SEARCH

SEARCH BY CITATION

Keywords:

  • atmospheric aerosols;
  • aerosol and cloud;
  • tropical precipitation

Abstract

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Model and Data
  5. 3. Results
  6. 4. Conclusions
  7. Acknowledgments
  8. References

[1] The direct radiative forcing of black carbon (BC) aerosols is able to cause a significant change in tropical convective precipitation ranging from the Pacific and Indian Ocean to the Atlantic Ocean. This change occurs often well away from emission centers, demonstrating a “remote climate impact.” The detailed mechanism of this change has been analyzed in this study. In the tropical Pacific region, the pattern of BC caused precipitation change is found to be similar to the pattern of precipitation anomaly corresponding to the El Niño/Southern Oscillation (ENSO) activities. The BC forced changes in the atmospheric circulation are represented by a strengthened Hadley cell in the Northern Hemisphere, a weakened one in the Southern Hemisphere, an enhancement of the Indian summer monsoon circulation, and a reduction of the lower level easterly wind in the central and east equatorial Pacific. The latter dynamic effect of BC is specifically similar to that of an El Niño event.

1. Introduction

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Model and Data
  5. 3. Results
  6. 4. Conclusions
  7. Acknowledgments
  8. References

[2] Black carbon (BC) aerosol is a strong absorber of solar radiation in the atmosphere. This characteristic of BC leads to a direct radiative heating to the atmosphere and a strong cooling at the Earth's surface particularly of the land [e.g., Hansen et al., 1998; Haywood and Ramaswamy, 1998; Satheesh and Ramanathan, 2000; Ramanathan et al., 2001; Wang, 2004]. The former heating effect of BC aerosols to the atmosphere is very substantial as measured by its radiative forcing (∼1 W/m2 in global mean, Table 3 by Wang [2004]; note that the atmospheric forcing is defined as TOA - surface forcing). However, the accompanying surface cooling, along with other combined responses of the climate system to the unique radiative-thermodynamic effect of BC, makes this heating relatively inefficient in elevating the global-mean surface air temperature compared to carbon dioxide [e.g., Penner et al., 2003; Wang, 2004; Roberts and Jones, 2004; Chung and Seinfeld, 2005; Hansen et al., 2005].

[3] In a previous study, it has been demonstrated that the climatic effect of the direct radiative forcing of BC are more significant at the regional scale than global scale [Wang, 2004]. The most substantial change of the regional scale caused by the direct forcing of BC appears in tropical convective precipitation, exemplified by an enhancement of up to 15% in the northern precipitation band of the Intertropical Convergence Zone (ITCZ) and a nearly 30% reduction in the southern band, both on a zonally mean basis. This modeled change caused by BC aerosols is most substantial over the Pacific Ocean particularly in the central and eastern part, followed by the northern Indian Ocean; the Atlantic ITCZ is also clearly affected by BC aerosols but the strength is less significant in comparison (Figure 1d). This result has been reproduced by two independent studies using different climate models [Roberts and Jones, 2004; Chung and Seinfeld, 2005]. The changing patterns of precipitation and circulation over several specific regions revealed in these studies are consistent with the results from studies using regional-scale aerosol loadings [Chung et al., 2002; Menon et al., 2002]. The signs of the enhancement/reduction in tropical convective precipitation caused by BC aerosols are opposite to those induced by sulphate aerosol's indirect forcing that introduces an overall cooling effect to both the atmosphere and Earth's surface mainly over the Northern Hemisphere [e.g., Rotstayn and Lohmann, 2002].

image

Figure 1. Surface precipitation measurements of the DMSP satellites: (a) anomaly of the 5-year average of 2001–2005 from the 17-year mean of 1987–2005 (with 1991 and 1995 missing); (b) same as Figure 1a except for the 5-year average of 1996–2000; and (c) difference between the 5-year means of 2001–2005 and 1996–2000, derived as the former minus the latter. (d) Model simulated convective precipitation change caused by BC direct radiative forcing; shown here is the last 20-year mean derived from two 60-year long model runs. The spatial resolution is 1 degree for the DMSP data and 2.8 degree for the model. The unit of precipitation in Figure 1 is mm/year.

Download figure to PowerPoint

[4] The direct radiative forcing of BC aerosols is mainly concentrated over source regions, predominantly in the Northern Hemisphere over land owing to their rather short lifetime (about a week when averaged globally) [Wang, 2004; Haywood et al., 1997; Penner et al., 1998; Myhre et al., 1998; Cooke et al., 1999; Tegen et al., 2000; Koch, 2001; Jacobson, 2001; Chung and Seinfeld, 2002]. The ITCZ is distant from many emission centers and this suggests that the effect of BC on tropical precipitation must have been implemented first through an alternation to the atmospheric general circulation and then propagated into remote oceanic regions. Tropical convection is an extremely important part of the atmospheric general circulation. Thus to identify the impact mechanism of BC aerosols on tropical convection is highly desirable. Such a mechanism, however, has not been analyzed in detail in the previous studies. This paper presents the results of an effort using global aerosol-climate model, satellite data, and historical SST data to address this issue.

2. Model and Data

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Model and Data
  5. 3. Results
  6. 4. Conclusions
  7. Acknowledgments
  8. References

[5] The satellite data used in this study are from the Defence Meteorological Satellite Program (DMSP). These data include the monthly precipitation and surface wind speed (the latter only covers 1996–2005) at a 1-degree resolution from 1987 to 2005 (missing 1991 and 1995 in the dataset) retrieved from the measurements of the Special Sensor Microwave/Imager (SSM/I) instrument on board the satellites. Also used is the sea surface temperature (SST) data from NOAA Earth System Research Laboratory with a 2 degree resolution and covering a time period from 1854 to present. The SST data from 1987 to 2005 excluding 1991 and 1995 were used to be consistent with the DMSP data analyses.

[6] The interactive aerosol-climate model used in this study was developed based on the Community Climate Model version 3 (CCM3) of the National Center for Atmospheric Researches [Kiehl et al., 1998] and first used by Wang [2004]. The horizontal resolution of the model is 2.8 × 2.8 degree. There are 18 vertical layers from the Earth's surface to ∼3 hPa. The distribution of BC aerosols in the atmosphere are predicted undergone processes including emissions, transport and mixing, dry deposition and gravitational sedimentation, and wet removal. The calculation of these processes utilizes the predicted winds, temperature, air density and pressure, cloud cover, and precipitation by the climate dynamics model. The emissions from both fossil fuel use and biomass burning were included, the annual emissions from these two sources were ∼8 and ∼6 TgC, respectively. These emissions were first mapped onto the surface grids of the model using a population density map (Y.-F. Li, Global population distribution database, available at http://www.na.unep.net/globalpop/1-degree/) for fossil fuel emissions and biomass burning in households or the land use maps [Olivier et al., 1995] for non-households biomass burning, and then supplied to the aerosol model at every transport time step [Wang, 2004; Mayer et al., 2000]. The inclusion of BC aerosols in the 18-band solar radiation module of CCM3 [Briegleb, 1992] was formulated following Kiehl and Briegleb [1993]. The needed optical parameters of BC aerosols were derived based on the Mie scattering theory [Wang, 2004], specifically, the Model 1 of BC optical parameter set by Wang [2004] is used which has a specific extinction coefficient of 7.93 m2/g, a single scattering albedo of 0.26, and an asymmetry factor of 0.75 for the 0.35–0.64 μm waveband [Wang, 2004, Table 1]. The direct effect of BC on long-wave radiation was omitted, and so was the much more complicated and poorly-understood indirect radiative forcing of BC aerosols. Above calculations were processed at the same time step of the climate model to achieve a fully interactive coupling between the aerosol module and the CCM3 climate model. Two model runs, one (REF) which excludes and the other (BCRAD) which includes the radiative effect of BC as well as associated influences on climate dynamics that are otherwise identical, are needed to isolates the climate effect of BC (e.g., through BCRAD –REF). Coupled with a slab ocean model each of these model runs lasts 60 years to reach quasi-equilibrium. The last 20-year means of model outputs were used in the analyses in this study. Many calculations carried out in this study are based on the same output of these two model runs by Wang [2004]. In addition, 2 model runs with new diagnostic variables including BC solar forcings and 3D shortwave heating were launched. The values of these diagnostic variables were derived by repeating calls to relevant modules of the model with prognostic BC profiles without actual interruption to the real model prediction. All configurations of these 2 additional runs as well as the model results are identical to the REF run by Wang [2004]. Details of the model and model configuration are given by Wang [2004].

3. Results

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Model and Data
  5. 3. Results
  6. 4. Conclusions
  7. Acknowledgments
  8. References

[7] The statistical analyses are done by using the student t-test for the satellite data for the time period of DMSP data set and using the paired t-test for the modeled results of the last 20 years. All the results discussed are statistical significant with a significant level of 0.1 unless otherwise indicated.

[8] It has been found that in the tropical Pacific regions the model predicted changes in convective precipitation caused by BC coincide in pattern with the satellite measured precipitation anomaly of a multiple year period that is dominated by the warm events of the El Niño/Southern Oscillation (ENSO) (Figure 1). The DMSP SSM/I precipitation data have been averaged over three 5-year periods, namely 1989 to 1994 (missing 1991), 1996 to 2000 (note that 1995 data are missing), and 2001 to 2005. Measured by the Multivariate ENSO Index (MEI; see www.cdc.noaa.gov/ENSO/), the above three time periods represent approximately a warm, cold, and then warm phase of ENSO activities, respectively. In the recent half decade (2001–2005; i.e., a warm phase of ENSO), the averaged precipitation measured by the DMSP instrument is higher than the long term average of the same precipitation data through the entire 17 years of the DMSP dataset in the regions north of the equator in the Pacific ITCZ, extending from the west to central and east Pacific, and also in the south branch of the South Pacific Convergence Zone (SPCZ) (Figure 1a). Over the Pacific and south of the equator, however, the precipitation anomaly of this 5-year period exhibits a negative sign in several places including the region between the ITCZ and SPCZ and that north of the Australia. The same pattern of precipitation is also found in the time period of 1989–1994 corresponding to another warm phase of ENSO within the time frame of the DMSP dataset (not shown). This pattern of precipitation anomaly during the warm phases of ENSO activity is quite similar to the model isolated BC effect on precipitation except for the strength of the negative centers (Figure 1d). Most interestingly, a better similarity in both pattern and strength can be achieved by comparing the difference in surface precipitation between a warm and a cold phase of ENSO (here 2001–2005 and 1996–2000) with the model isolated BC effect on precipitation (Figures 1c and 1d). In contrast, precipitation anomaly during a relative cold phase of 1996–2000 (Figure 1b) displays an almost mirror image to that of Figure 1a, curiously similar to the result of indirect radiative forcing of sulphate aerosols [e.g., Rotstayn and Lohmann, 2002]. Over the north Indian Ocean and Atlantic Ocean, the similarity between modelled results and the observed precipitation anomaly derived from the satellite measurements is not as clear as in the equatorial Pacific Ocean.

[9] The above similarities between the effects on precipitation in equatorial Pacific by BC aerosols and by ENSO activities will potentially provide us with a unique method to study the climate effect of BC aerosols by taking an advantage from our cumulated knowledge about ENSO phenomena.

[10] The perturbation in zonally mean circulation caused by BC is generally symmetric across the equator in a global and year mean base, represented by an enhancement of the convection in the north as a primary effect and a consequent weakening in the south of the equator (Figure 2). It peaks during the northern hemispheric summer and weakens in the winter. Clearly, such a pattern implies an alternation to the Hadley circulation. On the other hand, BC caused solar forcing and heating mainly concentrate in the northern hemispheric midlatitudes largely over the land (Figure 2; also Figures 5 and 6 of Wang [2004] for geographic distribution of BC solar forcings), implying thus a remote impact.

image

Figure 2. Colour map shows the zonal means of shortwave heating rate of BC aerosols in Kelvin per year. This rate is derived diagnostically by launching a repeat call to the radiation module with predicted BC profile without sending the output to the real model prediction procedure and thus does not reflect any response and feedback to the BC direct forcing. Black dashed lines represent the zonal means of BC caused perturbation in stream function as counter-clock wise circulations in 109 kg/s. The zonal means of BC caused vertical velocity change (here actually measured by pressure change, i.e., ω = dp/dt) are shown in red (negative values and thus upward perturbations) and blue lines (positive values and thus downward perturbations). The contours of this quantity shows −0.5, −0.25, 0.25, and 0.5 in 10−4 hPa/s. All quantities are the last 20-year means from the 60-year model runs. (top) The yearly means, (middle) the June-July-August (JJA) means, and (bottom) the December-January-February (DJF) means.

Download figure to PowerPoint

[11] Detailed analysis has further revealed that BC aerosols are able to cause at least three alternations to the lower level atmospheric circulation (Figure 3, top). In the east tropical Pacific (here from 180 W to 100 W), such alternation is represented by a strong gradient of temperature anomaly between regions north of the equator and those south of the equator along with a strong dynamic perturbation in the north that reduces the easterly wind. Over the western and central Pacific (120 E to 180 E), BC direct forcing causes an eastward enhancement in the low level wind accompanying a positive temperature change in the central Pacific. The third alternation in the atmospheric circulation by BC is an enhancement in the summer monsoon circulation over the northern Indian Ocean, driven by a BC-forced positive temperature gradient in the lower troposphere from the Arabian Sea and Bay of Bengal to the equatorial Indian Ocean. Because the difficulty in searching the observations of lower level temperature, the sea surface temperature data (SST) are used to compare with the model results. Also used is the DMSP SSM/I wind speed data. These data were averaged into the same three 5-year periods as did for the SSM/I precipitation data. Shown in Figure 3 (bottom) are their anomalies in the 5-year period of 2001–2005. Again, the dynamic and temperature effects of BC aerosols over the tropical Pacific Ocean are similar in many aspects to the anomalies of the SST and surface wind speed during a warm phase of ENSO activity (Figure 3, bottom), as if the atmospheric temperature anomaly forced by BC was caused by the perturbation of SST attributed to ENSO especially from the central to the east tropical Pacific.

image

Figure 3. (top) Temperature (colour map) and wind (vector) changes in the lowermost troposphere caused by BC aerosols. The length of the wind vector between two grids represents 1 m/s. Data shown are the last 20-year means averaged in the lowermost 3 model layers ranging from the surface to ∼1 km above the sea level. (bottom) Colour map shows the sea surface temperature (SST) anomaly of the 5-year average of 2001–2005 from the 17-year mean SST of 1987–2005 without 1991 and 1995 in order to be consistent with the DMSP precipitation data. The black lines show the surface wind anomaly larger than 0.15 m/s (with an incremental interval of 0.15 m/s) of the 5-year period of 2001–2005 from the 10-year mean surface wind speed (1996–2005). The SST data are NOAA-ERSST_v2 in 2 degree resolution. Surface wind speed data are from the DMSP retrievals in a 1-degree resolution, note that it only covers 1996–2005.

Download figure to PowerPoint

[12] For the Pacific ITCZ, the pair of enhancement/reduction of convection and thus the meridional Hadley circulation distributes across the equator in the months of June, July, and August (JJA) but moves into the Northern Hemisphere during the months of December, January, and February (DJF) with a reduced magnitude compared to the case of JJA (Figures 4a, 4b, and 4c). Similar to the ENSO effect, the effect of BC aerosols is to form an anomalous temperature and pressure gradient from the west to central and east in equatorial Pacific. This gradient consequently leads to a reduction of the easterly surface wind in the central and east Pacific and an eastward enhancement in the surface wind over the west to central Pacific, enabling the convection and precipitation to enter the central and eastern Pacific more easily (Figure 4d). This mechanism resembles the effect of an El Niño event on Pacific convection and precipitation distribution [Rasmusson and Carpenter, 1982]. In terms of circulation, the above effect is equivalent to a reduction in the Walker circulation. Several previous modeling studies have suggested the similar effect on the Walker circulation by the global warming caused by the greenhouse gases [e.g., Knutson and Manabe, 1995; Vecchi et al., 2006]. Nevertheless, the result of this study demonstrates that BC aerosols are also efficient in causing the changes to the atmospheric circulation over the Pacific Ocean. Note that the excluded indirect forcing of BC in this study might quantitatively change the results described here, presumably the addition from BC aerosols to the number concentration of other existing hygroscopic aerosols in affected region is significant. The indirect forcing of BC would be a negative TOA forcing in the solar band that is opposite in sign to its direct forcing. This would, however, likely be compensated by the consequent longwave forcing with a warming effect.

image

Figure 4. (a) The annual, (b) JJA, and (c) DJF means of BC caused changes in stream function and vertical velocity over the Pacific, averaged from 120 E to 90 W in latitude. The lines and colours have the same representations as in Figure 2. Note that the DJF vertical velocity change in Figure 4c is not statistical significance based on paired t-test. (d) The latitudinal means of BC caused changes in zonal wind speed averaged over the lowermost 3 model layers (dU, the red line) and sea level pressure (dPSL, the blue line), averaged from 5 S to 5 N in latitude, each accompanied with a pair of dotted lines representing respectively the 5% and 95% percentile that show the range of the 0.1 significance level from the t-test. Note that the hypothesis of E{dU} = 0 over the Pacific is not rejected by the 0.1 but 0.3 significant level (not shown) due to the low variability nature of this averaged property. All data are the last 20-year means derived from two 60-year long model runs.

Download figure to PowerPoint

4. Conclusions

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Model and Data
  5. 3. Results
  6. 4. Conclusions
  7. Acknowledgments
  8. References

[13] In summary, the direct radiative effect of BC aerosols can force a significant change in atmospheric circulation and precipitation in the tropics, ranging from the Pacific and Indian Ocean to the Atlantic Ocean. This change often occurs away from the source regions of BC, demonstrating a “remote climate impact” of BC aerosols that perhaps can be extended to all types of aerosols. The changes in convective precipitation over the Pacific Ocean are found to be very similar to the difference in tropical Pacific precipitation between a warm and a cold period of ENSO activity. Based on this result, it appears to be a reasonable conjecture that the detailed forcing mechanism of BC on atmospheric circulation and thus the tropical precipitation over Pacific is similar to that of an (perhaps strong) El Niño in comparison to a La Niña event, that is the BC caused changes in convection and precipitation over the entire tropical Pacific can be triggered by the BC heating to the lower troposphere only in a few specific areas such as the east most and also the west tropical Pacific regions. This BC heating over above regions can form a temperature and thus pressure anomaly over the equatorial Pacific favoring the propagation of convection from the western to the central and eastern Pacific. The enhancement of convection in the Northern Hemisphere can also extend the BC effect to the Southern Hemisphere by consequently suppressing the convection and thus precipitation there. Such a conjecture is apparently supported by the modeled distribution of BC aerosols that concentrates in both ends of Pacific rather than being transported into the remote equatorial Pacific. It can be also hypothesized that the direct radiative forcing of absorbing aerosols might have enhanced the warm phase while lessened the cold phase of ENSO. However, this cannot be examined until we are able to separate the signal of aerosol forcing from natural variability of tropical precipitation.

Acknowledgments

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Model and Data
  5. 3. Results
  6. 4. Conclusions
  7. Acknowledgments
  8. References

[14] I thank C. Zhang, A. Sokolov, and R. Prinn for useful discussions. I also appreciate the effort of M. Mayer, J. Reilly, and M. Babiker that led to the derivation of the BC emissions data, the help offered by J. Kiehl, P. Rasch, and W. Collins during the model development, and the constructive comments from two anonymous reviewers. The DMSP satellite data were kindly provided by the National Climate Data Center of US National Oceanic and Atmospheric Administration (NOAA). NOAA_ERSST_V2 data provided by the NOAA/OAR/ESRL PSD, Boulder, Colorado, USA, from their Web site at http://www.cdc.noaa.gov/. This research was supported by the US National Science Foundation (grant ATM-0329759), by NASA (grant NNG04GP30G), and by the MIT Global Change Joint Program.

References

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Model and Data
  5. 3. Results
  6. 4. Conclusions
  7. Acknowledgments
  8. References