Aerosol loading over the Indian Ocean and its possible impact on regional climate
Chul Eddy Chung* & V Ramanathan
Center for Clouds, Chemistry and Climate (C4), Scripps Institution of Oceanography, La Jolla CA. 92093, U.S.A.
*[E-mail: cchung@fiji.ucsd.edu]
Received 29 August 2003
This paper provides a review of aerosol forcing results from the Indian Ocean Experiment (INDOEX) and also summarizes the follow-on modeling studies that examine the impact of the haze on regional climate. Every dry season from November to May, anthropogenic haze spreads over most of the northern Indian Ocean, and South and Southeast Asia. The INDOEX documented this Indo-Asian haze at various scales during 1995-2001. The haze particles consisted of several inorganic and carbonaceous species, including absorbing black carbon clusters, fly ash and mineral dust. Because of black carbon contributing as much as about 14% to the fine particle mass, the single-scattering albedo estimated by several independent methods was consistently around 0.9 both inland and over the open ocean. Anthropogenic sources contributed as much as 75% (±10%) to the aerosol loading and the optical depth. The regional aerosol forcing resulting from the direct and indirect effects was derived by integrating the multi-platform observations of satellites, aircraft, ships, surface stations and balloons with 1- and 4-D models. The haze layer reduces the net solar flux at the surface by as much as 20 to 40 Wm−2 on a monthly mean basis and heats the lowest 3 km atmosphere by as much as 0.4 to 0.8 K/day, which enhances the solar heating of this layer by 50 to 100%.
The INDOEX also documented year-to-year fluctuations of the haze forcing. For instance, the southernmost extent of the haze varied from about 10°S to about 5°N. In assessing the haze impacts on the cold dry-season regional climate, we conducted two CCM3 experiments with two extreme locations of the forcing: 1) extended haze forcing (EHF) and 2) shrunk haze forcing (SHF). Over India where the forcing is centered, the simulated climate changes are very similar between EHF and SHF. The most important effect of the haze is a surface cooling, and a strengthening of the inversion in the lower troposphere. The surface cooling has been confirmed by observations. The stabilization of the boundary layer results in a reduction of evaporation and sensible heat flux from the land. Rainfall patterns get substantially disrupted in local and remote regions, with the disruption being very sensitive to the southern extent of the imposed haze forcing. Both forcings lead to global circulation/precipitation perturbations; and the EHF produces about an order of magnitude larger responses.
One key remote response to the haze is the suppression of convection in the western equatorial Pacific, which has implications for ENSO variability. Since the western Pacific convection suppression would weaken the trade winds over the Pacific and induce warm anomalies in the eastern basin, we speculate that the Great Indo-Asian haze might have an important role in the amplitude and frequency of El Niño events during the recent decades. The haze-ENSO connection is further demonstrated by the Cane-Zebiak Pacific Ocean/atmosphere model. The focus of the studies thus far has been on the dry season (November to May) aerosols. The role of anthropogenic aerosols during the wet season from June to September is to be explored.
[Key words: Absorbing aerosols, INDOEX, climate, Indian Ocean, haze, ENSO, modelling, Asian Brown Cloud]
[IPC Code: Int.Cl.7 C09K3/30, G01B11/22]
1. Introduction
Aerosols change the climate directly by scattering and absorption of the solar radiation. Figure 1 depicts the aerial extent of the aerosols in the Asian continent and adjacent ocean areas. The aerosols in these regions are brownish, and the brownish color results from a significant amount of black carbon aerosol, which enhances the absorption of the solar radiation in the atmosphere. As Fig. 1 shows, the brownish hazy layer covers the Asian regions extensively, and is now called the “Asian Brown Cloud (ABC)”. In view of the widespread extent and the brownishness, the climatic effects must be large and significant. It has been estimated recently1 that black carbon from fossil
fuel combustion increased from about 2 Tg/year in the 1950s to as high as 6 Tg/year in the 1990s, with Asia contributing as much as 60% to this increase. The ABC is certainly a global environmental issue. The ABC Project2 is a new major international effort to address both the science and policy issues related to the widespread air pollution in Asia. The ABC project is an extension of the Indian Ocean Experiment (INDOEX), which documented3 the haze in South Asia and the tropical Indian Ocean from 1995 to 2001.
Aerosols can change the climate indirectly as well, e.g., by producing more cloud drops (with smaller effective radius) that in turn makes the cloud brighter
(first indirect effect4) and longer lasting (second indirect effect5). The INDOEX was conducted to quantify the direct and the indirect aerosol forcing from observations in South Asia and the tropical Indian Ocean3. The INDOEX team monitored the year-to-year fluctuation of the haze as well6,7. Figure 2 depicts the Aerosol Optical Depths (AODs) of the South Asian (SA) haze for the months of February 1997 and March 19987. Both months had a widespread presence of the haze, and yet the southward extent varied substantially from one year to another. In February 1997, the haze reached as far as 10°S, but in March 1998 it stayed north of the equator. The regional concentration and the temporal fluctuation render the climatic effects of the aerosols very difficult to estimate.
In assessing the aerosol effects on climate, the approach so far has been to use a global distribution of anthropogenic aerosols in climate models. The 3D aerosol distribution was estimated by source inventories and reanalyzed/modeled winds, in which
global sulfate aerosol effects were added in the GCMs8,9. These studies8,9 demonstrated that the sulfate aerosol cooling compensated the CO2-induced warming considerably. While this approach is conceptually appealing, its primary limitation is the lack of global scale observations for aerosol properties, in particular the absorption, the vertical distribution and the spatial distribution away from the source regions. With respect to absorbing aerosols, even the sign of the net effect, one of cooling or warming, is dependent on these parameters as well as on the distribution of clouds in the model10-13. As a result climate assessment studies that rely solely on models to generate the aerosol forcing may have large uncertainties. In order to avoid these pitfalls, we focus here on the South Asian region and the tropical Indian Ocean for which the aerosol radiative forcing has been determined from observations during the INDOEX and we employ the observationally determined forcing in a global climate model. Aerosol radiative forcing is defined as the effect of aerosol,
Fig. 1—Aerosol Optical Depths (AODs) at 550 nm derived from Moderate Resolution Imaging Spectroradiometer (MODIS) instrument onboard the TERRA satellite. The AODs were averaged from November 2000 to May 2001. AOD is the vertical integral of aerosol concentration weighted with the effective cross-sectional area of the particles intercepting (by scattering and absorption) the solar radiation at the wavelength of interest. The higher AOD is, the more there is aerosol.
both natural and anthropogenic, on the radiative fluxes at the top of the atmosphere (TOA), at the surface and on the absorption of solar (and long-wave in the case of dust and sea salt particles) radiation within the atmosphere.
The INDOEX observations were made during the Indian dry monsoon season (December to April) from 1995 to 2000. An international group of scientists collected aerosol, chemical and radiation data from ships, satellites and surface stations3,14,15 which culminated in an intensive field experiment (IFP) with five aircrafts, two ships, several satellites and numerous surface observations conducted during January through March, 1999. In addition to these platforms, the European Meteorological Satellite Organization moved the geostationary satellite, METEOSAT-5, over the Indian Ocean to support the INDOEX campaign. This haze, about the size of
USA, extends from the south Asian continent to the Arabian Sea, and from the Bay of Bengal to the Indian Ocean ITCZ. In the vertical direction3, it extends from the surface to about 3~4 km.
The regional-scale aerosol forcing was derived from a Monte Carlo Aerosol-Cloud Radiation (MACR) model, which integrated all the observations, satellite-retrieved aerosol regional distributions over the ocean and modeled aerosol information for the land regions (using a 4-D assimilation model16. According to the INDOEX estimates3,17, the haze directly decreases the surface solar absorption by about 20 Wm-2 (averaged over 0–20ºN and 40–100ºE), and indirectly by about 6 Wm-2 during the January–March 1999. The atmospheric absorption by the aerosols was as much as 19 Wm-2 regionally, because of a significant amount of black carbon. The aerosol climate forcing over this area is one order of magnitude greater than the greenhouse gas forcing. Furthermore, the haze forcing is regionally concentrated with its magnitude diminishing towards the ITCZ. Close to the source, the haze layer generates an atmospheric diabatic heating of up to +0.8K/day in the lower troposphere regionally, which is comparable to a deep heating anomaly of 2 K/day during moderate ENSO events (diagnosed by Nigam et al.18). How would the monsoon respond to the sharp gradient of the anthropogenic climate forcing?
In this paper we summarize how the haze climate forcing was derived from the INDOEX, and analyse the numerical experiments that were conducted with the derived forcing. As the INDOEX was carried out during the NH winter and spring times, the numerical simulation analysis will be centered in the winter/spring monsoon. In the climate modeling study, we assess the significance of the interannual variations in the aerosol forcing through two numerical sensitivity experiments with the National Center for Atmospheric Research (NCAR) Community Climate Model version 3 (CCM3)-1) extended haze forcing (EHF) and 2) shrunk haze forcing (SHF). The extended haze forcing (EHF) case has the forcing extending up to 10°S, and for the shrunk haze forcing (SHF) case the forcing is restricted to the area north of the equator. These two cases essentially bracket the interannual variability of the haze spatial extent. For the EHF case, the haze decreases the surface solar absorption by 24.5 Wm-2 and increases the atmospheric solar heating by 18 Wm-2, and the respective values for SHF is
Fig. 2—Monthly-mean AOD at 630 nm over the Indian Ocean. (a) February 1997 and (b) March 1998. The AODs were retrieved from the satellite AVHRR observations7.
13.5 Wm-2 and 10.6 Wm-2 (averaged over 0–20ºN and 40–100ºE) from January to March. During the summertime, the haze forcing is not added to the CCM3, since there is no INDOEX observation during this period. Implementing year-round forcings is proposed for the future study.
Here, the simulation analysis includes assessing the remote impact of the aerosols. This is a particularly important issue with the South Asian (SA) haze, since the haze creates a large atmospheric diabatic heating.
Our focus here is on El Niño/Southern Oscillation (ENSO). We will show that the haze impacts differ substantially between the EHF and the SHF cases while the two case results do not differ much over India. Since the EHF and the SHF depict two extreme cases of the interannual variability of the haze extent, these two sensitivity simulations by the CCM3 offers insights into the potential impacts of the haze on climate variability. When ENSO variability is addressed, we will combine the CCM3 with a tropical Pacific coupled model (designed by Zebiak &
Cane19). We caution, however, that the present modeling study ignores coupling with the Indian Ocean climate (since SSTs are prescribed in the CCM3) and hence is just a first step towards an understanding of this important phenomenon.
This paper is organized into 6 sections. In section 2, the methodology and results of theINDOEX are summarized3. Section 3 introduces the two numerical experiments of the South Asian haze effects with the NCAR/CCM3. The Indian climate change due to the South Asian haze effects is discussed in section 4.
The remote impact of the haze on ENSO variability is discussed in section 5. Discussions and suggestions follow in section 6.
2. Indian Ocean Experiment (INDOEX)
Aerosol optical and chemical measurements were performed on board the C-130 (aircraft), the R/V Ronald H. Brown, R/V Sagar Kanya20 and from KCO.
The Kaashidhoo Climate Observatory (KCO) is a surface site established in 1998 on the island of Kaashidhoo in the Republic of the Maldives, about 500 to 1000 km downwind of major cities in the subcontinent21. The KCO continually measured the size distribution and chemical composition of fine particles, collected on filters and cascade impactors22. The C-130 aerial observations, for instance, show that the fine aerosol (dry mass at diameters<1 µm) was typically composed of 32% sulfate, 26% organic compounds, 14% black carbon (BC), 10% mineral
dust, 8% ammonium, 5% fly ash, 2% potassium, 1%
sea salt and trace amounts of methyl sulfonic acid (MSA), nitrate and minor insoluble species3,23. The BC aerosol and the fly ash, as observed during INDOEX, are unquestionably human-produced since natural sources are negligible1,24. The in-situ data for size resolved aerosol chemical composition;
microphysics and vertical profile were used to develop an aerosol microphysical model21.
In order to understand the low-level cloud, the geo- stationary METEOSAT and the AVHRR satellite data were used. The low-level cloud has a strong diurnal cycle, and so the METEOSAT (which can resolve the diurnal cycle) was much needed. On other hand, the METEOSAT has only 5 km resolution, while the AVHRR has a 1 km resolution without the diurnal sampling capability. The presence of the low level clouds can exert a significant influence on the radiative forcing when the aerosols are highly absorptive. The average North Indian Ocean (NIO) diurnal mean low cloud cover estimated during INDOEX is 29%, but the actual value can be as low as 15% (depending on the threshold temperature used to distinguish clear from cloudy pixels in the satellite radiances). Uncertainty in the satellite retrieved low cloud fraction is the largest source of errors in the cloudy sky aerosol forcing estimates.
A) Direct forcing in clear skies
During INDOEX, the direct aerosol radiative forcing for clear skies was estimated solely with observations by correlating surface and TOA [top of atmosphere] radiation fluxes with column aerosol optical depth (AOD). AOD was measured by sun photometers, the surface flux by pyranometers, and the TOA flux by CERES (Clouds and Earth’s Radiant Energy System) budget instrument. The aerosol forcing at the surface, when subtracted from the forcing efficiency at the TOA, yields atmospheric absorption. This approach works mainly where surface measurements are available. Figure 3 shows how Satheesh & Ramanathan25 used the KCO observations to calculate the clear-sky aerosol forcing at the surface, TOA and in the atmosphere. Their25 finding was striking in that the surface forcing is about 3 times as large as the TOA forcing, indicative of a significant atmospheric absorption.
B) MACR
The Monte-Carlo Aerosol Cloud Radiation (MACR) model was developed26, and used in two modes3-1) to retrieve aerosol optical depths (AODs)
from satellite data, thus extending in-situ data to regional scales; and 2) in conjunction with satellite low-cloud data to obtain the regional direct and indirect forcing. As Fig. 4 shows, the MACR ingests in-situ (C-130 and Male Raman Lidar) and surface (KCO and R/V “Ron Brown”) chemical and microphysical data (the left box and arrow leading into MACR) to develop an aerosol optical model.
MACR adopts three types of sky conditions, specifically, clear skies, low clouds (including cirrus), and convective clouds including anvils, fully accounting for the cloud-aerosol effects on the radiation. It is through these effects on radiation that
MACR accurately produces either AODs or aerosol radiative forcings.
C) Aerosol forcing and low-level cloud
The aerosol forcing depends on the altitude level at which the aerosol is located and whether the aerosol is above or below clouds10. While the role of absorbing aerosols is enhanced at higher altitudes, particularly if above clouds, the role of sulfate decreases due to humidity effects11. A key parameter in this regard is the single-scattering albedo (SSA), ω = Cs/(Cs + Ca), where Cs is the scattering coefficient and Ca is the absorption coefficient. In the visible spectrum, the
Fig. 3—a) Diurnal aerosol forcing at the surface versus AOD at 0.5 µm. The circles correspond to the sum of direct and diffuse radiation and triangles correspond to global radiation measured by pyranometers. Open and solid symbols represent 1998 and 1999, respectively.
Model-0.9 and Model_087 correspond to SSA of 0.9 and 0.87. The vertical and horizontal bars give the uncertainties in the forcing and AOD. b) As in a), but at the TOA. The vertical bar represents the pixel-to-pixel variability of the TOA fluxes. (modified from Satheesh &
Ramanathan25)
SSA ranges from about 0.99 for pure sulfates to 0.96 for fly ash, 0.8 for dust and 0.23 for soot (black carbon cluster). Thus soot, only at 14% of the total fine particle mass, upholds the haze SSA around 0.9 and can enhance the aerosol absorption by an order of magnitude. During INDOEX, the boundary layer aerosol profile occurred about 1/3 of the time and the elevated profile about 2/3 of the time. The solar radiation reflected by low-level cloud enhances the haze absorption in the atmosphere significantly in case of the elevated profile.
When the low-level cloud fraction exceeds a certain value, the TOA forcing of this absorbing haze can even be positive28 (Fig. 5). The direct forcing of aerosols can change from large positive values over land and over extensive marine stratus decks (fraction
> 50%) to negative over sub tropical oceans with nearly cloud free or broken clouds with fraction
< 20%. The fundamental inference is that global average forcing depends critically on the specification of cloud fraction, albedo and altitude.
D) Indirect effects
Two fundamental relationships characterize the indirect effect: a) dependence of cloud drop number density (Nc) on aerosol number density (Na); and b) dependence of effective radii (Re) on Na. These parameters in turn depend on the vertical velocity, the cloud liquid water content, and the chemical composition of aerosol, which influences the CCN activation spectra. The vertical velocity and the CCN- supersaturation spectrum regulate the supersaturation, which determines the number of activated CCN; in addition the vertical velocity plays some role in determining the cloud liquid water content.
Entrainment and cloud thickness, two macroscopic parameters of cloud dynamics, also govern cloud liquid water content.
During INDOEX, the cloud and aerosol microphysical data taken at KCO and from the C-130 are used to develop a composite scheme for the first indirect effect (see details in appendix of ref. 3). This scheme estimates cloud optical depth as a function of
Fig. 4—INDOEX data integration scheme modified from Ramanathan et al.3 Ground based data are measured by the Kaashidhoo Climate Observatory (KCO) and the R/Vs “Ron Brown” and “Sagar Kanya”. C-130 and Citation are aircraft. CERES (Clouds and Earth’s Radiant Energy System), a radiation budget instrument onboard the Tropical Rainfall Mapping Mission (TRMM) provides the TOA forcing. The second indirect effect and the semi-direct effect are obtained by 3D cloud model27 (Ackerman et al.27).
aerosol optical depth. The composite scheme is inserted in MACR, in conjunction with the satellite data for cloud cover and aerosol optical depth, to estimate the regional forcing for the first indirect effect. However, INDOEX observations do not include data on the second indirect effect and the semi-direct effect. Also, these observation data were used to develop a cloud optical parameterization scheme that relates the aerosol microphysics to cloud drop and size distribution from which the first indirect effect can be theoretically derived in the post- INDOEX period. As Fig. 6 shows, this relation differs somewhat from those in different areas.
E) Aerosol assimilation over land
The land surface has spatially varying albedoes, differing from the ocean surface having a fixed albedo. Since the surface albedoes needed in retrieving the AODs and calculating the aerosol forcing are not well known for land surfaces, a 4-dimensional global aerosol assimilation model was developed especially for INDOEX16. The model of Collins et al.16’s combines assimilation of satellite
aerosol retrievals and a chemical transport model.
This model initializes daily values of satellite retrieved AODs for the ocean and the aerosol surface emission data for land, and uses model winds to advect the aerosol horizontally and vertically. In another sense, the modeled aerosol forcing in this fashion is constrained and corrected by those derived over the ocean with direct observations and MACR.
The model predictions were extensively evaluated against data collected during the field campaign.
F) Overall description of the SA haze radiative forcing
First, we describe the January–March 1999 averaged aerosol (natural + anthropogenic) direct radiative forcing averaged over 0–20ºN and 40–100ºE shown in Fig. 7. Before the area averaging, the forcing was estimated as a function of latitude and longitude. For clear skies, the TOA forcing [F(T)] is – 7 ±1 W m-2. For average cloudy skies, the TOA direct forcing ranges from –0 to –4 W m-2, with a mean of –2 W m-2. Inclusion of clouds introduces a net heating of about 5 W m-2 (difference between clear and cloudy sky forcing) largely because of the aerosol absorption of the reflected radiation from clouds. Roughly 50% of the ± 2 W m-2 estimated error in the cloudy sky direct forcing is due to the uncertainty in cloud fraction; about 25% is due to the uncertainty in the vertical profile of aerosols and the balance of 25% is due to the uncertainty in clear sky forcing. Figure 7 also shows the forcing for the surface [F(S)] and the atmosphere [F(A)]. In summary, the TOA, atmosphere and surface forcing values are respectively: F(T) = –2 ± 2.0 W m-2; F(A) = 18 ± 3 W m-2; F(S) = –20 ± 3 W m-2. The cloudy sky surface direct forcing is a factor of 10 larger than the TOA forcing, as opposed to a factor of 3 for clear skies reported by Satheesh &
Ramanathan25. About 60% of the large atmospheric forcing, F(A), is due to soot and the balance is due to fly ash, dust and water vapor absorption of the radiation scattered by aerosols.
The regional average forcing due to the first indirect effect is –5 ± 2 W m-2 (TOA), +1 ± 0.5 W m-2 (atmosphere), and −6 ± 3 W m-2 (surface). The large error bars are due mostly to the uncertainties in the satellite retrieved cloud fraction. It is interesting to note that the TOA negative forcing due to the indirect effect is the same magnitude but opposite in sign to the positive forcing due to the direct effect of clouds;
i.e., clouds on one hand decrease the negative direct forcing by 5 W m-2 (from the clear sky
forcing of
Fig. 5—Diurnal average aerosol direct forcing at the TOA (upper panel) and at the surface (lower panel) versus low level cloud fraction and SSA (modified from Podnorny and Ramanathan28).
Note that the low-level cloud fraction can change even the sign of the TOA aerosol forcing.
Fig. 6—Aircraft data illustrating the increase in cloud drops with anthropogenic aerosol concentration in case of INDOEX observations and other observations. (modified from Ramanathan et al.17)
Fig. 7—Aerosol direct radiative forcing for the North Indian Ocean and adjacent subcontinents (0° to 20°N; 40° to 100°E). The values include the effects of natural and anthropogenic aerosols. The values on top of each panel reflects TOA forcing; within the box shows the atmospheric forcing and below the box is the surface forcing. (modified from Ramanathan et al.3)
−7 W m-2 to –2 W m-2 in cloudy skies), while on the other hand they enhance the negative forcing by –5 W m-2 through the first indirect effect. This important new result is largely because the absorbing aerosol layer is located within and above the tops of low clouds.
The total aerosol forcing (direct + indirect) is one order of magnitude larger than the greenhouse gas forcing at the surface or in the atmosphere17. The greenhouse forcing, though much homogeneous spatially, is about +1 W m-2 (surface), +1.6 W m-2 (atmosphere) and +2.6 W m-2 (TOA). The greenhouse forcing is only comparable to the aerosol forcing, when the TOA values are compared or the global mean is taken. Therefore, the aerosol climate forcing should be dealt with at regional scales, and at the surface or in the atmosphere (as opposed to at the TOA).
3. INDOEX Aerosol Forcing into NCAR/CCM3 The atmospheric component of CCM3 model29 we have used, is a spectral model with a triangular truncation at wavenumber 42 (T42) and with 18 hybrid sigma-pressure vertical levels. The CCM3 includes a land surface model30, and is forced by prescribed SSTs. In our numerical experiments, the model is run with a seasonally varying climatological SST, as generated with the observed 1950–79 SSTs.
In the control run, the model was run without the haze effects for 85 years with the repeating SST seasonal cycle. In each INDOEX haze experiment, the aerosol forcing is applied every winter/spring season repeatedly during the model integration; all the experiments are of 60 model years or more. A lengthy integration was undertaken to ensure the statistical significance of the results presented here relative to model internal variability. More details of the experiments are found in Chung et al.31 and Chung &
Ramanathan32.
Inserting the aerosol effects in a climate model can be done in two ways. In one way, the observed aerosol properties and optical depths are prescribed, and the model radiation code generates the aerosol radiative forcing using model generated cloud fields.
We adopt the other approach here, in which we impose a regional distribution of the aerosol forcing due to the haze as summarized in section 2. The rationale for directly prescribing the heating field is to keep the radiative forcing as close to the observed values as possible, given that the aerosol forcing depends very strongly on cloud fraction (see section
2C). If the model cloud fraction is very different from the observed values, the aerosol forcing (even its sign at the TOA) will depart drastically from the observed values.
As discussed in section 1, the spatial extent of the South Asian haze varies from one year to another. In view of this, we designed two cases: 1) EHF (Extended Haze Forcing) case has the forcing applied northward of 10°S, and 2) Shrunk Haze Forcing (SHF) is restricted to the regions north of the equator (as shown in Fig.8). For the EHF case, the aerosol radiative forcing at the surface is −24.5 Wm-2, and for SHF –13.5 Wm-2 (averaged over 0–20ºN and 40–100ºE) from January to March. The INDOEX estimate during January–March 1999 is −17 to −23 Wm-2 for the direct effects, and so these two cases can be understood as essentially bracketing the uncertain haze forcing estimate as well. In both experiments, the ratio “R” of the prescribed surface forcing to the atmospheric forcing was kept as estimated during INDOEX. The “R” varies spatially because of its dependence upon the cloud fraction and the surface albedo, and each experiment achieved this spatial variation as well. The effective value of R averaged over land is about −0.9 in both experiments. To understand the significance of the R, we conducted a 3rd experiment by altering the EHF experiment. This experiment is identical to the EHF, except that the
“R” is strictly −1.5 everywhere. The third case is referred to as R=−1.5 experiment, and essentially the clear-sky radiative forcing was prescribed to the CCM3 in this experiment.
Figure 8 shows the imposed solar heating rates (upper panels) and the computed reduction in the solar flux at the surface (lower panels) for the EHF and SHF cases. As shown in this figure, the haze forcing has two opposing effects: a cooling effect at the surface and a warming effect in the lower atmosphere.
4. South Asian Climate Response
The local impacts refer to the changes in the region within the domain of the imposed haze forcing.
Changes outside this region are referred to as remote effects. We find that the local climate changes are very similar between the two haze forcings, particularly over India where the forcing is centered (not shown). The local climate changes are similar perhaps because the changes are driven radiatively. In what follows, we mainly discuss the India-area mean climate change in the EHF case, and its sensitivity to
the “R” – the ratio of the surface forcing to the atmospheric forcing.
Figure 9a displays the India-area mean temperature changes at the surface and at 2 km altitude. The dominating feature is the surface cooling and the low- level warming during the seasons when the haze forcing was imposed. The surface cooling is largest in February with a magnitude of 0.9 K, while the lower atmospheric warming is largest in March. The imposed haze forcing is maximum in March–April.
The surface cooling maximum precedes the atmospheric warming peak (and imposed forcing peak), perhaps because the wintertime has a more stable boundary layer and thus enables the surface to respond more to the surface haze forcing than the springtime when the surface is climatologically warmer and the boundary layer mixing is more active.
Figure 9a also indicates more stable boundary layers due to a decreasing lapse rate. The stablization
of the boundary layer is visualized in more detail in Fig. 9b, where the vertical temperature profile change is shown. This panel also includes the temperature change for R=−1.5 experiment (same as EHF experiment except with a differing “R”). In case of the EHF (where R≅−0.9), the layers below 0.5 km cool, and in case of “R=−1.5” experiment the layers below 1.2 km cool. The surface cooling amplitude and the vertical extent of the near-surface coolings depend crucially on the ratio “R”.
Krishnan & Ramanathan33 looked at the observed long-term India-mean surface temperature, and verified our finding qualitatively. They used the aerosol forcing characteristics of being seasonally asymmetric and regionally confined in the analysis of the surface temperature data. As Fig. 10 shows, the Indian subcontinent has a cooling trend during the dry season in the recent decades, when contrasted to the wet season value or to the global mean value. Their
Fig. 8—Two CCM3 experiments of the South Asian (SA) haze effects using the NCAR/CCM3. (left panels) Extended Haze Forcing (EHF) experiment, and (right panels) Shrunk Haze Forcing (SHF) experiment (modified from Chung & Ramanathan32). All the panels show the Jan.–Mar. difference between an experiment and the control run (no haze effect). The CCM3 was integrated with the climatological seasonal cycle of SSTs in all runs. In each experiment, the low-level atmospheric solar heating was increased and the net surface solar flux was reduced.
idea is to filter out the greenhouse driven warming by using the fact that the south Asian haze is much more intense during the dry season due to the subsidence and the haze forcing is very negligible far outside of the Indian region. Without such a contrast, the observed data does not show a conspicuous cooling trend.
It is very important to note that the haze stablizes the boundary layer, because the stabilization will trap the pollutions longer, thus making the pollution removal harder. As a result of the stablization, the sensible heat flux from the surface decreases significantly (by 16 W/m2 in March; see Fig. 11) and this decrease balances most of the solar flux reduction in surface energy budget. The remaining balance of the surface solar radiation reduction is compensated by decreased surface evaporation (by about 6 W/m2) and a decrease in net (up minus down) longwave radiation. Suppressed evaporation indicates a slowed- down hydrological cycle in this area. The haze effects on the summertime climate are to be explored in the future.
Fig. 10—Observed surface temperature analysis. (a) The thin green curve is the time series difference of temperature variations between the dry season (Jan.–Mar.) and the rest of the year (June–
Dec.) over Indian subcontinent. The think blue curve is the difference between the variations of temperature over India and the global mean for the (Jan.–May) season. (b) The thick yellow curve is for the summer monsoon (June–Sep.) season; the thin red curve is for the post-monsoon (Oct.–Dec.) season. (modified from Krishnan & Ramanathan33)
Fig. 11—Changes in the surface heat fluxes (“R≅−0.9”
experiment, i.e., EHF case). Shown is the average over India where the imposed haze forcing is centered.
5. Impact on ENSO
Figures 12 shows the change in the velocity potential at the 200 hPa and 850 hPa levels to give an overview of the local and remote effects. One conspicuous feature in all panels is a global response to the haze effects. The response is the wave number one type, i.e., planetary scale response to the imposed heating, but the spatial extent and the magnitude are much smaller and weaker for the SHF effects. As can be inferred from this figure the EHF drives one order of magnitude bigger global impacts than the SHF as far as the mean climate change is concerned. The
Fig. 9—Simulated temperature changes averaged over India: a) EHF experiment (i.e., “R≅−0.9” experiment), and b) “R=−1.5”
experiment. This figure shows that during the period of the imposed haze forcing the boundary layer becomes stabilized.
reason for much greater global impacts by the EHF is to be explored. As we look at the details of the haze effects, we find that the South Asian haze impacts the tropical Pacific trade wind and the NH midlatitude jet stream, and here we briefly discuss its impacts on ENSO variability.
The tropical Pacific surface has warmed up in the recent decades. The warming pattern is El Niño like, that is, associated with a weakening of the zonal SST gradient34. Recent studies have attempted to attribute this El Niño like warming to a greenhouse warming.
Meehl et al.35 summarized the coupled model responses to the CO2level increase, indicating that the relative warming of the eastern Pacific was simulated only with some of the models, while some models showed the opposite response, leaving the underlying mechanism very unclear. Recently, Cai & Whetton36 proposed that the greenhouse gas induced warming is initially La Niña like and later becomes El Niño like.
The effects of greenhouse gases on the zonal SST gradient of the tropical Pacific seem very controversial. Clement et al.37 demonstrated that a uniformly increased heat forcing into the ocean leads to La Niña like warming (i.e., strengthened zonal SST gradient) because the vertical heat advection in the eastern basin becomes greater and cancels part of the warming forcing. Their calculation was later supported by an ocean GCM study (Seager &
Murtugudde38). On the other hand, Meehl &
Washington34 argued that the cloud feedback would act to reduce the zonal SST gradient in the tropical Pacific.
Our CCM3 experiments show that the Indo-Asian haze can remotely suppress the convection in the equatorial western Pacific. The convection suppression in the western Pacific leads to a weaker zonal gradient of deep diabatic heating in the tropical Pacific, and thus more relaxed trade winds. The trade
Fig. 12—Velocity potential change for the period January–March (modified from Chung & Ramanathan32). All the panels use the same contour interval.
Fig. 13—December–February precipitation change with the SHF experiment (a), and with the EHF (b). Diabatic heating (Q) change averaged over 400–500 hPa layers is displayed for the EHF experiment in (c). (modified from Chung & Ramanathan32)
wind relaxing (i.e., positive zonal wind anomaly) deepens the thermocline in the eastern basin and warms the overlying ocean. The weakened zonal gradient of SST further weakens the trade wind.
Figures 13 a and b depict the wintertime precipitation changes for the two CCM3 experiments. Both experiments show large perturbations far beyond the haze forcing area. The rainfall change in the SHF experiment is characterized by a northward migration towards the Bay of Bengal, while the EHF experiment shows an additional eastward shift. Most of the precipitation change is due to changes in convective precipitation. The convection in the western equatorial Pacific is greatly suppressed in EHF, apparently as a result of more extensive enhanced convections over the Indian Ocean. Conversely, the SHF simulation largely lacks this feature. Figure 13c shows changes in the deep diabatic heating (Q averaged over 400–500hPa layers), which resembles the precipitation changes shown in Fig.
13b. This similarity is not surprising since the latent heat component dominates the total diabatic heating in the tropics. Our explanation for the simulated convection suppression in the western Pacific in EHF is based on the constraint that convergence in one region has to be balanced by divergence elsewhere (as indeed shown in Fig. 12). As one Walker circulation cell encompasses the western Pacific (associated with the rising portion of the cell) and the Indian (sinking portion), the low-level convergence enhancement (thus precipitation enhancement) in the Indian Ocean region would adversely affect the low-level convection in the western Pacific.
Now that we have demonstrated the linkage of the haze to the convection suppression with CCM3, we explore the impact of the latter on the Pacific SST to see if the observed El Niño like warming can be accurately simulated. For this quantitative assessment, another simulation was conducted with the Cane- Zebiak coupled anomaly model19. This model is a tropical Pacific domain model exclusively employing atmosphere/ocean dynamics of relevance to ENSO.
The Cane-Zebiak model simulates mainly interannual variability, and has been popularly used for ENSO prediction. We expanded the deep heating domain of the model to include the Indian longitudes. Over this Indian-Pacific domain, the December-February mean CCM3 simulated deep heating anomalies (Fig. 13c) were added to the Cane-Zebiak model heating during December-February. The model was run with this heating anomaly for 1000 model years, and was
contrasted with the run without the heating anomaly (see Fig. 14b). We followed Chung and Nigam’s39 study, in inserting externally derived deep heating anomalies into the CZ atmosphere model.
Figure 14a shows our estimate of the observed El Niño like warming (details in Chung and Ramanathan32). When Fig. 14a is compared to Fig.
14b, one can see that the coupled model simulates a warming in the central/eastern Pacific (Fig. 14b), and the warming amplitude is very close to that of the observed warming (Fig. 14a). Figure 14c depicts the CZ model response to the deep heating perturbation from the SHF experiment. The SHF also contributes
Fig.14—Quantitatively assessing the linkage of the simulated precipitation change over the western equatorial Pacific (Fig. 12) to the observed El Niño like warming in the tropical Pacific in the last 3 decades, using the Cane-Zebiak coupled model. (a) estimate of the observed SST change up to ~2000 due to the increasing bias towards El Niño (warm) phase, as obtained by multiplying the ENSO loading vector by 0.5, (b) Cane-Zebiak coupled model response to Fig. 12c (400-500 hPa heating anomaly from the EHF experiment), and (c) CZ model response to 400-500 hPa Q anomaly from the SHF experiment. The coupled model was run for 1000 model years with this deep heating anomaly during December–February (experiment run) and without it (control run), and the annually-averaged difference between the two runs is displayed in (b)-(c). The similarity between (a) with (b) indicates that the SA haze is a partial reason for the recent El Niño like warming in the tropical Pacific. (modified from Chung &
Ramanathan32)
to El Niño like warming but only by slightly less than half of the observed magnitude. The results shown in Figs 13 and 14 suggest that the Indo-Asian haze may be partially responsible for the observed bias towards the warm state in the tropical Pacific.
6. Discussion and Suggestions
The South Asian (SA) haze is wide spread in the northern Indian Ocean and the South Asian continent during the dry season. The Indian Ocean Experiment (INDOEX) documented3,17 this haze from 1995 till 2001. Using the INDOEX estimates of the haze radiative forcing, this paper has demonstrated the potential of substantial aerosol impacts on regional and global climate through climate modeling. We note however that the specific details of the regional changes can be very model dependent and we need to repeat this study with other climate models. The potential for regional changes posed by the regionally confined SA haze prompts the need to understand the roles of other regional aerosols in the tropics. In particular, to be explored are the east Asian dust traveling across the Pacific; the Saharan dust mixed with the Sahelian biomass burning plume spreading over most of the sub-tropical Atlantic; and mostly biomass plumes from Indonesia, Brazil and Southern Africa.
The INDOEX was a major experimental effort to derive aerosol radiative forcings from observations.
We first hope for an international effort to accurately estimate the global aerosol forcing. When the INDOEX estimate of the aerosol forcing in the N.
Indian Ocean area was compared with previously- calculated values (calculated by models), differences were often at a factor of two or three. Currently, an international project to address the aerosols in the entire Asian region is underway. The Atmospheric Brown Cloud (ABC) Project2 is a new major international effort sponsored by the United Nations Environment Programme (UNEP) to address both the science and policy issues related to Asian air pollution. The ABC project is an extension of the INDOEX, and has a broad participation of the international scientific community and regional policymakers.
In this paper, we focused on the winter-spring times, and did not address the summertime forcing for the anthropogenic SA haze or the Indian summer monsoon. The summertime aerosol forcing is still unknown, and has to be estimated.
The climatic effects of the SA haze forcing are further complicated by the temporal fluctuation of the
forcing. The temporal variability aspect is particularly important, since the absorption of the solar radiation by the SA haze is another source of atmospheric diabatic heating. Atmospheric diabatic heating perturbation can impact the climate in remote regions, and furthermore such impacts might be quite different if the heating perturbation varies temporally. In particular, we would like to ask questions such as
“how has the SA haze disrupted the India summer monsoon—ENSO connection which seems behaving abnormally in the recent periods?” We treated the aerosol forcing as stationary patterns and conducted two experiments differing in the spatial extent in this study. If the SA haze varies interactively with the surrounding meteorological fields in the climate model, its climate effects may be quite different and surprising. This interactive haze forcing experiment is being planned for the next phase of the study. A lot of work needs to be done in this area. Lastly, this paper suggested a potential that the SA haze might remotely drive El Niño like warming in the tropical Pacific.
This points to the role of the aerosols in warming the surface.
Acknowledgement
We thank Dr. F. Li of Scripps Institution for providing AOD data in generating Figs 1 and 2. This work was funded by a NSF grant (ATM-0201946) and Vetlesen Foundation support to the Scripps Institution of Oceanography.
References
1 Novakov T, Ramanathan V, Hansen J E, Kirchstetter T W, Sato M, Sinton J E & Sathaye J A, Large historical changes of fossil-fuel black carbon aerosols, Geophys Res Lett, 30 (2003) 2002GL016345.
2 Ramanathan V & Crutzen P J, http://www-abc- asia.ucsd.edu/, (2002).
3 Ramanathan V, Crutzen P J, Lelieveld J, Mitra A P, et al.
Indian Ocean Experiment: An integrated analysis of the climate forcing and effects of the great Indo-Asian haze, J Geophys Res, 106 (2001) 28369-28370.
4 Twomey S, Atmospheric aerosols (Elsevier, New York) 1977, pp.302.
5 Albrecht B A, Aerosols, cloud, mircophysics, and fractional cloudiness, Science, 245 (1989) 1227-1230.
6 Rajeev K, Ramanathan V & Meywerk J, Regional aerosol distribution and its long-range transport over the Indian Ocean, J Geophys Res, 105 (2000) 2029-2043.
7 Li F & Ramanathan V, Winter to summer monsoon variation of aerosol optical depth over the tropical Indian Ocean, J Geophys Res, 107 (2002) 2001JD000949.
8 Haywood J M, Stouffer R J, Wetherald R T, Manabe S &
Ramaswamy V, Transient response of a coupled model to estimated changes in greenhouse gas and sulfate concentrations, Geophys Res Lett, 24 (1997) 1335–1338.
9 Tett S F B, Stott P A, Allen M R, Ingram W J & Mitchell J F B, Causes of twentieth century temperature change near the earth’s surface, Nature, 399 (1999) 569–572.
10 Haywood J M & Shine K P. Multi-spectral calculations of the direct radiative forcing of tropospheric sulphate and soot aerosols using a column model, Q J Roy Met Soc, 123 (1997) 1907-1930.
11 Heintzenberg J, Charlson R J, Clarke A D, Liousse C, Ramaswamy V, Shine K P, Wendisch M & Helas G.
Measurements and modeling of aerosol single-scattering albedo: progress, problems and prospects, Beitr Phys Atmosph, 70 (1997) 249-263.
12 Liao H & Seinfeld J H. Effect of clouds on direct aerosol radiative forcing of climate, J Geophys Res, 103, (1998) 3781-3788.
13 Haywood J M & Ramaswamy V, Global sensitivity studies of the direct radiative forcing due to anthropogenic sulfate and black carbon aerosols, J Geophys Res, 103 (1998) 6043-6058.
14 Krishnamurti T N, Jha B, Prospero J, Jayaraman A &
Ramanathan V, Aerosol and pollutant transport and their impact on radiative forcing over the tropical Indian Ocean during the January-February 1996 pre-INDOEX cruise, Tellus, 50B (1998) 521-542.
15 Jayaraman A, Lubin D, Ramachandran S, Ramanathan V, Woodbridge E, Collins W D & Zalpuri K S, Direct observations of aerosol radiative forcing over the tropical Indian Ocean during the January-February 1996 pre- INDOEX cruise, J Geophys Res, 103 (1998) 13,827-13,836.
16 Collins W D, Rasch P J, Eaton B E, Khattatov B, Lamarque J F & Zender C S, Simulating aerosols using a chemical transport model with assimilation of satellite aerosol retrievals: Methodology for INDOEX, J Geophys Res, 106 (2001) 7313-7336.
17 Ramanathan V, Crutzen P J, Kiehl J T & Rosenfeld D, Aerosol, climate and the global environment: A hazy future for the blue planet? Science, 294 (2001) 2041-2236.
18 Nigam S, Chung C & DeWeaver E, ENSO a diabatic heating in ECMWF and NCEP-NCA R Reanalyses, and NCAR CCM3 simulation, J Climate, 13 (2000) 3152–3171.
19 Zebiak S E & Cane M A. A model El Niño—Southern Oscillation, Mon Wea Rev, 115 (1987) 2262–2278.
20 Jayaraman A, Satheesh S K, Mitra A P & Ramanathan V, Latitude gradient in aerosol properties across the inter tropical convergence zone: Results from the joint Indo-US study onboard Sagar Kanya, Curr Sci, 80, (2001) 128-137.
21 Satheesh S K, Ramanathan V, Li-Jones X, Lobert J M, Podgorny I A, Prospero J M, Holben B N & Loeb N G, A model for the natural and anthropogenic aerosols over the tropical Indian Ocean derived from Indian Ocean Experiment data, J Geophys Res, 104 (1999) 27,421-27,440.
22 Chowdhury Z, Hughes L S, Salman L G & Cass G R, Atmopsheric particle size and composition measurements to support light extinction calculations over the Indian Ocean, J Geophys Res, 106 (2001) 28,597-28,607.
23 Lelieveld J et al., The Indian Ocean Experiment: Widespread Air Pollution from South and Southeast Asia, Science, 291 (2001) 1031-1036.
24 Cooke W F, Liousse C, Cachier H & Feichter J, Construction of a 1°×1° degree fossil fuel emission data set for carbonaceous aerosol and implementation and radiative impact in the ECHAM4 model, J Geophys Res, 104 (1999) 22,137-22,162.
25 Satheesh S K & Ramanathan V, Large differences in the tropical aerosol forcing at the top of the atmosphere and Earth’s surface, Nature, 405 (2000) 60- 63.
26 Podgorny I A, Conant W C, Ramanathan V & Satheesh S K, Aerosol modulation of atmospheric and surface solar heating rates over the tropical Indian Ocean, Tellus B, 52 (2000) 947–958.
27 Ackerman A S, Toon O B, Stevens D E, Heymsfield A J, Ramanathan V & Welton E J, Reduction of tropical cloudiness by soot, Science, 288 (2000) 1042-1047.
28 Podgorny I A & Ramanathan V, A modeling study of the direct effect of aerosols over the Tropical Indian Ocean, J Geophys Res, 106 (2001) 24097–24105.
29 Kiehl J T, Hack J J, Bonan G B, Boville B A, Williamson D L & Rasch P J, The National Center for Atmospheric Research Community Climate Model: CCM3, J Climate, 11 (1998) 1131–1149.
30 Bonan G B, A land surface model (LSM version 1.0) for ecological, hydrological, and atmospheric studies: technical description and user’s guide, NCAR Tech. Rep. TN- 417+STR, (1996) pp. 122.
31 Chung E C, Ramanathan V & Kiehl J T, Effects of the South-Asian absorbing haze on the Northeast monsoon and surface-air heat exchange, J Climate, 15 (2002) 2462–2476.
32 Chung E C & Ramanathan V, South Asian haze forcing:
Remote impacts withimplications to ENSO and AO, J Climate, 16 (2003) 1791–1806.
33 Krishnan R & Ramanathan V, Evidence of surface cooling from absorbing aerosols. Geophys. Res Lett, 29 (2002) 54-1 to 54-4.
34 Meehl G A & Washington W M, El Niño-like climate change in a model with increased atmospheric CO2
concentrations, Nature, 382 (1996) 56–60.
35 Meehl G A, Boer G J, Covey C, Latif M & Stouffer R J, The Coupled Model Intercomparison Project (CMIP), Bull Amer Soc, 81 (2000) 313-318.
36 Cai W & Whetton P H, A time-varying greenhouse warming pattern and the tropical-extratropical circulation linkage in the Pacific Ocean, J Climate, 14 (2001) 3337–3355.
37 Clement A C, Seager R, Cane M A & Zebiak S E, An ocean dynamical thermostat. J Climate, 9 (1996) 2190–2196.
38 Seager R & Murtugudde R, Ocean dynamics, thermocline adjustment, and regulation of tropical SST, J Climate, 10 (1997) 521–534.
39 Chung E C & Nigam S, Asian summer monsoon - ENSO feedback on the Cane-Zebiakmodel ENSO, J Climate, 12 (1999) 2787–2807.
