Search

We use the CAMx (Comprehensive Air Quality Model with Extensions) chemical transport model (CTM) with 4-km horizontal resolution over the Korean Peninsula to investigate source contributions to PM2.5 in Korea from domestic and upwind sources. We modeled 2015 and 2016 to account for meteorological variation with Korean emissions from the Clean Air Policy Supporting System (CAPSS), meteorology from WRF (Weather, Research, and Forecasting) model, and regional boundary concentrations from the GEOS-Chem global CTM. The CAMx particulate source apportionment technology (PSAT) provided PM2.5 source contributions from 5 source sectors and 6 geographic regions within Korea, international sources, and boundary concentrations. PM2.5 contributions from outside Korea are important with boundary concentrations plus the “other” emissions sector (includes marine shipping, agricultural ammonia, and international emissions from North Korea and Japan within the CAMx domain) contributing 67% of annual average PM2.5 in Seoul in 2016 and 71% in 2015. The boundary concentrations contributed between 30% and 50% of PM2.5 at different Korean cities with contributions generally lower in 2016 than in 2015. For Korean sources, PM2.5 contributions from Electric Generating Unit (EGU) emissions were smaller than contributions from mobile and industrial emissions sources although there is considerable day-to-day variation in contributions. On an annual basis in 2016, the “other” category contributed 25% followed by mobile sources at 23%, industrial sources at 6%, and EGU sources at 3%. For 2015, the contributions were similar. Focusing on March when PM2.5 concentrations were higher than other months, the contributions from other, mobile, industrial, and EGUs were 21%, 18%, 4%, and 4%, respectively in 2016. For 2015, contributions from these four categories were 18%, 15%, 3%, and 3%, respectively.

High concentrations of PM2.5 have become a serious environmental issue in South Korea, which ranked 1st or 2nd among OECD countries in terms of population exposure to PM2.5. Quantitative understanding of PM2.5 source attribution is thus crucial for developing efficient air quality mitigation strategies. Here we use a suite of extensive observations of PM2.5 and its precursors concentrations during the international KORea-US cooperative Air Quality field study in Korea (KORUS-AQ) in May–June 2016 to investigate source contributions to PM2.5 in South Korea under various meteorological conditions. For the quantitative analysis, we updated a 3-D chemical transport model, GEOS-Chem, and its adjoint with the latest regional emission inventory and other recent findings. The updated model is evaluated by comparing against observed daily PM2.5 and its component concentrations from six ground sites (Bangnyung, Bulkwang, Olympic park, Gwangju, Ulsan, and Jeju). Overall, simulated concentrations of daily PM2.5 and its components are in a good agreement with observations over the peninsula. We conduct an adjoint sensitivity analysis for simulated surface level PM2.5 concentrations at five ground sites (except for Bangnyung because of its small population) under four different meteorological conditions: dynamic weather, stagnant, extreme pollution, and blocking periods. Source contributions by regions vary greatly depending on synoptic meteorological conditions. Chinese contribution accounts for almost 68% of PM2.5 in surface air in South Korea during the extreme pollution period of the campaign, whereas an enhanced contribution from domestic sources (57%) occurs for the blocking period. Results from our sensitivity analysis suggest that the reduction of domestic anthropogenic NH3 emissions could be most effective in reducing population exposure to PM2.5 in South Korea (effectiveness = 14%) followed by anthropogenic SO2 emissions from Shandong region (effectiveness = 11%), domestic anthropogenic NOx emissions (effectiveness = 10%), anthropogenic NH3 emissions from Shandong region (effectiveness = 8%), anthropogenic NOx emissions from Shandong region (effectiveness = 7%), domestic anthropogenic OC emissions (effectiveness = 7%), and domestic anthropogenic BC emissions (effectiveness = 5%).

Anthropogenic emissions in East Asia have changed dramatically in recent years. To measure these changing emissions in support of air-quality modeling, we developed a top-down emission update system using surface observations and a geographical information system spatial-allocation technique. We deploy a data-processing system to construct adjustment factors to prefecture-level SO2 emissions by comparing surface and modeled observations. A case study is conducted over East Asia for 2016 in which we update Chinese SO2 emissions using measurements from around 1500 surface monitoring sites. Model simulations using updated SO2 emissions are improved relative to the existing simulation system (e.g., R = 0.23 to R = 0.8), suggesting that the newly designed system provides an efficient, practical forecast solution. Finally, estimated SO2 emissions are compared with existing emission inventories, agreeing well with recent reports of reduced SO2 emissions from Chinese anthropogenic sources.

We have developed the Integrated Climate and Air Quality Modeling System (ICAMS) through the one-way nesting of global–regional models to examine the changes in the surface ozone concentrations over East Asia under future climate scenarios. Model simulations have been conducted for the present period of 1996–2005 to evaluate the performance of ICAMS. The simulated surface ozone concentrations reproduced the observed monthly mean concentrations at sites in East Asia with high R2 values (0.4–0.9), indicating a successful simulation to capture both spatial and temporal variability. We then performed several model simulations with the six IPCC SRES scenarios (A2, A1B, A1FI, A1T, B1, and B2) for the next three periods, 2016–2025 (the 2020s), 2046–2055 (the 2050s), and 2091–2100 (the 2090s). The model results show that the projected changes of the annual daily mean maximum eight-hour (DM8H) surface ozone concentrations in summertime for East Asia are in the range of 2–8 ppb, −3 to 8 ppb, and −7 to 9 ppb for the 2020s, the 2050s, and the 2090s, respectively, and are primarily determined based on the emission changes of NOx and NMVOC. The maximum increases in the annual DM8H surface ozone and high-ozone events occur in the 2020s for all scenarios except for A2, implying that the air quality over East Asia is likely to get worse in the near future period (the 2020s) than in the far future periods (the 2050s and the 2090s). The changes in the future environment based on IPCC SRES scenarios would also influence the change in the occurrences of high-concentrations events more greatly than that of the annual DM8H surface ozone concentrations. Sensitivity simulations show that the emissions increase is the key factor in determining future regional surface ozone concentrations in the case of a developing country, China, whereas a developed country, Japan would be influenced more greatly by effects of the regional climate change than the increase in emissions.

We investigate ozone air quality changes in 2050 caused by global changes in climate and anthropogenic emissions of ozone precursors by using a global chemical transport model driven by meteorological fields from a general circulation model. We use projected emissions based on the Representative Concentration Pathway (RCP) scenarios and conduct model simulations to quantify the effects of climate and emission changes on future air quality, focusing on ozone in surface air. Our model results show that annual mean concentrations of surface ozone will be lower in 2050 relative to 2000 by −3.3, −3.7, and −4.2 ppbv under RCP6.0, RCP4.5, and RCP2.6, respectively. In contrast, the RCP8.5 projection results in a slight increase of 2.1 ppbv caused by a methane increase. The ozone reductions are driven primarily by decreases in NOx emission, which dominate the climate penalty on ozone driven by temperature increases. We also estimate the effect of 21st century climate change on ozone air quality, assuming no changes in anthropogenic emissions of ozone precursors in the future. We further use a statistical method to analyze the results in order to quantify the effect of each meteorological variable change on ozone concentration in summer. Temperature increase is found to result in ozone increases of up to 2.2 ppbv over land. Ozone over the oceans, however, is largely reduced with specific humidity increase, particularly in the Northern Hemisphere, where the ozone concentration decreases by 0.8 ppbv. We find that future increases in natural NOx emissions from lightning and soil make an important contribution to the formation of nitric acid and might seriously offset future decreases in nitrogen deposition caused by anthropogenic NOx emission reduction.

Ozone concentrations in East Asia were simulated using the Community Multi-scale Air Quality (CMAQ) model, and its source contributions were estimated by multiple source–receptor modeling techniques. To study relationships between ozone concentrations and precursor emission sources, three approaches were applied to four months (January, April, July, and October 2009) to represent seasonal characteristics and compare results, with a particular focus on South Korea. Brute force (BF) is a traditional sensitivity analysis method used to estimate model output response to an input change. The high-order decoupled direct method (HDDM), a computational method, is an efficient and accurate alternative to the BF method for sensitivity. The Ozone and Particulate Precursor Tagging Methodology (OPTM) provides contribution information quantified by tracking emissions from selected sources throughout the simulation period. The approaches generally show that most of the receptor regions were substantially influenced by emissions from central China, which is the largest anthropogenic emissions source region in East Asia. Local emissions were still major contributors, especially South Korea and Japan during July 2009. On the other hand, a case study of maximum 8-h ozone concentrations derived from CMAQ–OPTM on April 9 in South Korea shows that the NOx and VOCs emissions from China contributed approximately 82% and 91%, respectively, to maximum 8-h ozone in Region 4 (South Korea) without boundary inflow, which indicates that Chinese emissions are the dominant contributor in this episode. A comparison study of the three approaches shows that HDDM tends to estimate biogenic source contributions lower than that from OPTM in China but similar to OPTM in South Korea and Japan. When comparing the BF method and HDDM, the sensitivity results show a reasonably good agreement during a given period. The location- and time-dependent maximum 8-h ozone isopleths over South Korea as a receptor region created by HDDM suggest that most ozone was being transported from central China, whereas almost no ozone was formed locally during April 2009, and local conditions were heavily VOC limited. On the other hand, local emissions were the dominant contributor during July 2009, and every source region showed a NOx-limited regime, which indicates that ozone concentrations in South Korea strongly depend on NOx emissions during this month.

Significant differences were found between two particulate sulfate size-distributions measured using a MOUDI impactor at Gosan, Jeju Island, Korea, and simulated via the Lagrangian photochemical model under a condition of aerosol internal mixing between mineral dust and urban pollution particles. It was suspected that these differences might have resulted from: (1) the assumption of aerosol internal mixing and (2) the uses of identical reaction probabilities (γ) of the gaseous sulfate precursors (SO2 and H2SO4) onto both urban pollution particles and mineral dust in the Lagrangian photochemical model simulations. In this study; therefore, some cases of aerosol external mixing between urban-derived pollution aerosols and mineral dust were investigated, with different magnitudes of γ for the gas-phase sulfate precursors onto the two different condensing media. The model simulations under the external mixing condition, with different magnitudes of the reaction probabilities (γi,urban and γi,dust) onto urban pollution particles and mineral dust, successfully reproduced the size-dependent particulate sulfate formation measured at the Gosan station. Further attempts were made to approximate the magnitudes of γSO2,urban and γH2SO4,dust under external mixing state conditions with the fixed γSO2,dust and γH2SO2,urban values of 10−4 and 1.0. The best-estimates of γSO2,urban and γH2SO4,dust found in this study were in the orders of 10−4–10−5 and 10−2–10−3, respectively.

Air quality modeling is a useful methodology to investigate air quality degradation in various locations and to analyze effectiveness of emission reduction plans. A comprehensive air quality model usually requires a coordinated set of emissions input of all necessary chemical species. We have developed an anthropogenic emissions processing system for Asia in support of air quality modeling and analysis over Asia (named SMOKE-Asia). The SMOKE (Sparse Matrix Operator kernel Emissions) system, which was developed by U.S. EPA and has been maintained by the Carolina Environmental Program (CEP) of the University of North Carolina, was used to develop our emissions processing system. A merged version of INTEX 2006 and TRACE-P 2000 inventories was used as an initial Asian emissions inventory. The IDA (Inventory Data Analyzer) format was used to create SMOKE-ready emissions. Source Classification Codes (SCCs) and country/state/county (FIPS) code, which are the two key data fields of SMOKE IDA data structure, were created for Asia. The 38 SCCs and 2752 FIPS codes were allocated to our SMOKE-ready emissions for more comprehensive processing. US EPA’s MIMS (Multimedia Integrated Modeling System) Spatial Allocator software, along with many global and regional GIS shapes, were used to create spatial allocation profiles for Asia. Temporal allocation and chemical speciation profiles were partly regionalized using Asia-based studies. Initial data production using the developed SMOKE-Asia system was successfully performed. NOx and VOC emissions for the year 2009 were projected to be increased by 50% from those of 1997. The emission hotspots, such as large cities and large point sources, are distinguished in the domain due to spatial allocation. Regional emission peaks were distinguished due to temporally resolved emission information. The PAR (Paraffin carbon bond) and XYL (Xylene and other polyalkyl aromatics) showed the first and second largest emission rate among VOC species. Most of point source emissions are located in layers 3 to 4, which the altitude range reaches 310–550 m AGL. Qualitative inter-comparison between model output and ground/satellite measurement showed good agreements in terms of spatial and temporal patterns. We expect that the result of this study will provide better air quality modeling inputs, which will act as a major step to improve our understanding of Asian air quality.

We examine the effects of below-cloud scavenging on regional aerosol simulations over East Asia using wet deposition fluxes observed at Acid Deposition Monitoring Network in East Asia (EANET) sites and the Community Multiscale Air Quality (CMAQ) model together with a new below-cloud-scavenging scheme. Typical air quality models, including CMAQ, assume below-cloud scavenging as a simple first-order process with a constant or simple form depending on rain intensity. The scheme used here accounts for the collection efficiency, terminal velocity of raindrops, raindrop-size distributions, and particle-size distributions, which are important factors affecting below-cloud scavenging. We conduct model simulations for spring 2001, including baseline and sensitivity simulations. Our analysis mainly focuses on May 2001 to rule out the effect of dust aerosols. Simulated wet deposition fluxes of SO42−, NO3−, and NH4+ by the new scheme are increased by 103, 16, and 108%, respectively, relative to the baseline simulation and show better agreement with observations. The effect of below-cloud scavenging on coarse particles is even greater, producing wet deposition fluxes two orders of magnitude higher than the baseline. The resulting changes in the model indicate the considerable impacts of below-cloud scavenging on regional aerosol simulations over East Asia, where both anthropogenic emissions and natural sources of aerosols are present throughout the year. An accurate wet scavenging simulation is critical to simulate the atmospheric burden and wet deposition fluxes of both fine-mode and coarse-mode aerosols over East Asia.

We examine the impacts of aerosols on regional meteorology due to intense Siberian forest fires occurred in May 2003 using both reanalysis data and global model simulations. Our analysis of the NCEP-DOE reanalysis data shows 99% statistical significant changes in meteorological variables over East Asia in May 2003 relative to the 30 years climatology. In particular a significant surface cooling was observed up to −3.5 K over Siberia and extended to the North Pacific region with the surface pressure increases up to 5.6 hPa. Whereas, smoke aerosols affected the large-scale circulations and resulted in the increases in rainfall rates of 2.9 mm day −1 averaged over the NW Pacific (10–35°N, 130–170°E). We use the climate model simulations with and without biomass burning emissions over Siberia to examine the effects of smoke aerosols on the regional meteorology. The simulated results show consistent changes in meteorological variables including surface temperature, surface pressure and precipitation rates with the observations over East Asia and the NW Pacific, which support that the observed changes are likely due to smoke aerosols from the Siberian forest fires. The implication is that smoke aerosols from the forest fires should be properly considered to correctly simulate both regional climate and synoptic scale weather patterns.

We estimated the source contributions to carbonaceous aerosol concentration in Korea on the basis of Intercontinental Chemical Transport Experiment Phase B (INTEX-B) anthropogenic emissions and satellite-derived biomass burning emissions by using a nested version of GEOS-Chem with a spatial resolution of 0.5 � � 0.667 � for the period March 2006eFebruary 2007. First, we evaluated the model by comparing the simulated and observed aerosol concentrations at East Asia Network (EANET) sites and at a site in Korea. The results indicate that the model reproduces the variability and magnitudes of the observed SO4 2� ,N O 3 � , and NH 4 þ concentrations in Korea and those of the observed PM 10 concentrations in East Asia. However, the organic carbon (OC) and black carbon (BC) aerosol concentrations estimated by the model are lower than those observed in Korea by a factor of 2, especially in winter. This underestimation is likely due to extremely low domestic anthropogenic emissions and lack in seasonal variation. Source adjustments using a simple fitting and the Emission Database for Global Atmospheric Research (EDGAR) monthly allocation factors for seasonal variation yield significantly improved model results (R 2

Household fuel use in developing countries, particularly as biomass and coal, is a major source of carbonaceous aerosols and other air pollutants affecting health and climate. Using state-of-the-art emission inventories, a global three-dimensional photochemical tracer/transport model of the troposphere, and a global radiative transfer model based on methods presented in the latest IPCC Assessment Report (2007-AR4), we estimate the radiative forcing (RF) attributable to household fuel combustion in Asia in terms of current global annual-mean RF and future global integrated RF for a one-year pulse of emissions (2000) over two time horizons (100 and 20 years). Despite the significant emissions of black carbon (BC) aerosols, these estimates indicate that shorter-lived (non-Kyoto) air pollutants from household fuel use in the region overall seem to exert a small net negative RF because of the strong influence of reflective aerosols. There are, however, major uncertainties in emission estimates for solid fuel burning, and about the sustainability of household fuel wood harvesting in Asia (the carbon neutrality of harvesting). In addition, there is still substantial uncertainty associated with the BC radiative forcing. As a result we find that the sign of the RF from household biomass burning in the region cannot be established. While recognizing the value of integrating climate change and air pollution policies, we are concerned that for a ‘Kyoto style’ post-Kyoto treaty (with global cap-and-trade and the Global Warming Potential as the metric) expanding the basket of components with a selection of short-lived species without also including the wider range of co-emitted species may lead to unintended consequences for global-scale climate. Additional measurement, modelling, and policy research is urgently needed to reduce the uncertainties so that the net impact on climate of emissions and mitigation measures in this sector can be accurately assessed.

In this study, background concentration sites of Deokjeok and Gosan, which were deemed suitable for monitoring the impact of long-range transported air pollutants, were selected. An investigation of the source types of pollutants, their locations, and relative quantitative contributions to the particulate concentrations at both sites using appropriate methodologies to make initial estimations was conducted. Episodic measurements of PM2.5, PM10, and size distribution, along with its ion and carbon components were performed from 2005 to 2007, and a comprehensive analysis of the results was conducted utilizing back trajectory analysis. As for frequency of wind direction, it was quite apparent that the two sites are heavily influenced by air masses originating from the eastern and northern regions of China. For PM2.5 and PM10, the mass concentrations from north and east China were higher than other cases, originating from the ocean. In the northerly-wind case, meteorological properties for Deokjeok and Gosan and the influence of carbon emissions from northwest Korea resulted in a changing of air mass properties during transport. As was the case with mass concentration, the highest contribution for ionic and carbon components of PM2.5 and PM10 for both sites appeared for the westerly wind case. A specially high relative contribution, greater than 1.4 times, was apparent in the secondary aerosol case because of a large influence of long-range transported pollutants from east China. Carbon components exhibited different behaviors for the northerly and westerly wind cases compared with secondary aerosol. The major reason for this discrepancy appears to be the carbon emissions from northwest Korea.

The fate and transport of mercury in East Asia is evaluated using the Sulfur Transport and dEposition Model (STEM)-Hg 3-D model. The model calculates mercury transport, transformation and deposition in East Asia during April 2001, the period of the ACE-Asia field campaign. Model predictions of dry and wet deposition are compared with the observations from 10 sampling sites in Japan. The model results are consistent with the observations, but tend to over-predict dry deposition. Sensitivity analysis of predicted results to uncertainties in the mercury reaction rates suggests that the oxidation of Hg0 to Hg(II) in the gas phase is the dominant pathway for atmospheric mercury removal processes. Simulation based on the most recently published reaction rate constants for gas phase oxidation of Hg0 to Hg(II) overestimates the production of Hg(II) in the gas phase. The regional mercury budget is calculated which shows that most of the reactive gas phase mercury (RGM) and particulate mercury (Hgp) are deposited around the source region while 28.5 Mg of Hg0 is exported out of East Asia during April 2001. The sensitivity analysis in the regional Hg budget to major uncertainties associated with Hg emission estimates and Hg chemistry are also evaluated.

An inverse modeling method using the four-dimensional variational data assimilation approach is developed to provide a top-down estimate of mercury emission inventory in China. The mercury observations on board the C130 aircraft during the Asian Pacific Regional Aerosol Characterization Experiment (ACE-Asia) campaign in April 2001 are assimilated into a regional chemical transport model, STEM. Using a 340 Mg of elemental mercury emitted in 1999, the assimilation results in an increase in Hg 0 emissions for China to 1140 Mg in 2001. This is an upper limit amount of the elemental mercury required in China. The average emission-scaling factor is ∼3.4 in China. The spatial changes in the mercury emissions after the assimilation are also evaluated. The largest changes are estimated on the China north-east coastal areas and the areas of north-center China. The influences of the observation and inventory uncertainties and the initial and boundary conditions on the emission estimates are discussed. Increasing the boundary conditions of Hg from 1.2 to 1.5 ng m −3 , results in a top-down estimate of Hg 0 emissions for China of 718 Mg, and leads the average scaling factor from 3.4 to 2.1.

In this study, the accuracy of NH3 emissions in East Asia is evaluated by a comparison of measured and predicted NH4NO3 concentrations. For the prediction of particulate NH4NO3 concentrations, this study uses the USEPA Models-3/Community Multi-scale Air Quality (Models-3/CMAQ) model. The measurement data are collected from 4 ground-based stations (Beijing, Shanghai, Qingdao, and Kangwha) and ACE-ASIA C130 Flights #6, #7, and #10. The four ground-based stations and air flight tracks are located in and around the Yellow Sea in East Asia. According to the results, the predicted fine-mode NH4+concentrations are 1.2–3.8 times overestimated compared to those measured on the mass basis, and fine-mode NO3− concentrations are 1.9–7.6 times overestimated on the mass basis. It is also revealed in this study that aberrantly high predicted fine-mode NO3− concentrations in East Asia can be attributed to the overestimated NH3 emissions. The predicted gas-phase NH3 concentrations and estimated NH3 emissions are likely to be 1.2–3.8 times overestimated in East Asia on the mass basis. Therefore, it is recommended that approximately 20–75% reduced NH3 emission fields should be used in future modeling studies for East Asia.

Long-range transport of pollutants influenced by anthropogenic and natural emission sources in East Asia is investigated by using backward trajectory analysis along the NASA TRACE-P flight tracks and a numerical simulation with the three-dimensional chemical transport model (STEM-2k1). Observation-based regional distributions of trace gases are reconstructed using the observations obtained by measurements on board the DC-8 and P3-B aircrafts. Systematic features of the spatial distribution for each species are identified. It is found that the observed concentrations of CO and some NMHCs, and the ratios between these species, are highly associated with the source distribution features and their regional characteristics. Reconstructed fields of the observed and modeled ethane/CO and ethane/propane are found to reproduce well the estimated emission ratios in East Asia. We also investigated the time rate of change of the concentration of species and their ratio along the trajectory. From this analysis the propane/ethane and propane/acetylene ratios are shown to preserve their emission ratios during regional transport. However systematic differences in the propane vs. acetylene/CO relationships are found between the model and observation values. This analysis suggests that further efforts are needed to improve the estimates of biomass burning emissions in SE Asia. The results presented in this paper also suggest ways to further extend the capabilities to derive observation-based inventories.

Asia is undergoing rapid urbanization resulting in increasing air pollution threats in its cities. The contribution of megacities to sulfur emissions and pollution in Asia is studied over a 25-year period (1975–2000) using a multi-layer Lagrangian puff transport model. Asian megacities cover 10% in 2000. Two future emission scenarios are evaluated for 2020—“business as usual (BAU)” and “maximum feasible controls (MAXF)” to establish the range of reductions possible for these cities. The MAXF scenario would result in 2020 S-emissions that are ∼80% lower than those in 2000, at an estimated control cost of US $87 billion per year (1995 US$) for all of Asia. An urban scale analysis of sulfur pollution for four megacities—Shanghai, and Chongqing in China; Seoul in South Korea; and Mumbai (formerly Bombay) in India is presented. If pollution levels were allowed to increase under BAU, over 30 million people in these cities alone would be exposed to levels in excess of the WHO guidelines.

An intercomparison study involving eight long-range transport models for sulfur deposition in East Asia has been initiated. The participating models included Eulerian and Lagrangian frameworks, with a wide variety of vertical resolutions and numerical approaches. Results from this study, in which models used common data sets for emissions, meteorology, and dry, wet and chemical conversion rates, are reported and discussed. Model results for sulfur dioxide and sulfate concentrations, wet deposition amounts, for the period January and May 1993, are compared with observed quantities at 18 surface sites in East Asia. At many sites the ensemble of models is found to have high skill in predicting observed quantities. At other sites all models show poor predictive capabilities. Source–receptor relationships estimated by the models are also compared. The models show a high degree of consistency in identifying the main source–receptor relationships, as well as in the relative contributions of wet/dry pathways for removal. But at some locations estimated deposition amounts can vary by a factor or 5. The influence of model structure and parameters on model performance is discussed. The main factors determining the deposition fields are the emissions and underlying meteorological fields. Model structure in terms of vertical resolution is found to be more important than the parameterizations used for chemical conversion and removal, as these processes are highly coupled and often work in compensating directions.