Global biogenic volatile organic compound (BVOC) emission inventory during 2001 to 2020

1) ModelWe use the latest Model of Emission of Gases and Aerosols from Nature (MEGANv3.2) to estimate the BVOC emissions from 2001 to 2020 with the input of time-varying satellite-retrieved vegetation and reanalysis meteorology data. Compared to the earlier version MEGANv2.1 (Guenther et al., 2012), MEGANv3.2 estimates vegetation emission factors based on variable plant species measurements instead of on fixed plant functional type (PFT, Guenther et al., 2020). Specifically, MEGANv3.2 uses the so-called Emission Factor Processor (EFP), to estimate the landscape average emission factors, which are based on the following three databases: (1) Growth form datasets for four PFTs: tree, shrub, grass, and crops; (2) Ecotype datasets: composed of a mix of emission-specific tree species/grass associated with specific emission capacities; and (3) Updated tree species/grass datasets corresponding to the biogenic emission classes. These updates can distinguish the differences in vegetation emission factors in regions with the same PFT but with varying plant species. The new version also considers the additional stress factors of emissions by using the simple threshold function, including high/low temperature, and strong wind.2) Model input data (Time-varying vegetation datasets, meteorological datasets, and CO2 concentration)The vegetation parameters driving MEGANv3.2 include LAI (leaf area index), VCF (vegetation cover fraction), and PFT. In this study, the Moderate-resolution Imaging Spectroradiometer (MODIS) vegetation retrievals from 2001 to 2020 were used. LAI data was obtained from Yuan et al. (2011), which improved the MODIS version 6 product MCD15A2H (Myneni et al., 2015) with a temporal resolution of 8 days and a spatial resolution of 0.5°×0.5°. The LAIv calculated in MEGANv3.2 is defined as LAI divided by VCF, representing the leaf area index per unit vegetation area. The VCF was from the yearly MODIS MOD44B version 6 dataset (DiMiceli et al., 2015).The PFT was obtained from the yearly MODIS MCD12C1 product with a spatial resolution of 0.05° (Friedl and Sulla-Menashe, 2015). The selected 17 MODIS IGBP (International Geosphere Biosphere Programme) global vegetation classification types were mapped to four main PFT classification types (i.e., tree, shrub, grass, and crop) in MEGANv3.2 based on methods from Sulla-Menashe and Friedl (2018). The reprocessed datasets were conservatively interpolated to a spatial resolution of 0.5°×0.5° as model inputs.The meteorological parameters driving the MEGANv3.2 model were from the Modern-Era Retrospective analysis for Research and Applications, Version 2 (MERRA-2) (Gelaro et al., 2017). The selected variables used in MEGANv3.2 include 2 m temperature, surface downward shortwave radiation, surface soil moisture, water vapor mixing ratio, 10 m wind speed, precipitation, surface air pressure, low-level wind speed, cloud cover, and snow cover. Photosynthetically active radiation (PAR) in MEGANv3.2 was obtained by dividing the surface downward shortwave radiation by two. The temporal resolution of these variables is either 1-hourly or 3-hourly, and the 3-hourly data are linearly interpolated to the uniform 1-hourly data. All selected parameters were further interpolated from the original 0.5° × 0.625° to a spatial resolution of 0.5° × 0.5° (consistent with the resolution of the vegetation datasets) for driving the MEGANv3.2 model. In our study, MERRA-2 data from 2001 to 2020 were used.In addition, the global annual averaged CO2 concentration data are obtained from https://gml.noaa.gov/ccgg/trends/gl_data.html3) SimulationsTo isolate the contribution of different influencing factors (vegetation, meteorology, and CO2) to BVOC emission trends from 2001 to 2020, we have performed nine sensitivity experiments. These experiments consist of two groups.The first group contains four experiments:EMIT_ALL is the control experiment that considers the historical changes of all factors.EMIT_VEG, EMIT_MET, and EMIT_CO2 consider only the historical changes of vegetation parameters, meteorological factors, and CO2 concentration, respectively, while the other factors are fixed as those in 2001.In the second group, five experiments were conducted to isolate the contributions of individual vegetation parameters (i.e., PFT, LAIv) and meteorological factors (i.e., temperature, light, and soil moisture):For vegetation parameters, the experimental setup is the same as EMIT_VEG but with PFT (EMIT_VEG_FIX_PFT) or LAIv (EMIT_VEG_FIX_LAIv) fixed as that in 2001.For meteorological factors, the experimental setup is the same as EMIT_MET but with temperature (EMIT_MET_FIX_T2m), light (EMIT_MET_FIX_RAD), or soil moisture (EMIT_MET_FIX_SM) fixed as that in 2001.The model horizontal resolution is 0.5° × 0.5°, the temporal resolution is 1 hour, and the simulation period is 2001-2020. The input variables include the satellite-retrieved vegetation parameters and MERRA-2 reanalysis data as described above.4) BVOC emission inventory This BVOC emission data are available as monthly mean emission fluxes as well as monthly averaged daily profiles of emissions.Spatial coverage: Global (latmin:-90 latmax:90 lonmin:-180 lonmax:180)Spatial-resolution: 0.5°x0.5°Temporal coverage: 2001-2020Dimensions and their names: monthly:  parameter (20) x month (12) x lat (180) x lon (360)monthly24h: parameter (20) x month (12) x hour (24) x lat (180) x lon (360)Data Format: NetCDFUnits: μg m-2 s-1Biogenic - 20 parameters'ISOP' = isoprene'MBO' = 2-methyl-3-buten-2-ol'MT_PINE' = monoterpenes: pines (alpha and beta)'MT_ACYC' = monoterpenes: acyclic (e.g., myrcene, ocimenes)'MT_CAMP' = monoterpenes: carene, camphene, others'MT_SABI' = monoterpenes: sabinene, limonene, terpinenes, others'MT_AROM' = C10 aromatic: cymenes, cymenenes & C8-C13 oxygenated (e.g., camphor)'MT_OXY' = oxygenated monoterpenes and monoterpenoid-related compounds (e.g., estragole) 'SQT_HR' = highly reactive sesquiterpenes (e.g., caryophyllene)'SQT_LR' = less reactive sesquiterpenes (e.g., longifolene, copaene) and salates'METOH' = methanol'ACTO' = acetone'ETOH' = acetaldehyde and ethanol'ACID' = organic acids: formic acid, acetic acid, pyruvic acid'LVOC' = C2 to C4 HC (e.g., ethene, ethane)'OXPROD' = oxidation products: aldehydes 'STRESS' = stress compounds (e.g., linalool)'OTHER' = other VOC (e.g., indole, pentane, methyl bromide)'CO' = carbon monoxide'NO' = nitric oxideReference:Wang, H., Liu, X., Wu, C., and Lin, G.: Regional to global distributions, trends, and drivers of biogenic volatile organic compound emission from 2001 to 2020, EGUsphere [preprint], https://doi.org/10.5194/egusphere-2023-1830, 2023.

1) 模型
我们采用最新版的天然源气体与气溶胶排放模型（Model of Emission of Gases and Aerosols from Nature, MEGANv3.2），以时变卫星反演植被数据与再分析气象数据为输入，估算2001至2020年的生物挥发性有机化合物（Biogenic Volatile Organic Compound, BVOC）排放量。相较于早期版本MEGANv2.1（Guenther等，2012），MEGANv3.2基于可变植物物种实测数据而非固定植物功能型（Plant Functional Type, PFT，Guenther等，2020）来估算植被排放因子。具体而言，MEGANv3.2采用所谓的排放因子处理器（Emission Factor Processor, EFP）估算区域平均排放因子，其依据以下三类数据库：(1) 四类植物功能型的生长型数据集：乔木、灌木、草本与农作物；(2) 生态型数据集：由与特定排放能力相关的排放特异性乔木/草本物种混合组成；(3) 更新后的对应生物排放类别的乔木/草本物种数据集。这些更新可区分相同植物功能型但植物物种不同的区域间植被排放因子差异。新版模型还通过简单阈值函数纳入了额外的排放胁迫因子，包括高低温与强风。

2) 模型输入数据（时变植被数据集、气象数据集与CO₂浓度）
驱动MEGANv3.2的植被参数包括叶面积指数（Leaf Area Index, LAI）、植被覆盖度（Vegetation Cover Fraction, VCF）与植物功能型。本研究采用2001至2020年的中分辨率成像光谱仪（Moderate-resolution Imaging Spectroradiometer, MODIS）植被反演数据。LAI数据源自Yuan等（2011），该数据集对MODIS第6版产品MCD15A2H（Myneni等，2015）进行了优化，时间分辨率为8天，空间分辨率为0.5°×0.5°。MEGANv3.2中计算的LAIv定义为LAI除以VCF，代表单位植被面积的叶面积指数。VCF数据来自年度MODIS MOD44B第6版数据集（DiMiceli等，2015）。
植物功能型数据源自空间分辨率为0.05°的年度MODIS MCD12C1产品（Friedl与Sulla-Menashe，2015）。依据Sulla-Menashe与Friedl（2018）的方法，本研究将筛选出的17种MODIS IGBP（国际地圈生物圈计划，International Geosphere Biosphere Programme, IGBP）全球植被分类类型映射为MEGANv3.2中的四类主要植物功能型：乔木、灌木、草本与农作物。经重新处理的数据集被保守插值至0.5°×0.5°的空间分辨率，作为模型输入。
驱动MEGANv3.2的气象参数源自现代回顾性分析与研究应用版本2（Modern-Era Retrospective analysis for Research and Applications, Version 2, MERRA-2，Gelaro等，2017）。MEGANv3.2中选用的变量包括2米气温、地表向下短波辐射、表层土壤湿度、水汽混合比、10米风速、降水量、地表气压、低层风速、云量与积雪覆盖。MEGANv3.2中的光合有效辐射（Photosynthetically Active Radiation, PAR）通过将地表向下短波辐射除以2得到。上述变量的时间分辨率为1小时或3小时，其中3小时分辨率数据被线性插值为统一的1小时分辨率数据。所有筛选参数均从原始的0.5°×0.625°空间分辨率进一步插值至0.5°×0.5°（与植被数据集分辨率一致），以驱动MEGANv3.2模型。本研究采用2001至2020年的MERRA-2数据。
此外，全球年度平均CO₂浓度数据源自https://gml.noaa.gov/ccgg/trends/gl_data.html。

3) 模拟实验
为分离不同影响因子（植被、气象与CO₂）对2001至2020年BVOC排放趋势的贡献，我们开展了9组敏感性试验。这些试验分为两组。
第一组包含4组试验：
EMIT_ALL为对照试验，考虑所有因子的历史变化。
EMIT_VEG、EMIT_MET与EMIT_CO2分别仅考虑植被参数、气象因子与CO₂浓度的历史变化，其余因子固定为2001年的基准值。
第二组开展了5组试验，以分离单个植被参数（即植物功能型、LAIv）与气象因子（即气温、光照与土壤湿度）的贡献：
对于植被参数，试验设置与EMIT_VEG一致，但将植物功能型（EMIT_VEG_FIX_PFT）或LAIv（EMIT_VEG_FIX_LAIv）固定为2001年的数值。
对于气象因子，试验设置与EMIT_MET一致，但将气温（EMIT_MET_FIX_T2m）、光照（EMIT_MET_FIX_RAD）或土壤湿度（EMIT_MET_FIX_SM）固定为2001年的数值。
模型的水平分辨率为0.5°×0.5°，时间分辨率为1小时，模拟时段为2001-2020年。输入变量包括前述的卫星反演植被参数与MERRA-2再分析数据。

4) BVOC排放清单
本BVOC排放数据以月平均排放通量与月平均每日排放廓线的形式提供。
空间覆盖范围：全球（纬度范围：-90°至90°，经度范围：-180°至180°）
空间分辨率：0.5°×0.5°
时间覆盖范围：2001-2020年
维度与名称：
monthly：参数（20个）× 月份（12个）× 纬度（180个）× 经度（360个）
monthly24h：参数（20个）× 月份（12个）× 小时（24个）× 纬度（180个）× 经度（360个）
数据格式：网络通用数据格式（Network Common Data Form, NetCDF）
单位：μg·m⁻²·s⁻¹
生物源VOC的20个参数：
'ISOP' = 异戊二烯
'MBO' = 2-甲基-3-丁烯-2-醇
'MT_PINE' = 单萜烯：松树源（α-蒎烯与β-蒎烯）
'MT_ACYC' = 单萜烯：无环类（如月桂烯、罗勒烯）
'MT_CAMP' = 单萜烯：蒈烯、莰烯及其他
'MT_SABI' = 单萜烯：桧烯、柠檬烯、萜品烯及其他
'MT_AROM' = C10芳香族：伞花烃、伞花烃衍生物及C8-C13含氧化合物（如樟脑）
'MT_OXY' = 含氧单萜烯及单萜类相关化合物（如草蒿脑）
'SQT_HR' = 高反应性倍半萜烯（如石竹烯）
'SQT_LR' = 低反应性倍半萜烯（如长叶烯、α-古巴烯）及水杨酸酯
'METOH' = 甲醇
'ACTO' = 丙酮
'ETOH' = 乙醛与乙醇
'ACID' = 有机酸：甲酸、乙酸、丙酮酸
'LVOC' = C2-C4烃类（如乙烯、乙烷）
'OXPROD' = 氧化产物：醛类
'STRESS' = 胁迫相关化合物（如芳樟醇）
'OTHER' = 其他VOC（如吲哚、戊烷、溴甲烷）
'CO' = 一氧化碳
'NO' = 一氧化氮
参考文献：Wang, H., Liu, X., Wu, C., and Lin, G.: Regional to global distributions, trends, and drivers of biogenic volatile organic compound emission from 2001 to 2020, EGUsphere [预印本], https://doi.org/10.5194/egusphere-2023-1830, 2023.

搜集汇总

背景与挑战

背景概述

该数据集是一个全球生物源挥发性有机化合物（BVOC）排放清单，覆盖2001年至2020年，基于MEGANv3.2模型生成，整合了卫星植被数据和再分析气象数据，空间分辨率为0.5°×0.5°。它包含20种BVOC参数的月均排放通量和日剖面数据，并通过敏感性实验分析了植被、气象和CO2对排放趋势的贡献，适用于气候变化和空气质量研究。