechanism driving the SMC by altering model parameterization schemes (PBL, cumulus and radiation), the solar heating rate, and the SST gradient. A radiative-driving model for subsidence is tested and found to quantitatively predict the altitude and strength of the SMC in all experiments. This includes experiments with a changed SST gradient, in which the changes in temperature profile and clouds near PBL top in the subsiding region altered radiative cooling so as to make the SMC vary in proportion to the SST gradient. These results indicate that the SMC is strongly controlled by radiative cooling rather than independently by SST patterns or convective processes, at least in this model setup. Experiments with altered solar heating of the atmosphere affected the strength of the SMC strongly, while changes in cumulus parameterization had a weaker effect. Parameterization changes influenced the SMC mainly by altering shallow cloud amount and humidity in subsidence regions, but vertical gradients of temperature also affect the radiative driving of the SMC in at least one case. The results highlight that atmospheric radiative heating is of prime importance for determining the character of deep and shallow overturning.

1. Introduction The large-scale Hadley-like circulation in the eastern Pacific Inter-Tropical Convergence Zone (ITCZ) in many studies has exhibited an additional shallow flow between 2 and 5 km of altitude, termed the shallow meridional circulation (SMC) [Zhang et al., 2004; Nolan et al., 2007]. An SMC over the eastern Pacific has been reported from in situ data [Yin and Albrecht, 2000; Zhang et al., 2004; Raymond et al., 2004], reanalysis data [Nicholson and Grist, 2003; Zhang et al., 2008], and idealized model simulations [Grabowski et al., 2000; Nolan et al., 2007]. The existence of SMCs has also been reported over regions like the Atlantic, western Africa [Zhang et al., 2006], the Indian subcontinent [Schott et al., 2002], and the central and western Pacific [Zhang et al., 2008]. Over the eastern Pacific ITCZ, Zhang et al. [2008] analyzed the seasonal cycle of the SMC using reanalysis data and reported that the SMC was the strongest during boreal fall. Yokoyama et al. [2014] over the same region showed that the JRA reanalysis indicates less shallow return flow (deeper overall circulation) than the ERA Interim and MERRA [Rienecker et al., 2011] reanalyses. Handlos and Back [2014] reported a strong shallow return flow from an independent method based on energy conservation.

C 2016. The Authors. V

This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.

NISHANT ET AL.

Numerical models have successfully simulated SMCs in both real and idealized frameworks. Nolan et al. [2007] simulated an SMC over the ITCZ and tested its sensitivity to the SST gradient and model physics. They reported a deepening and strengthening of the SMC when the SST gradient between the equator and the tropics was increased. They also suggested that the best representation of the SMC in the WRF model is achieved by using Yonsei University (YSU) [Hong et al., 2006] and Grell-Devenyi ensemble (GRELL) [Grell and Dvnyi, 2002] schemes for planetary boundary layer (PBL) and cumulus convection, respectively. However, this and other studies did not fully explain how cumulus schemes affect the SMC. SMCs, due to their moisture transport, can enhance the development of stratocumulus clouds [Nolan et al., 2007]. The radiative feedback associated with these clouds is a major source of spread for climate projections [Bony and Dufresne, 2005]. Results of Zhang et al. [2004] anticipated the participation of low level return flows associated with the SMC in air-sea interactions crucial for climate variability in the eastern

RADIATIVE DRIVING OF SHALLOW FLOWS

831

Journal of Advances in Modeling Earth Systems

10.1002/2015MS000606

Pacific. More recently, Sherwood et al. [2014] showed a connection of these shallow circulations and their net upward moisture transport to Earths equilibrium climate sensitivity and found that global climate models (GCMs) mostly underestimate them in comparison to reanalysis data. Looking at the climatic implications of the SMCs, it is crucial to understand the mechanism behind them, which will help in their better representation in GCMs. In spite of the above advancements, a quantitative or even qualitative explanation of the SMC remains lacking. The main cause hypothesized for the occurrence of SMCs is gradients in SST and surface pressure, a mechanism analogous to a sea breeze; however, some studies have implicated radiative cooling within the lower troposphere in driving SMC-like circulation features. Nolan et al. [2007] interpreted the SMC as a sea-breeze-type circulation caused by the strong Laplacian of the SST field, plus the fact that deep convection cannot always be sustained in the relatively cool and stable eastern Pacific part of the ITCZ. The variation in insolation over the eastern Pacific, together with ocean dynamics, leads to the observed temperature difference of approximately 8–128C between the equator and 308N (NOAA/ESRL http://www.esrl.noaa.gov/psd/). High temperature and low pressure near the ITCZ, and vice versa in the subtropics, creates greater geopotential thickness and lower surface pressure near the ITCZ as compared to subtropics, causing a reversal of the pressure gradient between the regions above the boundary layer, in turn driving shallow return flow out of ITCZ. Bretherton et al. [2005] and Nolan et al. [2007] reported the occurrence of the strongest SMCs away from the deep convection. They further added that the SMCs are suppressed or even eliminated when deep convection develops in the vicinity of the ITCZ, suggesting a competition between deep and shallow return flows. Circulations in general are responsive to changes in cloud cover and water vapor via radiative driving [Betts and Ridgway, 1988; Slingo and Slingo, 1988; Randall et al., 1989; Hartmann and Larson, 2002; Li et al., 2015; Voigt and Shaw, 2015]. In a similar way, shallow return flows could be radiatively driven. This can be taken as an alternate hypothesis to explain the SMCs. Due to the presence of clouds and/or a sharp gradient of humidity or temperature near the top of the boundary layer, the top of the moist layer cools strongly to the space. Such cooling must be balanced, on average, by warming from advection or convection. In regions with weak convection and rainfall, and given that horizontal variations of temperature in the tropics are small, the radiative cooling is approximately balanced by subsidence warming against a stable lapse rate. By continuity, air descending at a given level requires horizontal inflow of air above the descending region and outflow below it. If descent occurs near the PBL top in the subtropics, this can result in the observed SMC. Wang et al. [2005] concluded that the radiative cooling due to shallow clouds strengthens the SMC, but failed to quantify if the radiative effects dominate over the SST and pressure gradient effect, and if the SMC can be predicted by radiation. Results of Muller and Held [2012] emphasized atmospheric radiative cooling associated with shallow clouds as the cause for transient shallow circulations and self-aggregation of convection, but since their study was based on an idealized small-domain simulation of radiativeconvective equilibrium with neither rotation nor surface temperature gradients, it is not clear how relevant their results are to the observed steady cross-equatorial shallow circulation. In summary, studies emphasizing the role of SST and surface pressure gradients in driving the SMCs failed to evaluate the role of radiation, whereas the studies noting the influence of radiative cooling due to shallow clouds on the SMC did not inspect the roles of radiation and SST gradients more generally. Vertical gradients of humidity or temperature in the lower troposphere, in addition to shallow clouds, influence the radiative cooling and hence could help drive the SMC. Also, the overall roles of radiative versus SST driving have not been quantitatively compared. Here we simulate the SMC in an idealized model configuration, compare the results with observations, and conduct a suite of experiments with the model to identify the roles of various processes in driving the SMC.

2. Methodology To test our hypotheses, we choose to vary the PBL and convective parameterizations in a model in order to test the degree of control these local physical processes have on the final outcome. These parameterizations not only work strongly in the ascending/convective region but are also capable of changing the moisture and temperature fields at low levels throughout the domain, which may have an indirect effect on radiation. In order to independently verify the roles of radiation and SST gradient, we additionally test the

NISHANT ET AL.


832


10.1002/2015MS000606

sensitivity of the SMC to the radiation parameterizations, the solar heating, and the SST gradient, for a fixed set of physical schemes. We now present the details of this. 2.1. WRF Setup and Experiments Flows are simulated using the Weather Research and Forecasting (WRF) model [Skamarock et al., 2005] version 3.5.1. The idealized Hadley-cell setup of Nolan et al. [2007] is adopted, in which the boundaries of the domain are defined as the equator and 308S. The surface is ocean with fixed SST, which linearly decreases from 308C on the equator to 208C at 308S latitude. While this is not the most realistic setup one could envision, we will argue that it captures the essential aspects of the SMCs. A zonally periodic and meridionally symmetric domain of 3200 3 4000 km 3 19 km is used, with a horizontal resolution of 20 km. In the vertical, 40 model levels are used with the model top at 19 km. The WRF double-moment six-class scheme (WDM6) is used for microphysics, and second-order horizontal diffusion is used for turbulent diffusion. The simulations include the effects of Earth’s rotation. The model is run for a period of 120 days. After approximately 20–30 days, the model appears to reach a quasi-steady state, but to be sure, the initial 90 days are considered as spin-up time. In order to understand the roles of SST gradient and radiative cooling on the mechanism of the SMC separately, we do two types of experiments. In the first type, we change the SST gradient between the north and the south boundaries of the domain, whereas in the second we change the model physics and solar constant. For the first type, three experiments are done. In the first two experiments, the SST difference between the equator and 308S is set to be 108C and 68C respectively with SST 5 308C at the equator, whereas for the third experiment the weaker gradient is employed but with temperature at equator maintained at 268C. For the second experiment type, three sets of experiments are set up to investigate the sensitivity of the SMC to (1) PBL and cumulus parameterizations, (2) shortwave (SW) radiation parameterizations, and (3) the magnitude of solar constant. In the first set, six experiments are done using all combinations among two PBL and three cumulus parameterizations. The YSU and Mellor-Yamada-Janjic (MYJ) [Janji, 1990] schemes are used for the PBL, and the Kain-Fritsch (KF) [Kain, 2004], Betts-Miller-Janiac (BMJ) [Betts and Miller, 1993], and GRELL ensemble schemes are used for cumulus convection. These six experiments are all performed with the Dudhia [Dudhia, 1989] and Rapid Radiative Transfer Model (RRTM) [Mlawer et al., 1997] schemes for shortwave (SW) and longwave radiation, respectively. In the second set of experiments, a ‘‘Base’’ pair of convection and PBL parameterizations is chosen which yields the closest representation of SMC and cloud cover to the observations, and further runs are then conducted with these using three alternate SW radiation parameterizations: Goddard [Shi et al., 2010], RRTM for Global Climate models (RRTMG), and Geophysical fluid Dynamics Laboratory (GFDL) [Gu et al., 2011]. Experiments were also done with two different longwave radiation parameterizations, but they are not reported here because they showed negligible changes in radiative heating or the SMC. Finally, for the third set of experiments, the magnitude of the solar constant S is changed, again using the Base physics set, from its default value of 1368.22 W/m2 to values 30%, or 50%, above or below this (four experiments in total). All the experiments are listed in Table 1. 2.2. Reanalysis Data In order to ensure that the model-simulated SMC is relevant to the real world, we compare several model fields to observations. For wind and moisture, we compare to ERA-Interim reanalysis data [Dee et al., 2011]. Most of the observational studies noted in section 1 were done using older data sets, making it useful also to confirm that the SMC is present in a newer reanalysis. Multiple studies have reported the existence of the SMC over the eastern Pacific, and since the ITCZ is relatively zonally invariant there over a wide range of longitudes, we select this study area which we take from 208N to 208S and 958W to 1608W. Wind and humidity fields associated with the SMC are analyzed using 6-hourly data over the 14 year period from January 2000 to December 2014. We find that the SMC can be observed in all four seasons. However, during MAM, a shallow flow appears to originate between 2 and 5 km but later merges with the deep circulation thus becoming difficult to disentangle and clearly identify as an SMC. For this reason, MAM data are omitted from our climatology. Since the ITCZ in the eastern Pacific study region sits north of the equator, the main

NISHANT ET AL.


833


10.1002/2015MS000606

Table 1. List of Experiments PBL Scheme Name of Experiment 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

YSU 1 BM/Base YSU 1 KF YSU 1 GRELL MYJ 1 BM MYJ 1 KF MYJ 1 GRELL Goddard RRTMG GFDL S 1 30% S 2 30% S 1 50% S 2 50% SST6a SST6b

YSU 冑冑冑

冑冑冑冑冑冑冑冑冑

MYJ

Cumulus Scheme BM

GRELL

冑

冑冑冑

冑

冑冑冑冑冑冑冑冑冑

冑

冑

SW Radiation Scheme KF 冑

冑

Dudhia

RRTMG

冑冑冑冑冑冑冑

Goddard

GFDL

冑

冑冑冑冑冑冑

冑

Solar Constant

SST Gradient

Equator SST (8C)

S S S S S S S S S S 1 30% S 2 30% S 1 50% S 2 50% S S

10 10 10 10 10 10 10 10 10 10 10 10 10 6 6

30 30 30 30 30 30 30 30 30 30 30 30 30 26 30

overturning circulation in the observations extends south of the ITCZ and will be compared with the model circulation where the upwelling is on the model equator. The GCM-Oriented Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observation (CALIPSO) cloud product (GOCCP) [Chepfer et al., 2010] data from one year (2007) is used to evaluate the credibility of modelsimulated clouds. CALIPSO flies as part of NASA’s A-Train constellation. It was launched to retrieve cloud and aerosol data in order to understand Earth’s radiation budget and climate. GOCCP is a cloud diagnostic designed to be consistent with the ones simulated by GCMs for model-observation comparisons.

3. Results 3.1. Comparison of the SMC Structure With Observations and the Reanalysis Our first step was to conduct the first six simulations of Table 1. The combination of YSU boundary layer and BM cumulus convection parameterizations yielded the best comparison to reanalysis data and also best represented the cloud amount. As a result, we chose this simulation as our Base setup for subsequent testing. Figure 1 shows meridional wind (V), relative humidity (RH) and cloud fraction (CF) from the WRF Base case, and from the observational data sets. For WRF, these atmospheric fields are zonal and temporal means over the last 30 days of the simulation. An SMC can be seen both in WRF (Figure 1a) and the observations (Figure 1b). The simulated flow includes a deep circulation characterized by rising air (at the equator), which after ascending flows poleward mainly at altitudes between 9 and 13 km. After this, sinking motion takes place (158S–208S), followed by equatorward flow near the surface. Along with this deep circulation, an SMC can also be observed, with a second poleward flow between 4 and 6 km originating from the same location as the deep circulation. Similarly, both a deep circulation (10–14 km) and SMC (2–5 km) can be observed in ERA-Interim reanalysis (Figure 1b); however, the structures of the two simulations are not identical. Convergence in the ERA-Interim reanalysis remains below 2 km altitude after which divergence begins, whereas in WRF the convergence reaches approximately 4 km. Also, the peak magnitude of V in WRF simulation is twice as strong as in the ERA-Interim reanalysis and occurs nearly 2 km higher, possibly because of the high SST gradient specified in the simulation following Nolan et al. [2007]. The observed SST difference between the equator and 308S in the eastern Pacific is closer to 5–68C, adopted below in our first sensitivity test. We also note that the deep outflow is 1–2 km lower in WRF than the reanalysis. The atmospheric boundary layer (within 2 km of the surface) shows the highest values of RH in both ERAInterim reanalysis (Figure 1d) and WRF (Figure 1c) compared to higher levels. In regions of deep convection, moisture is transported to higher altitudes leading to relatively high RH aloft there, but low humidity elsewhere above the boundary layer. The high values of RH within the boundary layer are associated with a

NISHANT ET AL.


834


10.1002/2015MS000606

Figure 1. Zonal and temporal means of (a, b) meridional wind (V), (c, d) relative humidity (RH), and (e, f) cloud fraction. Here ‘‘a/c/e’’ represent WRF, ‘‘b/d’’ represent ERA-Interim, and ‘‘f’’ represents CALIPSO-GOCCP in the central/east Pacific. CALIPSO-GOCCP shows the total annual mean, whereas ERA-Interim shows annual mean without MAM. For WRF, the means are calculated for the last 30 days of the simulation.

sharp vertical gradient near 3–4 km altitude away from the equator in WRF (Figure 1c). This gradient in RH contributes to radiative cooling near the top of the PBL. At the southern boundary of the domain, we get a strong subtropical jet, which gives rise to instabilities. When viewed in time lapse, storms can be seen passing in the zonal direction through this part of the domain during the entire simulation, analogous to behavior in the midlatitude storm tracks (which in nature occur significantly farther poleward). These storms produce convection, and a secondary region of upward transport of moisture resulting in the formation of stratocumulus clouds in this region (Figure 2). Similar instabilities and RH pattern were also reported by Nolan et al. [2007] in their simulations. The observed RH pattern in Figure 1d is similar to that simulated with WRF (Figure 1c), except for the secondary moist region which is not seen in the observations at these latitudes and the intensity of the vertical RH gradient, which is approximately two times higher in the WRF simulation. Since shallow clouds are expected to influence the SMC [Muller and Held, 2012; Wang et al., 2005], it is particularly important to ensure that the model simulates similar shallow cloud amount seen in observations. Cloud fraction follows a pattern similar to that of RH: most clouds are located near the equator in the WRF simulation (Figure 1e) but near 58N in the satellite data (Figure 1f) due to the different location of the SST maximum. Also the low level clouds in WRF are vertically and meridionally confined as compared to the satellite data. In the respective convergence zones, the model cloud fraction is much higher than observed, likely because of the strong and linear SST gradient specified in the simulation, which produces a very narrow ITCZ.

NISHANT ET AL.


835


10.1002/2015MS000606

In spite of these problems, modelsimulated shallow cloud cover for the Base case averaged below 2.5 km altitude was comparable with that observed (Figures 1e and 1f). Since the model-simulated SMC and other related fields qualitatively resemble those observed, we conclude that the simulations are adequate to explore the mechanisms controlling SMC circulations via the remaining experiments listed in Table 1. 3.2. Role of Radiative Cooling in Driving Circulations In the tropics, as a consequence of the smallness of the Coriolis parameter, horizontal temperature gradients are Figure 2. Snapshot of spatial cloud distribution over the model domain. small above the boundary layer [Sobel et al., 2001]. Due to this, the temperature profiles in deep convective and nonprecipitating regions are linked and approximately follow a moist adiabat. However, temperature in the subsiding region can only obey this if diabatic cooling reduces the potential temperature of the air at precisely the rate demanded by a moist adiabat (which also balances the latent heat released in the ascending region). In the subtropics of Earth, this diabatic cooling consists largely of net radiative cooling, but with local turbulent fluxes entering the budget at least within the planetary boundary layer, and horizontal divergence of energy fluxes by large-scale eddies also playing a role. In our experiments, the latter process may not be realistically represented due to the limited domain size. To close the mass budget, divergence from the convective zone must occur in order to compensate for the nonprecipitating region subsidence, which results in deep convection [Hartmann and Larson, 2002]. Changes in radiative cooling could thus have a considerable effect on the deep atmospheric circulation [Voigt and Shaw, 2015]. Similarly, shallow circulations can be driven by the vertical profile of radiative cooling above the boundary layer [Muller and Held, 2012]. In order to test the radiative-driving idea for the SMC, we compare the divergence of meridional wind simulated by the model to an estimated, radiatively driven mass divergence in which eddy heat fluxes are neglected. Radiative calculations of the subsidence and mass divergence have been adapted from Sherwood and Meyer [2006] [see also Folkins et al., 2002]. Time-averaged subsidence xrad in the nonconvective region is constrained by H dH 21 xrad 5QR : (1) T dp Here H, QR , T, and p denote potential temperature, radiative cooling rate, temperature, and pressure, respectively. The time-averaged upward mass flux in the convective region and downward flux (subsidence xrad) in the clear-sky region are taken to be equal, such that the radiatively driven mass divergence (DIVRAD) from the convective region can be represented as a derivative of xrad: DIVRAD 5
; dp

(2)

where < : > is the mean over the subsidence region. 3.3. The SMC in the Simulations and the Radiative Model Figure 3 compares the profiles of average divergence of meridional wind (DIV) with those of radiatively driven divergence (DIVRAD) calculated from (1)–(2), for all the experiments. Here DIV and DIVRAD can be interpreted as the true and predicted V, respectively. DIV is calculated as the mean of the derivative of meridional wind (dV/dy) over the ascending region (equator-108S), whereas DIVRAD is calculated over the

NISHANT ET AL.


836


10.1002/2015MS000606

Figure 3. (a, c, e, g) Divergence of meridional wind DIV and (b, d, f, h) radiative divergence DIVRAD, horizontally averaged within the study region for experiments with different: (a, b) SST and ERA-Interim, (c, d) PBL and cumulus parameterizations, (e, f) SW radiation parameterizations, and (g, h) insolation.

NISHANT ET AL.


837


10.1002/2015MS000606

subsidence region (108S–308S). All the simulations, irrespective of the parameterization schemes, SST gradient or solar constant, produce both deep and shallow circulations. However, the strength and altitude of the SMC varies among the simulations. Importantly, satisfactory agreement can be observed (Figure 3, left and right) between true and radiatively predicted V, indicating the assumptions underlying (1)–(2) are approximately valid. This means that, to first approximation, the variations in SMC among the WRF runs require corresponding variations in radiative cooling in the subsiding part of the SMC circulation. DIV calculated from the ERA-Interim data (Figure 3a) shows a similar pattern as the Base case, but both the deep and shallow circulations are weaker in magnitude. As expected, a reduction of the SST gradient (Figures 3a and 3b, SST6a and SST6b) leads to weakening of the SMC. A halving of the SST gradient produced roughly a halving of the SMC. SST at the equator also had an impact on the altitude of the shallow circulation. Higher SST at the equator resulted in higher altitude of the SMC and vice versa (Figures 3a and 3b). The altitude of the shallow circulation is generally lower when the MYJ PBL scheme is used (Figures 3c and 3d), in both the actual and radiative profiles, than with the YSU scheme. This may reflect differences in the depth of the boundary layer transport in the two schemes. There is not much difference among the magnitudes of the SMC in the six profiles from the first set of experiments (Figures 3c and 3d), which indicates relatively small changes in radiative cooling when the PBL or CU schemes are changed. The simulations with YSU 1 GR and YSU 1 KF schemes showed weaker SMC than the other combinations, the reason being an additional much shallower weak outflow (near 2 km, see Figure 3c) in these two simulations. This made the circulation trimodal, leading to the overall weakening of the SMC at the measured level. In general, stronger insolation leads to more warming of the atmosphere; however, this warming is unevenly distributed in the atmosphere due to clouds, moisture, and attenuation of incoming sunlight. Therefore, to test the effect of heating on the SMC, we did one set of four experiments in which we varied the solar constant, and another set with different solar radiation schemes (shown in Figures 3g and 3h, and 3e and 3f). In the subsiding region, experiments with increased insolation experience greater warming due to the absorption by water vapor in the lower troposphere. This decreases the net radiative cooling, resulting in net reduction of both shallow and deep circulations. The opposite effect is seen with a reduction in the amount of insolation, and both changes are consistently larger when the solar constant is changed by 650% as compared to 630%. When the SW radiation schemes are changed, this leads to negligible change in deep circulations but significant changes to the SMC. The SMC slows by approximately 50% at 5 km altitude (Figures 3e and 3f), when the Goddard, RRTMG, or GFDL schemes are used in place of Dudhia scheme. These experiments therefore underline the potential impact of radiative transfer errors on the SMC, although the Dudhia scheme may be an outlier in this respect, as the others produce similar results. Detailed examination of the behavior of various schemes is beyond the scope of this study, but differences could have arisen from different spectroscopy or band models [DeAngelis et al., 2015] or different treatment of clouds. In order to investigate if the weakening of the SMC is followed by the intensification of flows at the other levels, the meridional mass stream functions for the reduced-SST-gradient experiments are calculated and compared with those of the Base case (Figure 4). A clear weakening of both shallow and deep meridional circulations can be observed in both experiments (Figures 4b and 4e) compared to the Base. A slight increase in the anomalous mean meridional mass stream function (Figure 4c) can be noticed in the SST6a experiment at the midtropospheric flows (8 km), however, its magnitude is much too weak to compensate for the weakening of the SMC. DIVRAD and DIV calculated near the SMC level are highly correlated (r 5 0.86, Figure 5a). To obtain them we average the nonnegative data at heights from 3 to 6 km, and from 08S to 108S for DIV and 108S to 308S for DIVRAD. Investigation of the relative importance of radiative cooling and static stability on DIVRAD revealed a stronger control by radiative cooling (Figure 6). Considering this along with the strong correlation between DIV and DIVRAD (Figure 5a), we infer a dominant role for radiative cooling in modulating the simulated SMC. Even the weak SMCs of the low-SST-gradient experiments each show a commensurate reduction in the radiative cooling, indicating that the control of the circulation by SST is mediated by radiative cooling. The correlation between DIV and DIVRAD is almost as strong (r 5 0.81) when the subsidence region is taken only from 108S to 208S, where the mass and energy budgets can be influenced by eddy fluxes associated

NISHANT ET AL.


838


10.1002/2015MS000606

Figure 4. Mean meridional mass stream function: (a, d) Base, (b) SST6a, (e) SST6b, and anomalous meridional mass stream function: (c) SST6a-Base and (f) SST6b-Base.

with the strongly baroclinic region near the southern end of the domain. While it is not clear how similar these modeled fluxes are to the real ones, this result nonetheless suggests that such fluxes do not disturb the primary balance embodied in equations (1)–(2). These results beg the question of what controls the radiative cooling rate in the key region. Radiative cooling at shallow levels is influenced by the vertical humidity and temperature gradients, as well as by cloud cover [Lilly and Schubert, 1980; Mapes and Zuidema, 1996]. The relationships between the relative humidity (RH) gradient and cloud fraction (CF) and DIVRAD in the SMC altitude range are shown in Figures 5b and 5c. The RH gradient is defined as the difference between RH averaged between 1–2 km and 4–5 km, and CF is

Figure 5. Relationship between radiative divergence DIVRAD and (a) divergence of meridional wind DIV, (b) relative humidity gradient dRH/dZ, and (c) cloud fraction.

NISHANT ET AL.


839


10.1002/2015MS000606

Figure 6. True radiative divergence DIVRAD for each simulation versus DIVRAD calculated with (a) h fixed to the mean profile across all the simulations, and (b) QR fixed to the mean profile.

defined as the mean cloud below 4 km altitude. Both a sharp RH gradient (Figure 5b) and shallow cloud amount (Figure 5c) are correlated with the strength of radiative driving of the SMC. A stronger moisture gradient at shallow levels, or greater cloud cover, induces radiative cooling which aids divergence. These relationships hold in all the simulations except for one outlier, experiment MYJ 1 BM, which shows more cloud but less SMC and radiative driving. Also, the relationship of RH gradient and clouds to DIVRAD is evident in SST6b simulation but not in the SST6a simulation, which shows an RH gradient similar to Base simulation, but still shows much lower DIVRAD and DIV. The lower DIVRAD in the SST6a simulation as compared to those of SST6b and Base likely results from the different temperature profiles, associated with the different SST in the ascending region for these simulations. To verify this, we estimate the impact of changes in air temperature between the Base and SST6a experiments on radiative cooling and DIVRAD using a simple grey body model of the infrared radiation flows in and out of a layer from 3 to 6 km altitude in the subsidence region. First, rough estimates of the fluxes in and out of the layer through the top and bottom are calculated using the MODTRAN [Berk et al., 1998] model with a default tropical sounding. From these, using the fact that each outgoing flux equals ð12ÞFin 1rT 4 where Fin is the incoming flux on the opposite side and the emissivity, we infer 50:33 and T 5 269 K, which is sufficiently close to the mean layer temperature of 3–6 km from the WRF Base case. We then compute the effective radiating temperatures of the two incoming fluxes and assign corresponding effective radiating altitudes where these temperatures match the WRF Base profile, at 2.2 and 12.8 km. Finally, we perturb the three temperatures according to the WRF-simulated changes SST6a-Base and compute the change in fluxes and flux convergence; the latter yields the anticipated reduction in DIVRAD due to changed temperature at fixed composition. The result is a reduction of 0.074 s21 [SST6a-Base], which compares reasonably well with the actual reduction in DIVRAD of 0.094 s21. A similar calculation for case SST6b yields a much smaller change, in accord with the fact that most of the slowdown in this experiment could be explained by reduced cloud cover. Thus, while (at least in our experiments) changes in model physics, SST gradient and incoming solar radiation affect the SMC primarily through cloud and water vapor fields via their radiative effects, changes in surface temperature may strongly affect it through the air temperature field in the absence of changes to relative humidity or cloud cover.

4. Conclusion and Discussion While we have largely duplicated the experimental setup of Nolan et al. [2007] to examine shallow meridional circulations (SMC), our study reinterprets those of this and other previous studies in emphasizing the role of radiation. We find that radiative cooling associated with vertical moisture and temperature gradients are the principal driving mechanism behind the SMC.

NISHANT ET AL.


840


10.1002/2015MS000606

Several lines of reasoning lead to this conclusion. First, experiments with changes in the solar constant but with fixed surface temperature, in which only differential heating and cooling of the atmosphere was changed, produced significant changes in the strength of the SMC. Increased insolation warms the moist shallow layers and thus reduces the net infrared cooling there. This weakens the SMC, and vice versa when insolation is weakened. Changes in SMC caused by varying parameterization schemes were also consistent with radiative control. These changes were at least as large with changes to radiation schemes (changing by up to 50%) as with changes to cumulus convection or boundary layer schemes; indeed with the MYJ boundary layer scheme, changes in cumulus scheme had almost no effect. These results explicitly show that atmospheric radiative cooling plays an important role in driving the meridional SMC. In general, most changes to the radiative driving of the SMC occurred due to changes in subsidence-region humidity or cloud cover, even in experiments where the radiation scheme or insolation was changed. In particular, the Goddard, RRTMG, and GFDL radiation schemes led to weaker humidity gradients and less shallow cloud than did the Dudhia scheme, accounting for most of the SMC difference. This suggests that the changes to the SMC begin with a subtle alteration of temperature profiles which then affects humidity and cloud profiles near and above the subtropical boundary layer, finally driving the change in circulation. While it cannot be ruled out that for some experiments the SMC changes were driven from the convecting region, with changes in subsidence-region radiation responding cooperatively, it seems unlikely that cloud and humidity profiles would consistently cooperate in this way. In any case, our results are not generally consistent with the concept of simple Newtonian cooling anomalies driven by temperature changes, which would allow radiation to passively respond to dynamical forcing, but instead require an active role for radiation. The experiments with a nearly halved SST gradient showed a roughly commensurate drop in SMC strength, at first suggesting strong SST control of the circulation. However, there was also a near halving of radiative driving, due to the change in the shallow cloud cover and changed temperature structure in the subsidence region. The decrease in radiative cooling fully explained the decrease in SMC in this case. The vertical gradients in temperature and moisture above the boundary layer had a significant effect on radiative cooling, in addition to shallow clouds which have been the focus of past studies considering radiation. Our study thus demonstrates more conclusively the role of radiative cooling in causing a steady meridional SMC. A limitation is that we have considered only one model configuration, with parameterized physics; if boundary layer energy transports reach higher into the SMC in nature than in the model, radiative driving in nature might be less dominant. Nonetheless this result help explain variations in the strength of the SMC in models, which appear connected to climate sensitivity [Sherwood et al., 2014]. Our result does not support or refute the suggested mechanism of that study but does suggest that the differences they found in large-scale lower tropospheric mixing among GCMs may have been driven radiatively by differences in climatological low cloud or humidity fields. This would provide a pathway for cloud amounts to affect cloud feedback by driving mixing. However, if cloud amount could modulate feedback strength in some other more direct way, this could offer a different explanation for the correlation between SMC and feedback noted by that study. These issues deserve further investigation. In any case, most GCMs lack a realistic representation of moisture and clouds at the shallow levels, due to unreliability of the boundary layer and shallow cumulus schemes. The present results indicate that to improve these circulations in the models will require attention to those aspects of model physics that control radiation above the top of the planetary boundary layer.

Acknowledgments We acknowledge the European Centre for Medium-Range Weather Forecasts and National Center for Atmospheric Research, for providing the ERAInterim Reanalysis data, WRF model and CALIPSO-GOCCP data set, as well as support for the WRF model. In addition, we would also like to thank the Australian NCI and ARC Discovery Project DP140101104 for providing computational resources for this work.

NISHANT ET AL.

References Berk, A., L. S. Bernstein, G. P. Anderson, P. K. Acharya, D. C. Robertson, J. H. Chetwynd, and S. M. Adler-Golden (1998), MODTRAN cloud and multiple scattering upgrades with application to AVIRIS, Remote Sens. Environ., 65, 367–375, doi:10.1016/S0034-4257(98)00045-5. Betts, A. K., and W. Ridgway (1988), Coupling of the radiative, convective, and surface fluxes over the equatorial pacific, J. Atmos. Sci., 45(3), 522–536, doi:10.1175/1520-0469(1988)0452.0.CO;2. Betts, A. K., and M. J. Miller (1986), A new convective adjustment scheme, Part II: Single column tests using GATE wave, BOMEX, ATEX, and Arctic Airmass data sets, Quart. J. Roy. Meteor. Soc., 112, 693–710. Bony, S., and J.-L. Dufresne (2005), Marine boundary layer clouds at the heart of tropical cloud feedback uncertainties in climate models, Geophys. Res. Lett., 32, L20806, doi:10.1029/2005GL023851. Bretherton, C. S., P. N. Blossey, and M. Khairoutdinov (2005), An energy-balance analysis of deep convective self-aggregation above uniform SST, J. Atmos. Sci., 62(12), 4273–4292, doi:10.1175/JAS3614.1. 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, doi:10.1029/2009JD012251.


841


10.1002/2015MS000606

DeAngelis, A. M., X. Qu, M. D. Zelinka, and A. Hall (2015), An observational radiative constraint on hydrologic cycle intensification, Nature, 528(7581), 249–253, doi:10.1038/nature15770. Dee, D. P., et al. (2011), The ERA-interim reanalysis: Configuration and performance of the data assimilation system, Q. J. R. Meteorol. Soc., 137(656), 553–597, doi:10.1002/qj.828. Dudhia, J. (1989), Numerical study of convection observed during the winter monsoon experiment using a mesoscale two-dimensional model, J. Atmos. Sci., 46(20), 3077–3107, doi:10.1175/1520-0469(1989)0462.0.CO;2. Folkins, I., K. K. Kelly, and E. M. Weinstock (2002), A simple explanation for the increase in relative humidity between 11 and 14 km in the tropics, J. Geophys. Res., 107(D23), 4736, doi:10.1029/2002JD002185. Grabowski, W. W., J.-I. Yano, and M. W. Moncrieff (2000), Cloud resolving modeling of tropical circulations driven by large-scale SST gradients, J. Atmos. Sci., 57(13), 2022–2040, doi:10.1175/1520-0469(2000)0572.0.CO;2. Grell, G. A., and D. Dvnyi (2002), A generalized approach to parameterizing convection combining ensemble and data assimilation techniques, Geophys. Res. Lett., 29(14), 1693, doi:10.1029/2002GL015311. Gu, Y., K. N. Liou, S. C. Ou, and R. Fovell (2011), Cirrus cloud simulations using WRF with improved radiation parameterization and increased vertical resolution, J. Geophys. Res., 116, D06119, doi:10.1029/2010JD014574. Handlos, Z. J., and L. E. Back (2014), Estimating vertical motion profile shape within tropical weather states over the oceans, J. Clim., 27(20), 7667–7686, doi:10.1175/JCLI-D-13-00602.1. Hartmann, D. L., and K. Larson (2002), An important constraint on tropical cloud - climate feedback, Geophys. Res. Lett., 29(20), 1951, doi: 10.1029/2002GL015835. Hong, S.-Y., Y. Noh, and J. Dudhia (2006), A new vertical diffusion package with an explicit treatment of entrainment processes, Mon. Weather Rev., 134(9), 2318–2341, doi:10.1175/MWR3199.1. Janji, Z. I. (1990), The step-mountain coordinate: Physical package, Mon. Weather Rev., 118(7), 1429–1443, doi:10.1175/15200493(1990)1182.0.CO;2. Kain, J. S. (2004), The Kain-Fritsch convective parameterization: An update, J. Appl. Meteorol., 43(1), 170–181, doi:10.1175/15200450(2004)0432.0.CO;2. Li, Y., D. W. J. Thompson, and S. Bony (2015), The influence of atmospheric cloud radiative effects on the large-scale atmospheric circulation, J. Clim., 28(18), 7263–7278, doi:10.1175/JCLI-D-14-00825.1. Lilly, D. K., and W. H. Schubert (1980), The effects of radiative cooling in a cloud-topped mixed layer, J. Atmos. Sci., 37(2), 482–487, doi: 10.1175/1520-0469(1980)0372.0.CO;2. Mapes, B. E., and P. Zuidema (1996), Radiative-dynamical consequences of dry tongues in the tropical troposphere, J. Atmos. Sci., 53(4), 620–638, doi:10.1175/1520-0469(1996)0532.0.CO;2. Mlawer, E. J., S. J. Taubman, P. D. Brown, M. J. Iacono, and S. A. Clough (1997), Radiative transfer for inhomogeneous atmospheres: RRTM, a validated correlated-k model for the longwave, J. Geophys. Res., 102(D14), 16,663–16,682, doi:10.1029/97JD00237. Muller, C. J., and I. M. Held (2012), Detailed investigation of the self-aggregation of convection in cloud-resolving simulations, J. Atmos. Sci., 69(8), 2551–2565, doi:10.1175/JAS-D-11-0257.1. Nicholson, S. E., and J. P. Grist (2003), The seasonal evolution of the atmospheric circulation over West Africa and Equatorial Africa, J. Clim., 16(7), 1013–1030, doi:10.1175/1520-0442(2003)0162.0.CO;2. Nolan, D. S., C. Zhang, and S.-h. Chen (2007), Dynamics of the shallow meridional circulation around intertropical convergence zones, J. Atmos. Sci., 64(7), 2262–2285, doi:10.1175/JAS3964.1. Randall, D. A., Harshvardhan, D. A. Dazlich, and T. G. Corsetti (1989), Interactions among radiation, convection, and large-scale dynamics in a general circulation model, J. Atmos. Sci., 46(13), 1943–1970, doi:10.1175/1520-0469(1989)0462.0.CO;2. Raymond, D. J., S. K. Esbensen, C. Paulson, M. Gregg, C. S. Bretherton, W. A. Petersen, R. Cifelli, L. K. Shay, C. Ohlmann, and P. Zuidema (2004), Epic2001 and the coupled ocean–atmosphere system of the tropical east pacific, Bull. Am. Meteorol. Soc., 85(9), 1341–1354, doi: 10.1175/BAMS-85-9-1341. Rienecker, M. M., et al. (2011), MERRA: NASA’s modern-era retrospective analysis for research and applications, J. Clim., 24(14), 3624–3648, doi:10.1175/JCLI-D-11-00015.1. Schott, F. A., M. Dengler, and R. Schoenefeldt (2002), The shallow overturning circulation of the Indian Ocean, Prog. Oceanogr., 53(1), 57–103. Skamarock, W. C., J. B. Klemp, J. Dudhia, D. O. Gill, D. M. Barker, W. Wang, and J. G. Powers (2008), A description of the Advanced Research WRF version 3, NCAR Tech, Note NCAR/TN-475+STR, 113 pp. Sherwood, S. C., and C. L. Meyer (2006), The general circulation and robust relative humidity, J. Clim., 19(24), 6278–6290, doi:10.1175/ JCLI3979.1. Sherwood, S. C., S. Bony, and J.-L. Dufresne (2014), Spread in model climate sensitivity traced to atmospheric convective mixing, Nature, 505(7481), 37–42, doi:10.1038/nature12829. Shi, J. J., et al. (2010), WRF simulations of the 20–22 January 2007 snow events over Eastern Canada: Comparison with in situ and satellite observations, J. Appl. Meteorol. Climatol., 49(11), 2246–2266, doi:10.1175/2010JAC2282.1. Slingo, A., and J. M. Slingo (1988), The response of a general circulation model to cloud longwave radiative forcing. I: Introduction and initial experiments. Q. J. R. Meteorol. Soc., 114(482), 1027–1062, doi:10.1002/qj.49711448209. Sobel, A. H., J. Nilsson, and L. M. Polvani (2001), The weak temperature gradient approximation and balanced tropical moisture waves, J. Atmos. Sci., 58(23), 3650–3665, doi:10.1175/1520-0469(2001)0582.0.CO;2. Voigt, A., and T. A. Shaw (2015), Circulation response to warming shaped by radiative changes of clouds and water vapour, Nat. Geosci., 8(2), 102–106, doi:10.1038/ngeo2345. Wang, Y., S.-P. Xie, B. Wang, and H. Xu (2005), Large-scale atmospheric forcing by southeast pacific boundary layer clouds: A regional model study, J. Clim., 18(7), 934–951, doi:10.1175/JCLI3302.1. Yin, B., and B. A. Albrecht (2000), Spatial variability of atmospheric boundary layer structure over the eastern equatorial pacific, J. Clim., 13(9), 1574–1592, doi:10.1175/1520-0442(2000)0132.0.CO;2. Yokoyama, C., E. J. Zipser, and C. Liu (2014), TRMM-observed shallow versus deep convection in the eastern pacific related to large-scale circulations in reanalysis datasets, J. Clim., 27(14), 5575–5592, doi:10.1175/JCLI-D-13-00315.1. Zhang, C., M. McGauley, and N. A. Bond (2004), Shallow meridional circulation in the tropical eastern pacific, J. Clim., 17(1), 133–139, doi: 10.1175/1520-0442(2004)0172.0.CO;2. Zhang, C., P. Woodworth, and G. Gu (2006), The seasonal cycle in the lower troposphere over West Africa from sounding observations, Q. J. R. Meteorol. Soc., 132(621), 2559–2582, doi:10.1256/qj.06.23. Zhang, C., D. S. Nolan, C. D. Thorncroft, and H. Nguyen (2008), Shallow meridional circulations in the tropical atmosphere, J. Clim., 21(14), 3453–3470, doi:10.1175/2007JCLI1870.1.

NISHANT ET AL.


842