key: cord-0858061-dzu9vv45 authors: Wang, Nan; Huang, Xin; Xu, Jiawei; Wang, Tong; Tan, Zhe-min; Ding, Aijun title: Typhoon-boosted biogenic emission aggravates cross-regional ozone pollution in China date: 2022-01-14 journal: Science advances DOI: 10.1126/sciadv.abl6166 sha: 80f071ba45fc84eb7f21ca29ff9f4de6f5f689f6 doc_id: 858061 cord_uid: dzu9vv45 Ozone pollution that threatens human health and the ecosystem is a global environmental challenge. In megacities, ozone pollution has long been mainly attributed to anthropogenic sources. However, the processes and mechanisms of cross-regional transport of ozone and its precursors under interactions between mixed sources remain unclear. Here, we show that Northwest Pacific typhoons could intensify the chemical interactions between anthropogenic and biogenic emissions, resulting in extreme ozone pollution in two main city clusters in China. By integrating field and satellite observations together with model simulations, we show that biogenic emission and cross-regional ozone transport are greatly enhanced by approaching typhoons, with the increments reaching up to 78.0 and 22.5%, respectively. Ozone formation efficiency has more than doubled because of abundant precursors and active photochemistry. This study highlights the importance of natural emissions in areas with intensive human activity, which needs to be considered in future air pollution control in China. Ground-level ozone (O 3 ) is an air pollutant that poses detrimental impacts on human health and crop productivity (1) (2) (3) (4) . O 3 also drives the production of the hydroxyl radical, which controls the photochemical lifetime of many atmospheric pollutants and reactive greenhouse gases (5) (6) (7) (8) (9) . As a secondary pollutant, tropospheric O 3 is predominantly formed in situ by photochemical reactions involving nitrogen oxides (NO x ) and volatile organic compounds (VOCs) (7, (10) (11) (12) . Thus, quantitatively understanding the sources and formation process of O 3 is of great importance for the mitigation policy of air pollution. Emissions of NO x are dominated by anthropogenic combustion processes, whereas VOCs originate from both human activities and biogenic emissions especially from forest ecosystems (5, (13) (14) (15) (16) . Severe O 3 pollution and its deterioration have long been attributed to intensive anthropogenic activities (6, 17, 18) . However, despite that strict measures have been widely implemented to cut down anthropogenic emissions in the past decades, increases in ground-level O 3 pollution have been witnessed in many places of the Northern Hemisphere (19, 20) . The highest O 3 concentrations and largest increment predominantly concentrate in the mid-latitude (21) , particularly evident in China. Specifically, O 3 pollution days in China are 93 to 575% more than other industrialized countries, jeopardizing the health of the world's largest population (22) . Meteorological parameters including incident sunlight and air temperature modulate the O 3 formation, accumulation, and transport (13, 23, 24) . The extreme O 3 pollution generally outbreaks in summer and has been proven to be closely linked to high air temperature and strong solar radiation together with stagnant weather conditions (24) (25) (26) (27) (28) . Recent observations have identified O 3 mixing ratios in cities of eastern China well in excess of present air quality standards under the influence of Northwest Pacific typhoon, a frequent weather activity in warm seasons (29) (30) (31) (32) (33) . Because of the fact that the typhoon weather is highly variable and the precursor emissions in this region are very complex, quantitatively understanding O 3 pollution under this circumstance is particularly challenging. Here, using in situ and satellite observations and chemical transport model for 6 years from 2013 to 2018, we provide a quantitative assessment of how the Northwest Pacific typhoon leads to these extreme O 3 pollution events in eastern China and also disentangle the key factors that control O 3 production and its transport under this unique environment. China's O 3 features highest levels in summer with 90th percentile of maximum daily 8-hour average (MDA8) of O 3 concentration reaching up to approximately 90 parts per billion (ppb) (5, 34) . Summer and autumn are also the seasons when China is subject to the most frequent typhoon activities in the Western Pacific. Long-term measurements from 2013 to 2018 in Fig. 1A and fig. S1 indicate that, under the typhoon-associated weather systems, O 3 pollution would get even worse, especially in two most densely populated city clusters-Yangtze River Delta (YRD) region and Pearl River Delta (PRD) region. Quantitatively, MDA8 O 3 concentrations in these two regions rise by 28.3 and 53.8%, respectively, under the influence of typhoons, comparing to the summertime climatology. Shanghai and Guangzhou, the two largest megacities located in YRD and PRD, respectively, have witnessed 30.3-and 39.1-ppb high anomalies as typhoons approach. Such an O 3 increment is particularly remarkable when typhoon tracks are nearby the island of Taiwan. Even more noteworthily, O 3 soars in succession with YRD polluted a few days earlier (Fig. 1, C and D) . The coastal eastern and southern China, including both YRD and PRD, is characterized by a very typical subtropical monsoon climate and western pacific subtropical high (WPSH) in summer (35) (36) (37) . The intervention of a typhoon tends to weaken and stretch westward WPSH (Fig. 1B) . The change in large-scale circulation and synoptic weather is highly sensitive to the specific position of typhoon (38) . To understand the typhoon tracks in the Western Pacific, we applied 2 of 8 an objective method to classify typhoons in the past decade. It is found that typhoons passing by Taiwan are the most conducive ones, bringing about O 3 pollution episodes in both YRD and PRD city clusters (Fig. 1 , C and D, and figs. S2 and S3). This kind of typhoon notably weakens the subtropical high, retracting the WPSH around 30°N (Fig. 1, C and D, and fig. S4 ) and greatly modifying the weather system in both regions. As the typhoons reshape the circulation, O 3 pollution starts to cover YRD and PRD in succession. Statistically, the typhoon-induced O 3 pollution features two distinct phases, that is, O 3 pollution tends to first deteriorate in YRD (the YRD phase) and then followed by PRD (the PRD phase) (Fig. 1 , C and D, and fig. S5 ). Spatially, it is found that the YRD region is the first to suffer from severe O 3 pollution when the typhoon center is located around 15°N to 20°N. In the subsequent phase of PRD pollution, typhoon center ranges from 20°N to 25°N ( fig. S5 ). The summertime extreme O 3 pollution in eastern China has been long attributed to fast photochemical production of intensive anthropogenic NO x and VOC emissions under intense solar radiation (13, 20, 27) . Moreover, southern China settles a large area of forest, just between YRD and PRD, which releases tremendous biogenic VOCs (BVOCs) as additional O 3 precursors ( Fig. 1A ) (35) . The spatial heterogeneity of these O 3 precursors and nonlinearity of O 3 chemistry make the regional transport and mixing processes substantially modulate O 3 pollution. When a typhoon approaches, the circulation in eastern China shows great shift from the typical summer monsoon to prevailing northerly winds (Fig. 2, A and B) , corresponded to the aforementioned YRD and PRD phases. During the YRD phase, the mainland China prevails southeast wind, which is mainly affected by the summer monsoon system ( Fig. 2A) . The air flow is relatively clean for PRD as it comes from the South China Sea. Quite the reverse, the low-pressure system of the cyclone makes the WPSH move northward to the mid-latitude and then cover the YRD region, leading to a stagnant condition there. As the cyclonic center moves northwest, the external air flow of the typhoon sweeps the southern China, the mainland prevails northerly winds (Fig. 2B) , and subsequently, PRD steps into pollution. The strong atmospheric subsidence of a typhoon periphery favors an intense solar radiation and stagnant condition. By comparing the meteorological conditions under the influence of typhoon with the summer average, it is clear that the noontime air temperature and incident solar radiation are increased by 4.6 K and 9.1 W/m 2 , together with the noticeable drop of relative humidity in both YRD and PRD (Fig. 2 , C to E). These parameters are all conducive to vigorous photochemistry and the resultant O 3 formation. In addition, biogenic emission itself is also highly dependent on the meteorology (15, 39) . The broad area of forest in southern China does undergo marked variations under typhoon-associated weather system based on the meteorology-driven calculation using the Model of Emissions of Gases and Aerosols from Nature (MEGAN) (see Materials and Methods). Take the Typhoon Nesat in 2017 as an example, isoprene, the most abundant BVOC species (15, 40) , witnesses a remarkable rise of 94% over the forest area in southern China. Regarding the spatial distribution, isoprene emission enhancement is found in central China and the northern part of PRD during the YRD phase and in most areas of PRD and its adjoining northern areas during the PRD phase (Fig. 3, A and B ). The significant increase in BVOC emissions correlates well with the rising air temperature, indicating the predominant impact of weather conditions (16, 38) . Statistically, the averaged isoprene emission of historical typhoon events from 2013 to 2018 is 78.0% greater than the summer climatology, providing solid evidence that BVOC emissions are prominently intensified by the typhoon-associated weather systems (Fig. 3D ). To identify the impact of the enhanced BVOCs on O 3 pollution in city clusters, we further adopt the backward Lagrangian particle dispersion modeling (LPDM) to simulate the potential source region of air masses for YRD and PRD. The 48-hour backward potential source regions of YRD and PRD point to the high vegetable coverage area in the southern China (Fig. 3C ). Considering the typical circulation of typhoon periphery, it is indicated that the intensified BVOCs are prone to be transported to city clusters with dense anthropogenic emission and contribute to O 3 pollution there. In addition to meteorological variations, O 3 formation sensitivity regime features great regional disparities over the eastern China. As clearly demonstrated in fig. S6 , in megacities with high NO x environments like YRD and PRD, the formation of O 3 is typically VOC-limited, while the O 3 production over the forest area is highly sensitive to the ambient NO x availability (41) . In normal days without intense cross-regional transport in warm seasons, the urban emissions or biogenic emissions alone are insufficient to build up O 3 pollution as the production efficiency maintains at a relatively low level. Conversely, the typhoon-associated weather tends to intensely mix anthropogenic and biogenic emissions by strong rotational winds. To quantitatively understand the interactions therein, we conduct parallel simulations using chemical transport model Weather Research and Forecasting (WRF) coupled Community Multiscale Air Quality (CMAQ) (WRF-CMAQ) (see Materials and Methods). The modeling results reveal that typhoon-boosted BVOC emissions enhance photochemical reactions in city clusters with contributions to O 3 formations by more than 30 ppb in both regions. Because both the city clusters are in the VOC-limited or transitional regime, the intervention of BVOCs plays an "adding fuel" role in urban photochemical pollution. Specifically, O 3 pollution is enhanced by anthropogenic and biogenic interactions via two pathways. One is the reaction with primary biogenic emissions, especially isoprene near the ground surface. It is worth noting that the reactivity of BVOCs often doubles or even triples that of anthropogenic VOCs (42, 43) . Once released, isoprene reacts with atmospheric oxidants and subsequently generates O 3 . This contribution is drastic near the ground surface, with the maxima contribution more than 30 ppb (Fig. 4, A and C) . The other is via the reaction with the transported biogenic intermediate products, including methyl vinyl ketone, methacrolein, methyl glyoxal, etc., in the upper planetary boundary layer (PBL) even during nighttime (44) . In general, isoprene itself predominately involves in local photochemistry due to its high reactivity; however, its intermediate product can be long-rangely transported hundreds of kilometers. The model results demonstrate that a huge amount of biogenic intermediate products inburst to downwind city clusters following typhoon circulation ( fig. S7 ) reacts with local oxidants, such as OH radical and O 3 and NO 3 radicals, and then generates RO 2 radical, thereby enhancing O 3 by almost Besides the vigorous interactions of BVOCs and anthropogenic emissions, cross-regional transport also matters because of the typical circulation of typhoon periphery. As shown in fig. S8 , during the first phase (YRD pollution), intensive anthropogenic emission in PRD could substantially affect the downwind areas. The YRD region, even further away, is still influenced with O 3 contribution more than 10 ppb. As a secondary pollutant with relatively long lifetime, O 3 could be either directly transported or locally produced by precursors from upwind areas (18, 45) . Our modeling results indicate that photochemically produced O 3 from PRD emissions is mainly distributed locally, and hardly can the precursors from PRD be transported to YRD. Instead, it is the direct advection of O 3 in the upper PBL and daytime convective mixing that make O 3 enhancement in PRD greatly contribute to O 3 pollution in YRD ( fig. S8, A and B) . In the PRD phase, the outflow of typhoon dominates the southern China, and the prevailing wind becomes from YRD to PRD. The accumulated O 3 in YRD is, in turn, horizontally transported back and lastly influences PRD with a contribution more than 20 ppb (fig. S8C ). By analyzing O 3 diurnal patterns along the transport pathway from tens of monitoring stations, the O 3 peaks depict a "lagging effect" from the upwind area to the downwind regions, which further supports the importance of cross-regional interaction based on direct in situ observations ( fig. S9 ). To identify the contribution of typhoon-enhanced O 3 pollution statistically, we then carried out the exactly same simulations for all the historical events from 2013 to 2018, as shown in fig. S10 . By integrating the findings, we propose a conceptual scheme illustrating the enhancement of cross-regional O 3 pollution by typhoonintensified anthropogenic and biogenic interaction and quantify the contributions from key factors in Fig. 5 . As shown, NO x and other pollutants first emitted in PRD pass the BVOC-concentrated forest area, and then the mixed anthropogenic and biogenic plume results in O 3 pollution when arriving YRD. In the second phase, pollutants accumulated in YRD are then swept back under the influence of typhoon outflow, passing the BVOC-concentrated forest area again, and give rise to PRD O 3 pollution. Comparisons of O 3 source apportionment between typhoon events and the summer average are analyzed using WRF-CMAQ (Fig. 4, B and D) . In YRD, the contributions of BVOC emissions are significantly increased from 16.9 to 33.6%. The cross-regional transport by PRD anthropogenic emissions also increases from 4.8 to 8.2%. In terms of PRD pollution, the regional transport dominates O 3 formation with an increase from 31 to 53.5%, among which the YRD contribution increases by approximately five times from 1.7 to 9.9%. The contribution of BVOC emission also gets more vital, increasing from 16.9 to 26.6%. Because that it is the anthropogenic and biogenic interaction that greatly enhances O 3 pollution, we further explored the potential benefits of a precedent control of anthropogenic emission before typhoon in pollution mitigation by using the model tools ( fig. S11 ). Promisingly, with the help of meteorology and air quality forecast, regional coordinated emission control in short term could help us to alleviate the O 3 enhancement due to approaching typhoons. O 3 pollution features the nonlinear response to emission intensity and involves multiple precursors and is also greatly modulated by synoptic weather. Southern China, characterized by two emissionintensive city clusters and a broad vegetation coverage, provides a unique lens to dig into the interaction of biogenic and anthropogenic emissions, particularly under typhoon when O 3 pollution peaks. Our study shows that a typhoon-affected weather system is the trigger to intensify the natural emissions of BVOCs and creates favorable meteorological conditions for photochemical reactions. Furthermore, the circulation of typhoon periphery enhances cross-regional transport and interactions of anthropogenic and biogenic emissions, facilitating O 3 pollution in urban areas. This O 3 enhancement is expected in other regions featuring both mixed emissions and frequent typhoon or hurricane activity, like the southeastern United States and northern India (46, 47) . Our results emphasize the importance of natural sources on tropospheric O 3 pollution even in areas with dense anthropogenic activity and replenish the understanding on typhoonintensified interactions of anthropogenic and biogenic emissions on tropospheric O 3 pollution. Even in highly polluted regions like eastern China, more consideration ought to be given to the coordinated cross-regional joint emission control. Moreover, the typhoon activities are strongly linked with climate variability and climate change (48, 49) , and biogenic emission could be also enhanced because of the global warming or carbon neutral-induced greening (50) . Therefore, these interactions between biogenic and anthropogenic activities may be further intensified and play a more vital role in O 3 pollution in the future. Clustering analysis is introduced to distinguish historical typhoon tracks. Because the number of clusters is already known, K-means clustering method is used. This method classifies n points of data X into initializing k clusters where u is the center points of an individual cluster and J is the minimum distance from every point to its center point. This classification is repeated until each cluster gets stable. This algorithm has been used in many aspects (28, 51) . In this study, we The 90th percentile of MDA8 O 3 is used to represent the degree of O 3 pollution in a specific region. Because the O 3 network data are only available after 2013, we then integrate O 3 data with the classified tracks from 2013 to 2018. LPDM was adopted with the aim to investigate the potential source region affecting air masses in YRD and PRD, respectively. Global Data Assimilation System (GDAS) data are used as the meteorological input, and the model is conducted using hybrid single-particle Lagrangian-integrated trajectory (HYSPLITY). In this study, we released 3000 particulates at 100 m over the receptor region, and the model was traced backward for 48 hours with a time resolution of 1 hour. After a calculation of mean wind and a turbulence transport component, the position of each particulate was identified in spatial. The footprint "retroplume," representing the distribution of the surface probability or the residence time of the simulated air mass, was used to identify the contribution from potential source regions. BVOC emissions by terrestrial ecosystems are estimated with the aid of MEGAN model (version 2.10). As a biogenic emission model, MEGAN can estimate global and regional BVOC emissions, with horizontal resolution ranging from 1 × 1 km to around 100 × 100 km. The model can estimate up to 146 BVOC species, for example, isoprene, monoterpene, sesquiterpene, and so on, and lump them according to a specific gas-phase chemistry mechanism. In this study, WRF-simulated results provide meteorological conditions including temperature, solar radiation, and relative humidity for MEGAN. Leaf area index and plant functional types retrieved from moderate resolution imaging spectroradiometer (MODIS) products are used to provide basic vegetation information. All the calculated BVOC species are mapped to Carbon Bond 05 (CB05) gas-phase mechanism. Major BVOC emissions, i.e., isoprene emissions, are identified and compared between typhoon-associated events and the summer averages. To illustrate the responses of O 3 to its precursors, the O 3 -NO x -VOC sensitivities are explored by using the ratios of satellite-retrieved formaldehyde (HCHO) and NO 2 . Daily HCHO and NO 2 column densities are obtained from the resolution of the GOME-2 retrievals. The data control is based on the principle by previous studies (52) . In this study, the area with HCHO/NO 2 ratio below 1 is considered as NO x -saturated (VOC-limited) regime, the area with HCHO/NO 2 ratio above 2 is identified as the NO x -limited regime, and the HCHO/ NO 2 ratio between 1 and 2 indicates the mixed-limited regime (53) . 4) ]. Anthropogenic emission inventories are obtained from the Multi-resolution Inventory for China, which has a grid resolution of 0.25° and covers anthropogenic emissions from human activities of power plant, transportation, industry, agriculture, and residential. Natural emissions including BVOCs and soil-emitted NO are calculated using the MEGAN (version 2.10). The WRF-CMAQ spins up for 10 days before starting a simulation. The key configurations of WRF-CMAQ include the Rapid Radiative Transfer Model (RRTM) for short-and long-wave radiation scheme, the Noah Land Surface Model for land-atmospheric interactions, the Yonsei University boundary layer scheme, the Lin microphysics scheme, the Kain-Fritsch scheme for cumulus parameterization, and the CB05 combined with AERO6 (aerosol 6 module) for gas-phase and aerosol chemistry, respectively. The model performance is then validated by the comparisons with available observations. Statistical metrics are introduced, including the averages (Obs mean and Sim mean ), mean bias (MB; Obs mean − Sim mean ), . The statistical results are summarized in table S1. In general, the evaluated results show that the WRF-CMAQ model can well reproduce the historical O 3 pollution events. The integrated process rate (IPR) module in CMAQ is used to calculate the contributions from each individual physical and chemical process to O 3 formation. The considered atmospheric processes include the photochemical process (the net O 3 formation rate via gas-phase reactions), the horizontal transport (including horizontal convection and diffusion), the vertical a transport (including vertical convection and diffusion), the emissions, the deposition (including dry and wet deposition), and the cloud processes with the aqueous chemistry. The individual contribution is calculated at each time step during the simulation. In this work, we mainly use the technique to investigate the key photochemical species including O 3 , CO, NO x , anthropogenic VOCs, BVOCs, and the intermediate BVOC product. To quantify all the individual impact of anthropogenic emissions and BVOC emissions to O 3 pollution, all the historical cases with typhoon crossing the Taiwan from 2013 to 2018 are simulated with WRF-CMAQ model. Parallel simulations are conducted under different emission scenarios. The scenarios include considering/ excluding BVOC emissions, anthropogenic emissions of PRD, anthropogenic emissions of YRD, and anthropogenic emissions out of YRD and PRD. For example, to identify the impact from PRD anthropogenic emissions on YRD, parallel simulations are conducted with and without PRD anthropogenic emissions. Then, the difference in O 3 between the two simulations is identified as the impact from PRD anthropogenic emissions. The influence from the emissions within the region itself is considered as the local contribution; otherwise, the one out of the border is regarded as the regional transport. By using this method, O 3 source apportionment is analyzed during historical typhoon events and the summer season. Moreover, by combining with the IPR technique, the changes of net O 3 formation rate due to gas-phase reactions (including photochemically production and consumption processes) and atmospheric physical process (i.e., advective transport) driven by a given set of emissions are recorded at each time step. Accordingly, the individual impact of anthropogenic emissions and BVOCs on O 3 formation can be identified. Supplementary material for this article is available at https://science.org/doi/10.1126/ sciadv.abl6166 Tropospheric ozone assessment report: Global ozone metrics for climate change Threat to future global food security from climate change and ozone air pollution Ground-level O3 pollution and its impacts on food crops in China: A review An integrated view of the causes and impacts of atmospheric changes Ground-level ozone in four Chinese cities: Precursors, regional transport and heterogeneous processes Increased polar stratospheric ozone losses and delayed eventual recovery owing to increasing greenhouse-gas concentrations Heterogeneous chemistry and tropospheric ozone Enhanced secondary pollution offset reduction of primary emissions during COVID-19 lockdown in China Formation of secondary organic aerosols through photooxidation of isoprene Distribution and budget of O3 in the troposphere calculated with a chemistry general circulation model Chapter 11 Ozone and other secondary photochemical pollutants: Chemical processes governing their formation in the planetary boundary layer OH radical formation from the gas-phase reactions of O3 with a series of terpenes Ozone pollution in China: A review of concentrations, meteorological influences, chemical precursors, and effects Estimates of global terrestrial isoprene emissions using MEGAN (Model of Emissions of Gases and Aerosols from Nature) The role of biogenic hydrocarbons in urban photochemical smog: Atlanta as a case study Tropospheric ozone change from 1980 to 2010 dominated by equatorward redistribution of emissions Role of export industries on ozone pollution and its precursors in China Tropospheric Ozone Assessment Report: Present-day distribution and trends of tropospheric ozone Anthropogenic drivers of 2013-2017 trends in summer surface ozone in China Global distribution and trends of tropospheric ozone: An observation-based review Ground-level PM2.5 concentration estimation from satellite data in the Beijing area using a specific particle swarm extinction mass conversion algorithm ENSO and Southeast Asian biomass burning modulate subtropical trans-Pacific ozone transport Simulation of sea-land breezes and a discussion of their implications on the transport of air pollution during a multi-day ozone episode in the Pearl River Delta of China Vegetation feedbacks during drought exacerbate ozone air pollution extremes in Europe NO2, PM10 and O3 urban concentrations and its association with circulation weather types in Portugal Air Quality during COVID-19 lockdown in the yangtze river delta and the pearl river delta: Two different responsive mechanisms to emission reductions in Understanding ozone pollution in the Yangtze River Delta of eastern China from the perspective of diurnal cycles Effect of meteorology and air pollutant transport on ozone episodes at a subtropical coastal Asian city, Hong Kong Integrated studies of a regional ozone pollution synthetically affected by subtropical high and typhoon system in the Yangtze River Delta region, China Dynamics of the typhoon haitang related high ozone episode over Hong Kong Numerical simulation and process analysis of typhoon-related ozone episodes in Hong Kong An overview of the China meteorological administration tropical cyclone database Severe surface ozone pollution in China: A global perspective Impacts of weather conditions modified by urban expansion on surface ozone: Comparison between the Pearl River Delta and Yangtze River Delta regions Impacts of compound extreme weather events on ozone in the present and future The impact of climate change on the Western Pacific subtropical high and the related ozone pollution in Shanghai Ozone affected by a succession of four landfall typhoons in the Yangtze River Delta, China: Major processes and health impacts Substantial ozone enhancement over the North China Plain from increased biogenic emissions due to heat waves and land cover in summer 2017 Temporal, spatial characteristics and uncertainty of biogenic VOC emissions in the Pearl River Delta region Impacts of seasonal and regional variability in biogenic VOC emissions on surface ozone in the Pearl river delta region, China Condensed atmospheric photooxidation mechanisms for isoprene Estimate of initial isoprene contribution to ozone formation potential in Beijing, China Nighttime chemistry at a high altitude site above Hong Kong Long-range transport of photochemical ozone in north-western Europe Global data set of biogenic VOC emissions calculated by the MEGAN model over the last 30 years Generation of a global synthetic tropical cyclone hazard dataset using STORM Westward migration of tropical cyclone rapid-intensification over the Northwestern Pacific during short duration El Niño Future change in tropical cyclone activity over the western north pacific in CORDEX-East Asia Multi-RCMs Forced by HadGEM2-AO A long-term estimation of biogenic volatile organic compound (BVOC) emission in China from 2001-2016: The roles of land cover change and climate variability A k-mean clustering algorithm for mixed numeric and categorical data Twelve years of global observations of formaldehyde in the troposphere using GOME and SCIAMACHY sensors Spatial and temporal variability of ozone sensitivity over China observed from the Ozone Monitoring Instrument