Modelling the Impacts of Iodine Chemistry on the Northern Indian Ocean Marine 1 Boundary Layer 2

Modelling the Impacts of Iodine Chemistry on the Northern Indian Ocean Marine 1 Boundary Layer 2 Anoop S. Mahajan, Qinyi Li, Swaleha Inamdar, Kirpa Ram, Alba Badia and Alfonso 3 Saiz-Lopez 4 5 Indian Institute of Tropical Meteorology, Ministry of Earth Sciences, Pune, 411016, India 6 Department of Atmospheric Chemistry and Climate, Institute of Physical Chemistry 7 Rocasolano, CSIC, Madrid, 28006, Spain 8 Institute of Environment and Sustainable Development, Banaras Hindu University, Varanasi, 9 221 005, India 10 Institute of Environmental Science and Technology (ICTA), Universitat Autònoma de 11 Barcelona (UAB), Barcelona, Spain 12 13 14


Introduction
Iodine compounds, emitted from the ocean surface, have been implicated in causing changes 41 to the chemical composition of the marine boundary layer (MBL (Carpenter, 2003;Platt and 42 Honninger, 2003;Saiz-Lopez et al., 2012a;Saiz-Lopez and von Glasow, 2012;Simpson et al.,

116
The WRF-Chem model (version 3.7.1), which included a full halogen scheme (Cl,Br,and I) 117 was used in the present study. The halogen chemistry scheme used in WRF-Chem and a 118 detailed description of the model setup are described in past studies (Badia et al., 2019;Li et 119 al., 2020). Sources of reactive iodine species considered in this study are an oceanic source of 120 organic iodine compounds (CH2ICl, CH2IBr, CH2I2, and CH3I) and inorganic compounds from 121 the ocean surface (HOI and I2). The sea-to-air fluxes of organic compounds were calculated 122 online (Liss and Slater, 1974). The oceanic emission of inorganic iodine (HOI and I2), which 123 is dependent on the deposition of O3 to the surface ocean and reaction with iodide (I -) was 124 calculated online using a parameterisation based on Badia et al. (2019), which was computed 125 using the empirical laboratory measured parameterisations by Carpenter et al. (2013) and 126 MacDonald et al. (2014). These emissions produced much higher than observed levels of IO 127 in the northern Indian Ocean MBL. The reasons for the overestimation are discussed further in 128 Section 3.2. Hence for the rest of the analysis, the emissions of I2 and HOI were reduced by 129 40% (i.e. 60% of the standard emission parameterisation). 130 The domain for the simulations was selected to cover the Indian subcontinent and the northern 131 Indian Ocean (as shown in Figure 1). We used a spatial resolution of 27 km and 30 vertical 132 layers (sigma levels of 1. 00, 0.99, 0.98, 0.97, 0.96,0.95, 0.94, 0.93, 0.92, 0.91,0.89, 0.85, 0.78, 133 0. 70, 0.60,0.51, 0.43, 0.36, 0.31, 0.27,0.23, 0.20, 0.17, 0.14, 0.11,0.08, 0.05, 0.03, 0.02, 0.01, 134 0.00) with the surface layer ~20 m above ground level and 10 layers within the boundary layer. 135 The simulation period included three seasons in the year of 2015 (pre-monsoon in April; 136 https://doi.org/10.5194/acp-2020-1219 Preprint. Discussion started: 10 December 2020 c Author(s) 2020. CC BY 4.0 License. summer monsoon in July; and the winter monsoon period in January). We ran the WRF-Chem Indian Southern Ocean Expedition (ISOE-8) (January 2015) (Mahajan et al., 2019b(Mahajan et al., , 2019a.  The MBL in the model results was defined as the lowest 10 layers (1.0 km above sea level). 156 The domain chosen for the model simulations, along with the tracks of the cruises from which 3.1 Model validation 161 A comparison between the model simulated IO and O3 from the HAL scenario and observations 162 made during the IIOE-2 and 8 th ISOE expeditions is shown in Figure 3 (Badia et al., 2019) and in the South China sea (Li et al., 2020) but lower than the modelled 174 and observed values of ~1.5 pptv in the tropical Atlantic MBL (Mahajan et al., 2010b).

175
The bottom panel of Figure 3 shows a comparison between the model simulated O3 with the 176 observations. Although the match between the model and observations is good in the northern   (Table 1). Even if only the MBL is considered after applying a land mask, the 194 mean IO mixing ratio is only 0.015±0.009 pptv in January, 0.011±0.006 pptv in April and 195 0.015±0.008 pptv in July (Table 1). At such low concentrations, iodine chemistry would not 196 have any measurable impact on atmospheric chemistry. The values closer to the Indian 197 subcontinent are negligible, although a high of ~0.04 pptv is seen close to the western Indian 198 and Pakistani coast during the summer monsoon period (July). It is well known that this region 199 experiences strong mixing in the northern Arabian sea during the summer monsoon period, 200 which triggers plankton blooms resulting in high productivity (Qasim, 1982). For the current 201 model runs, emissions of organic iodine are based on a climatology concentration of organic 202 halogens in the sea water (Ziska et al., 2013), which show high organoiodides emissions in this 203 region. However, despite the being an area of high productivity, the values of IO predicted in 204 the orgI scenarios are significantly lower than the observations (by a factor of 10-20; Figure 2) 205 and show the need for an inorganic iodine flux. Such a flux has been suggested to be ubiquitous 206 and dependent on the ozone deposition and seawater iodide concentrations (Carpenter et al., 207 2013;MacDonald et al., 2014). 208 https://doi.org/10.5194/acp-2020-1219 Preprint. Discussion started: 10 December 2020 c Author(s) 2020. CC BY 4.0 License.
The middle panels in Figure 4 show the distribution of IO for the HAL scenario, which includes incorporated in the available model mechanisms (Lewis et al., 2020). A further uncertainty on 247 the IO concentration calculation is that most chemical transport models tend to underestimate  than other regions studied hitherto using regional models, implying a reduced impact of iodine 278 chemistry on the atmosphere in the northern Indian Ocean environment.

279
The bottom panels in Figure 4 show the difference in IO between the HAL and orgI scenarios.   19.49±5.97 ppbv in July. This shows that the advection of anthropogenic sources from the 325 continent affects the ozone in the remote MBL too (Table 2).

326
The middle panels of Figure 5 show the absolute difference in O3 over the model domain.

327
During January and April significant ozone destruction is observed in the MBL, with as much 328 as 3.5 ppbv lower O3 in the HAL scenario as compared to the BASE scenario. During January 329 relatively larger destruction is observed in the Bay of Bengal as compared to the Arabian Sea.  Table 3). The reason for larger absolute differences as compared to mean differences is that 349 there are both increases and decreases seen through the domain, and hence the absolute 350 differences gives us an idea of the total impact of iodine chemistry.

351
The bottom panels in Figure 5 show the percentage change in O3 between the BASE and HAL destruction/production because of all halogens was -10 to +5% (Li et al., 2020). Over Europe, 378 combined halogen chemistry, which includes I, Br and Cl, significantly reduces the 379 concentrations of and O3 by as much as 10 ppbv. The contribution because of only iodine is 380 also larger than in the current domain, which is expected because of the higher IO 381 concentrations simulated in Europe (Li et al., 2019).

383
Halogen oxides interact with nitrogen oxides to change the NO/NO2 ratio by reacting with NO 384 to form NO2. Additionally, iodine oxides can also react with NOx to form iodine nitrate 385 (IONO2), which can be taken up on aerosol surfaces to act as a sink or recycle both nitrogen 386 compounds and iodine compounds (Atkinson et al., 2007). Thus, the resultant increase or  (Table   405 2). Over only the MBL, the mean NO2 mixing ratios are 0.10±0.46 ppbv in January, 0.06±0.30 406 https://doi.org/10.5194/acp-2020-1219 Preprint. Discussion started: 10 December 2020 c Author(s) 2020. CC BY 4.0 License. ppbv in April, and 0.07±0.29 ppbv in July. This shows that the MBL is much cleaner than the 407 air above the Indian subcontinent (Table 2).  (Table 2).

426
The middle panels of Figures 6 and 7 show the absolute difference in NO2 and NO for the HAL  (Table 2). This is because NO2 shows an increase in some  (Table 3).  (Table 2). When 472 we consider the mean absolute change to see the actual impact of iodine chemistry, the values 473 of the means are much higher, with as much as ~3.5% change in NO2 and 7% change in NO 474 observed over the MBL (Table 3). This change in NOx is smaller than simulated in Europe with 475 NO2 predicted to increase across most of Europe with most regions showing an increase 476 between 50 -200 pptv. However, this was the increase reported due to the inclusion of all the 477 halogens, and the impact of only iodine would be lower, even though higher levels were 478 simulated across Europe (Li et al., 2019).  (Table 2). Over only the 496 MBL, the mean OH mixing ratios are 0.15±0.05 pptv in January, 0.27±0.08 pptv in April and 497 0.27±0.08 pptv in July (Table 2).

498
HO2 shows much higher concentrations over the Indian subcontinent as compared to the 499 surrounding ocean MBL, with HO2 mixing ratios peaking over 15 pptv in the subcontinent as 500 compared to mixing ratios less than 10 pptv over most of the MBL (Figure 9). There is a 501 correlation between the hotspots for NOx, and low concentrations of HO2 over the Indian 502 subcontinent. This is due to the titration of HO2 by NO, which forms NO2 and leads to an 503 increase in O3 formation. A gradual decrease in the HO2 mixing ratios is observed from the lanes are lower than the surrounding areas. This is due to the earlier mentioned titration of HO2 509 by ship emitted NO, which leads to an increase in OH but a decrease in HO2. When averaged 510 over the entire domain, the mean HO2 mixing ratios are 7.10±1.49 pptv, 10.18±1.64 pptv and 511 9.24±1.97 pptv in January, April, and July, respectively (Table 2). Over only the MBL, the 512 mean HO2 mixing ratios are higher at 7.32±1.12 pptv, in January but lower in April and July 513 at 9.80±1.36 and 8.67±1.53 pptv (Table 2).

514
The middle panels of Figures 8 and 9 show the absolute difference in OH and HO2. For OH, a 515 small increase in the OH concentration is observed in most of the MBL during the months of 516 January and April, with the largest increase of about 0.03 pptv observed in the Arabian Sea.

517
However, for most of the domain, the increase in OH is small, with differences of 0.01 pptv 518 compared to the BASE scenario. During the monsoon month of July, a small decrease is 519 observed over most of the domain with an increase observed further south close to the equator. to clean air-masses reducing the ozone deposition driven emissions and hence the difference 528 between the HAL and BASE scenarios is also the lowest during July. When averaged over the 529 whole domain, the mean change in OH mixing ratios is negligible at 0.001±0.006 pptv in January, 0.006±0.007 pptv in April and -0.003±0.006 in July. In the case of HO2, the average 531 difference over the whole domain is also small at -0.48±0.43 pptv in January, -0.35±0.38 pptv 532 in April and -0.19±0.16 in July. Over the MBL too, the differences are larger with the largest 533 difference being -0.67±0.36 pptv in January (Table 2).

534
The bottom panels in Figures 8 and 9 show the percentage changes in OH and HO2 between 535 the BASE and HAL scenarios. Significant differences are observed with an increase in OH and 536 a decrease in the HO2 over most of the MBL. The largest change in OH is observed in the 537 northern Arabian Sea MBL, with a difference of more than 15% between the HAL and BASE  (Table 2), but the absolute percentage 548 change in OH is higher at 3.6 % in January, while the HO2 absolute percentage change (Table   549 3) is about ~8.4 % showing the large impact of iodine chemistry on the oxidation capacity of −3.7 to 0.73 pptv) in Europe (Li et al., 2019) and enhancing the total atmospheric oxidation 556 capacity in polluted areas of China, typically 10% to 20% (up to 87% in winter) and mainly by 557 significantly increasing OH levels (Li et al., 2020). The moderate increase in the oxidation 558 capacity over the northern Indian Ocean and the Indian subcontinent is due to the lower 559 concentrations of IO in the domain, along with the fact that this number is calculated only for 560 the impact of iodine chemistry, while the past studies have reported the impact of all halogens.

561
Globally the average increase in OH because of the inclusion of iodine chemistry has been 562 estimated to be 1.8 %, which is comparable to the current domain (Sherwen et al., 2016).  (Table 2).

587
The middle panels of Figure 10 show the absolute difference in NO3 over the model domain.

588
During the months of January and April, a significant reduction of up to -1.5 pptv is observed 589 in the MBL. During January, a reduction is observed in the Bay of Bengal as well as the Arabian 590 Sea, but in April the reduction in NO3 is largely observed in the Arabian Sea. This correlates 591 well with the IO distribution which also shows more iodine activity in the Arabian Sea during 592 April. A reduction in NO3 is also visible over the Indian subcontinent, and like O3 show that 593 the effects of iodine chemistry are not just limited to the MBL. Indeed, there are pockets of an 594 increase in NO3 observed over the subcontinent. During July, negligible difference is observed 595 between the HAL and BASE case, with a smaller than 0.5 pptv decrease seen across the MBL.

596
However, during the same period, an increase of up to 1.5 pptv can be seen over the NOx 597 hotspots over the Indian subcontinent. Decreases of up to -1.5 pptv are also observed along the 598 shipping lanes, showing the strong interaction between iodine and NOx chemistry. Over the 599 whole domain, the inclusion of iodine chemistry results in a mean decrease of about ~-0.4 pptv, 600 which is slightly higher when a mean is taken only for the MBL ( Table 2). The absolute change 601 in NO3 is even higher, with NO3 values changing by an average of 0.5 pptv across the whole 602 domain in July (Table 3). This value is however lower than the effect of all the halogens, as 603 https://doi.org/10.5194/acp-2020-1219 Preprint. Discussion started: 10 December 2020 c Author(s) 2020. CC BY 4.0 License.
shown by Li et al. (2019) in Europe, where halogens significantly reduced the concentrations 604 of NO3 by as much as 20 pptv.

605
The bottom panels in Figure 10 show the percentage change in NO3 between the BASE and 606 HAL scenarios. As much as a 50% reduction in the NO3 concentrations is observed in the MBL 607 when iodine chemistry is included, with the largest differences observed in the Arabian Sea, 608 close to the Indian subcontinent, further west closer to the equator and in the Bay of Bengal.

609
For most of the other domain, the change in NO3 is <20%. Over the Indian subcontinent, the 610 relative change in NO3 is small, due to larger absolute concentrations and in some places a 611 small increase (<5%) is predicted, especially in July when iodine chemistry is not highly active.

612
The relative change in the shipping lanes is smaller than the surrounding areas due to the higher  (Table 3).

619
In this study, we used the WRF-Chem regional model to quantify the impacts of the observed     Table 1: Monthly mean of IO concentration in parts per trillion by volume (pptv) over the 918 domain region for model simulations in January, April, and July 2015, and simulation scenarios 919 orgI, HAL and difference between HAL-orgI, before and after applying a land mask over the 920 model domain.