An Inventory on Black Carbon and Organic Carbon Emission by the Different Vegetative Ecosystem Over India

Fires provoke land degradation, deforestation, imbalance in the ecosystem, and promote changes in land use. To add more vein aerosols such as Black Carbon (BC) and Organic Carbon (OC) were emitted during the combustion event which plays a major role in climate change, pollution, and health. Hence this study aims to estimate the emission, residence, dry deposition ux, and sequestering ability of deposited BC and OC from different vegetative res across India using MODIS satellite data from 2013 to 2019. It was observed that the mean OC and BC emission were about 5.08 × 10 7 tonnes 4.48 × 10 6 tonnes during the re season across India. On a national scale, cropland res contributed the largest portion (80%) of total carbonaceous aerosol emissions from open res. In co-emission of species, forest shares a maximum relationship (> 92 percent) among carbonaceous aerosols. Estimation of deposition ux of emitted species showed cropland has higher deposition rates with a residence period ranging between 7hours to 23days. From the observed results, it is evident that higher aerosol emission combined with negligible deposition will be a potential threat to the environment. Waste utilization promoting strategies has to be adopted in India since agriculture contributes to major aerosols emission.


Introduction
Since the rise of vascular plants on the earth, res became a natural occurrence and a major threat to the terrestrial ecosystem. In the earth system, re is an integral component, and it has a complex relationship with climate and in uencing changes in the environmental attributes, particularly the carbon cycle 1,2,3,4 . Emission from a forest re is considered one of the major sources of air pollution, and these emissions generate trace gases and aerosols that have both immediate and long-term effects on the atmosphere 5 . The global carbon cycle receives potential feedback from terrestrial ecosystem res, and carbonaceous aerosols such as black carbon (BC) and organic carbon (OC) are found to be a major portion of such wild re emissions 6 .
BC is formed out of incomplete combustion. BC is a short-lived climate forcer considered second only to carbon dioxide, forcing +1.1W/m 2 7,8 . In contrast, OC formed during gas to the particle conversion process 9 can also impact climate by forming cloud droplets by acting as cloud condensation nuclei and by direct scattering of light 10 . The source of BC and OC is highly variable with time and space and is also regionally dependent based on the region's characteristics 11 . Notably, these aerosols can play a vital role in climate mitigation by serving as a long-term sink for carbon dioxide and cryosphere evolution when deposited into the terrestrial carbon pool 12 .
In Asia regions, the characterization of aerosol and Greenhouse Gases (GHG) emissions in the environment and climate have been studied by conducting several international and national eld experiments such as INDOEX 13,14 , ACE-Asia 15 , TRACE-P 16 , APEX 17, and EAREX 18 . However, aerosol deposition studies have not focused on these regions, even though vital in climate mitigation and policymaking. The deposition is a complex process that depends on aerosol properties, land surface characteristics, and micrometeorological conditions. Modelling studies are predominant in the aerosol deposition ux and lifetime studies. However, modelled studies show limited skill in simulating aerosols' spatial distribution, especially in mid to upper tropics. It has resulted in faster aging, higher scavenging ratio by both dry and wet deposition in the Asian continent 19,20,21 . The deposition of BC and OC occurs by wet or dry deposition mechanism into the terrestrial or aquatic ecosystem is known as a sink.
A carbon sink is an important factor in the climatic negotiations globally as they have adopted markets for carbon trading since the Kyoto protocol 22 . Globally, the land ecosystem dominates the anthropogenic sink of carbon by 30%, followed by the ocean accounting for 25%, and the remaining emanated particles and gases stay in the atmosphere 23 . Carbon stock (drawn down carbon) varies across forest types, with tropical forest accounting for a stock of 303 tonnes of carbon per hectare, followed by temperate (66 ton/ha) and boreal forests (44 ton/ha) 24 . However, to date, quanti cation of the tropical land sink and sources were not spatially and temporally explicit 25 . The Indian forest's Carbon stock is 6941 million tonnes as of 2011 26 .
Estimating the burned area is essential in quantifying the impact of emissions on climate-carbon cycle feedback, especially during the climate emergency era of the twenty-rst century 27 . Historically, burned area information is based on ground estimates of res reported by the re management team 28 .
Countries' methods to calculate emissions are based on data collected, which were unreliable on global and continental scales. Unearthing the burned area through satellite served as a better alternative to older techniques, especially after satellite-based observation in 1974. Fire emission estimates have been continuously progressing since then, and the values used were best-guessed values of average annual area burned, the e ciency of burning, and biomes density 29,30 .
India has been adopting strategies to increase the carbon sink by approximately three billion tonnes to combat climate change 31 because the annual mean temperature will rise by 2.5°C to 4°C in India by the end of the century 32 . However, a lack of knowledge prevails in understanding the long-term and inter-annual interaction between re and carbon ux as the effects of altered re regime on soil carbon, nutrient storage, or limited plant productivity exist. The potential for utilizing satellite-based observation to overcome the aspects mentioned above is widely recognized. Based on our literature for the rst time across India, this study addresses the deposition ux and residence time of carbonaceous aerosol. Hence, this study aims to address the following objectives 1) To address re trends across different land cover types and major contributors to the highest vegetation re concentration. 2) To estimate carbon stock in burned land cover across study region 3) Understanding the deposition ux of biomass burning aerosol during the study period. These results help us understand re management, carbon dynamics, climate mitigation, and other related issues.

Page 4/25
The aerosols emitted during the re event is majorly composed of BC and OC. Emission of BC and OC from different vegetation types is studied from the year 2013-2019, respectively.

BC and OC emission of different vegetative types
The emission scenario of different plant cover types is shown in Supplementary Fig. (1). From this Fig., we observe that the major components burned in res are croplands (agricultural and natural crops), which share 80% of the total emission. Forest was next to it with a share of 3-14% of the total burned area, contributing to aerosols emission. While shrubland involves emissions from burning, which is not signi cant and shows a shrinking trend from 2017 to 2019. The major contribution of cropland to the overall burned area is that India is one of the largest producing countries globally and also evident with its agricultural re accounting for 43-57% of all cropland res 50 . A study on Greenhouse gas inventory in Thailand also indicated similar results of more than 80% of emissions by the agricultural sector 51 . The decrease in forest res contribution maybe because of land clearing for collecting Mahua (Madhuca Indica) owers and the shifting cultivation by the tribal community, which accounts for 23% of annual deforestation in India 52 .

Cropland ecosystem
Cropland emission of BC ranges from 5.  55 . A similar satellite study conducted between 1950-2015 across Asia has estimated mean BC (3.89Tg) and OC (6.92Tg), which vary by a factor of ten to the power of ve compared to our results 56 . The relationship between BC and OC emission is studied using multiple regression analysis. The regression coe cient (R 2 ) returns a value of fteen percent showing a lesser signi cant contribution by OC towards BC emission, which is a characteristic of residential biofuels and agricultural is burning 57 . ANOVA test was carried out to verify the signi cance of the difference between groups. In the cropland ecosystem, there was a statistically signi cant difference between the groups of BC and OC groups during the study period as determined by one-way ANOVA (F (6, 14615) =79.702, p=0.001. The emission trend of mean BC and OC in the cropland ecosystem is shown in Fig 1. Forest ecosystem BC emission ranges from 2.16 ×10 7 Kg to 2.9×10 9 kg, and OC is from 1.86×10 8 kg to 2.5×10 10 kg during the emission events in the study period. The mean BC emission in forests is increasing consistently with each year from 2.16×10 7 kg (2013) to 2.9×10 9 kg (2019), respectively. While mean OC emission is found to be 1.31×10 10 kg with a high concentration of 2.5×10 10 kg (2019) and minimum concentration in 1.86 ×10 8 kg (2013), respectively. Forest cover is observed to decline from 2013 (1.83%) -2015(1.2%), which serves as a major reason for the decline in re-related emissions 58 . Fire incidents in the forests tended to increase during 2015-2019 due to low priority in managing anthropogenic re as funds to be spent on res are reduced 59 . As per the Indian State of Forest Report, the major contributors of a forest during our study period were people who indulge in clearing activities for purposes like cultivation, non-timber forest produces collection, and hunting/poaching purposes. The annual average aerosol emission from forest res in China was 2.7 Gg, and 27.4 Gg for BC and OC is signi cantly less than our dry seasonal average emissions 60 .In a continental context, a satellite study put forward the emission of BC (2.3×10 7 kg/year) and OC (2.3×10 7 kg/year) from open non-agricultural res, which is signi cantly less than our obtained results of BC (4.16×108 kg/year) and OC (4.84×10 9 kg/year) respectively 61 . The emission returns a signi cant relationship between the organic and BC with a regression coe cient of R 2 =0.92, p=<0.01. It suggests that the co-emission ratio of OC and BC is high in India's forests. Further, the mean annual OC (1.31×10 9 kg) is found to be higher than BC (1.50×10 8 kg), and this ratio is a general characteristic of forest emission 62 . A study ANOVA test was carried out to verify the signi cance of the difference between group and in forest ecosystem there was a statically signi cant difference in the emission of BC and OC is observed during the study period as determined by one-way ANOVA (F (6, 1499) =149.026, p=0.001.The emission trend of mean BC and OC in a forest ecosystem is shown in Fig A similar inventory study on shrubland emission using a smaller grid satellite data observed BC (19.4×10 3 kg) and OC (13.76×10 5 kg) as their annual emissions, which are found to be ten times lesser than our emission 63 . The mean BC and OC emission have a healthy relationship of seventeen percent at a con dence level of 95 percent. We hypothesize that due to much lesser natural res and a lack of human intervention in forest bush res during the dry season in the study period. ANOVA test was carried out to verify the signi cance of the difference between groups and in the shrubland ecosystem. There was a statistically signi cant difference between the emission of BC and OC during the study period as determined by one-way ANOVA (F (6, 1499) =24.501, p=0.001. The emission trend of mean BC and OC in the shrubland ecosystem is shown in Fig. 3. A satellite-based global estimated value for Southeast Asia was 5.14×10 20 kg/year (BC) and 5.3 ×10 21 kg /year (OC), respectively 64 . While our results show 3.04×10 10 kg/year (BC) and 3.53×10 11 kg/year (OC), which is ten times less than the estimates as it covers the entire region of southeast Asia, but in individual terms, it offers a signi cant contribution.

A regional analysis of emissions A Geospatial analysis
The spatial analysis of OC and BC emission in kilograms during 2013-2019 is done using the Geographical Information System (GIS), which is presented as Fig. 4 and 5. The emission is segregated into grids, with each grid covering 27.75 km across the country. The emission is segregated into low (1.09-2.6), medium (2.69-6.11) maximum (greater than 71.2), and no emission (0-1.9). The estimated mean BC and OC emission for India from total biomass burning is around 1.98×10 7 Kg and 1.59×10 8 Kg for the base year 2013. A signi cant increase in species emission is observed in the advancing years with 5.73×10 13 Kg of BC and then 3.06×10 14 Kg of OC until 2017, which then showed a declining trend. Since 2016 stringent enforcement of a ban on stubble burning by the government made major agricultural regions reduce its stubble count from leading to a reduction in emission, which is estimated based on a satellite study by NASA (https://www.downtoearth.org.in/news/air/stubble-burning-down-in-punjabharyana-up-since-2016-nasa-maps-68331). Supplementary Table 1. Overall state-wise analysis of BC and OC indicates that Andhra Pradesh has the maximum amount of emission, followed by Arunachal Pradesh, Assam, Bihar, Chhattisgarh, Gujarat, Jharkhand, Karnataka, and Kerala. Cropland res are high in states like Gujarat, Chhattisgarh, Odisha, Madhya Pradesh, Bihar, Tripura, Uttar Pradesh, and Andhra Pradesh. These regions are cultivating six major crops produced in India 65 . The states that accounted for the high cropland emission are the states with high population density and hold the most fertile land for agriculture 66 . Forests share the next major emission of BC and OC in India with the increasing wild res during the dry season every year. Manipur, Meghalaya, Uttarkhand, and Tripura shares maximum emissions during the study period. These regions come under moist deciduous forest types that are subjected to frequent res 67 . Uttarkhand is the youngest mountain region of the Himalayas, mainly experiencing annual forest res, which have worsened during our study period, especially in the year 2016 68 . In a study, the highest concentration of emerging intensi ed re hotspots is found in the northeast and central India, which is substantiated by our results 69  In the Indian forest ecosystem, the carbon stock is found to be reduced after 2017, which is due to drought experienced for three years consecutive years in major forest regions 75,76 . Besides, a reduction in dense forest cover of over 20,000 hectares, which is diverted for mining, thermal power projects, and wild res, also serves as a major causative agent for the reduction in carbon stock after 2015 58 . Similarly, in shrubland ecosystem carbon sink is found to increase consistently from 2013 Grasslands stock lays majorly on above-ground biomass, which dips down mainly because of re, poor management, deforestation (direct effect), and shifting agriculture (indirect effect). For instance, 31% of shrublands are lost in a decade by re, deforestation, and fodder production as demand for feeding livestock has grown by 65% 77 . Besides these, more than eighty percent of India's shrubland is in the poor range class that it is easily affected by weather and soil erosion 78 . The carbon sink of different biomes is represented in supplementary Fig. 6. Carbon sequestration of the burned area accounts for enhancing or degrading the carbon deposits of a region. The estimated current annual carbon stock (CACS) varied between different ecosystems, plotted in Supplementary Table 2. Cropland deposits of carbon decrease consistently from the base year 2013 until 2015 and then the stock seems to improve gradually until 2019 from 1.25×10 11 t of C per hectare to 1.13×10 13 t of C per hectare. Similarly, in the shrubland ecosystem, carbon stock increased from 6.70×10 9 t C/ha to 3.78×10 11 t of C /ha until 2018 and then reduced to 137% in 2019. A higher variability is observed in forest ecosystem carbon stock reduced from 5.29×109 t of C per hectare to -8.16×10 10 t of C per hectare during 2013-2019. The total mean carbon stock of the vegetative ecosystem increased from 48% to 234% during 2013-2019, with dips in 2014-2015 (236%) and 2016-2017 (244%). Overall, the average carbon sink variability differs from yearly in our study, which agrees with a global study conducted in the terrestrial ecosystem using models 79 .

Deposition ux of BC and OC
The size and properties of the BC particles determine the residence time and deposition ux. The deposition ux of BC particles is estimated for different land cover emissions and is calculated with a range of upper and lower limits. The estimated lower limit means deposition ux was observed to be highest in the cropland ecosystem with 0.47 kg of BC deposited per kilometre and 11.9 kg/km (OC) in a day for an atmospheric mean concentration 5.85 ×10 10 kg. While upper limit extents to 0.55 kg/km/day (BC) and 11.9 kg/km/day (OC). Interestingly, minimum means deposition ux is observed in shrublands for an average BC concentration (1.11×10 5 kg/km/day) and OC concentration (6.28×10 6 kg/km/day), which has a mean ux of 1.04×10 -6 kg/km/day for BC, OC has 5.86×10 -6 kg/km/day as the lower limit and 1.93×10 -5 (BC) and 1.09 ×10 -4 kg/km/day (OC) as the upper limit. Dry deposition ux is most suitable for particles having a higher aerodynamic size (<2.5µm), and the washout mechanism majorly removes sub-micron particles. The dry deposition value of small micron particles in the range of 0-2.5 and 1-2.5µm was 35±3% and 21% 80 . The BC and OC coarse particles are not considered.
Mean uxes of BC deposition for seven years from 2013-2019 were 2.77×10 1 Tg/year (BC) and 1.57×10 2 Tg/year (OC) for shrubland, 9.1×10 1 Tg/year (BC) and 1.41×10 3 Tg/year (OC) for the forest, and 1.41×10 6 Tg/year (BC) and 3.1×10 7 Tg/year (OC) for cropland ecosystem respectively. Overall, the mean deposition ux of BC and OC decreases with years with an exception in 2014. As our study is carried out during the summer season, where the low frequency of rainfall is observed leading to lesser deposition of carbonaceous aerosols and such a declining trend in dry deposition rate is observed across the globe.
Global annually averaged dry deposited BC-based on model experiments vary from 0.66 to 1.66 Tg/year, which signi cantly differs with a ratio of 17.3:1 for shrubland, 56.8:1 for the forest more than a percent difference in cropland. This large variation may be attributed because discrepancies exist in modelling the submicron particles 47 .
Aerosol particles may have individual varying residence time, and it uctuates with space and time. BC's mean residence time constantly stood for the study period with 0.7 days for upper estimates in the cropland ecosystem while lower estimates uctuated between 0.1day (4.4×10 3 kg/km/day) to 23 days (6.4×10 3 kg/km/day). However, shrubland and forest ecosystem emitted aerosols had a similar mean atmospheric residence time of 8.6 days. Deposition ux of BC over grassland and shrubland is estimated to be between 7-11 days, which best ts our results 81 . However, the global BC residence time to be 7.85 days, which is very much less than our prediction 82,83 . The deposition ux of BC and OC is presented in Tables 1 and 2.

Conclusion
In this study, a high spatiotemporal resolution (monthly data in a 0.25º by 0.25º grid) emissions inventory was developed from biomass burning across India based on Fire CCI burned area data for a of total agricultural emission. the co-emission of species, forest shares a maximum BC and OC emission with a relationship greater than 92%. Temporally maximum total aerosol emission is observed in 2014 and a minimum in 2018 with the frequent re events. The monthly variation of black and OC emission had its peak from March to May of the dry season. The resultant means dry deposition uxes of BC and OC emission for seven years from 2013-2019 were 1.01×10 -5 kg/km/day and 5.74×10 -5 kg/km/day for shrubland, similarly, 3.32×10 -5 kg/km/day and 5.13×10 -4 kg/km/day, for forest 5.14×10 -1 kg/km/day and 11 kg/km/day for cropland ecosystem respectively. The average residence/lifetime of BC emitted from other vegetative ecosystem res is between 8.6 and 11.6 days. The above estimations were performed by employing appropriate emission factors for carbonaceous aerosols, which are in good agreement with the previous emission inventory studies conducted in India shown in Table 2. Human activities are the major cause of res in the study region, coupled with natural events like lightning, climate-induced change, and meteorology. Therefore, OC and BC emissions must be mitigated by proper management of anthropogenic interventions leading to wild re in forests and proper administration of land use, especially in croplands that cover most of India's land as it will lead to land degradation and erosion.
Uncertainties and future recommendation This work-integrated many different datasets and formulae to estimate the emission of carbonaceous aerosols from wild res and their carbon dynamics. First, appropriate emission factors were employed to calculate the emission from satellite observations, which may suffer limitations as it is not based on real-time activity data and uncertainties in India's monitoring of res prevail. However, the factors used were in good agreement with the previous studies conducted in India. Earlier emission inventories studies in India are presented in table 2. Carbon stock and sink of the burned area are also estimated which are not a representation of the individual species, region, or type of re. Further, estimated carbonaceous aerosol deposition is unlikely to contribute to the carbon stock especially in immediate years. This is a new learning process, conducted in terms of biomass emissions-based lifetime and deposition of particles across India. In general, the deposition velocity of the particle varies with space and time. Hence, we recommend carrying out continuous in-situ measurement of BC concentration and nd out representative deposition velocity for that region, which will help reduce the limiting factors in radiative forcing and climate change studies. This study also recommends carrying out activity-based emission inventories across India to reduce the uncertainties in estimating carbonaceous aerosol emission.

Study area
India is a subcontinent covering a diversity of terrestrial ecosystems. It is located between 8°4', 37°6 'N latitudes and 68°7', 97°25' E longitude. One of the rich biodiversity hotspots and the seventh-largest country in the world with a land area of 12 lakh square miles (3,286,592 Km2). It has the highest mountains in the north, the Thar desert to its west, the Gangetic delta to its east, and the Deccan plateau (agro-ecological diversity) to the south. It experiences drastic land-use changes, climate, topography, ora (11%), fauna (6.5%), and socio-economic conditions due to the ever-increasing population 33  Burned biomass (Bb) =A×B×β×EF×10 -3 - (1) Where Bb is the biomass consumption amount (kg); A is the burned area measure (m 2 ); B is the biomass amount inside the burned area (kg m -2 ); β is the combustion e ciency or the fraction of fuel that burned, and EF is an emission factor of species (g/kg). An emission factor is a representative value that attempts to relate the quantity of a pollutant released into the atmosphere with an activity associated with the release of that pollutant 43 . In this study, each land-cover type's emission factors were obtained from a publication 44 , which has an integrated emission factor data of 370 published studies. Emission factor values used in this study were 3 (OC), 0.53 (BC) for Shrublands and grasslands, 4.4 (OC), 0.51 (BC) for Forests, and 4.9 (OC) and 0.42 (BC) for agricultural res is given in gram species per kilogram of dry matter burned.

Estimation of carbon sequestration and BC ux
Soil OC (SOC) pool of the burned area can be derived from estimating the soil's OC content to a depth of 30cm or by calculating above and below groundmass carbon stocks. The summation of carbon stock in different segments gives us the total carbon pool of each land-use type of region. A typical carbon stock for a speci c region and land use can be derived by multiplying the carbon stock per unit area (t/ha) with the total area covered by speci c land use. In our case, we used the reference stock of tropical dry area (38t C ha -1 ) which is derived from mean estimates of different studies by IPCC 45  Deposition ux and Residence time The dry deposition ux of atmospheric BC was estimated by multiplying the BC concentration and dry deposition velocity.

Flux=C× Vd -(5)
Where C is the atmospheric BC concentration (kg/km 2 ) and Vd is the dry deposition velocity. The dry deposition velocity is obtained by summation of the inverse of atmospheric aerodynamic, quasi-laminar sublayer, and surface or canopy resistance. A lower and upper estimate for forest and grasslands is selected for this study, which is 7×10-7 and 1.5×10-5 kilometers per second respectively. These values were derived from global modeled studies, and some uncertainties exist by order of two magnitude due to particle size and microclimatic conditions 46,47 . These values were derived with certain assumptions for grassland and forest deposition which is found in a study by Pryor and reference cited therein 48 . The settling velocity is generally higher than 1.5×10 -5 km/s in the terrestrial ecosystem, and hence it is considered as a lower bound in-residence time estimation.
The residence time of an aerosol species can be obtained from the ratio of aerosol concentration or load to the total load of the species by its removal rates 49 . The satellite-derived aerosols were considered as atmospheric load in the calculation. We know of no competing of interest associated with this publication, As Corresponding Author, I con rm that the manuscript has been read and approved for submission by all the named authors.    Spatial distribution of re derived OC in kilograms per grid in India. Note: The designations employed and the presentation of the material on this map do not imply the expression of any opinion whatsoever on the part of Research Square concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors.