Application of REMOTE to study of the spreading of forest fire smoke in the atmosphere over Indonesia 1996-1998

Kadarsah*, Edvin Aldrian**, Manabu Kanda***,* UNESCO Researcher at Kanda Laboratory, Tokyo Institute of Technology

** BPPT, Indonesia***Department of International Development Engineering 、 Tokyo Institute of Technology

　

Scenario ModelResults

1000 mb

We investigate the role of different meteorological condition in setting up the smoke haze distribution. We used a control scenario of forest fire activity during severe 1997 El Nino and compare with the presumably a normal year of 1996 and La Nina year of 1998. During the latter two periods we used similar forest fire emission inventory as the control run but with the latter meteorological periods. We simulated those scenariosusing REMOTE under a linux operating system with main forcing data from the European Centre for Medium range Weather temperature anomaly in other years that consequently brings less distributed area in the normal and in La Nina period.

Introduction Model DescriptionThe Climate-Chemistry/Aerosol Model REMOTE (Regional Model with Tracer Extension)

Conclusions

[1] Aldrian, Edvin. (2003), Simulations of Indonesian Rainfall with a Hierarchy of Climate Models, Dissertation of Max-Planck-Institut fur Meteorologie, Universitas Hamburg[2] Chevillard, A., P. Cias, U. Kartens, M. Heimann, M. Schmidt, I. Levin, D. Jacob, R. Podzun, V. Kazan, H. Sartorius dan E. Weingartner. (2002), Transport of Rn-222 using the regional model REMO : a detailed comparison with measurements over Europe, Tellus-B, 54, 850-871[3] Chevillard, A., U. Kartens, P. Cias, S. Lafont dan M. Heimann (2002), Simulation of atmospheric CO2 over Europe and western Siberia using the regional scale model REMO, Tellus-B, 54, 872-894[4] Heil, A., Goldammer.(2001), Smoke-haze pollution: a review of the 1997 episode in Southeast Asia, Reg. Environ Change (2001) 2, 24-37 [8] Langmann, B dan A.Heil (2004), Release and dispersion of vegetation and peat fire emission in the atmosphere over Indonesia 1997/1998. Atmos. Chem. Phys. Discuss., 4, 2117–2159, 2004[6] Langmann, B., S.E. Bauer dan I.Bey. (2003), The Influence of the global photochemical composition of the troposphere on European summer smog, Part I: Aplication of a global to mesoscale model chain,J. Geophys. Res., 108 (D4), 4146, doi : 10.1029/2002JD002072[5] Langmann, B. (2000), Numerical modeling of regional scale transfort and photochemistry directly together with meteorological processes, Atmos .Environ. 34, 3585-3598.[7] Langmann, B., M.Herzog, dan H.F.Graf. (1998), Radiative forcing of climate by sulfate aerosols as determined by a regional circulation chemistry transport model, Atmos.Environ., 32, 2757-2768[9] Nicholls N. (1989), ENSO, drought and flooding rain in South-East Asia, dalam H. Brookfield and Y. Byron (eds.), United Nation University Press, Tokyo[10] Strum, K., W. Stichler, G. Hoffmann dan B. Langmann., (2003), Water isotopes in precipitation : Model result versus station measurements, 　 Atmos. Environ., 32, 2857-2869

[Acknowledgements]【 References 】

Monthly mean PM10 concentration modelled for the scenario runs :Normal Period (left column), El Nino Period (middle column) and La Nina (right column) and for every level (1000 mb, 850 mb, 750 mb and 500 mb)

Validation

•Based on REMO5.0 (Jacob, 2001)•ECHAM4 physics (Roeckner et al., 1996)

•1/2° - 1/6° horizontal resolution, 20 vertical layers •Tracer species implemented on-line): emissions, advection, convective up- and downdrafts, vertical diffusion, chemical reactions, dry and wet deposition (Langmann, 2000 ） •Modelling of photochemical (e.g. O3, H2O2), acidic (e.g. SO42-) and radioactive (e.g. 222Rn, 210Pb) trace species, stable water isotopes, CO2•Focus areas: Europe, Indonesia, tropical Andes and (China)

Langmann,2000

Smoke levels over Indonesia on September 29, 1997. Source:NASA-TOMS( Langmann dan Heil, 2004)

Data

Some important surface libraries concerning on the vegetation type.

• Three scenario : normal (1996), El Niño(1997) , La Niña(1998)

Setting model:• Sistem forecast mode (restart every 30 hours) • initializing and driven (nudged at the lateral

boundary every 6 hours) by the European Centre for Medium-Range Weather Forecasts (ECMWF) analysis while aerosol processes are calculated continuously (cf. Langmann and Heil, 2004).

• Emission data Jul ｙ -December 1997.• Single CPU Linux (OS Suse 9 )PC • programming Fortran 90 (2 step). • Step I : initialitation 6 hours• Step II : computation 24 hours.

Fire Locations (1997-1998)

Number of total hotspot July-Dec 1997. Notice the significant overlapping between fraction peat soil and hot spot locations

850 mb 750 mb 500 mb

This research has been supported by the Indonesian smoke induced by drought episodes (INSIDE). Project, funded by the European Commision, Framework 6 ECO-ASIA Project No: ASIA Pro ECO-ASI/B7-301/2598/20-2004/79071 with a joint research between BPPT Indonesia, Max Planck Institut für Meteorologie of Germany and Univ. Cambridge UK. and Kanda Laboratory.

This poster is presented in Meteorological Society of Japan spring, 2007

The major parameter measured is PM10 (particulate with a diameter below 10 μm). The PM10 concentration is reduced exponentially in vertical direction, which sharply reduced below 850 mb (2 km). The emission data inventory emission and the simulation result shows that Kalimantan, Sumatra, and Papua are major source of smoke haze of forest fire in Indonesia.

The research shows that the distribution of the smoke haze will be larger during the La Niña, Normal, and El Niño period, consecutively. The extent of the smoke distribution reaches the maximum on September 1997 in Kalimantan of about 2.904.000 km2. Meanwhile the concentration becomes larger from the El Niño, La Niña, and normal periods, consecutively. The concentration of PM10 reaches maximum on October 1996 and 1998 as much as 35000 μg/m3 or during La Niña and Normal periods.

Differences on smoke haze distribution among those three periods are mainly influenced by the sea surface temperature conditions that supply water vapor to the atmosphere that consequently determine the liquid water content (LWC).

The water content level will bring implication on the precipitation processes and the wet deposition that is larger and the accumulated PM10 will reach the maximum during the La Nina period. Furthermore, the smoke haze distribution is influenced by the wind that reached maximum during the El Niño period, especially on September 1997.

The minor differences on wind vector especially in south Indonesia, while large differences occur in west Sumatera in southeast direction. The vector wind differences show that the dry period has similarity in wind strength especially in south Indonesia. As the result the different of smoke distribution occurs in the leeward side or after the source of forest fire as the Malaysian peninsula in July. From August to November the difference on smoke haze distribution occur near the source and in southeast of the source location. This means that El Nino will cause the distribution of smoke further northwest and the intensity will be much higher during the El Nino episode. In comparison to Figure 5, the El Nino minus La Nina case, there is much more difference than the El Nino minus normal, with almost similar spatial distribution and further southeast to the location of forest fire source. Furthermore, the difference on intensity is much clearer in El Nino minus normal case than the other case. This result is almost unexpected, since during La Nina, hopefully rainfall will have higher probability to fall. In summary of both cases, except in July of El Nino minus normal, all the major differences between the two scenarios against control run occurs near the source of forest fire. In December, during the two scenarios, there is no more emission but in El Nino case there is still some emission left.

For the efficacy reason, we display the simulation result as relative to the control run of 1997. That means we show the difference between the simulation result of 1996 (normal period) against 1997 (El Nino period) and the simulation result of 1998 (La Nina period) against 1997 (El Nino period). Figure left and right illustrate the monthly difference of smoke haze distribution (PM10) and the wind vectors between the two periods. Although this method is good in describing the differences, a better presentation is to show also the absolute smoke haze distribution on each month.

Concentration of maximum PM10Vertical of Smoke Distribution for El Nino Period

Monthly mean Liquid Water Content (LWC) modelled for the scenario runs :Normal Period (left column), El Nino Period (middle column) and

La Nina (right column) and for every level (1000 mb, 850 mb, 750 mb and 500 mb)

( Langmann dan Heil, 2004)

Monthly mean PM10 concentration and wind vector modelled for Normal Period (1000 mb)

Monthly mean PM10 concentration and wind vector modelled for El Nino Period 1000 mb)

Monthly mean PM10 concentration and wind vector modelled for La Nina Period (750 mb)

The wide of smoke haze distribution against against

height

Monthly mean PM10 concentration and wind vector modelled for Scenario run: Normal Period

(1000 mb)-El Nino Period ( 1000 mb)

Monthly mean PM10 concentration and wind vector modelled for Scenario run: La Nina Period

(1000 mb)-El Nino Period ( 1000 mb)