Clim Dyn (2015) 44:1685–1697 DOI 10.1007/s00382-014-2284-4

Impacts of aerosols on dynamics of Indian summer monsoon using a regional climate model Sushant Das · Sagnik Dey · S. K. Dash 

Received: 6 November 2013 / Accepted: 30 July 2014 / Published online: 19 August 2014 © Springer-Verlag Berlin Heidelberg 2014

Abstract  A regional climate model, RegCM has been utilized to examine the dynamic impacts of large aerosol radiative forcing on the atmospheric temperature and circulation in India during the monsoon (Jun–Sep) seasons of 2009 and 2010. Surface shortwave radiative forcing at the aerosol hot spots is in the range −25 to −60 W m−2 with the larger values observed during the summer monsoon season of 2010 (due to larger dust load) relative to that in 2009. It is important to note that the summer monsoon rainfall in 2010 was declared to be a normal monsoon as against the deficit rain in 2009. Changes in near surface air temperature show a spatial dipole pattern with the aerosol effect dampening out above 500 hPa with a larger change observed for natural aerosols relative to anthropogenic aerosols. The dipole pattern is characteristics of aerosol-induced change. Aerosols tend to strengthen the summer monsoon zonal mean wind at 850 hPa over the hotspots (larger effect in 2009 than in 2010) whereas there is negligible impact on the corresponding mean meridional wind component. This has resulted in a southward shift of the monsoon circulation during 2010 summer, leading to an increase in upward motion over the core monsoon region and thereby increasing the cloud fraction. This may also be facilitated by the aerosol induced heating in the lower troposphere. In 2009, the upward motion is enhanced to the south of the core monsoon region. The dynamic effects imply a positive feedback of the aerosol direct radiative forcing on the summer monsoon circulation over India. S. Das · S. Dey (*) · S. K. Dash  Centre for Atmospheric Sciences, Indian Institute of Technology Delhi, Hauz Khas, New Delhi 110016, India e-mail: [email protected]

Keywords  Aerosol forcing · Dynamic impact · Zonal and meridional circulation · Indian monsoon region

1 Introduction The Indian monsoon region has been identified to have very large aerosol loading and surface cooling (e.g. Ramanathan et al. 2005; Dey and Girolamo 2010; Ramachandran et al. 2012). Numerous studies were carried out in the last decade to quantify the columnar aerosol burden (in terms of aerosol optical depth, AOD), major aerosol types, vertical distribution and radiative forcing at the top-of-theatmosphere and surface in campaign modes (Moorthy et al. 2005; Tripathi et al. 2006; Satheesh et al. 2008 to name a few) as well as using ground-based continuous measurements (Singh et al. 2004; Ganguly et al. 2006; Dey and Tripathi 2008; Moorthy et al. 2013 to name a few). Several hypotheses were proposed to explain the possible impacts of this enormous aerosol load on the monsoon circulation at a longer time scale. Ramanathan et al. (2005) have proposed a weakening of the monsoon circulation in view of increased anthropogenic aerosol loading, supported later by Bollasina et al. (2011) and Ganguly et al. (2012). Lau and Kim (2006) have hypothesized a change in monsoon life cycle due to an advancement of monsoon onset driven by latitudinal gradient in heating from elevated aerosols. The complexity in aerosol–monsoon connection at various space–time scales is not fully resolved due to the combined effects of microphysical and dynamical impacts of aerosol radiative forcing intertwined with meteorological forcing. Krishnamurti et al. (2013) have shown that enhanced cloud condensation nuclei resulting from high aerosol load may reduce the number of monsoon depressions. In-situ and satellite measurements of aerosol and cloud

13

1686

properties cannot fully resolve this issue because they are better suited for understanding the microphysical connection (e.g. Heymsfeld and McFarquhar 2001; Konwar et al. 2012). The dynamic link can better be understood using climate models having capability of simulating the meteorological and aerosol fields and allowing their interaction. Here, we present the direct impacts of aerosols on the dynamics of atmospheric circulation over the Indian monsoon region using a regional climate model, RegCM. A regional climate model was used because of its ability to simulate the aerosol distribution and circulation at a finer resolution with respect to any global climate model (GCM). GCMs simulate the spatial distribution of atmospheric aerosols and the atmospheric processes that govern aerosol–climate interaction at coarse resolution. On the other hand, the high-resolution regional climate models like RegCM are advantageous tools for dynamical downscaling of the meteorological processes. Furthermore, coupling the regional climate model, RegCM with an online aerosol module in it provides an opportunity to better resolve aerosol spatial distribution and its effects on meteorology at regional scale compared to any GCM. It should also be noted that the regional climate models cannot account for teleconnection and global adjustments to regional perturbation, which is considered to be very important (e.g. Bollasina et al. 2011). Other studies have emphasized towards more regional adjustment as a primary factor (e.g. Ramanathan et al. 2005) for which the selected study region (CORDEX domain) is likely to be big enough, and hence, RegCM can provide added values particularly by resolving radiative forcing gradients with more detail. In addition, extreme aerosol events (Cavazos et al. 2009) occurring at local scale can also be better resolved by RegCM than the coarse resolution GCMs. Before using a regional climate model with aerosol loading, it is pertinent to evaluate the performance of that model in simulating the mean circulation pattern over the same region. RegCM model has been tested and extensively used for the simulations of Indian summer monsoon circulation and rainfall in the past (e.g. Dash et al. 2006, 2012). In this study, analysis has been carried out for the monsoon (Jun– Sep) seasons of the two successive years 2009 (deficit monsoon) and 2010 (normal monsoon) (Mazumdar et al. 2010, 2011). These results may help in advancing our understanding of the dynamic feedback of the observed large surface dimming and atmospheric heating due to aerosols during the monsoon.

2 RegCM and experiments For the present study, version 4.1 of the International Centre for Theoretical Physics (ICTP)- regional climate model

13

S. Das et al.

RegCM4.1 has been used. Details of the RegCM configuration are discussed in Giorgi et al. (2012). In brief, RegCM 4.1 is a hydrostatic sigma-p vertical coordinate model with a dynamical core based on NCAR MM5 model. The radiative transfer calculations are based on CCM3, land surface physics follows BATS scheme, and planetary boundary layer is represented by Holstag’s scheme. Grell scheme and Emanuel schemes closure are for the convection over ocean and land respectively and the large scale cloud and precipitation are computed by SUBEX scheme. RegCM4.1 is coupled with an aerosol module that considers both natural (Zakey et al. 2006) and anthropogenic (Solmon et al. 2006) aerosols. The model has been utilized to study dust transport over the Indian monsoon region at four size bins (in the range 0.01–20 um) recently (Das et al. 2013). The space–time variation and vertical distribution of dust were simulated reasonably well near the major source regions, while absence of an inventory of local dust (viz. road dust, emission from construction activities etc.) has led to an under-estimation away from the source regions. Anthropogenic aerosols include sulphate and hydrophilic and hydrophobic black carbon (BC) and organic carbon (OC) (Solmon et al. 2006). The aqueous and gaseous phase chemical processes for the conversion of SO2 to SO4 are described in Qian et al. (2001). EDGAR (Solmon et al. 2006) and Liousse (Junker and Liousse 2008) inventories were used for simulating suphate and carbonaceous aerosols respectively for the year 2000 at 1° × 1° resolution. More detail information about the biomass and agricultural burning emission can be obtained from Liousse et al. (1996). Together considering four bins of dust, hydrophobic and hydrophilic components of BC and OC and aqueous conversions of sulphate, total 10 tracers were configured in the model experiments. The model also follows different constant velocities of hydrophobic and hydrophilic components of OC and BC as well as that of SO2 and SO4 over land and ocean. Aerosols are removed from the atmosphere through dry and wet deposition. It is to be noted that the model does not consider indirect effects of aerosols. Previously, Nair et al. (2012) examined the performance of RegCM in simulating anthropogenic aerosols and observed an underestimation in black carbon concentration by a factor of 2–5 with respect to in situ measurements, primarily due to under-estimation in the emission inventory. In this study, simulations were carried out at 50 km horizontal resolution and 18 vertical sigma levels with the model top at 50 hPa over the South Asian domain (Fig. 1) defined by the international programme named the Coordinated Regional Climate Downscaling Experiment (CORDEX) (Giorgi et al. 2009). This programme has been initiated to study the regional climate features at finer resolutions by selecting domains of several important regions across the globe. Although the model was simulated over

1687

Aerosol direct impacts on Indian monsoon circulation Fig. 1  Spatial distribution of topography (m) and mean seasonal wind circulation (m/s) at 850 hPa during 2009 and 2010 monsoon seasons over the CORDEX South Asia Domain. The study region is the Indian subcontinent 60°E–100°E, 5°N–40°N shown by black box. The yellow box represents the Indo-Gangetic Basin (IGB) within which lies an urban site Kanpur marked with ‘star’. The white trapezoidal box represents the Great Indian Desert

larger South Asian CORDEX domain, here the results are analyzed and presented for the Indian monsoon region (60°E–100°E longitude and 5°N–40°N latitude). The initial and boundary conditions were considered from NCEP-NCAR reanalysis data. The lateral boundary conditions are updated every 6 h. Optimum interpolated weekly sea surface temperature gridded data at 1° × 1° resolution have been obtained from the National Oceanic and Atmospheric Administration. The surface elevation data used are taken from the United States Geological Survey. The model was run from Nov 1, 2008 to Dec 31, 2010 considering the first 2 months as spin up time. We performed six experiments to understand the dynamic impacts of aerosols on the regional circulation during the monsoon season. First experiment was carried out without (hereafter EXP1_CTRL) allowing the natural (all four size bins) and anthropogenic aerosols to interact with the meteorological field. In the second experiment (EXP_2), the interaction is allowed through the perturbation of the radiation field by dust and anthropogenic aerosols, so that the response of the atmospheric thermodynamic field to the direct aerosol radiative forcing can be examined. Similarly, third (EXP_3) and fourth (EXP_4) experiments were carried out considering only natural aerosols with and without allowing interaction of aerosols with the meteorological field to examine the influence of natural aerosols relative to the total aerosols. The model is not a coupled model and hence any change in the circulation due to change in SST

influenced by aerosol-induced surface dimming will not be captured. To address this issue, EXP_5 and EXP_6 idealized experiments were carried out with and without reducing SST by 0.2 °C uniformly over the CORDEX domain in the absence of aerosols. We note that uniform reduction of SST may not be the real case due to uneven aerosol distribution. This magnitude of reduction in SST was chosen following the study by Hsieh et al. (2013). Sensitivity study with modification of SST as function of AOD would have been better, but these idealized experiments are useful in assessing the internal variability of the model with respect to the perturbation by aerosols. In this paper, the changes in temperature profiles, zonal and meridional winds at lower (850 hPa) and upper (200 hPa) troposphere, vertical wind and cloud fraction due to inclusion of both natural and anthropogenic aerosols have been analyzed and presented in the following section.

3 Results Analysis of the model output shows that the model captures some of the important characteristics of Indian summer monsoon in both the years of 2009 and 2010 (Fig. 2). The simulated easterly jet at 200 hPa and south westerly at 850 hPa are well simulated as documented in the previous studies (e.g. Dash et al. 2006, 2012). As expected (e.g. Sinha et al. 2013), the 850 hPa winds are stronger in the

13

1688

S. Das et al.

Fig. 2  Spatial distributions of RegCM simulated (top panel) wind pattern at 850 hPa and (bottom panel) at 200 hPa during the monsoon season of 2009 (left) and 2010 (right). Colour shading indicates wind speed. Note the different scale for the wind vector in the two images

year 2010 (normal monsoon season) over the Great Indian Desert than in the year 2009 (deficit monsoon season). The stronger wind in 2010 has generated more dust and resulted in larger dust transport to the Indian landmass leading to a stronger radiative impact compared to 2009. The wind field at 200 hPa showed similar kind of characteristics but with stronger easterly jet over southern peninsular region during 2010 compared to 2009 monsoon season. We used MODIS—Terra level 3 gridded data at 550 nm wavelengths to evaluate model simulated AOD. MODIS retrieves aerosol properties at 10 km spatial resolution using two separate algorithms—one over ocean (Tanre et al. 1997) and the other over land based on the dark target approach (Kaufman et al. 1997). The MODIS retrieved AOD has been globally validated and utilized by numerous researchers to study the variability of aerosol characteristics at global as well as regional scales including India where an overestimation is observed with respect to AERONET (e.g. Ramachandran and Cherian 2008). The model reproduces the spatial variation of AOD over the Great Indian Desert and Indo-Gangetic Basin (IGB, marked in Fig. 1) as shown in Fig. 3 appreciably well. Comparison has been made with MODIS data as well as with the satellite-based

13

aerosol climatology available in the literature (e.g. Dey and Girolamo 2010). For example, the high values over the IGB and dust source regions, the north–south spatial gradient of AOD over the oceans, low AOD over the topographically high regions are noticeable in Fig. 3. The model underestimates AOD over coastal area of northern Bay of Bengal and some parts over eastern India. These regions are highly dominated by the outflow of anthropogenic aerosols especially due to biomass burning, which may be highly underestimated in the respective emission inventories. Additional factors, which are to be noted for differences in spatial AOD, are the inherent internal model variability and the uncertainties in MODIS retrieval over bright surfaces (Remer et al. 2008). One interesting observation is the larger spatial coverage of AOD hotspot over the IGB during the 2010 monsoon season (normal) than the 2009 season (deficit), which is attributed to stronger wind at 850 hPa (Fig. 2) over Indian land mass causing more emissions over desert source and transportation of larger dust load to the IGB. This difference cannot be captured by satellites (bottom panel of Fig. 3) and ground-based radiometers, because their measurements are limited to clear days only, while the model

Aerosol direct impacts on Indian monsoon circulation

1689

Fig. 3  Spatial distributions of RegCM simulated (top panel) AOD and MODIS-derived AOD (bottom panel) during the monsoon season of 2009 (left) and 2010 (right)

simulates aerosol load in clear as well as cloudy condition. Mean (±1 standard deviation, σ) seasonal AOD over Kanpur (an urban site in the IGB) as simulated by RegCM during the monsoon season (0.78 ± 0.03) is 30 % higher relative to the AERONET observation. Mean RegCM-simulated AOD agrees within 15 % of the AERONET-retrieved AOD for the clear days, suggesting that aerosol load is very high in the cloudy days too. Since the direct measurements are not possible in these days by passive sensors, any observation-based climatology for the monsoon season is biased low. We note that larger AOD is simulated over some parts of the central region, which is consistent with the observed AOD during 2009 as compared to 2010. This can be attributed due to lesser washout of aerosols through wet deposition in the model (figure not shown) and biases in simulated meteorological field (Das et al. 2013). The aerosol-induced shortwave radiative forcing at surface (SWRF) and at top-of-the-atmosphere (TOA) for the monsoon seasons of 2009 and 2010 are shown in Fig. 4. Surface SWRF is <−25 W m−2 at the hotspots like IGB and <−55 W m−2 over the dust source regions. Surface SWRF further shows a latitudinal gradient (high in the

north and low in the south) over the oceans, similar to the observations reported by Satheesh et al. (2010). The spatial patterns in surface SWRF are similar in both the years with larger dimming observed in the 2010 monsoon season due to higher AOD relative to the previous 2009 monsoon season. The SWRF at TOA is observed to be <−21 W m−2 over the IGB during 2010 compared to −15 to −20 W m−2 during 2009. Aerosol SWRF at TOA is negative over most parts of the subcontinent except at the desert sources and snow-covered Himalayan terrain, where SWRF switches to positive sign due to high surface albedo in both the years (Das et al. 2013). The net result (difference between TOA and surface SWRF) is large warming in the troposphere (in the range 10–50 W m−2) in both the monsoon seasons. It is interesting to note that the change (indicated by Δ) in near-surface air temperature at 2 m (hereafter referred to as T2m) due to aerosol radiative forcing (EXP_2-EXP1_ CTRL) shows a dipole pattern (top panel of Fig. 5) that do not directly correspond to spatial distributions of AOD and surface SWRF (Figs. 3, 4), thereby indicating the complexity and non-linearity in the dynamic impacts of aerosols. ΔT2m varies within a wide range +1.8 to −1.8 °C over the

13

1690

S. Das et al.

Fig. 4  Spatial distributions of RegCM simulated (top panel) aerosol surface SW radiative forcing (W m−2) and (bottom panel) TOA SW forcing during the monsoon season of 2009 (left) and 2010 (right)

study domain with a high (>0.6 °C and significant at 90 % CI) over the Central India during the 2009 monsoon season that shifts further southward during the 2010 monsoon season. The low ΔT2m (<−0.4 °C) over the Great Indian Desert during the 2009 season switches to positive value (>1 °C) during the 2010 season. On the other hand, larger AOD over the IGB during the 2010 season is associated with negative ΔT2m. Surface dimming by aerosols leads to a reduction in absorbed radiation at surface, which in turn may lead to the negative ΔT2m. However, ΔT2m is positive for SWRF <−25 W m−2 over the western India including dust source region in the year 2010 (Fig. 5). ΔT2m does not depend only on aerosol-induced dimming at surface (direct effect), but also depends on flux exchange between surface and atmosphere, dimming from the change in cloudiness and relative fraction of natural and anthropogenic aerosols. The spatial pattern of ΔT2m due to only natural aerosols (EXP_3-EXP_4) shown in the bottom panel of Fig. 5 is same as of total aerosols, but larger in magnitude implying an opposing effect of anthropogenic aerosols on ΔT2m.

13

The differences in concentration, vertical distribution and optical properties of the natural and anthropogenic aerosols contribute to the observed difference in ΔT2m. When only natural aerosols (dust) are considered, emitted radiation from the surface is again absorbed more efficiently by the aerosol layer (compared to anthropogenic aerosols, which are smaller in size), leading to a partial compensation of negative ΔT2m. This would explain the positive ΔT2m over the desert region where dust load is higher in 2010 than in 2009. Since anthropogenic aerosols are underestimated by the model (Nair et al. 2012), the response of T2m to aerosol forcing is biased towards natural aerosols. Negative ΔT2m over high albedo surface results from less radiation absorbed by the surface. The effect is stronger in the normal monsoon season of 2010 (more negative ΔT2m in the bottom panel of Fig. 5 than the upper panel) than in deficit season of 2009 because of larger dust load transported by stronger wind. The dynamic effect of aerosols on cloud fraction (as discussed in the subsequent section) further controls the observed dipole pattern in ΔT2m. Similar

Aerosol direct impacts on Indian monsoon circulation

1691

Fig. 5  Spatial distributions of changes in near surface air temperature (ΔT2m in the text) in  °C due to the direct impacts of both natural and anthropogenic aerosols (top panel) and only natural aerosols (bottom panel) over the Indian monsoon region during the monsoon season of (left) 2009 and (right) 2010. The purple dotted contour lines are regions with 90 % significance level

Fig. 6  Spatial distributions of ΔT2m (without aerosols) due to perturbed SST experiment during the monsoon season of (left) 2009 and (right) 2010

dipole patterns were also reported from Europe for anthropogenic aerosols (Zanis et al. 2012; Zanis 2009). The robustness of the attribution of observed ΔT2m to aerosols and inherent internal variability of the model

has been tested by perturbing the SST uniformly (Fig. 6). ΔT2m obtained from EXP_5-EXP_6 does not show much change of the similar magnitude due to the change in circulation when SST is reduced. For example, ΔT2m mostly

13

1692

S. Das et al.

Fig. 7  Latitudinal variations of aerosol-induced changes in vertical cross sections (averaged over 70°–85°E longitudes within the study domain) of temperature (°C) during 2009 (left) and 2010 (right). The contours represent the absolute values in the EXP1_CTRL experiment and the color shades represent the changes incurred by aerosols

varies within ±0.4 °C in 2009 and most parts of India in 2010 except the central IGB, where it is ~0.6–0.9 °C. More importantly, the dipole pattern is missing in these perturbation experiments. In addition, the spatial variation is very similar (e.g. negative over the desert sources and Peninsular India and positive over IGB) in both the years. Hence, the statistically significant ΔT2m observed in Fig. 5 indicates a robust aerosol signal. Since the aerosol signal can be distinguished, we proceed to analyze further results, which are derived considering the first two experiments (i.e. EXP_2-EXP1_CTRL). The vertical cross sections of ΔT2m attributed to both natural and anthropogenic aerosols across the latitudes (averaged over 70°–85°E longitudes, Fig. 7) reveal that the aerosol effect can be seen up to the mid-troposphere above which the aerosol signal fades away due to significantly less aerosol concentration. In the lower troposphere, latitudinal variation of ΔT2m reveals alternate cooling and heating in 2009 that reverses in 2010. ΔT2m reduces by 0.2°–0.3 °C in the mid-troposphere (~500 hPa) in the 2010 season, when the AOD is high. This is not the case in the 2009 season when AOD is relatively low. Moreover, the aerosol induced ΔT2m (Fig. 5) and surface pressure (Fig. 9) will influence changes in local circulation through updraft (downdraft) of warmer (cooler) air from (into) the lower troposphere. Next we looked into the changes in wind pattern at 850 hPa (upper panel of Fig. 8) and 200 hPa (bottom panel of Fig. 8) and the changes in the surface pressure due to aerosols (Fig. 9). The changes in surface pressure are seen to be consistent with changes in the wind circulation at 850 hPa for both the years. The associated increment in the surface pressure over the Bay of Bengal causes a diverging anticyclonic pattern. In 2010 monsoon season, aerosol induced changes led decrement in surface pressure over the IGB facilitates the shift in horizontal wind component. In 2009, zonal wind at 850 hPa (Fig. 10) is strengthened by 2–3 m/s in magnitude due to aerosols over the core

13

monsoon region (within 20°–25°N latitude), whereas the meridional wind shows negligible changes (Fig. 10), resulting in a stronger westerly flow (Fig. 8). The overall magnitude of the effect remains same in the 2010 monsoon season but a smaller increase in zonal component (Fig. 11) resulted in a southward shift of zonal circulation (Fig. 8). This has led to a stronger outflow of aerosols into northern Bay of Bengal in 2010 relative to the 2009 monsoon season (as also observed in the AOD map in Fig. 3). The zonal wind at 200 hPa has weakened over the north Indian Ocean, but strengthened over the northwestern part of the subcontinent (including Great Indian Desert) during the 2009 season. This, along with a weakened meridional component, led to a northwesterly flow at the upper troposphere over the northern and northwestern India due to aerosols. The signal of aerosol-induced changes in the upper tropospheric circulation is weak in 2010 (Fig. 11). The changes in meridional cross-section of vertical wind due to aerosols are depicted in Fig. 12 along with the changes in cloud fraction. Note that the microphysical connection between aerosols and clouds are not considered in the simulations and hence, the observed change in cloud fraction is solely attributed to the dynamic effect of aerosols. The contours (negative and positive values are for upward and downward motion in Pa/s) represent the mean seasonal vertical wind in the EXP1_CTRL values, while the color shadings represent the magnitude of changes due to aerosols. The strength of the upward motion is suppressed over the core monsoon region during the 2009 season, probably because of the strengthening of zonal wind. This resulted in a relative reduction of cloud fraction by 2–3 % up to 500 hPa altitude. A reduction in strength of zonal wind at the lower and mid-troposphere, mostly over the oceans due to aerosols in the 2009 monsoon season, has led to an increasing updraft (south of 15°N latitude) and thereby facilitating cloud formation. During 2010 monsoon season, the southwestward shift of circulation at the

1693

Aerosol direct impacts on Indian monsoon circulation Fig. 8  Spatial distributions of aerosol-induced changes of wind circulation (in m/s) at 850 hPa (top panel) and 200 hPa (bottom panel) during 2009 (left panel) and 2010 (right panel). The color shades represent the changes wind speed incurred by aerosols and the arrows indicate the changes in wind direction

Fig. 9  Spatial distributions of aerosol-induced changes of surface pressure (in hPa) during 2009 (left panel) and 2010 (right panel)

lower troposphere over the core monsoon region allowed the upward motion to strengthen. This has resulted in an increase in cloud fraction (bottom panel of Fig. 12), while cloud fraction reduces below 700 hPa over the oceans. The atmospheric diabatic heating (i.e. difference between TOA and surface forcing) due to aerosols also may influence the convection above the aerosol layer leading to the observed change in cloud fraction. The relative changes in cloud fraction due to aerosols are linked to the observed ΔT2m. For example, the decrease in cloud fraction over the core

monsoon region in 2009 (Fig. 12) may have increased the temperature up to ~500 hPa (Fig. 7). Similar connection between changes in cloud fraction and ΔT2m can be drawn for the 2010 monsoon season also.

4 Discussion and conclusions In this work, the direct impact of aerosols on the atmospheric circulation over India during the summer monsoon

13

1694 Fig. 10  Spatial variations of aerosol-induced changes in zonal component (left panel) and meridional component (right panel) at 850 hPa (top panel) and 200 hPa (bottom panel) of seasonal mean wind (m/s) fields during the monsoon season of 2009

Fig. 11  Same as Fig. 10, but for the monsoon season of 2010

13

S. Das et al.

Aerosol direct impacts on Indian monsoon circulation

1695

Fig. 12  Latitudinal variation of aerosol induced changes in vertical cross sections (averaged over 70°–85°E longitudes within the study domain) of omega (left panel) and cloud fraction (right panel) during the 2009 (top panel) and 2010 (bottom panel) monsoon seasons. The

contours represent the absolute values in the EXP1_CTRL run and the color shades represent the changes incurred by aerosols. The unit of omega is in 10−5 hPa/s and it is dimensionless for cloud fraction

season was examined using a regional climate model. The model simulates larger AOD during the normal monsoon season relative to deficit season due to higher aerosol load in cloudy condition, which is not captured by any measurements. Aerosols perturb the radiative fluxes within the atmospheric column and thereby induce changes in the thermodynamic structures of the atmosphere. Since, aerosols are mostly confined to the atmosphere below 450– 500 hPa, the effect fades away aloft. The aerosol induced heating is thus linked to changes in vertical wind shear, zonal and meridional circulation and more importantly, advection (Zanis et al. 2012). These short-term changes in turn redistribute aerosol load initiating a feedback mechanism. Four issues demand detailed discussion to comprehend the significance of the results obtained here. Firstly, the aerosol indirect effects were not included during the simulations. Indirect effect facilitates cloud formation and enhances its lifetime (Albrecht 1989) and thus the total SWRF (due to both aerosols and clouds) must be higher.

In other words, if microphysical effect of aerosols were considered, cloud fraction would have increased by a larger magnitude in 2010 than what is shown in Fig. 12. Sengupta et al. (2013) have shown that large aerosol load during the monsoon season actually facilitates invigoration of cloud in the Indian monsoon region. This would have resulted in a larger dimming effect at the surface, especially at the central India and IGB. The observed negative values of ΔT2m (e.g. the IGB in 2010) might have been underestimated, since only the direct effect of aerosols on the circulation is considered here. On the other hand, the large surface dimming and atmospheric warming would have suppressed convection (Ramanathan et al. 2001). Unless indirect effect is included in the model configuration, the net effect is difficult to be quantified. Secondly, since RegCM is an atmospheric model, the results might have been biased to the fact that SST is prescribed in the model and does not react to the aerosol surface radiative forcing (which is very likely to be strong over the Arabian Sea and quite important over

13

1696

the Bay of Bengal). Several studies have shown that this energy imbalance is very sensitive in terms of precipitation and circulation response (e.g. Yue et al. 2011; Solmon et al. 2012). However, since SST gradients and humidity fluxes are very important, the large South Asia CORDEX, which includes actual geographical locations of important aerosol sources (e.g. Arabian Peninsula), the response is adequate. We have performed additional idealized experiments to understand the robustness of the aerosol signal to the observed changes in T2m by uniformly reducing the SST following an earlier study by Hsieh et al. (2013). The net impact, in reality, should be the combined changes incurred due to SST dimming and aerosols. However, as discussed earlier, the aerosol signal (Fig. 5) is stronger than the perturbed SST signal (Fig. 6) and thus our interpretation of aerosol direct effects hold true. Thirdly, the simulated SWRF is underestimated due to two factors—underestimation in carbonaceous aerosols owing to uncertainty in emission inventory (Nair et al. 2012) and consideration of external mixing of dust and anthropogenic particles during the transport. It may be noted that internal mixing of dust and BC was reported from the IGB (Dey et al. 2008). Internal mixing enhances the diabatic heating; thereby the results presented here are conservative estimates. Lastly, the conclusions are based on the simulation for short period of 2 years. Simulations for longer period would have yielded more comprehensive results. The interpretation of the results obtained in this study are made in terms of the physical forcing induced by aerosols and it is important that even with 2 years of model simulations, strong aerosol signal is emerging out. In future, simulations for longer period will be undertaken to further examine the robustness of the aerosol signal. The major conclusions of this study are as follows: 1. Larger AOD and SWRF were observed during the 2010 monsoon season than in the previous season due to larger dust load transported by stronger monsoon winds. Observations fail to capture the high AOD under cloudy condition, thereby leading to a low bias in climatology of AOD during the monsoon season. 2. Aerosol induced changes in near surface air temperature (at 2 m) show a spatial dipole pattern with the value varying in the range +1.8 to −1.8 °C (change of more than ±0.6 °C is significant at 90 % CI). The magnitude reduces with altitude, eventually the effect fading away above 500 hPa. The impact is stronger for natural aerosols relative to the anthropogenic aerosols. The dipole pattern is not observed in the simulations where SST is perturbed without aerosols suggesting the robustness in aerosol signal in the present study.

13

S. Das et al.

3. Zonal wind at the lower troposphere is strengthened by aerosols resulting in a southward shift of the circulation, which in turn, increases the upward motion of the air facilitating cloud formation. The effect is prominent to the south of the core monsoon region in the 2009 monsoon season whereas the same has been shifted northward over the IGB in the 2010 monsoon season. The convective process might have been facilitated above the aerosol layer by large diabatic heating of aerosols. Acknowledgments  This work is supported by financial grant from Ministry of Earth Sciences, Govt. of India under CTCZ Programme (MoES/CTCZ/16/28/10) through a research project operational at IITD (IITD/IRD/RP02479). The first author is thankful to CSIR for providing scholarship to carry out research work in IIT Delhi. The authors acknowledge ICTP for providing the RegCM4.1 model (http://eforge.ictp.it). The efforts of PIs of Kanpur AERONET site (Drs. Brent Holben, R. P. Singh and S. N. Tripathi) are acknowledged. We acknowledge the comments by the anonymous reviewers who helped us improving the quality of the original version of the manuscript.

References Albrecht BA (1989) Aerosols, cloud microphysics and fractional cloudiness. Science 245:1227–1230 Bollasina MA, Ming Y, Ramaswamy V (2011) Anthropogenic aerosols and the weakening of the South Asian summer monsoon. Science 334:502–505 Cavazos C, Todd MC, Schepanski K (2009) Numerical model simulation of the Saharan dust event of 6–11 March 2006 using the Regional Climate Model version 3 (RegCM3). J Geophys Res 114:D12109. doi:10.1029/2008JD011078 Das S, Dey S, Dash SK, Basil G (2013) Examining mineral dust transport over the Indian subcontinent using the regional climate model, RegCM4.1. Atmos Res 134:64–76 Dash SK, Shekhar MS, Singh GP (2006) Simulation of Indian summer monsoon circulation and rainfall using RegCM3. Theor Appl Climatol 86:161–172 Dash SK, Mamgain A, Pattnayak KC, Giorgi F (2012) Spatial and temporal variations in Indian summer monsoon rainfall and temperature: an analysis based on RegCM3 simulations. Pure Appl Geophys 1–20. doi:10.1007/s00024-012-0567-4 Dey S, Girolamo LD (2010) A climatology of aerosol optical and microphysical properties over the Indian subcontinent from 9 years (2000–2008) of Multiangle Imaging SpectroRadiometer (MISR) data. J Geophys Res 115:D15204. doi:10.1029/200 9JD013395 Dey S, Tripathi SN (2008) Aerosol direct radiative effects over Kanpur in the Indo-Gangetic basin, northern India: long-term (2001– 2005) observations and implications to regional climate. J Geophys Res 113:D04212. doi:10.1029/2007JD009029 Dey S, Tripathi SN, Mishra SK (2008) Probable mixing state of aerosols in the Indo-Gangetic Basin, northern India. Geophys Res Lett 35:L03808. doi:10.1029/2007GL032622 Ganguly D, Jayaraman A, Gadhavi H (2006) Physical and optical properties of aerosols over an urban location in western India: seasonal variabilities. J Geophys Res 111:D24206. doi:10.1029/ 2006JD007392 Ganguly D, Rasch PJ, Wang H, Yoon J (2012) Fast and slow responses of the South Asian monsoon system to anthropogenic aerosols. Geophys Res Lett 39:L18804. doi:10.1029/2012GL053043

Aerosol direct impacts on Indian monsoon circulation Giorgi F, Jones C, Asrar GR (2009) Addressing climate information needs at the regional level: the CORDEX framework. WMO Bull 58(3):175–183 Giorgi F, Coppola E, Solmon F, Mariotti L, Sylla MB, Bi X, Elguindi N, Diro GY, Nair V, Giuliani G, Turuncoglu UU, Cozzini S, Guttler I, O’Brien TA, Tawfik AB, Shalaby A, Zakey AS, Steiner F, Stordal F, Sloan LC, Brankovic C (2012) RegCM4: model description and preliminary tests over multiple CORDEX domains. Clim Res 52:7–29 Heymsfeld AJ, McFarquhar GM (2001) Microphysics of INDOEX clean and polluted trade cumulus clouds. J Geophys Res D22:28653–28673 Hsieh WC, Collins WD, Liu Y, Chiang JCH, Shie CL, Caldeira K, Cao L (2013) Climate response due to carbonaceous aerosols and aerosol-induced SST effects in NCAR community atmospheric model CAM3.5. Atmos Chem Phys 13:7489–7510 Junker C, Liousse C (2008) A global emission inventory of carbonaceous aerosol from historic records of fossil fuel and biofuel consumption for the period 1860–1997. Atmos Chem Phys 8:1195– 1207. doi:10.5194/acp-8-1195-2008 Kaufman YJ, Tanre D, Remer LA, Vermote EF, Chu A, Holben BN (1997) Operational remote sensing of tropospheric aerosol over land from EOS moderate resolution imaging spectroradiometer. J Geophys Res 102:17051–17067 Konwar M, Maheshkumar RS, Kulkarni JR, Freud E, Goswami BN, Rosenfeld D (2012) Aerosol control on depth of warm rain in convective clouds. J Geophys Res 117:D13204. doi:10.1029/20 12JD017585 Krishnamurti TN, Martin A, Krishnamurti R, Simon A, Thomas A, Kumar V (2013) Impacts of enhanced CCN on the organization of convection and recent reduced counts of monsoon depressions. Clim Dyn 41:117–134 Lau KM, Kim KM (2006) Observational relationships between aerosol and Asian monsoon rainfall, and circulation. Geophys Res Lett 33:L21810. doi:10.1029/2006GL027546 Liousse C, Penner JE, Chuang C, Walton JJ, Eddleman H, Cachier H (1996) A global three-dimensional model study of carbonaceous aerosols. J Geophys Res 101(D14):19411–19432. doi:10.1029/95JD03426 Mazumdar AB, Khole M, Devi SS (2010) Weather in India, Monsoon season (June to September 2009). Mausam 61(3):411–454 Mazumdar AB, Khole M, Devi SS (2011) Weather in India, Monsoon season (June to September 2010). Mausam 62(3):463–512 Moorthy KK, Sunilkumar SV, Pillai PS, Parameswaran K, Nair PR, Ahmed YN, Ramgopal K, Narasimhulu N, Reddy RR, Vinoj V, Satheesh SK, Niranjan K, Rao BM, Brahmanandam PS, Saha A, Badarinath KVS, Kiranchand TR, Latha KM (2005) Wintertime spatial characteristics of boundary layer aerosols over peninsular India. J Geophys Res 110:D08207. doi:10.1029/2004JD005520 Moorthy KK, Babu SS, Manoj MR, Satheesh SK (2013) Buildup of aerosols over the Indian region. Geophys Res Lett 40:1011–1014. doi:10.1002/grl.50165 Nair VS, Solmon F, Giorgi F, Mariotti L, Babu SS, Moorthy KK (2012) Simulation of South Asian aerosols for regional climate studies. J Geophys Res 117(D4):1–17 Qian Y, Giorgi F, Huang Y (2001) Regional simulation of anthropogenic sulfur over east Asia and its sensitivity to model parameters. Tellus 53:171–191 Ramachandran S, Cherian R (2008) Regional and seasonal variations in aerosol optical characteristics and their frequency distributions over India during 2001–2005. J Geophys Res 113:D08207. doi:1 0.1029/2007JD008560 Ramachandran S, Kedia S, Srivastava R (2012) Aerosol optical depth trends over different regions of India. Atmos Environ 49:338–347 Ramanathan V, Crutzen PJ, Lelieveld J, Mitra AP, Althausen D, Anderson J, Andreae MO, Cantrell W, Cass GR, Chung CE,

1697 Clarke AD, Coakley JA, Collins WD, Conant WC, Dulac F, Heintzenberg J, Heymsfeld AJ, Holben B, Howell S, Hudson J, Jayaraman A, Kiehl JT, Krishnamurti TN, Lubin D, McFarquhar G, Novakov T, Ogren JA, Podgorny IA, Prather K, Priestley K, Prospero JM, Quinn PK, Rajeev K, Pash P, Rupert S, Sadourny R, Satheesh SK, Shaw GE, Sheridan P, Valero FPJ (2001) Indian Ocean experiment: an integrated assessment of the climate forcing and effects of the great Indo-Asian haze. J Geophys Res 106(D22), 28371–28398 Ramanathan V, Chung C, Kim D, Bettge T, Buja L, Kiehl JT, Washington WM, Fu Q, Sikka DR, Wild M (2005) Atmospheric brown clouds: impacts on South Asian climate and hydrological cycle. PNAS 102:5326–5333 Remer LA, Kleidman RG, Levy R, Kaufman YJ, Tanre D, Mattoo S, Martins JV, Ichoku C, Koren I, Hongbin YU, Holben BN (2008) Global aerosol climatology from the MODIS satellite sensors. J Geophys Res 113:D14S07. doi:10.1029/2007JD009661 Satheesh SK, Moorthy KK, Babu SS, Vinoj V, Dutt CBS (2008) Climate implications of large warming by elevated aerosol over India. Geophys Res Lett 35:L19809. doi:10.1029/2008GL034944 Satheesh SK, Vinoj V, Krishnamoorthy K (2010) Assessment of aerosol radiative impact over oceanic regions adjacent to Indian subcontinent using multisatellite analysis. Meteorol Adv. doi:10.1155/2010/139186 Sengupta K, Dey S, Sarkar M (2013) Structural evolution of monsoon clouds in the Indian CTCZ. Geophys Res Lett 40:5295–5299 Singh RP, Dey S, Tripathi SN, Tare V, Holben BN (2004) Variability of aerosol parameters over Kanpur, northern India. J Geophys Res 109:D23206. doi:10.1029/2004JD004966 Sinha P, Mohanty UC, Kar SC, Dash SK, Robertson AW, Tippett MK (2013) Seasonal prediction of the Indian summer monsoon rainfall using canonical correlation analysis of the NCMRWF global model products. Int J Climatol 33:1601–1614 Solmon F, Giorgi F, Liousse C (2006) Aerosol modeling for regional climate studies: application to anthropogenic particles and evaluation over a European/African domain. Tellus B 58(1):51–72 Solmon F, Elguindi N, Mallet M (2012) Radiative and climatic effects of dust over West Africa, as simulated by a regional climate model. Clim Res 52:97–113 Tanre D, Kaufman YJ, Herman M, Mattoo S (1997) Remote sensing of aerosol properties over oceans using the MODIS/EOS spectral radiances. J Geophys Res 102:16971–16988 Tripathi SN, Tare V, Chinnam N, Srivastava AK, Dey S, Agarwal A, Kishore S, Lal RB, Manar M, Kanawade VP, Chauhan SSS, Sharma M, Reddy RR, Gopal KR, Narasimhulu K, Reddy LSS, Gupta S, Lal S (2006) Measurements of atmospheric parameters during Indian Space Research organization geosphere biosphere programme land campaign II at a typical location in the Ganga basin: 1. Physical and optical properties. J Geophys Res 111:D23209. doi:10.1029/2006JD007278 Yue X, Liao H, Wang HJ, Li SL, Tang JP (2011) Role of sea surface temperature responses in simulation of the climatic effect of mineral dust aerosol. Atmos Chem Phys Discuss 11:1121–1152 Zakey AS, Solmon F, Giorgi F (2006) Development and testing of a desert dust module in a regional climate model. Atmos Chem Phys 6:4687–4704 Zanis P (2009) A study on the direct effect of anthropogenic aerosols on near surface air temperature over Southeastern Europe during summer 2000 based on regional climate modeling. Ann. Geo. 27:3977–3988 Zanis P, Ntogras C, Zakey A, Pytharoulis I, Karacostas T (2012) Regional climate feedback of anthropogenic aerosols over Europe using RegCM3. Clim Res 52:267–278

13