Effects of the 2006 El Niño on tropospheric ozone and carbon monoxide: implications for dynamics and biomass burning

We have studied the effects of the 2006 El Ni˜ no on tropospheric O3 and CO at tropical and sub-tropical latitudes measured from the OMI and MLS instruments on the Aura satellite. The 2006 El Ni ˜ no-induced drought caused forest fires (largely set to clear land) to burn out of control dur- ing October and November in the Indonesian region. The effects of these fires are clearly seen in the enhancement of CO concentration measured from the MLS instrument. We have used a global model of atmospheric chemistry and transport (GMI CTM) to quantify the relative importance of biomass burning and large scale transport in producing ob- served changes in tropospheric O3 and CO. The model re- sults show that during October and November biomass burn- ing and meteorological changes contributed almost equally to the observed increase in tropospheric O3 in the Indone- sian region. The biomass component was 4-6 DU but it was limited to the Indonesian region where the fires were most in- tense. The dynamical component was 4-8 DU but it covered a much larger area in the Indian Ocean extending from South East Asia in the north to western Australia in the south. By December 2006, the effect of biomass burning was reduced to zero and the observed changes in tropospheric O3 were mostly due to dynamical effects. The model results show an increase of 2-3% in the global burden of tropospheric ozone. In comparison, the global burden of CO increased by 8-12%. Ni˜ no events are generally a change in convection in the tropi- cal troposphere associated with an eastward shift of the warm SST anomaly and large scale Walker circulation. This shift results in an increase in tropospheric column O3 in the west- ern Pacific and a decrease in the eastern Pacific relative to non-El Ni˜ no years. The effect of El Ni ˜ no on specific humid-

decreased significantly. The effects of forest fires in the Indonesian region during the 1997 El NiAo was studied by Cfiundru ct GI. [2002] using the GEOS-C'HEM global model of chemistry and transport. Their study suggested that about half of the increase in tropospheric colun~n 07one in the Indonesian region was due to biomass burning and the other half due to dy~lamical effects. A similar conclusion was arrited by Sndo crnd 7ukclhtr.shi [2001j from their modcl study.
Their conclusion about the contribution of bio~nass burning was indirect since they did not explicitly include the effect of biomass burning in their model study, The combined OMI and MLS instruments on the Aura satellite hace been prokiding ncar global rneasurernents of TCO from August 2004to present [e.g., Zicmke et ul., 2006Schocherl ct ul., 20071 as discussed in Section 2. In addition the hlLS instn~rrlent has been providing daily measurements of CO at seteral pressure letels in the troposphere and stratosphere [i,ive~cy ef ul., 20081. The purpose of this paper is to study the effects of the 2006 El Nifio on tropospheric composition measured f?om the OM1 and MLS instruments on the Aura satellite arid evaluate specific roles of the karious processes using a global model of chemistry and transport. 'The global model used in this study is the NASA GMI CTM (Global Modeling Initiative Chemical Transport Model). It successfully simulates a wide range of observations of chemical constituents in the troposphere and stratosphere, including data collected by instnltlients on board satellites. We provide a brief overview of the relevant tropospheric processes in the GMI CTM in section 3. In our study, the El NiAo related changes in O3 and CO fields based on OM1 and MLS measurements are compared with corresponding changes based on GMI C'TM simulation to delineate the relative importance of biomass burning and large scale transport.
In the hllowing, section 2 discusses the satellite data, cection 3 the (;MI nod el, and sections 4 and ti compare obsertcd and lneasured 0 1 and CO, respectively. Section 6 discusscs the impact of biomass burning emissions over Indonesia during the 2006 El Nifio, and section 7 compares the strength and frcclucncy of the 2006 and recent El NiCo etents with pretious El Nirlo events.
Finally, section 8 proc ides a summary.
OM1 is one out of a total of four instruments onboard the Aura spricecraft, which is llown in a sun-synclironous polar orbit at 705 km altitude uith a 98.2'' inclination. 'rhe spacecraft has an eclliatorial crossing time of 1 :45 pm (ascending node) with around 98.8 lnit~utes per orbit ( 14.6 orbits per day on average). Oh11 is a nadir-scanning instrument that for cisible (3.50-500 nm) and UV \tavelength channels :  Pickering, pprsoncrl c.oi?li?2~micution). Tirne-appropriate anthropogcnic and biomass burning emissions include, surface emissions from industryifossil fuel, biomass burning, biogenic and biofuel combustions and contributions from aircraFt emissions. They are based on the Global Fire Emission Database, version 2 (CiFEDv2) and discussed in the GMI CTM papers listcd above.
Monthly total biomass burning of CO and NOx are shown in Table 1 for the later part of 2006 when the Indonesian fires cvere most intense. The rneteorologicd fields that drive transport are from the (ioddatd hlc;ldeling and ,2ssimilation Office (CiMAO) 4iEOS-4 data assimilation system ((;EOS-4-ILdAS) [EElc~om gf LII,, 30051.
Feattarcc of the circttlation, strch as anticyclones. in the (;MI C TM arc;: realistically t-izpresentcd in the simulations because of the data constraints on rneteorological analyses. The CjEOS-4-IIAS fields have been regrictded to 42 kertical Ic\icls uith a lid at 0.01 hPa. The horizontal resolution is 2.5" x 2" (longitude by latitude).
14 validation of the global model tared in this study is; ghcn in the Appendix. Thic analysis compares the ~o n a l and seasonal variability of ' TC' O derived frorn the model uith obsened variability fiom OMIIfMLS.

Model comparison of El Nifio related changes in O3 with satellite measurements
As in C'h~intlru tlt (21. [2007] and Logcm ct crl. [2008], we hate cllosen October, Noventber and Decembcr 2005 as a baseline for estimating El Nifio-related changes in these months in the preceding and the following year.
Year 2005 was a neutral year from an El Nifio perspectike.
Maps of tropospheric O3 mean mixing ratio are sho-\vn in Figure  in Ilecember in the western Pacific remains elevated (Figure 2c). Logan et ul. [2008] attribute some of these increases in O3 due to lightning NOx in the Indonesian region.

Impact of biomass burning during the 200El Niiio
In sections 4 and 5 we analyzed the El NiAo related changes in tropospheric O3 and CO in 2006 using both the model and observations. In this section we analyze the model results to estimate the relatibe importance of biomass burning and large scale transport in producing the El Niiio related changes in tropospheric 0 3 . Our approach is similar to the one used by C'hundru et al.
[ 2002] in analyzing the effect of the 1997 El NiHo on tropospheric 0 3 . The model was run in two modes: The first mode explicitly included the NOx and CO emission rates associated with the Indonesian fires as given in Table 1     Simulations with ("Indonesia") and without ("No Indonesia") the 2006 Indonesia fire emissions.