1. Introduction
The impact of aerosol on cloud and precipitation is an important issue in climate change studies (Penner et al. 2001). However, assessing the effects of aerosol has been difficult because of the complexity of aerosol-related processes and the involvement of atmospheric dynamics and thermodynamics in the aerosol–cloud–precipitation relationship. The effects of aerosol include both direct radiative effects and indirect effects on cloud properties (Twomey et al. 1984; Albrecht 1989; Penner et al. 2001). Recent studies suggest that meteorological conditions may affect the aerosol–cloud relationship (Sekiguchi et al. 2003; Ackerman et al. 2004; Matsui et al. 2006; Fan et al. 2007; Mauger and Norris 2007). It is indicated that large-scale atmospheric circulation may modulate both aerosol and cloud–precipitation variations on intraseasonal (Tian et al. 2008; Beegum et al. 2009a) and on interannual (quasi biennial) time scales (Beegum et al. 2009b). In addition, uncertainty in satellite–aerosol products in the presence of clouds (Shi et al. 2011; Chand et al. 2012) increases the difficulty of unraveling the causal relationship between aerosol and cloud variations.
Different approaches have been employed to understand the effects of aerosol on cloud and precipitation. One is based on data limited in both the space coverage and the time period, such as field experiments over a certain region (e.g., Rosenfeld 1999, 2000; Andreae et al. 2004). The results obtained from these analyses may only be applicable to short-period variations and to specific regions. Another approach is to conduct numerical simulations with climate models with specified aerosols (e.g., Menon et al. 2002; Qian et al. 2003; Lau et al. 2006). In reality, the distribution and amount of aerosols depend on wind, precipitation, and stability and there are interactions between aerosol and atmospheric dynamics and thermodynamics. Thus, it is necessary to understand the effects of aerosol variations for a better application of model results of the aerosol's effects. In this study, we are concerned with factors for interannual variations of aerosols over the Maritime Continent and the western North Pacific Ocean.
The aerosol effects on cloud and precipitation depend on the cloud (Lee et al. 2008), aerosol (Fan et al. 2007), and precipitation (Wang et al. 2011) type, as well as the moisture availability and the vertical wind shear (Fan et al. 2007, 2009; Lee et al. 2008). As such, aerosol effects may display strong regional features. For example, Lin et al. (2006) indicated that aerosols increase precipitation over the Amazon region, whereas Rosenfeld (1999, 2000) found that aerosols inhibit precipitation over the Indonesian region. The aerosol effects may also change from time to time, which may be the reason for the contrasting relationship in the Amazon region identified in previous studies (Andreae et al. 2004; Koren et al. 2004; Lin et al. 2006). For example, Yu et al. (2007) identified an opposite smoke and warm cloud relationship in 2002 and 2003. As a result of the temporal change of the aerosol effects and the impacts of atmospheric processes on the aerosol concentration, it is possible that the aerosol–cloud–precipitation relationship may depend on the time scale.
In the present study, we focus on the relationship between interannual variations of monthly-mean aerosol, cloud, and precipitation on large spatial scales, which is relevant to the year-to-year climate variability. This interannual relationship may differ from short-period relationships that likely vary with time and from small spatial scale relationships that are of local features. In particular, we are concerned with the effects of large-scale circulation and precipitation changes on the interannual variations of aerosols. Such large-scale effects are manifestations of different processes under the modulation of various meteorological conditions. Our approach is similar to that employed by Jin and Shepherd (2008), but their analysis is confined to coastal regions of eastern China. Our analysis covers the global domain for the purpose of unraveling the different types of aerosol–cloud–precipitation relationships. In addition, the length of data in the present analysis is about double of that used by Jin and Shepherd (2008).
In the following, the datasets used in the present analysis are described in section 2. Section 3 examines the local aerosol–cloud–precipitation relationship on the global domain. Section 4 analyzes aerosol variations over the Maritime Continent and the western North Pacific and the impacts of tropical Pacific and Indian Ocean sea surface temperature (SST) anomalies. Section 5 examines the spatial and temporal evolution of composite anomalies to demonstrate the large-scale control of biennial aerosol variations during 2001–07. A summary is given in section 6 along with discussions.
2. Datasets and methods
The datasets used in the present study include the following:
Monthly-mean aerosol optical depth (AOD) and cloud fraction from the Moderate Resolution Imaging Spectroradiometer (MODIS) dataset (Pincus et al. 2012) at 0.55 μm for the period of March 2000–December 2011, which has a horizontal resolution of 1° latitude by 1° longitude.
Monthly-mean SST from the Tropical Rainfall Measuring Mission (TRMM) Microwave Imaging (TMI), version 4, which has been available since November 1997 and has a horizontal resolution of 0.25° latitude by 0.25° longitude (Wentz et al. 2000).
Monthly-mean precipitation from the Global Precipitation Climatology Project (GPCP; Adler et al. 2003) on a 2.5° latitude by 2.5° longitude grid and available from January 1979 to December 2010.
Monthly-mean surface winds from the National Centers for Environmental Prediction–Department of Energy (NCEP–DOE) reanalysis, version 2 (Kanamitsu et al. 2002), which has been available since January 1979 on the T62 Gauss grid.
In addition to the above datasets, the present study uses the Multiangle Imaging SpectroRadiometer (MISR) AOD at 555 nm (Marchand et al. 2010) on a 0.5° × 0.5° grid covering the period of February 2002–June 2012 and the Sea-viewing Wide Field-of-view Sensor (SeaWiFS) AOD at 550 nm (Sayer et al. 2012) on a 1° × 1° grid for the period of September 1979–December 2010. We also use cloud products of CloudSat (Mace et al. 2009; Zhang et al. 2010) with a resolution of 2° × 2° and GCM-Oriented Cloud–Aerosol Lidar and Infrared Pathfinder Satellite Observation (CALIPSO) Cloud Product (GOCCP; Chepfer et al. 2010) with a resolution of 2.5° × 2.5°, both of which are available after June 2006, and International Satellite Cloud Climatology Project (ISCCP) cloud dataset (Rossow and Schiffer 1999) on a 2.5° × 2.5° grid available for the period of July 1983–December 2009. These additional AOD and cloud fraction datasets are mainly utilized to confirm the results derived based on the MODIS dataset.
The present analysis focuses on monthly-mean anomalies that are obtained by removing climatological monthly means constructed based on a multiyear average. The analysis is performed for the common period of March 2002–December 2010 during which all the variables are available. Pointwise correlation is used to document the local relationship among different variables. For the purpose of analyzing the local correlation, AOD and cloud fraction are interpolated to the precipitation grids.
Lag–lead correlation analysis is conducted to reveal the relation of variations in different variables and the relation of aerosol variations in different regions. Composite analysis is performed to unravel common features in the spatial and temporal evolution of different variables. In this study, the composite fields are constructed with respect to October as the month when the AOD over the Maritime Continent displays peak values and the largest anomalies. Thus, October is denoted as lag = 0 and positive and negative lags refer to months after and before the peak AOD value over the Maritime Continent, respectively.
3. Types of aerosol–cloud–precipitation relationships
In this section, we examine the local correlation between interannual aerosol and cloud–precipitation variations. The spatial variations of the local correlation may suggest different types of aerosol–cloud–precipitation relationships. Figure 1 shows the gridpoint simultaneous correlation calculated based on all the monthly-mean anomalies during the analysis period.
Pointwise correlation of monthly-mean anomalies of AOD with (a) cloud fraction and (b) precipitation during the period of March 2000–December 2010. The contour interval is 0.1 with shading denoting regions where the value of the correlation coefficient exceeds 0.3. Solid and dashed lines are for positive and negative correlations, respectively. The rectangular boxes denote regions for area average in Fig. 4.
Citation: Journal of Climate 26, 17; 10.1175/JCLI-D-12-00253.1
The aerosol–cloud correlation displays obvious spatial variation, which is more pronounced than that based on short-duration data (Sekiguchi et al. 2003; Matsui et al. 2006). The overall distribution indicates a control of climatology. In the subtropical oceanic regions where descent motion dominates, a large positive aerosol–cloud correlation is seen over the three oceans (Fig. 1a). The aerosol–precipitation correlation, however, is weak in the subtropics (Fig. 1b). In the equatorial regions, a weak or negative correlation is observed. In particular, a large negative aerosol–cloud and aerosol–precipitation is seen over the Maritime Continent where convection is active and near the date line where there is large interannual variability in convection from the impacts of El Niño–Southern Oscillation (ENSO). A notable negative aerosol–precipitation is also seen over the equatorial Atlantic Ocean (Fig. 1b). The correlation is weak over most land regions except for the western part of Australia where a moderate positive aerosol–cloud correlation is observed (Fig. 1a).
The negative correlation over the Maritime Continent may include both the inhibition of aerosol on precipitation (Rosenfeld 1999) and the effect of precipitation on aerosol. Above-normal precipitation may reduce aerosol particles through the wet deposition. Dry conditions may lead to an accumulation of aerosol particles from an increase in the residence time of aerosols in the air. In addition, the dry conditions set up the fire burning during the dry season in the Maritime Continent region (e.g., Field et al. 2009).
The positive aerosol–cloud correlation over the subtropical North Atlantic Ocean has been obtained by previous studies at different time scales (Sekiguchi et al. 2003; Koren et al. 2005; Matsui et al. 2006). This suggests the possibility that the impacts of aerosol on clouds may exist on different time scales in this region. As such, the relationship appears to be robust. This positive correlation may be caused by the indirect effect of aerosol, by which the increase in aerosol particles prolongs the cloud lifetime (e.g., Myhre et al. 2007). This interpretation may apply to the other subtropical oceanic regions as well.
One region of interest to us is the region to the southeast of Japan where a positive correlation is seen between aerosol and cloud variations (Fig. 1a). The aerosol–precipitation correlation is also positive in this region (Fig. 1b), though weaker when compared to the aerosol–cloud correlation. Bao et al. (2009) indicated a possible influence of aerosol on cloud and surface shortwave radiation in this region that, in turn, contributes to in situ SST change. There is a moderate positive aerosol–cloud correlation over northern China where both the mean and variability of AOD are large (Bao et al. 2009).
We confirmed the above results, which were based on data with limited temporal coverage, by performing additional analyses with other available AOD and cloud datasets. Figure 2 shows the local correlation of MODIS AOD with the cloud fraction of ISCCP, CloudSat, and CALIPSO. The distribution of the correlations based on ISCCP cloud resembles that of Fig. 1a. A similar distribution of correlations is also seen for CloudSat and CALIPSO clouds, though the magnitude of correlations over the subtropics appears weaker, which may be because of the difference in the data periods. Table 1 presents the correlation coefficients of the MODIS cloud fraction with MISR and SeaWiFS AODs calculated based on area means over the Maritime Continent (5°S–5°N, 95°–135°E) where a negative aerosol–cloud and aerosol–precipitation correlation is observed, the southeast of Japan (30°–35°N, 135°–165°E) where a positive correlation is seen both between aerosol and cloud and between aerosol and precipitation, and the subtropical eastern North Pacific (15°–25°N, 160°–130°W) where there is a positive aerosol–cloud correlation but a weak aerosol–precipitation correlation (Fig. 1). The negative correlation over the Maritime Continent and the positive correlation over the region to the southeast of Japan and the subtropical eastern North Pacific are reproduced, though the magnitude of the correlation coefficient display differences, which may be related to the lower sampling frequency of MISR and SeaWiFS compared to MODIS. The correlation coefficients of area-mean monthly AOD variations over the regions of 5°S–5°N, 95°–135°E and 30°–35°N, 135°–165°E reach 0.87 (0.91) and 0.70 (0.80), respectively, between MODIS and MISR (SeaWiFS).
Pointwise correlation of monthly-mean anomalies of MODIS AOD with (a) ISCCP, (b) CloudSat, and (c) CALIPSO cloud fraction. The correlation is calculated over the period of March 2000–December 2009 in (a) and June 2006–December 2010 in (b) and (c). The contour interval is 0.1 with shading denoting regions where the value of the correlation coefficient exceeds 0.3. Solid and dashed lines are for positive and negative correlations, respectively.
Citation: Journal of Climate 26, 17; 10.1175/JCLI-D-12-00253.1
Correlation coefficients of monthly-mean anomalies of MODIS cloud fraction with MODIS, MISR, and SeaWiFS AOD averaged over the Maritime Continent (5°S–5°N, 95°–135°E), the region to the southeast of Japan (30°–35°N, 135°–165°E), and the subtropical eastern North Pacific Ocean (15°–25°N, 160°–130°W) during the period of March 2000–December 2010.
The ground-based aerosol observations from the Aerosol Robotic Network (AERONET; Holben et al. 1998) provide AOD measurements against which the satellite retrievals may be validated (e.g., Kleidman et al. 2005; Remer et al. 2005; Kahn et al. 2007; Shi et al. 2011). In the context of this study, we identified one station (Singapore at 1°N, 103°E) in the region of 5°S–5°N, 95°–135°E and two stations in the region of 30°–35°N, 135°–165°E that have multiyear observations (starting in November 2006 for Singapore, October 2000 for Shirahama, and September 2000 for Osaka). While there are several other stations in the above regions, the temporal coverage is short (only available after August 2011 at the best). We have compared the daily-mean AOD variations at Singapore and Shirahama at the 500 nm of level 1.5 AERONET observations with the corresponding MODIS AOD variations averaged over the 1° × 1° boxes centered at these two stations. The correlation coefficient at Singapore is 0.40 for daily means in October 2007–11 and at Shirahama is 0.61 for daily means in June 2001–09. Note that the correlation for Singapore is relatively low likely because the large-scale control (e.g., by ENSO) is less pronounced during 2007–11 than during 2001–07 (see Fig. 6, described in greater detail below).
The ISCCP and CALIPSO cloud products include different levels of clouds. This provides an opportunity to examine the aerosol–cloud relationship for different types of clouds [for the classification of cloud types, refer to Rossow and Schiffer (1999)]. Figure 3 displays the correlation of MODIS AOD with ISCCP deep convective cloud, low-level cloud, and midlevel cloud. The high-level cloud (including deep convective cloud) correlation shows a distribution similar to the deep convective cloud correlation with some slight differences in the magnitude of the correlation coefficient. According to Fig. 3, the negative AOD–cloud correlation over the Maritime Continent is mainly caused by the deep convective cloud, and the low-level cloud has an opposite correlation. The positive correlation over the subtropical oceanic regions appears to be contributed by clouds at different levels with a relatively large contribution from the midlevel cloud. The correlation of low-level, midlevel, and high-level clouds based on CALIPSO tends to display an overall similar feature (not shown). Here, we want to point out that the negative correlation over the Maritime Continent is indicative of physical connections and is not an artifact of the cloud contamination on AOD. This is because the cloud contamination is expected to lead to an increase in AOD with more clouds and thus a spurious positive correlation.
Pointwise correlation of monthly-mean anomalies of MODIS AOD with ISCCP (a) deep convective, (b) low-level, and (c) midlevel cloud fraction over the period of March 2000–December 2009. The contour interval is 0.1 with shading denoting regions where the value of the correlation coefficient exceeds 0.3. Solid and dashed lines are for positive and negative correlations, respectively.
Citation: Journal of Climate 26, 17; 10.1175/JCLI-D-12-00253.1
To further understand the relationship among aerosol, cloud, and precipitation variations, we examine the lag–lead correlation with respect to area-mean AOD for three regions of noticeable correlations (Fig. 4). Apparently, different correlations are seen in the above three regions. In the Maritime Continent region, positive AOD anomalies correspond to negative cloud and precipitation anomalies (Fig. 4a). The temporal evolution of these correlations displays a pronounced biennial feature. This feature suggests a large-scale control on the interannual variations of aerosol, cloud, and precipitation in this region, which will be demonstrated in the next section. In the region to the southeast of Japan, positive AOD anomalies correspond to positive cloud and precipitation anomalies (Fig. 4b). Clear biennial variations are visible, albeit not so pronounced as in the Maritime Continent region. In the subtropical eastern North Pacific, a large simultaneous aerosol–cloud correlation is only seen at lag 0 (Fig. 4c). At the other lag times, the correlation is generally weak. This suggests that in this region the aerosol variability is a high-frequency feature with no clear signal of large-scale control. A similar feature is seen over other subtropical oceanic regions.
Lag–lead correlation of monthly-mean anomalies of AOD (solid thick curves), cloud fraction (dashed thick curves), and precipitation (solid thin curves) with respect to AOD in the regions of (a) 5°S–5°N, 95°–135°E, (b) 30°–35°N, 135°–165°E, and (c) 15°–25°N, 160°–130°W based on the period of March 2000–December 2010.
Citation: Journal of Climate 26, 17; 10.1175/JCLI-D-12-00253.1
The satellite aerosols retrievals are subject to cloud contamination and cloud artifacts (Shi et al. 2011; Chand et al. 2012), though efforts have been made to reduce these cloud effects. The monthly-mean anomalies on a large gridbox average may have removed most of the systematic biases and reduced the uncertainties related to small spatial and temporal scale features. Here, we perform a sensitivity test to examine to what extent the cloud contamination may affect the aerosol anomalies and the interannual aerosol–cloud relationship. The sensitivity test is conducted for the Maritime Continent in October and the region to the southeast of Japan in June. The above two regions show opposite aerosol–cloud correlations (Fig. 1). The AOD in the above two regions has the largest interannual variability in October and June, respectively. In the test, we recalculate monthly-mean AOD values by excluding those days with the cloud fraction exceeding 80%. This threshold value is chosen following Shi et al. (2011) as the aerosol retrievals with reported cloud fraction less that 80% are likely “good” or “best” data. We note that the selection of this threshold is subjective. The actual contamination of clouds on AOD depends not only on the cloud cover, but more importantly on the cloud thickness. Deep (thick) clouds likely introduce more contamination than thin clouds even with a smaller cloud cover.
Figure 5 shows the ratio of total monthly-mean AOD and cloud fraction values, and the monthly-mean AOD anomalies before and after the removal of daily AOD corresponding to cloud cover over 80%. According to Fig. 5, the total monthly-mean AOD value over the Maritime Continent in October is reduced by 10%–20% and over the region to the southeast of Japan in June is reduced by 10%–30%. This percent reduction in AOD is smaller than that of cloud fraction that reaches 30%–40%. The monthly-mean AOD anomalies remain at the same sign except for the year 2008 over the region to the southeast of Japan. The aerosol–cloud correlation coefficient changes from −0.90 to −0.81 over the Maritime Continent in October and from +0.79 to +0.72 over the region to the southeast of Japan in June. Thus, it appears that the cloud contamination and cloud artifacts do not significantly alter the local aerosol–cloud correlation.
The ratio of total monthly-mean AOD and cloud fraction values before and after the exclusion of daily AOD corresponding to cloud fraction exceeding 80% (curves, the rhs y axis) and monthly-mean AOD anomalies before and after the exclusion of daily AOD corresponding to cloud fraction exceeding 80% (bars, the lhs y axis) (a) in October over 5°S–5°N, 95°–135°E and (b) in June over 30°–35°N, 135°–165°E.
Citation: Journal of Climate 26, 17; 10.1175/JCLI-D-12-00253.1
4. Aerosol variations over the Maritime Continent and the western North Pacific
The time series of normalized area–mean AOD, cloud fraction, and precipitation anomalies averaged over 5°S–5°N, 95°–135°E, and 30°–35°N, 135°–165°E are shown in Figs. 6a,b, respectively. The SST anomalies averaged over the Niño-3.4 region (5°S–5°N, 170°–120°W) and the equatorial southeastern Indian Ocean region (0°–10°S, 90°–105°E) are shown in Fig. 6c for the purpose of understanding the role of tropical Indo-Pacific SST. Several important features can be noticed for the AOD variations over the Maritime Continent. Note that Tosca et al. (2010) have examined the AOD and precipitation time series in the Maritime Continent region and their causal link for the period 2000–06.
Normalized monthly-mean anomalies of AOD (solid thick curves), cloud fraction (dashed thick curves), and precipitation (solid thin curves) averaged over (a) 5°S–5°N, 95°–135°E and (b) 30°–35°N, 135°–165°E, and (c) normalized monthly-mean SST anomalies averaged over 5°S–5°N, 170°–120°W (solid curve) and 0°–10°S, 90°–105°E (dashed curve). The vertical lines denote October in (a),(c) and June in (b). The numbers in the respective panels are the correlation coefficients between two variables during the analysis period for all months in (a),(b) and October in (c).
Citation: Journal of Climate 26, 17; 10.1175/JCLI-D-12-00253.1
First, large AOD anomalies are confined to fall (Fig. 6a). The strong phase locking of large AOD anomalies in the fall of each year results in a sharp peak in monthly-mean AOD standard deviation around October (Fig. 7a). The large variability of AOD in fall is related to the fact that fires are intentionally set during the dry season in Indonesia to clear the land (Chandra et al. 2009; Ziemke et al. 2009). The large AOD anomalies in fall correspond to large, but opposite cloud and precipitation anomalies except for in October 2009 (Fig. 6a), verifying the large simultaneous negative correlation in Figs. 1 and 2. The above result agrees with Reid et al. (2012).
Std dev of monthly-mean AOD averaged over (a) 5°S–5°N, 95°–135°E, (b) 30°–35°N, 135°–165°E, and (c) lag–lead correlation of AOD averaged over 30°–35°N, 135°–165°E (solid curve) and 30°–40°N, 110°–120°E (dashed curve) with respect to AOD averaged over 5°S–5°N, 95°–135°E. The calculation is based on the period of March 2000–December 2010. The time period in (a),(b) is repeated for 1 yr.
Citation: Journal of Climate 26, 17; 10.1175/JCLI-D-12-00253.1
Second, the AOD anomalies in fall tend to be positive in El Niño–developing years, but negative in La Niña–developing years. During the analysis period, three peaks appear in the fall of 2002, 2004, and 2006, all of which are El Niño–developing years (Figs. 6a,c). This is consistent with previous studies (van der Werf et al. 2008; Tosca et al. 2010; Reid et al. 2012). Minimum values are seen in the fall of 2000, 2007, 2008, and 2010, corresponding to negative SST anomalies in the Niño-3.4 region (Figs. 6a,c). Note, however, that not all of the low AOD years correspond to low Niño-3.4 SST (e.g., 2001, 2003, and 2005), indicating the impacts of other factors. The simultaneous correlation coefficient in fall between the AOD over 5°S–5°N, 95°–135°E and SST over 5°S–5°N, 170°–120°W is 0.58 during 2000–10, which is significant at the 90% confidence level.
The above correspondence indicates the impacts of ENSO on AOD variations over the Maritime Continent. It is well known that the Maritime Continent has large interannual variations of convection and precipitation related to ENSO (e.g., Hendon 2003) via the east–west Walker circulation. Warm (cold) SST in the equatorial central Pacific leads to a decrease (increase) of precipitation over the Maritime Continent (Wang et al. 2003; Juneng and Tangang 2005). Thus, in the fall of El Niño–developing years, fire burning is increased and more widespread (Field et al. 2009; Reid et al. 2012) because the dry conditions are sustained, releasing more aerosols into the atmosphere (Heil and Goldammer 2001; van der Werf et al. 2006, 2008; Field et al. 2009; Tosca et al. 2010; Reid et al. 2012). The suppressed precipitation also reduces the wet deposition of aerosols from the atmosphere so that aerosol particles are more likely to accumulate in the lower troposphere because of the increased residence time. In contrast, in the fall of La Niña–developing years, the Maritime Continent is wetter than average and the fire burning is reduced with less aerosols entering the atmosphere. The enhanced precipitation also leads to more wet deposition. These decrease aerosols in the atmosphere over the Maritime Continent. This wet deposition effect is confirmed by a partial correlation coefficient of −0.44 between precipitation and AOD over 5°S–5°N, 95°–135°E after excluding the effect of convection represented by 500-hPa vertical motion over the same region. Under the year-to-year change of the dry conditions controlled by ENSO, fire-burning activity causes large interannual variability of AOD in this region.
Third, the AOD variations appear to be associated with SST anomalies in the equatorial southeastern Indian Ocean. In the fall of 2001, 2005, 2009, and 2010, low AOD values correspond to positive SST anomalies in the region 0°–10°S, 90°–105°E (Figs. 6a,c). In the fall of 2002 and 2006, high AOD values correspond to negative SST anomalies in the above region. The simultaneous correlation coefficient in fall between the AOD over 5°S–5°N, 95°–135°E and SST over 0°–10°S, 90°–105°E is −0.73 during 2000–10, which reaches the 99% confidence level.
The above correspondence indicates that the AOD variations over the Maritime Continent are contributed by the Indian Ocean SST change, as pointed out by previous studies (van der Werf et al. 2008; Field et al. 2009). Negative SST anomalies in the equatorial southeastern Indian Ocean may suppress convection and precipitation over the Maritime Continent (e.g., Wang et al. 2003). On the one hand, this enhances the dry conditions favorable for fire burning. On the other hand, it reduces the wet deposition of aerosol particles from the atmosphere. Together, they lead to an increase in AOD. Opposite effects are expected when there are positive SST anomalies in the equatorial southeastern Indian Ocean.
Fourth, the AOD during 2001–07 features alternatively low and high values. This biennial AOD variation appears to be related to ENSO. During 2002–06, there are three El Niño events (2002, 2004, and 2006). Thus, ENSO displays a biennial feature in this period. As ENSO can affect precipitation over the Maritime Continent (Wang et al. 2003; Juneng and Tangang 2005), modulating the conditions for fire burning and wet deposition, ENSO appears to be an important factor for the biennial variability of aerosols over the Maritime Continent during the above period.
As discussed above, the year-to-year AOD variations over the Maritime Continent in fall are contributed by both equatorial central-eastern Pacific and equatorial southeastern Indian Oceans' SST anomalies. In the fall of 2002, 2006, and 2010, the SST anomalies in the above two regions are opposite (Fig. 6c) and thus it appears that they work together for the AOD anomalies over the Maritime Continent. In the fall of 2000, 2007, and 2008, SST anomalies in the equatorial southeastern Indian Ocean are small. In these years, equatorial central-eastern Pacific SST appears to be a main player. So it is in the fall of 2007 that there are negative SST anomalies in the equatorial southeastern Indian Ocean. In the fall of 2001, 2005, and 2009, the contribution of SST anomalies in the equatorial central-eastern Pacific is either small or opposite. In these years, the tropical southeastern Indian Ocean SST seems to be the main factor for AOD anomalies over the Maritime Continent. We have performed a partial correlation analysis to investigate the relative impacts of the equatorial Pacific and Indian Oceans' SST anomalies. The partial correlation coefficients of AOD over 5°S–5°N, 95°–135°E with SST over 5°S–5°N, 170°–120°W and 0°–10°S, 90°–105°E in fall are 0.50 and 0.69, respectively. Thus, it appears that the Indian Ocean's SST anomalies have a larger contribution to AOD variations over the Maritime Continent than the equatorial Pacific's SST anomalies.
To further demonstrate the individual impacts of the equatorial central-eastern Pacific and equatorial southeastern Indian Oceans' SST anomalies, we show in Fig. 8 composite AOD, precipitation, surface wind, and SST anomalies for the fall of 2001, 2005, and 2009 and the fall of 2000, 2007, and 2008, respectively. The former three cases represent the equatorial Indian Ocean SST influence and the latter three cases represent the equatorial Pacific SST influence. In both composites, AOD is below normal over the Maritime Continent (Figs. 8a,b). In the equatorial Indian Ocean SST influence cases, this is related to above-normal precipitation over the equatorial eastern Indian Ocean (Fig. 8c) that is induced by warmer SST and associated anomalous low-level wind convergence (Fig. 8e). In the equatorial Pacific SST influence cases, it is related to above-normal precipitation to the east of the Maritime Continent and below-normal precipitation over the equatorial western Pacific (Fig. 8d). These precipitation anomalies are caused by cooler SST in the equatorial central-eastern Pacific and associated easterly wind anomalies over the equatorial western Pacific (Fig. 8f). In both composites, although large precipitation anomalies are mainly over the oceanic regions where more moisture is available, positive precipitation anomalies extend to the land regions (Figs. 8c,d) and thus enhance the dry conditions favorable for fire burning.
Composite 3-month mean (a),(b) AOD, (c),(d) precipitation (mm day−1), and (e),(f) surface (10 m) wind (m s−1) and SST (°C) anomalies for (left) October 2001, 2005, and 2009 and (right) October 2000, 2007, and 2008. The wind scale is shown at the top of the respective panels.
Citation: Journal of Climate 26, 17; 10.1175/JCLI-D-12-00253.1
In October 2009, positive SST anomalies in the equatorial central-eastern Pacific correspond to negative AOD anomalies over the Maritime Continent, opposite to the situation in 2002, 2004, and 2006 (Figs. 6a,c). In addition, this is the only October when the AOD and precipitation anomalies are not opposite during the analysis period (Fig. 6a). We have examined the spatial distribution of AOD and precipitation anomalies in October 2009 (figures not shown). Low AOD dominates Sumatra and Kalimantan, which is consistent with an increase in precipitation that is induced by local warm SST anomalies and associated low-level wind convergence in the equatorial eastern Indian Ocean. Precipitation is below normal in oceanic regions to the southeast of the Philippines and in the tropical northeastern Indian Ocean, which is a response to equatorial central Pacific warming. The AOD anomalies averaged over 5°S–5°N, 95°–135°E are negative because of the dominance of large anomalies over the land regions, whereas the precipitation anomalies averaged over the same region are positive because of large negative anomalies over the oceanic regions and small positive anomalies over the land regions. This leads to an apparent inconsistent relationship between AOD and precipitation anomalies in this month.
The impacts of the tropical Pacific and Indian Oceans' SST anomalies on AOD variations over the Maritime Continent may be further amplified by a positive feedback in this region as suggested by Tosca et al. (2010). An increase in fire burning leads to more aerosols in the atmosphere, which may increase atmospheric stability and decrease surface temperature through scattering and absorbing solar radiation. This reduces precipitation and enhances the dry condition, which may lead to more fire emissions. One question is what the relative roles of fire burning and meteorological changes are to AOD variability in the Maritime Continent region. Previous modeling studies showed that both biomass burning and meteorological changes contributed almost equally to the tropospheric ozone increase in the Indonesian region during October and November of the 2006 (Chandra et al. 2009) and 1997 (Chandra and Ziemke 2002) El Niño years. Here, we perform a partial regression analysis based on area means to estimate AOD variations in the Maritime Continent region that are associated with regional precipitation variations. The obtained precipitation-related AOD variance accounts for about 41.5% of the total variance. The other part of AOD variance, which is attributed to the human activity, is about 58.5%.
In the region to the southeast of Japan, there appears to be a tendency for minimum and maximum values of AOD to occur alternatively in late spring–early summer during 2002–07 (Fig. 6b). For example, low values are observed in May 2002, June 2004, and May 2006 and high values are observed in May 2003, May–July 2005, and July 2007. Interestingly, these high and low values lag those in the Maritime Continent by about 7–9 months. The standard deviation of AOD in this region also displays a clear phase locking: the largest value is in June (Fig. 7b). The lag–lead correlation with respect to AOD in the Maritime Continent displays a notable positive correlation of 0.31 at lag = 8 months and a negative correlation of −0.24 at lag = −5 months (Fig. 7c). While in the above we emphasize the time lag between large and small AOD values in the region to the southeast of Japan and those over the Maritime Continent, there are some large AOD values in the region to the southeast of Japan that appear not to be related to those over the Maritime Continent, for example, in March 2001, June 2001, April 2006, and May 2008.
The AOD over eastern China (30°–40°N, 110°–120°E) shows a significant positive correlation with AOD over the Maritime Continent at lag = 8 months (Fig. 7c). A notable positive correlation is also seen at lag = 0 month. At lag = −4 months, there is a negative correlation. Thus, it appears that there is a connection between AOD variations over the Maritime Continent and eastern China.
5. Spatial–temporal evolution of composite anomalies during 2001–07
As pointed out in the previous section, the biennial variations of AOD over the Maritime Continent during 2001–07 may be linked to biennial ENSO during this period. To demonstrate the contribution of ENSO to biennial AOD variations, we make a composite analysis for the three large AOD anomaly cases (2002, 2004, and 2006) during the analysis period. We take October as the reference time because this is the month when the largest AOD variance is observed over the Maritime Continent. Figure 9 shows composite anomalies of AOD, precipitation, wind, and SST at lag = 0 (corresponding to October) and lag = 8 (corresponding to June in the next year). Figure 10 is similar except for lag = −12 (corresponding to October in the preceding years) and lag = −5 (corresponding to May). Note that only three cases are included in the composite because of the limited length of data. As such, the robustness of the results needs to be validated using long data when available.
Composite 3-month mean (a),(b) AOD, (c),(d) precipitation (mm day−1), and (e),(f) surface (10 m) wind (m s−1) and SST (°C) anomalies at lag = 0 (October) and lag = 8 (June of the next year). The wind scale is shown at the top of the respective panels.
Citation: Journal of Climate 26, 17; 10.1175/JCLI-D-12-00253.1
As in Fig. 9, but for lag = −12 (October of the preceding year) and lag = −5 (May).
Citation: Journal of Climate 26, 17; 10.1175/JCLI-D-12-00253.1
At lag = 0 (October), large positive AOD anomalies over the Maritime Continent (Fig. 9a) correspond to below-normal precipitation (Fig. 9c) and lower-level divergent wind anomalies (Fig. 9e). Lower-level divergence and associated descent suppresses precipitation, which may contribute to more aerosols through reducing wet deposition and increasing aerosol residence time. The dry conditions also set up for more fire burning in the Maritime Continent region (e.g., Field et al. 2009). Over the equatorial western-central Indian Ocean and the equatorial central Pacific, precipitation is above normal (Fig. 9c). There are lower-level westerlies over the equatorial western Pacific and easterlies over the equatorial Indian Ocean (Fig. 9e). These correspond to warm SST anomalies in the equatorial central-eastern Pacific, cold SST anomalies in the equatorial western Pacific, to the northeast of Australia, and in the equatorial southeastern Indian Ocean (Fig. 9e). These are typical features of El Niño events (e.g., Wang et al. 2003). Anomalous heating associated with warm SST anomalies in the equatorial central Pacific induces anomalous descent and suppresses convection over the Maritime Continent (Hendon 2003). This indicates the contribution of ENSO to AOD anomalies over the Maritime Continent and the surrounding regions. Wind and precipitation anomalies over the Maritime Continent may also be contributed by negative SST anomalies in the tropical southeastern Indian Ocean (Wang et al. 2003). Negative AOD anomalies are seen over the northern Indian Ocean and the equatorial central Pacific (Fig. 9a), corresponding to above-normal precipitation (Fig. 9c). Though the precipitation anomalies in these two regions are comparable to those over the Maritime Continent, the corresponding AOD anomalies are much smaller. This is because AOD variability in the Maritime Continent region is enhanced by human activity (Reid et al. 2012). Anticyclonic winds are seen over the South China Sea, featuring a response to El Niño events (Wang et al. 2003; Wu et al. 2003). In the region to the east of Japan, negative SST anomalies are accompanied by anticyclonic winds and below-normal precipitation to the northwest (Figs. 9c,e).
At lag = 8 (June of the next year), SST anomalies are weak in the equatorial Pacific (Fig. 9f), as are the wind and precipitation anomalies (Fig. 9d). The AOD anomalies are small over the Maritime Continent and positive AOD anomalies cover large areas from eastern China through the northwestern Pacific (Fig. 9b). Both wind and precipitation anomalies show a southwest–northeast contrast over the South China Sea–western North Pacific, with anticyclonic winds and below-normal precipitation over the Philippines and cyclonic winds and above-normal precipitation over the region to the southeast of Japan (Figs. 9d,f). Wind anomalies around the Philippines are considered a delayed response to El Niño through two different processes. One is local air–sea interaction process triggered by equatorial central Pacific warming (Wang et al. 2000). An anomalous anticyclone develops over the Philippine Sea in response to anomalous heating over the equatorial central Pacific. Northeasterly winds to the southeast of the anticyclone enhance local SST cooling through a wind–evaporation effect. The SST cooling, in turn, favors the maintenance of the anticyclone through a Rossby wave–type response. The other is the tropical Indian Ocean capacitor mechanism (Xie et al. 2009). El Niño–induced northern Indian Ocean warming persists into summer, exciting a Kelvin wave response over the western Pacific. The Kelvin wave–induced Ekman divergence contributes to the formation of an anomalous anticyclone over the Philippine Sea. The lower-level anomalous cyclone to the southeast of Japan is induced by anomalous heating over the Philippine Sea through atmospheric teleconnections (Nitta 1987; Huang and Sun 1992). On the one hand, the associated lower-level convergence induces anomalous ascent, leading to precipitation increase (Fig. 9d). On the other hand, the anomalous northerly winds in the northwestern part of the anomalous cyclone transport aerosol particles from the midlatitudes where the mean aerosol is large (Bao et al. 2009) to the subtropics, leading to an AOD increase (Fig. 9b). This may explain why the precipitation and AOD anomalies are of the same sign over the region to the southeast of Japan (Figs. 9b,d). Negative SST anomalies are seen in this region (Fig. 9f), which appears to be a response to increased cloudiness (not shown) associated with increased precipitation (Fig. 9d). As shown by Bao et al. (2009), the increase of AOD in this region may lead to an increase in cloudiness that reduces the shortwave radiation reaching the ocean surface and thus lowers the SST. In addition, oceanic processes such as anomalous upwelling associated with cyclonic winds may also contribute to the SST cooling.
ENSO-induced circulation and precipitation anomalies over the Maritime Continent are most pronounced in fall and winter around the ENSO mature phase (Fig. 6). The tropical southeastern Indian Ocean's SST anomalies attain a maximum in fall (Fig. 6), as do their influences on circulation and precipitation anomalies. As fall is the fire-burning season in the Maritime Continent (Chandra et al. 2009; Ziemke et al. 2009), large AOD anomalies are observed in this region because of a combined effect of ENSO and the tropical southeastern Indian Ocean's SST anomalies. On the other hand, ENSO's influence on circulation in the region to the southeast of Japan is most prominent in the early summer of the ENSO-decaying year, as are the associated AOD anomalies. This explains the 8-month time lag of AOD variations between the Maritime Continent and the region to the southeast of Japan.
At lag = −12 (October of the preceding year), the AOD anomalies are nearly opposite to those at lag = 0 over the Maritime Continent, tropical Indian Ocean, and East Asia (Figs. 10a, 9a). The anomalies are weak over the equatorial Pacific. The large positive precipitation anomalies and anomalous northwesterly winds over the equatorial southeastern Indian Ocean appear to be a response to warm SST anomalies along the coast of Sumatra (Figs. 10c,e). The enhanced precipitation then favors the decrease of AOD over the Maritime Continent and the surrounding regions (Fig. 10a). Thus, the equatorial Indian Ocean's SST anomalies could be an important factor for AOD anomalies at this time. This is consistent with Field et al. (2009) who indicated that the Indian Ocean's dipole mode is an important contributor to fire burning in Indonesia.
At lag = −5 (May), there is a tripole pattern of SST and precipitation anomalies as well as AOD anomalies from the equatorial central Pacific to East Asia (Figs. 10b,d,f). Positive SST anomalies develop over the equatorial central Pacific, accompanied by large cyclonic winds over the western North Pacific and above-normal precipitation over the equatorial central Pacific (Figs. 10d,f). Further to the northwest are anticyclonic winds and a band of positive rainfall anomalies extending from central China to the east of Japan. The AOD anomalies are negative from northeastern China through the midlatitude North Pacific (Fig. 10b). The wind anomalies over the western North Pacific feature a response to developing warm SST anomalies in the equatorial central Pacific. These warm SST anomalies induce Rossby wave–type responses with cyclonic winds to the northwest of warm SST anomalies.
The local and remote relationship is further demonstrated in the temporal evolution of composite area–mean anomalies for the selected regions shown in Fig. 11. Large AOD, cloud, and precipitation anomalies over the Maritime Continent are seen in fall (Fig. 11a), which corresponds to positive SST anomalies in the equatorial central-eastern Pacific and positive precipitation anomalies in the equatorial central Pacific (Figs. 9c,e). Over the region to the southeast of Japan, positive AOD anomalies appear in June of the next year and negative AOD anomalies in May of the current year with the same sign rainfall anomalies (Fig. 11b).
Composite AOD (solid thick curves), cloud fraction (%; dashed thick curves), and precipitation (mm day−1; solid thin curves) anomalies averaged over (a) 5°S–5°N, 95°–135°E, (b) 30°–35°N, 135°–165°E, and (c) 30°–40°N, 110°–120°E. The scale for AOD and cloud fraction anomalies is on the lhs y axis and that for precipitation anomalies is on the rhs y axis. Year 0 and year +1 refer to the current and the next year, respectively.
Citation: Journal of Climate 26, 17; 10.1175/JCLI-D-12-00253.1
Eastern China's AOD displays a tendency of in-phase variations with the Maritime Continent AOD in October (Fig. 11c). In addition, eastern China's AOD shows another period of large anomalies in June of the next year, the same as the region to the southeast of Japan (Fig. 11b). The temporal evolution of the composite anomalies displays a change in the relationship between AOD and precipitation in eastern China. Around May and September of the current year, precipitation anomalies lead opposite AOD anomalies by about 1 month, whereas around July of the next year, positive AOD anomalies lead positive precipitation anomalies by about 1–2 months. The opposite anomalies are also seen around October of the preceding year. This suggests that the local processes linking the AOD and precipitation anomalies in eastern China change with time. The changing relationship leads to a weak aerosol–precipitation correlation in eastern China (Fig. 1b).
6. Summary and discussion
The present study identifies three distinct relationships between interannual monthly anomalies of aerosol, cloud, and precipitation. Over the Maritime Continent, AOD is negatively correlated with both cloud and precipitation. Over the region to the southeast of Japan, AOD is positively correlated with both cloud and precipitation. In the subtropical oceanic regions, AOD is positively correlated with cloud, but has a weak correlation with precipitation. The above relationship is validated using several datasets and is confirmed by a sensitivity test conducted for cloud contamination. The results indicate that the aerosol–cloud–precipitation relationship varies from region to region.
The AOD variations over the Maritime Continent display a clear biennial feature during 2001–07. This is related to the biennial periodicity of ENSO during this period. The biennial variations are also found in AOD over the region to the southeast of Japan. The present analysis shows that the biennial variations in the above two regions display a time lag of about 8 months. This is attributed to the control of ENSO on large-scale circulation change over tropical and subtropical regions. Around October in El Niño–developing years, precipitation is suppressed over the Maritime Continent in response to anomalous heating over the equatorial central Pacific. This, on the one hand, reduces the wet deposition of aerosol particles, and on the other hand, sets up conditions favorable for fire burning in the local dry season. Both lead to an increase in AOD. During early summer of El Niño–decaying years, as a delayed response to El Niño, convection is suppressed around the Philippines, which induces an anomalous lower-level cyclone to the southeast of Japan through atmospheric teleconnection. The associated lower-level convergence and northerly winds to the northwest partly lead to both positive precipitation and AOD anomalies.
In addition to ENSO, the tropical Indian Ocean's SST anomalies may contribute to AOD variations over the Maritime Continent. Negative SST anomalies in the equatorial southeastern Indian Ocean suppress precipitation over the Maritime Continent, favoring an increase in AOD through reducing wet deposition and setting up dry conditions favorable for fire burning. The equatorial southeastern Indian Ocean's SST anomalies may work together with the equatorial central-eastern Pacific Ocean's SST anomalies or act independently in modulating AOD over the Maritime Continent.
The AOD variations over eastern China appear to be partly related to those over the Maritime Continent. In addition, the relationship between aerosol and precipitation variations over eastern China displays an obvious change with time. Specifically, aerosol and precipitation anomalies are of the same sign in some months, but are opposite in some other months. As such, the overall correlation between aerosol and precipitation is weak over eastern China. The reason for the changing relationship is speculated to relate to the modulation of large-scale meteorological conditions, but further investigation is needed to unravel the specific reasons.
An apparent weakness of the present study is the limited length of data. There are only three cases of biennial variations in the analysis period. Nevertheless, the control of ENSO on the large-scale conditions for the aerosol variations appears to be robust. This agrees with previous studies of the dynamical control of ENSO on precipitation and wind variability over the Maritime Continent and the western North Pacific. During the analysis period, ENSO is dominated by biennial periodicity leading to biennial variations of AOD over the Maritime Continent and over the region to the southeast of Japan. With the change in the ENSO periodicity, it is expected that the interannual time scale of AOD variations over the Maritime Continent and the western North Pacific may vary accordingly.
Another issue to bear in mind is that the present study used level-3 MODIS AOD data with 1° resolution. The level-3 products have higher uncertainties than the level-2 data with 10-km resolution over land because of the large uncertainty with varying surface reflectance. If the local-scale artifacts in AOD were removed based on the high spatial resolution of level-2 data and then averaging over spatial (1° × 1°) and temporal (daily and month) scales were performed, this would provide a more realistic and better AOD data. This, however, is not available for the present study. While monthly averaging on the 1° × 1° grids may have reduced some biases, the cloud contamination is still present in the level-3 MODIS AOD data, and thus the results of the present study should be taken with a pinch of salt.
In this study, we focus on the impacts of the equatorial central-eastern Pacific and equatorial southeastern Indian Oceans' SST anomalies on large-scale monthly-mean AOD variations over the Maritime Continent. There may be other factors that can contribute to the AOD variations, in particular, on small spatial and short time scales. In addition, it is humans who ignite the fire and set the extent of fire burning. The contribution of human activity may not be linearly related to oceanic and atmospheric forcing. This human component of the AOD anomalies in relation to fire burning relies largely on the population and the agriculture activity. As noted by Field et al. (2009), major fire events were absent in Kalimantan during the 1960s and 1970s despite several severe drought events, whereas drought conditions were accompanied by severe haze events from 1982 onward as a result of the population growth and the agriculture pattern change. Thus, the observed AOD anomalies depend on both the extent and severity of dry conditions set up by the equatorial Indo-Pacific's SST anomalies and the number and scale of fires ignited by humans. Further studies are necessary to understand how aerosol variations are influenced by the large-scale conditions and the aerosol feedback on convection and precipitation in this region.
The present study mainly focuses on the effect of large-scale circulation and precipitation changes on AOD variations. In reality, there are complex interactions among aerosol, cloud–precipitation, and circulation variations. Thus, the present empirical analysis may not account for the cause and effect unambiguously. Model experiments could be an approach to help demonstrate the physical processes in the cause–effect relationship. More studies are needed to unravel the complex aerosol–cloud–precipitation–circulation interactions.
Acknowledgments
The three anonymous reviewers' comments have led to a significant improvement of this paper. RW acknowledges the support of the Chinese University of Hong Kong direct grant (2021105), the Hong Kong Research Grants Council grant (CUHK403612), and the National Natural Science Foundation of China grants (412750851 and 41228006). ZW acknowledges the support of the National Key Basic Research Program of China grant (2009CB421404), the National Nature Science Foundation of China grant (40730951), and the Fundamental Research Funds for the Central Universities grant (11lgjc10). We thank Soo-Chin Liew and Santo V. Salinas Cortijo for their effort in establishing and maintaining the Singapore station and Brent Holden for his effort in establishing and maintaining the Shirahama station of AERONET.
REFERENCES
Ackerman, A. S., M. P. Kirkpatrick, D. E. Stevens, and O. B. Toon, 2004: The impact of humidity above stratiform clouds on indirect aerosol climate forcing. Nature, 432, 1014–1017.
Adler, R. F., and Coauthors, 2003: The Version-2 Global Precipitation Climatology Project (GPCP) monthly precipitation analysis (1979–present). J. Hydrometeor., 4, 1147–1167.
Albrecht, B., 1989: Aerosol, cloud microphysics, and fractional cloudiness. Science, 245, 1227–1230.
Andreae, M. O., D. Rosenfeld, P. Artaxo, A. A. Costa, G. P. Frank, K. M. Longo, and M. A. F. Silva-Dias, 2004: Smoking rain clouds over the Amazon. Science, 303, 1337–1342.
Bao, Z., Z.-P. Wen, and R. Wu, 2009: Variability of aerosol optical depth over East Asia and its possible impacts. J. Geophys. Res., 114, D05203, doi:10.1029/2008JD010603.
Beegum, S. N., K. K. Moorthy, S. S. Babu, R. R. Reddy, and K. R. Gopal, 2009a: Large scale modulations of spectral aerosol optical depths by atmospheric planetary waves. Geophys. Res. Lett., 36, L03810, doi:10.1029/2008GL036509.
Beegum, S. N., K. K. Moorthy, S. S. Babu, R. R. Reddy, K. R. Gopal, and Y. N. Ahmed, 2009b: Quasi-biennial oscillations in spectral aerosol optical depth. Atmos. Sci. Lett., 10, 279–284, doi:10.1002/asl.243.
Chand, D., and Coauthors, 2012: Aerosol optical depth increase in partly cloudy conditions. J. Geophys. Res., 117, D17207, doi:10.1029/2012JD017894.
Chandra, S., and J. R. Ziemke, 2002: Tropical tropospheric ozone: Implications for dynamics and biomass burning. J. Geophys. Res., 107 (D14), doi:10.1029/2001JD000447.
Chandra, S., J. R. Ziemke, B. N. Duncan, T. L. Diehl, N. J. Livesey, and L. Froidevaux, 2009: Effects of the 2006 El Niño on tropospheric ozone and carbon monoxide: Implications for dynamics and biomass burning. Atmos. Chem. Phys., 9, 4239–4249.
Chepfer, H., S. Bony, D. M. Winker, G. Cesana, J. L. Dufresne, P. Minnis, C. J. Stubenrauch, and S. Zeng, 2010: The GCM-Oriented CALIPSO Cloud Product (CALIPSO-GOCCP). J. Geophys. Res., 115, D00H16, doi:10.1029/2009JD012251.
Fan, J., R. Zhang, G. Li, and W.-K. Tao, 2007: Effects of aerosols and relative humidity on cumulus cloud. J. Geophys. Res., 112, D14204, doi:10.1029/2006JD008136.
Fan, J., and Coauthors, 2009: Dominant role by vertical wind shear in regulating aerosol effects on deep convective clouds. J. Geophys. Res., 114, D22206, doi:10.1029/2009JD012352.
Field, R. D., G. R. van der Werf, and S. S. P. Shen, 2009: Human amplification of drought-induced biomass burning in Indonesia since 1960. Nat. Geosci., 2, 185–188, doi:10.1038/NGEO443.
Heil, A., and J. G. Goldammer, 2001: Smoke-haze pollution: A review of the 1997 episode in Southeast Asia. Reg. Environ. Change, 2, 24–37, doi:10.1007/s101130100021.
Hendon, H. H., 2003: Indonesian rainfall variability: Impacts of ENSO and local air–sea interaction. J. Climate, 16, 1775–1790.
Holben, B. N., and Coauthors, 1998: AERONET—A federated instrument network and data archive for aerosol characterization. Remote Sens. Environ., 66, 1–16.
Huang, R.-H., and F.-Y. Sun, 1992: Impact of the tropical western Pacific on the East Asian summer monsoon. J. Meteor. Soc. Japan, 70, 243–256.
Jin, M., and J. M. Shepherd, 2008: Aerosol relationships to warm season clouds and rainfall at monthly scales over east China: Urban land versus ocean. J. Geophys. Res., 113, D24S90, doi:10.1029/2008JD010276.
Juneng, L., and F. T. Tangang, 2005: Evolution of ENSO-related rainfall anomalies in Southeast Asia region and its relationship with atmosphere–ocean variations in Indo-Pacific sector. Climate Dyn., 25, 337–350, doi:10.1007/s00382-005-0031-6.
Kahn, R. A., J. G. Michael, D. L. Nelson, K. K. Yau, M. A. Bull, B. J. Gaitley, J. V. Martonchik, and R. C. Levy, 2007: Satellite-derived aerosol optical depth over dark water from MISR and MODIS: Comparison with AERONET and implications for climatological studies. J. Geophys. Res., 112, D18205, doi:10.1029/2006JD008175.
Kanamitsu, M., W. Ebisuzaki, J. Woollen, S.-K. Yang, J. J. Slingo, M. Fiorino, and G. L. Potter, 2002: NCEP–DOE AMIP-II Reanalysis (R-2). Bull. Amer. Meteor. Soc., 83, 1631–1643.
Kleidman, R. G., N. T. O'Neil, L. A. Remer, Y. J. Yaufman, T. F. Eck, D. Tanré, O. Dubouik, and B. N. Holben, 2005: Comparison of Moderate Resolution Imaging Spectroradiometer (MODIS) and Aerosol Robotic Network (AERONET) remote-sensing retrievals of aerosol fine mode fraction over ocean. J. Geophys. Res., 110, D22205, doi:10.1029/2005JD005760.
Koren, I., Y. J. Kaufman, L. A. Remer, and J. V. Martins, 2004: Measurement of the effect of Amazon smoke on inhibition of cloud formation. Science, 303, 1342–1345.
Koren, I., Y. J. Kaufman, D. Rosenfeld, L. A. Remer, and Y. Rudich, 2005: Aerosol invigoration and restructuring of Atlantic convective clouds. Geophys. Res. Lett., 32, L14828, doi:10.1029/2005GL023187.
Lau, K.-M., M.-K. Kim, and K.-M. Kim, 2006: Asian summer monsoon anomalies induced by aerosol direct forcing: The role of the Tibetan Plateau. Climate Dyn., 26, 855–864, doi:10.1007/s00382-006-0114-z.
Lee, S. S., L. J. Donner, V. T. J. Phillips, and Y. Ming, 2008: The dependence of aerosol effects on clouds and precipitation on cloud-system organization, shear and stability. J. Geophys. Res., 113, D16202, doi:10.1029/2007JD009224.
Lin, J. C., T. Matsui, R. A. Pielke Sr., and C. Kummerow, 2006: Effects of biomass-burning-derived aerosols on precipitation and clouds in the Amazon basin: A satellite-based empirical study. J. Geophys. Res., 111, D19204, doi:10.1029/2005JD006884.
Mace, G. G., Q. Zhang, M. Vaughan, R. Marchand, G. Stephens, C. Trepte, and D. Winker, 2009: A description of hydrometeor layer occurrence statistics derived from the first year of merged CloudSat and CALIPSO data. J. Geophys. Res., 114, D00A26, doi:10.1029/2007JD009755.
Marchand, R., T. Ackerman, M. Smyth, and W. B. Rossow, 2010: A review of cloud top height and optical depth histograms from MISR, ISCCP, and MODIS. J. Geophys. Res., 115, D16206, doi:10.1029/2009JD013422.
Matsui, T., H. Masunaga, S. M. Kreidenweis, R. A. Pielke Sr., W.-K. Tao, M. Chin, and Y. J. Kaufman, 2006: Satellite-based assessment of marine low cloud variability associated with aerosol, atmospheric stability, and the diurnal cycle. J. Geophys. Res., 111, D17204, doi:10.1029/2005JD006097.
Mauger, G. S., and J. R. Norris, 2007: Meteorological bias in satellite estimates of aerosol-cloud relationships. Geophys. Res. Lett., 34, L16824, doi:10.1029/2007GL029952.
Menon, S., J. Hansen, L. Nazarenko, and Y. Luo, 2002: Climate effects of black carbon aerosols in China and India. Science, 297, 2250–2253.
Myhre, G., and Coauthors, 2007: Aerosol-cloud interaction inferred from MODIS satellite data and global aerosol models. Atmos. Chem. Phys., 7, 3081–3101.
Nitta, T., 1987: Convective activities in the tropical western Pacific and their impacts on the Northern Hemisphere summer circulation. J. Meteor. Soc. Japan, 65, 373–390.
Penner, J. E., and Coauthors, 2001: Aerosols, their direct and indirect effect in climate change. Climate Change 2001: The Scientific Basis, J. T. Houghton et al., Eds., Cambridge University Press, 289–348.
Pincus, R., S. Platnick, S. A. Ackerman, R. S. Hemler, and R. J. P. Hofmann, 2012: Reconciling simulated and observed views of clouds: MODIS, ISCCP, and the limits of instrument simulators. J. Climate, 25, 4699–4720.
Qian, Y., L. R. Leung, S. J. Ghan, and F. Giorgi, 2003: Regional climate effects of aerosols over China: Modeling and observation. Tellus,55B, 914–934.
Reid, J. S., and Coauthors, 2012: Multi-scale meteorological conceptual analysis of observed active fire hotspot activity and smoke optical depth in the Maritime Continent. Atmos. Chem. Phys., 12, 2117–2147.
Remer, L. A., and Coauthors, 2005: The MODIS aerosol algorithm, products, and validation. J. Atmos. Sci., 62, 947–973.
Rosenfeld, D., 1999: TRMM observed first direct evidence of smoke from forest fires inhibiting rainfall. Geophys. Res. Lett., 26, 3105–3108.
Rosenfeld, D., 2000: Suppression of rain and snow by urban and industrial air pollution. Science, 287, 1793–1796.
Rossow, W. B., and R. A. Schiffer, 1999: Advances in understanding clouds from ISCCP. Bull. Amer. Meteor. Soc., 80, 2261–2288.
Sayer, A. M., N. C. Hsu, C. Bettenhausen, Z. Ahmad, B. N. Holben, A. Smirnov, G. E. Thomas, and J. Zhang, 2012: SeaWiFS Ocean Aerosol Retrieval (SOAR): Algorithm, validation, and comparison with other data sets. J. Geophys. Res., 117, D03206, doi:10.1029/2011JD016599.
Sekiguchi, M., T. Nakajima, K. Suzuki, K. Kawamoto, A. Higurashi, D. Rosenfeld, I. Sano, and S. Mukai, 2003: A study of the direct and indirect effects of aerosols using global satellite data sets of aerosol and cloud parameters. J. Geophys. Res., 108, 4699, doi:10.1029/2002JD003359.
Shi, Y., J. Zhang, J. S. Reid, B. Holben, E. J. Hyer, and C. Curtis, 2011: An analysis of the collection 5 MODIS over-ocean aerosol optical depth product for its implication in aerosol assimilation. Atmos. Chem. Phys., 11, 557–565.
Tian, B., and Coauthors, 2008: Does the Madden-Julian oscillation influence aerosol variability? J. Geophys. Res., 113, D12215, doi:10.1029/2007JD009372.
Tosca, M. G., J. T. Randerson, C. S. Zender, M. G. Flanner, and P. J. Rasch, 2010: Do biomass burning aerosols intensify drought in equatorial Asia during El Niño? Atmos. Chem. Phys., 10, 3515–3528.
Twomey, S., M. Piepgrass, and T. Wolfe, 1984: An assessment of the impact of pollution on global cloud albedo. Tellus, 36B, 356–366.
van der Werf, G. R., J. T. Randerson, L. Giglio, G. J. Collatz, P. S. Kasibhatla, and A. S. Arellano Jr., 2006: Interannual variability in global biomass burning emissions from 1997 to 2004. Atmos. Chem. Phys., 6, 3423–2441.
van der Werf, G. R., and Coauthors, 2008: Climate regulation of fire emissions and deforestation in equatorial Asia. Proc. Natl. Acad. Sci. USA, 105, 20 350–20 355.
Wang, B., R. Wu, and X. Fu, 2000: Pacific–East Asian teleconnection: How does ENSO affect East Asian climate? J. Climate, 13, 1517–1536.
Wang, B., R. Wu, and T. Li, 2003: Atmosphere–warm ocean interaction and its impacts on Asian–Australian monsoon variation. J. Climate, 16, 1195–1121.
Wang, Y., Q. Wan, W. Meng, F. Liao, H. Tan, and R. Zhang, 2011: Long-term impacts of aerosols on precipitation and lightning over the Pearl River delta megacity area in China. Atmos. Chem. Phys., 11, 12 412–12 436.
Wentz, J., C. Gentemann, D. Smith, and D. Chelton, 2000: Satellite measurements of sea surface temperature through clouds. Science, 288, 847–850.
Wu, R., Z.-Z. Hu, and B. P. Kirtman, 2003: Evolution of ENSO-related rainfall anomalies in East Asia. J. Climate, 16, 3742–3758.
Xie, S.-P., K. Hu, J. Hafner, Y. Du, G. Huang, and H. Tokinaga, 2009: Indian Ocean capacitor effect on Indo-western Pacific climate during the summer following El Niño. J. Climate, 22, 730–747.
Yu, H., R. Fu, R. E. Dickinson, Y. Zhang, M. Chen, and H. Wang, 2007: Interannual variability of smoke and warm cloud relationships in the Amazon as inferred from MODIS retrievals. Remote Sens. Environ., 111, 435–449.
Zhang, Y., S. A. Klein, J. Boyle, and G. G. Mace, 2010: Evaluation of tropical cloud and precipitation statistics of Community Atmosphere Model version 3 using CloudSat and CALIPSO data. J. Geophys. Res., 115, D12205, doi:10.1029/2009JD012006.
Ziemke, J. R., S. Chandra, B. N. Duncan, M. R. Schoeberl, O. Torres, M. R. Damon, and P. K. Bhartia, 2009: Recent biomass burning in the tropics and related changes in tropospheric ozone. Geophys. Res. Lett., 36, L15819, doi:10.1029/2009GL039303.
