Sensitivity of population smoke exposure to fire locations in Equatorial Asia

CitationKim, Patrick S., Daniel J. Jacob, Loretta J. Mickley, Shannon N. Koplitz, Miriam E. Marlier, Ruth S. DeFries, Samuel S. Myers, Boon Ning Chew, and Yuhao H. Mao. 2015. “Sensitivity of Population Smoke Exposure to Fire Locations in Equatorial Asia.” Atmospheric Environment 102 (February): 11–17. doi:10.1016/j.atmosenv.2014.09.045.

Published Versiondoi:10.1016/j.atmosenv.2014.09.045

Permanent linkhttp://nrs.harvard.edu/urn-3:HUL.InstRepos:14550027

Terms of UseThis article was downloaded from Harvard University’s DASH repository, and is made available under the terms and conditions applicable to Open Access Policy Articles, as set forth at http://nrs.harvard.edu/urn-3:HUL.InstRepos:dash.current.terms-of-use#OAP

Share Your StoryThe Harvard community has made this article openly available.Please share how this access benefits you. Submit a story .

Accessibility

  1  

Sensitivity of population smoke exposure to fire locations in Equatorial Asia 1  

Patrick S. Kim1,*, Daniel J. Jacob1,2, Loretta Mickley2, Shannon Koplitz1, Miriam E. 2  

Marlier3, Ruth DeFries3, Samuel S. Myers4,5, Boon Ning Chew6 3  

4  

1Department of Earth and Planetary Sciences, Harvard University, Cambridge, MA, USA 5  

2School of Engineering and Applied Sciences, Harvard University, Cambridge, MA, 6  

USA 7  

3Department of Ecology, Evolution, and Environmental Biology, Columbia University, 8  

New York, NY, USA 9  

4Department of Environmental Health, Harvard School of Public Health, Harvard 10  

University, Cambridge, MA, USA 11  

5Harvard University Center for the Environment, Harvard University, Cambridge, MA, 12  

USA 13  

6Centre for Remote Imaging, Sensing and Processing, National University of Singapore, 14  

Singapore 15  

16  

Corresponding Author: Patrick S. Kim, Department of Earth and Planetary Sciences, 17  

Harvard University, Cambridge, MA 02138, USA. (kim68@fas.harvard.edu) 18  

19  

20  

21  

22  

23  

  2  

Key Points 24  

25  

1) Smoke exposure sensitivity to Equatorial Asian fires computed with adjoint model 26  

2) Protecting Sumatran peatswamp forests is key to future regional air quality 27  

3) GEOS-Chem adjoint can provide guidance for targeted land conservation 28  

29  

Abstract 30  

31  

Land clearing by fire in Equatorial Asia, a substantial fraction for conversion to 32  

oil palm plantations, can lead to high smoke concentrations across this densely populated 33  

region and represents a serious public health concern. Here we use the adjoint of the 34  

GEOS-Chem chemical transport model to show that population exposure to the smoke is 35  

highly dependent on where the fires from clearing (and other activities) take place. In 36  

2006, we find that Sumatran fires disproportionately contributed to the regional 37  

population-weighted smoke exposure (37% of regional fire emissions, 63% of the 38  

exposure). The information from the adjoint can provide guidance for targeted land 39  

conservation as oil palm agriculture expands. Protecting peatswamp forests in Southeast 40  

Sumatra emerges as a high priority. We present the adjoint as a tool that can be useful in 41  

a cost-benefit analysis to inform policymakers on the relative merits of targeting 42  

conservation in different regions of Equatorial Asia. 43  

44  

Index Terms 45  

0305 Aerosols and particles (0345, 4801, 4906) 46  

  3  

0345 Pollution: urban and regional (0305, 0478, 4251, 4325) 47  

0365 Troposphere: composition and chemistry 48  

3355 Regional modeling (4316) 49  

50  

Keywords 51  

Equatorial Asia, Adjoint, Palm Oil, Fire, Conservation, Population-Exposure 52  

53  

1. Introduction 54  

55  

Global palm oil production has more than doubled since 2000 in response to 56  

soaring demand [FAO, 2013]. Indonesia and Malaysia together account for 86% of the 57  

world’s production and Indonesia plans to double its output by 2020 [AFP, 2009; Koh 58  

and Ghazoul, 2010]. The demand for palm oil is driving rapid deforestation to clear land 59  

for new plantations [Koh et al., 2011; Miettinen et al. 2012]. As a result, the area covered 60  

by oil palm plantations in Indonesia tripled from 2000 to 2012 [FAO, 2013]. The land is 61  

mainly cleared by fire since it is the most economical method available. The resulting 62  

smoke pollutes the airshed of one of the most densely populated regions of the world. 63  

According to Marlier et al. [2013], the unusually large fire season of 1997 resulted in 64  

10,000 excess deaths from smoke exposure in Equatorial Asia. More recently, Sumatran 65  

fires in June 2013 caused a 24-h maximum smoke concentration of 300 µg m-3 in 66  

Singapore, far exceeding the 25 µg m-3 air quality guideline from the World Health 67  

Organization [WHO, 2005]. 68  

  4  

The fire season in Equatorial Asia is July - November (dry season), with large 69  

interannual variation in intensity, as illustrated in Figure 1 by comparison of 2006 (a high 70  

fire year) to the 2004 - 2010 average. This variability is largely driven by dry conditions 71  

related to the El Nino-Southern Oscillation. Dry conditions also increase the likelihood of 72  

escaped, or unintentional, fires to burn out of control [Reid et al., 2013]. There is less 73  

interannual variability in the location of the fires, as land conversion presently takes place 74  

mostly in the South Sumatran lowlands and the southern and western coasts of Borneo 75  

[Miettinen et al., 2010]. The fire plumes are transported by a prevailing southeasterly to 76  

southwesterly flow (Figure 1) such that some population centers are minimally affected 77  

by the smoke (e.g., Jakarta), while others are heavily affected (e.g., Singapore). This 78  

difference in impact has important political implications for land management, but also 79  

suggests opportunities, as some potential oil palm development regions (western Borneo, 80  

northern Sumatra) might lead to much smaller overall population exposure to smoke than 81  

other regions (southern Sumatra). The analysis below using the adjoint of the GEOS-82  

Chem chemical transport model (CTM) allows us to identify where fires could most 83  

effectively be restricted to reduce population exposure and to readily determine the air 84  

pollution exposure associated with any future land management scenario. 85  

86  

2. Methods 87  

88  

The GEOS-Chem CTM is driven by Goddard Earth Observing System (GEOS-5) 89  

assimilated meteorological data from the NASA Modeling and Assimilation Office 90  

(GMAO). The data have a native horizontal resolution of 0.50° x 0.67° with 72 pressure 91  

  5  

levels and 6-h temporal frequency (3-h for surface variables and mixing depths). Here we 92  

focus on simulating the impact of PM emitted by the fires, mainly organic and black 93  

carbon (OC/BC). The OC/BC simulation in GEOS-Chem is described by Wang et al. 94  

[2011]. Anthropogenic emissions are from Bond et al. [2007]. Open fire emissions are 95  

from the GFED3 inventory [Giglio et al., 2010; van der Werf et al., 2010] with monthly 96  

resolution. The standard GFED3 product does not consider small fires, which account for 97  

one third of total fire emissions in Equatorial Asia [Randerson et al., 2012]. We therefore 98  

increase the GFED3 emissions by 50%. 99  

The GEOS-Chem results presented here are for a simulation with the native 0.50o 100  

x 0.67o horizontal resolution over East Asia [70o – 150o E, 11o S – 55o N] [Chen et al., 101  

2009], nested within a global simulation with 4o x 5o horizontal resolution that provides 102  

dynamic boundary conditions. We focus on the 2006 fire season (July - November) 103  

following a 1-year initialization. As shown in Figure 1, the surface air flow for smoke 104  

transport in 2006 is typical of the 2004 - 2010 mean, even though 2006 was an El Nino 105  

year that led to dry conditions and high fire activity. 106  

Figure 2 shows the mean smoke concentrations in surface air simulated by GEOS-107  

Chem for the 2006 fire season. Measurements of 24-h mean mass concentrations of PM10 108  

(particulate matter finer than 10 µm aerodynamic diameter) are available from a network 109  

of 50 surface sites maintained by the Malaysian Department of the Environment 110  

(http://www.doe.gov.my/) and the Singaporean National Environment Agency 111  

(http://www.nea.gov.sg/) and are also shown in Figure 2. They are predominantly in 112  

urban areas and away from the main fire locations. We estimate the smoke concentration 113  

  6  

at each site in the observations by subtracting the mean concentration for the bracketing 114  

non-burning months (June and December). 115  

There are frequent events with smoke in excess of 100 µg m-3 and these are 116  

captured by the model, as illustrated in Figure 2 for Singapore and Sibu (Borneo). 117  

Observed July - November mean smoke concentrations are typically in the range of 10-118  

50 µg m-3 and this is again well captured by the model, with r = 0.84 for the spatial 119  

pattern shown in Figure 2 (n = 50). GEOS-Chem underestimates the mean smoke 120  

concentration on Malaysian Borneo (the northern coast), which may be due to the 121  

inability of the model to resolve fine-scale transport around the Tama Abu mountain 122  

range on the island [Wang et al., 2013]. The model simulates seasonal mean smoke 123  

concentrations in excess of 100 µg m-3 in southern Sumatra and southern Borneo but 124  

there are no observations there. This includes the city of Palembang (1.5 million people) 125  

in southern Sumatra (Figure 1). 126  

We use the adjoint of GEOS-Chem v34 [Henze et al., 2007] for source attribution 127  

of the simulated smoke concentrations at selected receptor sites and for the Equatorial 128  

Asian population as a whole over the July - November duration of the fire season [to, t1]. 129  

A single simulation with the adjoint model operating backward in time over [t1, to] yields 130  

the complete time-dependent footprint of sources contributing to the smoke 131  

concentrations at a particular receptor site and for any averaging time. 132  

Smoke concentrations in GEOS-Chem are proportional to the fire emissions so 133  

that the total smoke PM(xR, t’) concentration at receptor site xR and time t’ is given by 134  

135  

PM (xR ,t ') =∂PM (xR ,t ')∂E(x,t)

E(x,t)dxdtx!∫∫to

t '∫ [1]

136  

  7  

where E(x, t) is the emission flux (g m-2 s-1) at location x and time t < t’, to is the 137  

beginning of the fire season, and the spatial integration is over the entire emitting domain. 138  

One adjoint simulation conducted over the time period [t1, to] provides the ensemble of 139  

sensitivities ∂PM(xR, t’)/∂E(x, t) for the PM concentration at a selected receptor site xR 140  

and at all times t’ to the emissions for all domain grid squares (x) and all prior times (t < 141  

t’). Instead of a single location for a receptor site, the adjoint can also provide the same 142  

ensemble of sensitivities for the population-weighted mean smoke concentration (𝑃𝑀) 143  

over the entire Equatorial Asian domain (i.e. ∂𝑃𝑀(t’)/∂E(x, t)). 144  

 145  

3. Results and Discussion 146  

 147  

Figure 3 shows the sensitivities of mean smoke concentrations during July – 148  

November 2006 to the mean emissions over the same time period for four receptor sites 149  

(Singapore, Jakarta, Palembang, and the population-weighted region). The sensitivities 150  

highlight the difference in exposure to the same emissions, but in different locations, and 151  

they are very different for the receptor sites shown. Singapore is particularly sensitive to 152  

fires in southeastern Sumatra. Jakarta has no sensitivity to fires in Sumatra or Borneo but 153  

high sensitivity to fires in Java and islands to the east. The population-weighted smoke 154  

concentration is much more sensitive to fires in Sumatra than in Borneo. It is most 155  

sensitive to fires in Java but fire activity there is low. 156  

The sensitivities computed in this manner for selected receptor sites can be 157  

combined with knowledge of emission patterns to immediately deduce the smoke 158  

concentrations at the receptor site and the contributions from different source regions. 159  

  8  

The sensitivities are dependent on meteorological conditions, but the July-November 160  

2006 period used seems typical of climatological conditions (Figure 1). 161  

Present-day fires in Equatorial Asia are concentrated most in southeastern 162  

Sumatra and southern Borneo  (Figure 1). Figure 4 shows the resulting contributions to 163  

annual mean smoke concentrations for 2006 in Singapore and for all of Equatorial Asia 164  

weighted by population. Sumatra was responsible for 37% of total fire emissions in 165  

Equatorial Asia in 2006 but contributed to 64% of the smoke concentration in Singapore 166  

and 63% of the smoke concentration to which the ensemble of the Equatorial Asian 167  

population was exposed. Preventing fire emissions in Southeast Sumatra would be of 168  

considerable benefit to Singapore and also benefit the region as a whole. 169  

The adjoint sensitivities can be readily applied to any future emission scenario to 170  

derive the resulting population exposure to smoke. Marlier [2014] describes two possible 171  

2009-2032 scenarios for Sumatra, a “High Oil Palm” scenario with plantation expansion 172  

into peatswamp forest (where fuel loads are particularly high), and a “Peat Protection” 173  

scenario where all remaining peatswamp forests are preserved. Figure 5 shows their mean 174  

fire season emissions for present-day (2004-2010, 256 Gg a-1 for all of Sumatra), the 175  

High Oil Palm scenario (414 Gg a-1) and the Peat Protection scenario (186 Gg a-1). Their 176  

present-day emission estimates are 40% lower than ours for the same time period, likely 177  

due to their explicit treatment of small fires compared to our application of a scale factor 178  

as correction to the standard GFED3 product (Figure 1). We have some confidence in our 179  

estimate on the basis of better comparisons to observed smoke concentrations (Figure 2). 180  

The Marlier [2014] present-day Sumatran emissions nevertheless serve as a baseline 181  

against which their future projections can be compared. 182  

  9  

The annual mean smoke concentrations from Sumatran fires, as obtained by 183  

applying the adjoint sensitivities to the Marlier [2014] emissions, are shown inset in 184  

Figure 5 for Singapore and for population-weighted Equatorial Asia. Values are low 185  

compared to Figure 2 because they are annual means for an average fire year, consider 186  

only Sumatran fires, and assume a lower base emission estimate, as discussed above. The 187  

smoke concentration in Singapore drops for both future scenarios compared to present-188  

day due to a shift of emissions away from Southeast Sumatra where Singapore is most 189  

sensitive; by contrast, the population-weighted mean concentration increases by 60% 190  

under the High Oil Palm scenario. The Peat Protection scenario reduces mean smoke 191  

concentrations relative to the High Oil Palm scenario by a factor of 4 for Singapore and a 192  

factor of 2 for the population-weighted mean. This example illustrates how the adjoint 193  

sensitivities allow us to immediately assess the public health benefits of different land 194  

management decisions in the region. 195  

196  

Acknowledgments 197  

198  

Data supporting Figure 2 is available upon request to BNC. Input files necessary for 199  

GEOS-Chem are available at http://geos-chem.org/. 200  

201  

This work was funded by the Rockefeller Foundation and the Gordon and Betty Moore 202  

Foundation through the Health & Ecosystems: Analysis of Linkages (HEAL) program, 203  

and by a Department of Energy Office of Science Graduate Fellowship to PSK made 204  

possible in part by the American Recovery and Reinvestment Act of 2009, administered 205  

  10  

by ORISE-ORAU under contract no. DE-AC05-06OR23100. The authors would like to 206  

thank Singapore's National Environment Agency and Malaysia's Department of 207  

Environment for collecting and archiving the surface air quality data. We thank Zifeng 208  

Lu for his assistance with the population data. 209  

210  

References 211  

212  

AFP (May, 27, 2009), Indonesia to double palm oil production by 2020, AFP Asian 213  

Edition. Available at http://www.thefreelibrary.com/Indonesia+to+double+palm 214  

+oil+production+by+2020-a01611882134. 215  

Bond, T. C., E. Bhardwaj, R. Dong, R. Jogani, S. Jung, C. Roden, D. G. Streets, and N. 216  

M. Trautmann (2007), Historical emissions of black and organic carbon aerosol 217  

from energy-related combustion, 1850-2000, Global Biogeochem. Cycles, 21, 218  

GB2018, doi:10.1029/2006GB002840. 219  

Chen, D., Y. Wang, M. B. McElroy, K. He, R. M. Yantosca, and P. Le Sager (2009), 220  

Regional CO pollution and export in China simulated by the high-resolution 221  

nested-grid GEOS-Chem model, Atmos. Chem. Phys., 9, 3825-3839, 222  

doi:10.5194/acp-9-3825-2009. 223  

FAO (2013), FAOSTAT: Statistical Databases and Data-Sets, Food and Agriculture 224  

Organization of the United Nations (FAO), Rome, Italy. 225  

Giglio, L., J. T. Randerson, G. R. van der Werf, P. S. Kasibhatla, G. J. Collatz, D C. 226  

Morton, and R. S. DeFries (2010), Assessing variability and long-term trends in 227  

  11  

burned area by merging multiple satellite fire products, Biogeosciences, 7, 1171-228  

1186, doi:10.5194/bg-7-1171-2010. 229  

Henze, D. K., A. Hakami, and J. H. Seinfeld (2007), Development of the adjoint of 230  

GEOS-Chem, Atmos. Chem. Phys., 7, 2413-2433, doi:10.5194/acp-7-2413-2007. 231  

Koh, L. P. and J. Ghazoul (2010), Spatially explicit scenario analysis for reconciling 232  

agricultural expansion, forest protection, and carbon conservation in Indonesia, 233  

Proc. Natl. Acad. Sci., 107, 11140-11144, doi:10.1073/pnas.1000530107. 234  

Koh, L. P., J. Miettinen, S. C. Liew, and J. Ghazoul (2011), Remotely sensed evidence of 235  

tropical pealand conversion to oil palm, Proc. Natl. Acad. Sci., 108, 5127-5132, 236  

doi:10.1073/pnas.1018776108. 237  

Marlier, M. E., R. S. DeFries, A. Voulgarakis, P. L. Kinney, J. T. Randerson, D. T. 238  

Shindell, Y. Chen, and G. Faluvegi (2013), El Nino and health risks from 239  

landscape fire emissions in southeast Asia, Nature Climate Change, 3, 131-136, 240  

doi:10.1038/nclimate1658. 241  

Marlier, M. E. (2014), Public health impacts from fires in tropical landscapes, Ph.D. 242  

thesis, Department of Earth and Environmental Sciences, Columbia University, 243  

New York, NY, USA. 244  

Miettinen, J., C. Shi, and S. C. Liew (2010), Deforestation rates in insular Southeast Asia 245  

between 2000 and 2010, Global Change Biology, 17, 2261-2270, 246  

doi:10.1111/j.1365-2486.2011.02398.x. 247  

Mietinnen, J., A. Hooijer, C. Shi, D. Tollenaar, R. Vernimmen, S. C. Liew, C. Malins, 248  

and S. E. Page (2012), Extent of industrial plantations on Southeast Asian 249  

  12  

peatlands in 2010 with analysis of historical expansion and future projections, 250  

GCB Bioenergy, 4, 908-918, doi:10.1111/j.1757-1707.2012.01172.x. 251  

Oak Ridge National Laboratory (ORNL): LandScan High Resolution Global Population 252  

Data Set (http://web.ornl.gov/sci/landscan/landscan_documentation.shtml), Oak 253  

Ridge National Laboratory, Oak Ridge, TN, USA, 2011. 254  

Randerson, J. T., Y. Chen, G. R. van der Werf, B. M. Rogers, and D. C. Morton (2012), 255  

Global burned area and biomass burning emissions from small fires, J. Geophys. 256  

Res., 117, G04012, doi:10.1029/2012JG002128. 257  

Reid, J. S., E. J. Hyer, R. S. Johnson, B. N. Holben, R. J. Yokelson, J. Zhang, J. R. 258  

Campbell, S. A. Christopher, L. DiGirolamo, L. Giglio, R. E. Holz, C. Kearney, J. 259  

Miettinen, E. A. Reid, F. J. Turk, J. Wang, P. Xian, G. Zhao, R. Balasubramanian, 260  

B. N. Chew, S. Janjai, N. Lagrosas, P. Lestari, N.-H. Lin, M. Mahmud, A. X. 261  

Nguyen, B. Norris, N. T. K. Oanh, M. Oo, S. V. Salinas, E. J. Welton, S. C. Liew 262  

(2013), Observing and understanding the Southeast Asian aerosol system by 263  

remote sensing: An initial review and analysis for the Seven Southeast Asian 264  

Studies (7SEAS) program, Atmos. Res., 122, 403-468, 265  

doi:10.1016/j.atmosres.2012.06.005. 266  

van der Werf, G. R., J. T. Randerson, L. Giglio, G. J. Collatz, M. Mu, P. S. Kasibhatla, D. 267  

C. Morton, R. S. DeFries, Y. Jin, and T. T. Leeuwen (2010), Global fire emissions 268  

and the contribution of deforestation, savanna, forest, agricultural, and peat fires 269  

(1997-2009), Atmos. Chem. Phys., 10, 11707-11735, doi:10.5194/acp-10-11707-270  

2010. 271  

  13  

Wang, J., C. Ge, Z. Yang, E. J. Hyer, J. S. Reid, B.-N. Chew, M. Mahmud, Y. Zhang, and 272  

M. Zhang (2013), Mesoscale modeling of smoke transport over the Southeast 273  

Asian Continent: Interplay of sea breeze, trade wind, typhoon, and topography, 274  

Atmos. Res., 122, 486-503, doi:10.1016/j.atmosres.2012.05.009. 275  

Wang, Q., D. J. Jacob, J. A. Fisher, J. Mao, E. M. Leibensperger, C. C. Carouge, P. Le 276  

Sager, Y. Kondo, J. L. Jimenez, M. J. Cubison, S. J. Doherty (2011), Sources of 277  

carbonaceous aerosols and deposited black carbon in the Arctic in winter-spring: 278  

implications for radiative forcing, Atmos. Chem. Phys., 11, 12453-12473, 279  

doi:10.5194/acp-11-12453-2011. 280  

WHO (2006), WHO Air quality guidelines for particulate matter, ozone, nitrogen dioxide 281  

and sulfur dioxide. 282  

283  

Figure Captions 284  

285  

Figure 1: Fire emissions and mean 0-1 km vector winds in Equatorial Asia in July - 286  

November. The top panel is for 2006 (a high fire year) and the bottom panel is the 2004-287  

2010 mean. Fire emissions are from the GFED3 inventory, increased by 50% to account 288  

for small fires (see text). Winds are from GEOS-5 assimilated meteorological data. Also 289  

shown are the locations of cities with more than one million people. 290  

291  

Figure 2: Surface smoke concentration in Equatorial Asia, July – November 2006. The 292  

top panel shows seasonal mean concentrations observed at a network of sites (circles; see 293  

text for details) and simulated by GEOS-Chem. The bottom panels show the time series 294  

  14  

at two representative sites, Singapore and Sibu (Borneo), with locations indicated by 295  

arrows in the top panel. 296  

297  

Figure 3: Sensitivity of mean smoke concentrations in July - November 2006 to the 298  

location of fire emissions for three large cities and for all of Equatorial Asia weighted by 299  

population (city locations indicated by white circles, population in parentheses). Values 300  

are GEOS-Chem adjoint mean sensitivities. The simulated annual mean smoke 301  

concentration for each receptor is shown inset, assuming no smoke outside of the July - 302  

November fire season. 303  

304  

Figure 4: Spatially resolved contribution of fire emissions to annual average smoke 305  

concentrations in Singapore (top) and for all of Equatorial Asia weighted by the 306  

population distribution shown inset (bottom). 307  

308  

Figure 5: Mean July – November fire emissions in Sumatra from Marlier [2014] for the 309  

present (2004 – 2010) and for two future scenarios, “High Oil Palm” and “Peat 310  

Protection” (see text). The emissions are plotted on the same scale as Figure 1. The 311  

Sumatran contribution to simulated annual mean smoke concentration for Singapore and 312  

for population-weighted Equatorial Asia are shown inset, assuming no smoke outside of 313  

the July - November fire season. 314  

315  

316  

317  

  15  

Figures 318  

319  

Figure 1 320  

321  

322  

323  

324  

325  

326  

327  

328  

329  

1 M5 M

City Size

Fire Emissions ( )

5 m/s

10 S

5 S

0

5 N

10 N

100 E 120 E 140 E

Jul - Nov 2006

Jul - Nov 2004 - 2010

ug m s-2 -1

0.001 0.002 0.005 0.01 0.02 0.05 0.1 0.2 0.5 1.0

10 S

5 S

0

5 N

10 N

  16  

Figure 2 330  

331  

332  

333  

334  

335  

336  

337  

338  

Smoke Concentration in Surface Air (Jul - Nov 2006)

1 2 5 10 20 50 100 200

Smoke Concentration ( )ug m -3100 E 120 E 140 E

10 S

5 S

0

5 N

10 N

Singapore

Jul Aug Sep Oct Nov

0

50

100

150Sibu

Jul Aug Sep Oct NovSmok

e Co

ncen

tratio

n (

)

ug m

-3

ObservationsGEOS-Chem

  17  

Figure 3 339  

340  

341  

342  

343  

344  

345  

346  

347  

348  

349  

350  

351  

352  

353  

Singapore (5.13 million) Palembang (1.46 million)

Jakarta (9.61 million) Population-weighted Equatorial Asia (281 million)

0.2 -3ug m

7.2 -3ug m

6.2 -3ug m

126 -3ug m

Sensitivity of Smoke Concentration at Receptor Site to Location of Emissions ( )ug m s-2 -1ug m /-3

10 20 50 100 200 500 1000

100 E 120 E 140 E 100 E 120 E 140 E10 S

5 S

0

5 N

10 N

10 S

5 S

0

5 N

10 N

  18  

Figure 4 354  

355  

356  

357  

358  

359  

360  

361  

362  

363  

364  

10 S

5 S

5 N

10 N

0

1e-14 1e-13 1e-12 1e-11

Fire Emission Contribution to Smoke Concentration at Receptor ( )

120 E 140 E100 E10 S

5 S

0

5 N

10 N

1e3 1e4 1e5 1e6

Popu

latio

n

Singapore

Population-weighted Equatorial Asia

m -2ug m -3

  19  

Figure 5 365  

366  

367  

5 N

0

5 S

High Oil Palm Scenario Peat Protection Scenario

Fire Emissions ( )ug m s-2 -1

0.001 0.002 0.005 0.01 0.02 0.05 0.1 0.2 0.5 1.0

Marlier 2004 - 2010 Mean

95 E 100 E 105 E 95 E 100 E 105 E 95 E 100 E 105 E

ug m-3SingaporePopulation

: 0.68ug m-3: 0.95

ug m-3SingaporePopulation

: 0.83ug m-3: 0.60

ug m-3SingaporePopulation

: 0.18ug m-3: 0.45