The eruption of the Miyake-jima Volcano (34.08°N, 139.53°E) in the Izu Islands, Japan, 180 km SSW of Tokyo, began on 8 July 2000. A substantial amount of NH3 gas was found to be emitted from the Miyake-jima Volcano together with SO2 gas and that geochemically significant quantities of aerosol particles composed of ammonium sulfate form in the plume. Through the use of satellite images, the additional atmospheric deposition of ammonium sulfate caused an increase of phyto-plankton in the nutrient-deficient region south of the Kuroshio. The emission of volcanic gases from the Miyake-jima has likely been modifying the marine air quality as well as the open ocean ecosystem over parts of the western North Pacific for the past several years.
 The eruption of the Miyake-jima Volcano began on 8 July 2000 and was still ongoing as of September 2003. The total emission amount of SO2 was estimated to be over 18 × 1012g SO2 from the beginning of the eruption until September 2003 and the average flux of SO2 from the volcano calculated with vertical profiles of SO2 obtained by COSPEC (Correlation Spectrometer) and wind velocity [Kazahaya et al., 2001] was found to exceed 20 × 109g day−1 during this period (data are available from the World Wide Web server at http://staff.aist.go.jp/kazahaya-k/miyakegas/COSPEC.html). This represents approximately 25% of the emission of SO2 from China and more than 10 times as that from Japan [Streets et al., 2000].
 A self-cruising observation boat named SCOOP (Self Cruising Ocean Observation Platform) [Senga et al., 2000] was deployed to investigate the volcanic gases and aerosols emitted from Miyake-jima including SO2 over the ocean. SCOOP intercepted volcanic plumes from Miyake-jima three times during the cruises. Based on the chemical measurements in the marine atmosphere, we investigated the volcanic effect to the marine environment.
2.1. SCOOP Measurement
 The self-cruising mono-hull boat equipped with atmospheric and oceanic measurement systems is made of FRP (Fiber Reinforced Polymer) and has a length of 8.0 m, a maximum width of 2.8 m, and a maximum displacement of 3.5 tons. It travels at a speed of 2–4 knots. DC and AC dynamos are provided to supply electric power by a diesel engine. Sample air taken from the top of a 6-m high mast at approximately 0.5 m3 h−1 was introduced into analytical and sampling instruments for aerosols and gases. For in-situ measurements of gaseous SO2 and NH3, sodium formate and H2O respectively were used as absorption solutions (mist trap absorbing technique) after removing aerosols by filtration. The concentrated absorption solutions for SO2 and NH3 were determined by chemiluminescence (cerium sulphate) [Takeuchi and Ibusuki, 1986] and fluorometric (o-phthaldialdehyde) [Genfa and Dasgupta, 1989] techniques under micro-flow injection systems, respectively.
 Ambient aerosols segregated into two size fractions (d < 2.5 μm and d > 2.5 μm) were collected on a PTFE (Poly Tetra Fluoro Ethylene) fiber tape filter by using a dichotomous virtual impactor air sampler equipped with an optical black carbon monitor. Simultaneously, vertical profiles of temperature, salinity and chlorophyll fluorescence in the water column can be obtained by a yo-yo system [Senga et al., 2000].
2.2. Numerical Model
 To determine the area affected by the volcanic deposition, the processes of gas-to-particle conversion and the deposition of volcanic SO2 were simulated by use of a numerical model to determine the behavior of water-soluble particles (MSSP; regional-scale Eulerian Model System for Soluble Particles) [Kajino et al., 2003]. The MSSP model includes the emissions of gas and aerosol, photochemical reactions, transport, deposition, and gas-to-aerosol equilibrium of volatile components. The simulation of sulfate aerosol for the Miyake-jima eruption has been well validated by ground based measurement of SO2 and particulate sulfate at a mountain site in central Japan [Kajino et al., 2003].
2.3. SeaWIFS Satellite Image Analysis
 Monthly SeaWiFS satellite images were observed over the surrounding region of the Japanese islands (16–45°N, 115–155°E) for 48 months from September 1997 to August 2001. To clearly detect any effect caused by the atmospheric deposition of volcanic substances, the nutrient deficient oceanic region to the south of the Kuroshio was chosen. To eliminate the effect of the yearly variation of the Kuroshio position, the sea surface temperature above a threshold value between 19–29°C was used to determine its position for each month.
3. Results and Discussion
3.1. Emission of Ammonia Gas From the Miyake-Jima Volcano
 SCOOP intercepted a volcanic plume 80–90 km west of Miyake-jima on 25 May 2001 (Figure 1). A maximum SO2 concentration of 40 ppb was observed in the plume, compared with background concentrations of less than 5 ppb outside the plume. Simultaneously, the NH3 concentration in the volcanic plume increased to 5.7 ppb from less than 0.5 ppb. Following the observed peaks of the aerosol precursor gases, a peak in particulate SO42− was observed (6.1 μg m−3 for 4-hr sampling) in the fine particle size (PM2.5), and this was associated with a molar ratio relative to the NH4+ ion of 1–1.2. It is believed that during the several hours of transport from Miyake-jima to SCOOP, significant quantities of SO2 and NH3 were converted to particulate NH4HSO4 and (NH4)2SO4. On a subsequent SCOOP cruise on 3–7 December 2001, we confirmed that fine ammonium sulfate aerosol was co-incident with a peak in volcanic SO2.
 In situ optical measurement of black carbon was also made, and the anthropogenic sulfate associated with black carbon was distinguished from the less colored volcanic sulfate. The molar ratio of Cl−/Na+ in the fine particle sizes of the marine aerosols (0.18 ± 0.05) was affected by the volcanic plume and tended to be lower than that of the similar sized aerosols under normal conditions (0.57 ± 0.11) along the cruise track in the coastal region where the air has already been affected by anthropogenic gases. Acidic gases emitted from the eruption are assumed to have reacted with sea-salt particles and released HCl gas from the particles in the marine boundary layer, resulting in lower Cl/Na ratios. The contribution of volcanic HCl seemed to be rather small compared with SO2 in this case [Kazahaya et al., 2001]. The loss of chloride from the aerosol may cause ozone depletion [Chang et al., 2004].
 Few published data are available on the emission of NH3 gas from volcanoes. However, NH4Cl (salammoniac) evaporite has been reported to occur near the volcanic vents at Miyake-jima after the 1983 eruption [Hirabayashi et al., 1984]. If the Miyake-jima volcano has been emitting NH3 and SO2 more-or-less continuously since the 2000 eruption, we can estimate the emission flux of reduced nitrogen based on the molar ratios of N to S in the plume measured (NH4:SO4:NH3:SO2 = 1:1:1:12). For the first 12 months after September 2000, the estimated amount of SO2 emitted was 10 × 1012g SO2 yr−1 (data are available at http://staff.aist.go.jp/kazahaya-k/miyakegas/COSPEC.html). As the amount of reduced nitrogen (NH4 + NH3) is equivalent to 15% on a molar basis of the emitted SO2 from Miyake-jima, the corresponding flux of ammonia-nitrogen from the Miyake-jima volcano for the year starting from September 2000 would be 34 × 1010 g N yr−1. This N flux is approximately 6% of the estimated total deposition of the reduced nitrogen (590 × 1010 g N yr−1) for the entire North Pacific [Duce, 1991]. The additional supply of the volcanic nitrogen compounds to the oceanic environment could therefore affect the chemical composition of a vast area of the surrounding surface ocean and hence conceivably affect oceanic primary productivity. In particular, uptake of reduced nitrogen compounds by phyto-plankton is known to be faster than uptake by oxidized forms of nitrogen, such as NO3−. If the additional nitrogen input from the volcanic source enhanced primary production in this region, the increase of chlorophyll-a concentration should be able to be seen after the 2000 eruption.
3.2. Distribution of Volcanic Substances Deposited
 The distribution of volcanic SO2 deposition was simulated by use of the MSSP numerical model. However, the sulfate deposition flux has not been validated quantitatively over an oceanic region. According to the model simulation, most of volcanic particulate sulfate would be deposited by wet removal associated with heavy precipitation events over the eastern region off northern Japan, where high primary production is also occurring due to mixing between the Kuroshio and nutrient rich cold waters of the Oyashio. To evaluate the deposition of reduced nitrogen, particulate ammonium salts were assumed to be deposited to the ocean surface as (NH4)2SO4. Before an eruption, the additional reduced nitrogen flux from the volcanic source to the nutrient deficient oceanic region south of the Kuroshio (16–45°N, 115–155°E) probably contributed little percentage of the atmospheric reduced nitrogen flux, since at that time most of the reduced and oxidized nitrogen was from anthropogenic sources in the spring.
3.3. Enhanced Primary Productivity After the Eruption
 Chlorophyll-a concentrations derived from the SeaWiFS monthly composite images (spatial resolution: 4 km × 4 km) before and after the eruption show higher values after the eruption. These high values are correlated with high sea surface temperatures (indicating the waters located south of the Kuroshio, primarily in the region of 16–25°N, 129–155°E) and high atmospheric deposition flux of volcanic sulfate simulated by the model.
 For example, compared with the satellite image data before eruption, the chlorophyll-a concentration after eruption tended to be higher in the summer (Figure 2). During winter, the seasonal vertical mixing would supply nutrients to the sea surface layer from deeper layers, but strong stratification would reduce the mixing of water less than 30-m deep in summer. The enhanced chlorophyll-a concentration during the summer after the eruption confirmed the supply of additional nutrients from atmosphere. It is difficult to detect the atmospheric nutrient input and its effects in winter and/or in high primary productivity regions, where the high chlorophyll-a concentration have large temporal and geographical variability.
 Another atmospheric input that can enhance primary productivity is mineral dust aerosol transported from the Asian continent. This will supply iron as a micronutrient to this region, although the deposition flux of mineral dust is rather small over the region south of 30°N in the western North Pacific in summer [Uematsu et al., 2003]. It is possible to stimulate the nitrogen fixing organisms during the dusty season in spring, when the investigated region is deficient in nutrients, particularly nitrogen compounds [Karl et al., 1997]. The occurrence of dust storms over the Asian continent and the frequency of dust event reports over Japan increased significantly every spring from 2000 to 2002 compared with the rest of the year [Kurosaki and Mikami, 2003]. However, a comparison of the satellite images from the spring of 1997 through 2000 reveals no pronounced change of the chlorophyll-a concentration in this region, except for April 2001. Therefore, atmospheric iron deposition may not be a major factor for the enhancement of primary productivity in the region of south of the Kuroshio in summer.
 The relationship between chlorophyll-a and sea surface temperature was determined for anomalies in the investigated region (Figure 3). There is a pronounced anti-correlation between the chlorophyll-a and the sea surface temperature, i.e., the chlorophyll-a concentration is low in the waters with high sea surface temperature. Although the monthly mean sea surface temperature is in the same range before and after the eruption, the chlorophyll-a concentrations are clearly high in July and August of 2000 and 2001 after the eruption. The associated volcanic reduced nitrogen supply to the ocean surface would enhance the primary productivity in a large area of the western North Pacific.
 The sources of the reduced nitrogen from the Miyake-jima volcano have yet to be determined. One possibility is that organic material that has accumulated in the oceanic sediments reacts with rising magma beneath the volcano, producing NH3 gas. If this hypothesis is true, more NH3 or other nitrogen compound emissions will be detected from other volcanoes that similarly interact with sedimentary materials of the world's oceans.
 We thank T. Suzuki, Y. Suzaki, and M. Suemori for developing and operating the SCOOP system. This study was supported by the Core Research for Evolutional Science and Technology (CREST) of the Japan Science and Technology Agency (JST).