VALIDATION OF TOMS VOLCANIC AEROSOL AND SO₂ PRODUCTS USING MODIS AND AVHRR

(3) Abstract.

This proposed research stems from the unique capability of our research group to validate the TOMS SO₂ and aerosol retrievals from other (infrared) satellite-based instruments, previously with AVHRR and GOES and now with Terra's MODIS sensor. We believe we are in position to provide the Earth Enterprise project with a valuable validation capability and complementary data products for the study of volcanogenic sulfur dioxide and aerosols.

The goal of this project is to provide a better understanding of the formation and fate of volcanogenic tropospheric aerosols, through improved detection, accuracy and interpretations of TOMS data. We will focus on several objectives to accomplish our goal: (1) using the recent eruption of Hekla (and for other eruptions) derive SO₂, ash and aerosol products from MODIS and other infrared satellite data to compare with TOMS measurements; (2) use the TOMS 20-year database and wind trajectory modeling to analyze cases of ash, aerosol and gas separation in drifting clouds, in order to compare IR/UV satellite retrievals; and (3) quantify chemical reaction rates for sulfur dioxide removal, and aerosol formation and removal, under a broad range of volcano-atmosphere conditions.

Our proposal goal and objectives respond to the Earth Science Enterprise's NRA by providing a robust means of validating the TOMS aerosol and SO₂ retrievals of volcanic emissions, which is a component of the Earth Enterprise interests in understanding how changes in atmospheric chemistry can impact global climate. We further wish to express our interest in joining the TOMS Science Team, as outlined in the NRA. This would serve to solidify our decade-long collaboration between Michigan Tech and NASA/Goddard. Our work would be highly complementary to that of Dr. Arlin Krueger and colleagues, with their established research in TOMS SO₂ and Aerosol Index algorithms.

VALIDATION OF TOMS VOLCANIC AEROSOL AND SO₂ PRODUCTS USING MODIS AND AVHRR

(4) Project Description

a. Project Objectives

Volcanic emissions of sulfur dioxide have the potential to generate abundant and long-lived atmospheric aerosols. While the climatic effect of stratospheric aerosols generated by large, infrequent eruptions is relatively well known (e.g., McCormick et al., 1995), the impacts on the atmosphere and climate by more common but less explosive events are uncertain. For example, 11 major events by Nyamuragira volcano, Zaire have injected roughly 0.5 Tg/yr over the last two decades (Bluth and Oppenheimer, 2000), compared to roughly 4 Tg/yr by all explosive events (Bluth et al., 1993).

The detection and measurement of these aerosols is important in studying their potential climate forcing. There are a number of unknown factors, however, namely: (1) the magnitude and distribution of tropospheric aerosols produced by volcanoes; (2) the aerosol formation rate with respect to sulfur dioxide injection mass; (3) the lifetime of the aerosols. The Total Ozone Mapping Spectrometer (TOMS) has a unique role in understanding these processes, with an ability to measure the sulfur dioxide precursor (Krueger et al., 1995), and the aerosols themselves (Krotkov et al., 1997). The importance of this role depends on the ability of the TOMS sensor and retrieval method to accurately discern and measure sulfur dioxide and the resulting aerosols, under a variety of volcanological and environmental conditions.

The Moderate Resolution Imaging Spectroradiometer (MODIS) sensor on board the EOS Terra platform provides us the first possible opportunity to cross-validate both TOMS retrieval products. Using MODIS we can produce maps of SO₂ clouds (Realmuto et al., 1997) from eruptions also seen by TOMS. From standard MODIS aerosol products and the split window technique (Wen and Rose, 1994) and a new sulfate retrieval scheme under development (Yu and Rose, 2000), we can compare our results with the TOMS Aerosol Index.

The goal of this project is to: provide a better understanding of the formation and fate of volcanogenic tropospheric aerosols, by improving the accuracy and utility of TOMS data.

To accomplish this goal, we plan to focus on the following main objectives:

(1) using the recent eruption of Hekla (and other eruptions as they occur), derive SO₂ and aerosol products from MODIS data to compare with near-coincident TOMS measurements;

(2) to use the TOMS 20-year database and wind trajectory modeling (for discerning vertical dimensions of clouds) to document cases of ash/aerosol and gas separation, in order to use complementary AVHRR data and new abilities to derive ash products;

(3) for the range of eruptions we have documented in the TOMS historical database, quantify chemical reaction rates for sulfur dioxide removal, and aerosol formation and removal, under a variety of volcano-atmosphere conditions.

Our proposal goal and objectives respond to the Earth Science Enterprise's NRA by providing a means of validating the TOMS aerosol and SO₂ retrievals of volcanic emissions, which is a component of the Earth Enterprise interests in understanding how changes in atmospheric chemistry can impact global climate. We believe we are uniquely able to quantitatively compare TOMS retrievals with those from the MODIS instrument, in addition to AVHRR and GOES, and this capability will provide NASA with powerful and unique methods for validating TOMS datasets.

b. Expected Significance

This project has the potential to make a significant impact on the Earth Enterprise project goals of validating TOMS data, as well as improving our understanding of how perturbations to the Earth's atmosphere may subsequently produce climatic change.

(1) This work will provide spectrally-independent validation of TOMS aerosol and SO₂ retrievals, through MODIS/TOMS case studies. Validation of volcanic emission retrievals is extremely difficult, due to the size, location, and great number of localized geometric, atmospheric, and volcanologic conditions which affect the behavior of and ability to observe volcanic clouds. A further obstacle comes from the sporadic nature of volcanic eruptions, combined with the various spatial, temporal and spectral resolutions of satellite and ground-based instruments used to study volcanic clouds. A significant accomplishment has been made through the comparison of UV and IR retrievals (Krotkov et al., 1999), in the case of near-coincident measurements of a cloud using TOMS and AVHRR. This study will create important opportunities for validation of the TOMS UV retrieval method with the Terra platform's MODIS instrument, starting with the excellent opportunity provided by the recent eruption of Hekla.

(2) We will document and begin evaluation of cases in the TOMS historical database for comparison with AVHRR and GOES data, focusing on separation of ash, gas and aerosols (however, as future significant eruptions occur, we will prioritize our attention towards these to take advantage of the IR capabilities of MODIS and other sensors). The results from these quantitative comparisons and 3-D analyses with wind trajectory models will provide vital information for the TOMS algorithm development by the TOMS SO₂ group, led by Dr. A. Krueger (e.g., Krueger et al., 1995; Gurevich and Krueger, 1997; Krotkov et al., 1997).

This research will have a significant long-term impact because it provides quantitative information on volcanic cloud behavior, and documents the critical vertical component of volcanic cloud dynamics (e.g., Schoeberl et al., 1993). Current remote sensing methods require assumptions on cloud height, thickness, and geometry, and ash retrievals in particular are often difficult due to meteorological and background interferences (Krueger et al., 1995). This study improves on earlier work (Krotkov et al., 1997; 1999a), providing a better means to study detailed cloud characteristics by reducing geometric assumptions.

(3) We will provide quantitative information on volcanic plume behavior over their lifetime, from SO₂ injection, conversion to aerosol, and removal from the atmosphere. Each year volcanic eruptions inject significant but widely varying amounts of ash and gases into the atmosphere. The conversion of sulfur dioxide to sulfate aerosols impacts atmospheric chemistry and global climate. While volcanoes are the largest contributor to stratospheric sulfate aerosols, their potential for producing tropospheric aerosols is less well known. This study will provide better information on the conversion of sulfur dioxide gas to sulfate and the fate of sulfate in the atmosphere, through the combined use of UV and IR data to develop empirical mass-time relationships. Combining meteorological with remote sensing data allows the development of 3-dimensional perspectives to quantify processes of sedimentation, aggregation, and scavenging in the atmosphere, and the effects of these processes and background conditions on satellite sensor retrievals.

c. Background

TOMS-based studies. Ultraviolet (UV) remote sensing of volcanic clouds began with the discovery that volcanic sulfur dioxide could be mapped with TOMS (Krueger, 1983). Previous estimates of volcanic SO₂ emissions were based on extrapolation of emission rates measured at the vent, or on petrologic measurements of volatile phases in melt inclusions and matrix glass. The TOMS instrument allowed a measure of the mass of SO₂ in a widely dispersed and violently erupted cloud, which could then be used to help validate estimates of sulfur release from volcanoes. The Nimbus TOMS offered a nadir resolution of 50 km and contiguous sunlit coverage, whereas the current Earth Probe TOMS has a slightly better (39 km) nadir resolution at the cost of small gaps in equatorial coverage (McPeters et al., 1993; 1998).

Since that time, TOMS SO₂ scientific studies have made a number of valuable contributions to the volcano-atmosphere community (Table 1). While the initial uses for TOMS focused on individual eruptions, it was soon recognized that the growing TOMS database, and satellite perspective, provided an ideal application towards estimating global emissions by explosive eruptions (Bluth et al., 1993). Schnetzler et al. (1997) related the SO₂ emissions from a variety of eruptions to the Volcano Explosivity Index, and proposed a more robust index to include sulfur dioxide.

Collaboration with researchers at Michigan Tech produced essentially the first merged satellite studies of volcanic clouds (Rose et al., 1995). Prior work had largely focused on individual sensors and measurements, but this partnership allowed us to simultaneously look at ash and gas, and facilitated many future studies on the interactions of species within clouds. At this time, validation of TOMS became increasingly important (e.g., Krueger et al., 1995), as the data were becoming relied upon more and more for chemical and physical models in the atmosphere, as well as for studies of magmatic systems.

Schoeberl et al. (1992; 1993) first developed an isentropic wind trajectory model to take advantage of tracer (i.e., SO₂ clouds) positions located with the TOMS sensor, to study the composition of the Antarctic polar vortex. Bluth et al. (1995) used the trajectory model to help identify cloud altitudes of the Mt. Spurr SO₂ clouds, and with an M.S. student (Shannon, 1996) develop the first 3-D view of a drifting volcanic gas cloud. Operational and experimental models being applied specifically to the prediction of volcanic cloud movements include the Canadian Emergency Response Model (CANERM) (Pudykiewicz, 1989; D'Amours, 1994), NOAA's Volcanic Ash Forecast Transport and Dispersion (VAFTAD) model (Heffter and Stunder, 1993), the Alaskan Geophysical Institute's PUFF ash-tracking model (Searcy et al., 1998) and the German-based ATHAM plume simulation model (Oberhuber et al., 1998). However, the Goddard/Schoeberl model remains a straightforward means of determining cloud altitudes, used in conjunction with inputs of cloud positions derived from satellite sensors.

More recently TOMS UV data have been used to map and characterize ash particles and sulfate aerosols in volcanic clouds (Seftor et al., 1997; Krotkov et al., 1997). Ash and other absorbing aerosols can be detected by contrast with Rayleigh scattering at UV wavelengths as measured with an Aerosol Index (AI) parameter in TOMS data. The AI is near zero for water clouds and cloud-free areas, increasingly positive for absorbing aerosols and ash, while sulfate aerosols produce a relatively small, negative signal. These innovations have opened up many new volcanic applications of TOMS data, particularly in the ability to compare some of the retrievals directly with AVHRR and GOES infrared data (Krotkov et al., 1999a; Mayberry et al., in press). The UV ash detection method is very robust, working for opaque, very fresh eruption clouds, over cold meteorological clouds, and in humid conditions, all of which cause difficulties with the IR retrieval methods (Krotkov et al., 1997). However, only the IR method works at night, and the finer spatial resolution of IR sensors allows more detailed retrievals as well as applications to smaller plumes.

TOMS reflectance data can also be used in some cases to retrieve quantitative information about volcanic ash. For a known (or assumed) underlying reflectivity, the observed and calculated reflectances can be matched in two UV channels to derive both ash optical depth and particle size (Krotkov et al., 1997). Although developed for cloud-free scenes, the method also appears to work for thick underlying clouds with reflectance < 30%. The effects of sulfate aerosol on the ash retrievals are expected to be significant but have not yet been quantified. Other factors that influence the mass estimations include particle nonsphericity (Krotkov et al., 1999b) and cloud altitude.

Table 1. Synopsis of TOMS Volcano Science from NASA/Michigan Tech

Infrared techniques. Infrared (IR) satellite tools for volcanic cloud detection were developed using data from the Advanced Very High Resolution Radiometer (AVHRR), which has a resolution of 1.1 km (Prata, 1989a; 1989b). The infrared method contrasts a pair of thermal bands (approximately 11 and 12 mm); the characteristic absorption of silicate particles at these two bands allows discrimination from water droplets, or meteorological clouds. Wen and Rose (1994) developed a means to estimate particle sizes, and thus estimate the total mass in the clouds. Unlike the UV method, IR retrievals rely on measuring emitted thermal radiation passing through the volcanic cloud. However, detection of ash clouds with this method are often hindered by cloud opacity at IR wavelengths, interferences by water vapor, and complex background emissivities (e.g., Wen and Rose, 1994; Schneider et al., 1999). Mass estimates are also sensitive to the assumed size distribution, cloud top and base temperatures, and cloud thickness; all of these issues require testing in order to begin to document analytical uncertainties.

The two-band IR technique has recently been adapted for use with the Geostationary Operational Environmental System (GOES) to study the ongoing activity of Montserrat (Davies and Rose, 1998; Mayberry et al., in press) - its spatial (4 km at subpoint) and temporal resolution (images every 1/2 hour) provides much more information on cloud dynamics than is possible with polar orbiting sensors (repeat times as little as several hours for AVHRR because multiple sensors are currently operational, but 24 hours for the single TOMS sensor). We have completed work on a new method for making atmospheric corrections in two-band IR data to allow more accurate ash retrievals for small tropical eruption clouds (Yu, 2000) and have also completed the first studies of the shape effects of volcanic ash on retrievals (Krotkov et al., 1999b). Yu and Rose (2000) used HIRS/2 data in 6 infrared bands to retrieval optical depths and sizes for mixtures of sulfate and silicate particles in 1-4 day old volcanic clouds of El Chichon. The model described in the paper was designed to be directly applied to other detectors, such as MODIS.

MODIS. The MODIS (Moderate Resolution Imaging Spectroradiometer) instrument, on board the EOS Terra satellite, employs a wide range of bands applicable for remote sensing of volcanic emissions. This nadir-looking sensor has a spatial resolution of 1 km, and with a swath width of 2330 km, covers the Earth approximately every two days (Mouginis-Mark and Doumergue-Schmidt, 2000).

Table 2. Comparison of MODIS and AVHRR parameters

In the last decade several retrieval algorithms have been developed to utilize thermal infrared detectors observing volcanic clouds. These include detection of sulfur dioxide using an absorption feature at 8.2 - 9.2 mm (Realmuto et al., 1997), the split window (10.3 - 12.5 mm) technique for detecting the presence of silicate ash (Wen and Rose, 1994) and a modified multi-band retrieval of sulfate aerosol (Yu, 2000; Yu and Rose, 2000) currently being tested. However, most volcanic plumes detectable with current satellite-borne sensors contain species of all of the above types, which interact with each other both chemically and radiatively. For the first time, a satellite-based sensor, MODIS, has the capability to provide data at all the wavelengths necessary for a simultaneous retrieval of information about the volcanic plumes and clouds using all of the above techniques.

Merged studies. Merging UV and IR datasets of the same volcanic cloud provides additional insights than can be retrieved from any single sensor. As UV (TOMS) and IR-capable (MODIS, AVHRR, GOES) instruments reside on different platforms, near-simultaneous observations of the same cloud are uncommon, but each case study provides glimpses into the physical and chemical processes occurring as the volcanic clouds mix with the atmosphere. For example, the 1994 eruption clouds of Rabaul were examined with TOMS and AVHRR, and we found that the co-erupted water coated ash particles and froze in the rising plume, scavenging ash and gas, and masking the particles from standard satellite detection schemes (Rose et al., 1995). Other studies include the ash-gas separation in the 1982 El Chichon volcanic clouds (Schneider et al., 1999); the south polar-circling clouds of the 1991 Hudson eruption (Constantine et al., 2000); rapid species removal from Lascar, 1993 clouds (Shocker, 1996); and 1997 Montserrat clouds generated from the interaction of a pyroclastic flow interacting with seawater (Mayberry et al., in press). These detailed studies of highly varied, evolving volcanic clouds have led to the publication of a summary paper (Rose et al., 2000a) which integrates all the observations.

A near-ideal comparison of UV and IR ash cloud retrievals was made using the NOAA-12 AVHRR brightness temperature difference and the Nimbus-7 TOMS AI, which had simultaneously detected the ash cloud about 18 hours after the August 18, 1992 eruption of Mount Spurr, Alaska (Krotkov et al., 1999a). Excellent agreement was found in the spatial extent of the ash cloud derived at UV and IR wavelengths. Although the spatial resolution was ~1 km for AVHRR data and ~50 km for TOMS data, the main patterns of the ash spatial distribution can be clearly seen in both.

d. Ongoing Related Work

Hekla case study. Hekla volcano in Iceland began activity on February 26, 2000, shortly after 1800 UT. Although the eruption continued for several days, an explosive phase occurred in the first few hours which generated a volcanic plume that rose to a height of 11-13 km, into the lower stratosphere. This plume detached to form a drifting cloud which could be tracked for at least two more days. We have amassed a set of IR and UV satellite data for this eruption (Table 3), which will be used for the proposed study.

Table 3. Satellite Datasets for Study of the 2000 Hekla Eruption

From preliminary study (Rose et al., 2000b) the eruption appears to represent venting of an ash-poor cloud, perhaps largely composed of volcanogenic sulfur dioxide gas and sulfate aerosol. The initial AVHRR and TOMS AI retrievals strongly indicate non-absorbing aerosols, which is indicative of an ash-poor, ice-rich cloud (Figure 1). This type of volcano-atmosphere interaction has not been observed, much less studied, with satellite data. The data set includes at least six MODIS images, which will allow the first applications of greatly improved retrievals on ice, water vapor, and sulfur dioxide (Figure 2).

At 1514 UT on February 28 a NASA research aircraft (on the SOLVE field experiment), with a variety of advanced sampling equipment on board, unintentionally encountered this volcanic cloud. This has resulted in an important validation opportunity for remote sensing, because we can not only compare independent satellite data and trajectory model results (Figure 3), but we also have direct sampling data from the aircraft transects. Studies of the dynamic process of atmospheric injection are vital if we are to understand the potentially large effects of large eruptions on the Earth's atmosphere. The remote sensing approach, applied to eruptions of smaller scale, is very likely to illuminate the process.

PDC investigation of aerosol formation and mapping. The Pacific Disaster Center (PDC) is a new facility on the island of Maui, Hawaii devoted to the mitigation of hazards in the Pacific region. In this product-oriented project we are focusing on the prediction and mapping of low-level sulfate aerosol plumes (vog) generated by the Pu'u O'o vent of Kilauea volcano. Vog results from the conversion of sulfur dioxide gas emitted from Kilauea into an aerosol of sulfuric acid droplets and particles. Our goal is to provide the PDC with an operational version of an integrated sulfur dioxide and sulfate mapping software package, as well as instructions for the use of these tools. The tools will allow the PDC to track the formation and transport of vog, and thus predict the arrival and potential impact of the vog on population centers.

Most relevant to this proposal is that we have utilized and developed techniques to map tropospheric volcanogenic SO₂ plumes (Realmuto et al., 1997) and sulfate aerosol clouds (Yu and Rose, 2000) with thermal infrared remote sensing, particularly MODIS. We are presently working on a forward model that calculates optical depths an any wavelength given a particle size distribution. We have adapted the radiative transfer model of Prata and Grant (in press) to the problem of low altitude aerosols, and developed routines to call MODTRAN (a required input to several retrieval schemes) to radiatively correct for the presence of a given atmosphere.

EOS Volcanology Team. We are completing the final year of our participation in NASA's EOS Volcanology Team project, which comprises an interdisciplinary effort since 1992 to develop operational algorithms in preparation for the launch of the EOS Terra satellite platform. The Michigan Tech role (W. Rose, PI; G. Bluth Co-I) has been to investigate the spectral, thermal and physical characteristics of volcanic eruption plumes and drifting clouds. The EOS project involves 25 scientists from around the world, and is headquartered at the University of Hawaii (P. Mouginis-Mark, Project Director). The Team is a valuable resource for this proposed project as it includes experts in volcanology, remote sensing, meteorology and atmospheric physics and chemistry.

The EOS project has helped initiate and sustain a number of the data merging studies mentioned in this proposal. However, the main goals of the project pertinent to this proposal are the development of an integrated methodology for the retrieval of sulfur dioxide (adapted from the method of Realmuto et al., 1997), sulfate (from Yu, 2000; Yu and Rose, 2000), and silicate ash (after Wen and Rose, 1994). Although the codes are designed around the MODIS instrument, they retain their capabilities for AVHRR and GOES data.

We have learned how to ingest MODIS data to produce images for Hekla (Table 3; Figure 2) and Etna, and converted the raw data into radiances and brightness temperatures as precursors to the two-band retrieval method of Wen and Rose (1994). The next steps will be to produce sulfur dioxide maps and add an atmospheric correction to our Brightness Temperature Difference code using MODTRAN. We plan to routinely acquire MODIS data for future eruptions, and have purchased additional storage capabilities for this, and already have data from Etna, Hawaii, Lascar (Chile), and Miyake-jima (Japan).

Synthesized volcanic cloud studies. In the past few years, our research group has made significant progress in studies of volcano-atmosphere interactions, through a combination of remote sensing, laboratory, and theoretical modeling studies. Complete conversion of SO₂ to sulfate in the stratosphere occurs at an e-folding rate of about 120 days. SO₂ loss from stratospheric volcanic clouds occurs at an e-folding rate of about 35 days, and the SO₂ loss rate for volcanic clouds which only barely reach the stratosphere is much more rapid, on the order of a few days (Bluth et al., 1992; 1997; Barker et al., 1998; Oppenheimer et al., 1998). The latter limits the stratospheric aerosol buildup from smaller eruptions (Bluth et al., 1997) and removal processes are most effective in wet eruptions where ice nucleates and grows on ash particles (Rose et al., 1995).

We can now map and measure both ash and SO₂ using both infrared (Wen and Rose, 1994; Realmuto et al., 1997) and ultraviolet data (Krueger et al., 1995; Krotkov et al., 1997). The infrared retrieval scheme now includes atmospheric corrections (Yu, 2000) and this enables us to look at smaller volcanic clouds (Rose and Mayberry, 2000). We have developed a multispectral IR technique which retrieves information on sulfate aerosols as well as ash (Yu and Rose, 2000). Our initial work with modeling volcanic clouds (Riley et al., 2000; Guo et al., 2000) has helped us evaluate theories of volcanic cloud evolution. Laboratory measurements relating ash particle geometry to fallout characteristics (Riley Ph.D. thesis in progress) have helped us evaluate the effects on radiative transfer models (Krotkov et al., 1999b). Laboratory work on adsorption on fine volcanic ash (Gu et al., 1999) has helped quantify rates of uptake and release of sulfur dioxide gas under a range of atmospheric conditions.

e. Project Work Plan

We consider the project as composed of the three main objectives outlined earlier: (1) complete cross-comparisons of TOMS retrievals with MODIS (and AVHRR and GOES as applicable); (2) study cases of ash/gas separation in the TOMS 20-year database, in order to develop a suite of conditions composed of isolated regions as well as mixtures of ash, gas and aerosol species, and environmental conditions; (3) study the chemical reaction rates of the sulfur dioxide and sulfate aerosols under the wide range of conditions produced from tropical to near-polar regions. Table 4 provides a summary of the work plan, arranged by year, topic and key investigators.

Table 4. Summary of Work Planned*

*Note: 2 Ph.D. students will be directly involved with the Bluth, Rose and Watson tasks. Students are typically started with case study work to learn the retrieval schemes, data and image processing, before developing their doctoral research topics.

(1) TOMS - MODIS comparisons. We have assembled a suite of satellite data for the February eruption of Iceland's Hekla volcano (Table 3), and have begun to perform basic comparisons of retrievals (Figures 1 and 2). With high-latitude eruptions, we typically benefit from orbit overlap, which gives us the opportunity to retrieve additional views of portions of the drifting cloud from sequential orbits, spaced approximately 100 minutes apart. Note several opportunities for near-simultaneous analyses among the three sensors (Table 3).

Our research aims are to develop a budget of the various volcanic cloud components, with time, during the 2-3 days of initial residence based on remote sensing. We will make quantitative retrievals of the silicate ash, sulfur dioxide gas, and sulfate aerosols by the respective sensors. These will be used to help cross-validate the retrieval methods on a case-by-case basis. The ultimate objective of these studies is to use the quantitative comparisons to develop an understanding of processes and reactions in young volcanic clouds and in particular to focus on the role of ice, and ice/sulfur interactions.

(2) Ash/gas separation. The primary objective of this task is to identify eruption clouds which can be tracked by UV and IR sensors for at least several days. The gas, ash and aerosol species in these cloud will be examined in order to observe their behavior in terms of mean particle radii, mass decay rates and 3-dimensional positions. A number of potential cases are proposed below, based on a preliminary study described by Schneider et al. (2000). Case studies are needed to examine details of particular geographic regions, atmospheric conditions, or volcanological inputs. We plan on performing retrievals for any future eruptions that may occur during the proposal period, but in the event that few or no eruptions occur, we have prepared an extensive set of eruptions for study from the TOMS database.

Out of the hundreds of explosive eruptions that have occurred over the past 20 years, the initial eruptions targeted for study include those where separation is known to have occurred: Mt. St. Helens, U.S. 1980; Galunggung, Indonesia 1982; El Chichon, Mexico 1982; Pinatubo, Philippines 1991; Hudson, Chile 1991; Lascar, Chile 1993; Montserrat, 1997; and Shishaldin, U.S. 1999. These eruptions represent a wide range of volcanological settings, injection altitudes, masses, and climatology. Conversely, there are cases where separation does not occur (e.g., Spurr, U.S. 1992) and these eruptions will also be examined.

As with the volcanic cloud data, there are potentially large amounts of meteorological data which need to be carefully examined for availability during the time and locations of interest, and for reliability. Radiosonde data are available from meteorological stations (e.g., airports) around the world. They provide an important means of studying the structure of the atmosphere, such as defining the tropopause, and as a way to analyze volcanic cloud movement. For each eruption case study, radiosonde data (temperature, wind speeds and directions) will be retrieved from surrounding stations in order to determine localized atmospheric conditions and tropopause heights. Radiosonde and windfield data access from Goddard archives will be supervised by co-PI Lait.

As each data set is developed, it will be necessary to compare data sets to find the best fit for each case study. For example, Allen et al. (1999) compared United Kingdom Meteorological Office and National Centers for Meteorological Prediction wind data, and found significant differences in matching trajectories to actual cloud positions - particularly for parcels near the tropopause.

The wind data and cloud positions from satellite data will be used to drive a trajectory model, in order to derive the most likely three-dimension positions of the volcanic ash, aerosol and gas clouds. The Goddard-based trajectory model developed by Schoeberl et al. (1992; 1993) will be used to reconstruct the atmospheric transport pattern of the volcanic clouds. The modeling method we have used successfully involves matching cloud positions (using TOMS, or IR-based pixels) to wind trajectories in order to calculate the most likely vertical positions as functions of wind speed and directions (e.g., Shannon, 1996; Constantine et al., 2000; Mayberry et al., in press). The trajectory simulations will be run for each day an identifiable cloudmass can be discerned in the satellite data.

The final product will be a comprehensive database of eruptions, with the various components and geographic separation documented. This revised TOMS database will be used to help identify and organize case studies for TOMS AI retrieval studies (e.g., Krotkov et al., 1999a).

(3) Sulfate and sulfur reactions. The current volcanic cloud tracking techniques have been rather successful at detecting clouds and calculating particle size, optical depth, cloud mass and area. However, each method suffers from large uncertainties and retrieval capabilities vary with environmental conditions. Efforts to predict cloud motion have shown some success, but currently processes within the cloud can only be theorized. These processes, which include gas-particle interactions, sedimentation, and aggregation, are currently unconstrained by any type of quantitative datasets. The analyses we will conduct for this project will reduce uncertainties in the retrievals, and document dynamic processes within the clouds.

Cloud parameters of particle types, sizes, sulfur dioxide, ash and sulfate aerosol concentrations will be retrieved for each cloud pixel (the pixel size being sensor dependent). Retrievals for sulfur dioxide and ash particles will be performed, and data from these will be used to calculate optical depths, particle sizes and masses (after Krotkov et al., 1999a). Infrared retrievals will include MODIS, AVHRR and GOES as appropriate.

With the pixel locations calculated in three dimensions, and pixel components for the species calculated, the UV and IR data can be analyzed in timewise fashion to document and compare trends in particle size and concentration as the clouds drift through the atmosphere. Size/mass trends will be used to study processes of sedimentation and aggregation and details of species removal rates.

The longevity of volcanogenic sulfur dioxide in the stratosphere is often assumed be be approximately 30-40 days e-folding time, yet this is on the basis of only two stratospheric-level eruptions, El Chichon and Pinatubo. Sulfate removal is assumed to be dominated by gravitational sedimentation; however this removal rate is also dependent on the amount of water vapor (season, latitude) and the size of the eruption. Bluth et al.'s (1997) literature survey suggested that the largest eruptions demonstrated removal rates on the order of 12 months, while smaller eruptions (e.g., Mt. St. Helens, Alaid) yielded removal rates on the order of 6 months. Other studies have also suggested that the SO₂ removal rate is related to cloud mass and altitude (Oppenheimer et al., 1998, Barker et al., 1998). As the processes involved are complex combinations of meteorological conditions as well as eruption-dependent factors, it is necessary to acquire a large set of case studies from which to examine volcano-atmosphere sulfur cycling.

The main tasks will involve documenting the decay rates of sulfur dioxide with respect to meteorological and volcanological conditions, using the TOMS data. When available, the formation and subsequent decay of sulfate aerosol will be retrieved through the use of MODIS data.

References:

Allen, D.R., M.R. Schoeberl and J.R. Herman (1999) Trajectory modeling of aerosol clouds observed by TOMS. Journal of Geophysical Research, 104, 27,461-27,471.

Barker, S.L., G.J.S. Bluth and A.J. Krueger (1998) Examining the removal rate of volcanogenic SO2 in the atmosphere. Eos Transactions AGU, 79, F935.

Bluth, G.J.S., C.C. Schnetzler, A.J. Krueger, and L.S. Walter (1993) The contribution of explosive volcanism to global atmospheric sulphur dioxide concentrations. Nature, 366, 327-329.

Bluth, G.J.S., W.I. Rose, I.E. Sprod, and A.J. Krueger (1997) Stratospheric loading from explosive volcanic eruptions. Journal of Geology, 105, 671-683.

Bluth, G.J.S., and C.M. Oppenheimer (2000) Twenty years of TOMS observations of sulfur emission from Nyamuragira Volcano, Zaire (invited). Fall AGU Meeting.

Constantine, E.K., G.J.S. Bluth, and W.I. Rose, (2000) TOMS and AVHRR observations of drifting volcanic clouds from the August 1991 eruptions of Cerro Hudson. Remote Sensing of Active Volcanism, Geophysical Monograph 116, 45-64.

D'Amours, R. (1994) Current and future capabilities in forecasting the trajectories, transport, and dispersion of volcanic ash clouds at the Canadian Meteorological Centre, U.S. Geological Survey Bulletin 2047, 325-331.

Gu, Y., J. Gierke, G. Bluth and W. Rose (1999) A laboratory study of SO₂ adsorption on Mt. St. Helen's volcanic ashes. Eos Transactions AGU, December 1999.

Guo, S., W.I. Rose, G.J.S. Bluth, C. Textor, H-F. Graf (2000) ATHAM model simulation of the 15 June 1991 Pinatubo volcanic plume. Fall AGU Meeting.

Gurevich G. and A. Krueger (1997) Optimization of TOMS wavelength channels for ozone and sulfur dioxide retrievals. Geophysical Research Letters, v. 24, 2187-2190.

Heffter, J.L., and B.J. Stunder (1993) Volcanic ash forecast and dispersion (vaftad) model. Weather and Forecasting, 8, 533-541.

Krotkov, N.A., A.J. Krueger and P.K. Bhartia (1997) Ultraviolet model of volcanic clouds for remote sensing of ash and sulfur dioxide. Journal of Geophysical Research, v. 102, 21,891-21,904.

Krotkov, N.A., O. Torres, C. Seftor, A.J. Krueger, W. Rose, A. Kostinski, G. Bluth, D. Schneider, and S.J. Schaefer (1999a) Comparison of TOMS and AVHRR volcanic ash retrievals from the August 1992 eruption of Mt. Spurr. Geophysical Research Letters, 26, 455-458.

Krotkov, N.A., D.E. Flittner, A.J. Krueger, A. Kostinski, C. Riley, W. Rose, and O. Torres (1999b) Effect of particle non-sphericity on satellite monitoring of drifting volcanic ash clouds. Journal of Quantitative Spectroscopy & Radiative Transfer, 63, 613-630.

Krueger, A. (1983) Sighting of El Chichon sulfur dioxide clouds with the Nimbus 7 Total Ozone Mapping Spectrometer. Science, v. 220, 1377-1378.

Krueger, A.J., L.S. Walter, P.K. Bhartia, C.C. Schnetzler, N.A. Krotkov, I. Sprod, and G.J.S. Bluth (1995) Volcanic sulfur dioxide measurements from the Total Ozone Mapping Spectrometer (TOMS) instruments. Journal of Geophysical Research, 100, 14,057-14,076.

Krueger, A.J., S. Schaefer, N. Krotkov, G. Bluth, and S. Barker (2000) Ultraviolet remote sensing of volcanic emissions and applications to aviation hazard mitigation. Remote Sensing of Active Volcanism, Geophysical Monograph 116, 25-43.

Mayberry, G.C., W.I. Rose and G.J.S. Bluth (in press) Dynamics of the volcanic and meteorological clouds produced by the December 26, 1997 eruption of Soufriere Hills volcano, Montserrat, W.I. Special Publication of the Geological Society of London.

McCormick, M.P., L.W. Thomason, and C.R. Trepte (1995) Atmospheric effects of the Mt. Pinatubo eruption. Nature, 373, 399-404.

McPeters, R.D. et al. (1993) Nimbus-7 total ozone mapping spectrometer (TOMS) data products user's guide. NASA Reference Publication 1323.

McPeters, R.D. et al. (1998) Earth Probe Total Ozone Mapping Spectrometer (TOMS) data product user's guide. NASA/TP-1998-206895.

Mouginis-Mark, P. and Doumergue-Schmidt (2000) Acquisition of satellite data for volcano studies. Remote Sensing of Active Volcanism, Geophysical Monograph 116, 9-24.

Oberhuber, J.M., M. Herzog, H.-F. Graf, and K. Schwanke (1998) Volcanic plume simulation on large scales. Journal of Volcanology and Geothermal Research, 87, 29-53.

Prata, A. (1989a) Infrared radiative transfer calculations for volcanic ash clouds. Geophysical Research Letters, v. 16, 1293-1296.

Prata, A. (1989b) Observations of volcanic ash clouds in the 10-12 um window using AVHRR/2 data. International Journal of Remote Sensing, v. 10, 751-761.

Prata, A.J. And I.F. Grant (in press) Retrieval of microphysical and morphological properties of volcanic ash plumes form satellite data: Application to Mt Ruapehu, New Zealand. Submitted to Quarterly Journal of the Royal Meteorological Society.

Pudykiewicz, J. (1989) Simulation of the Chernobyl dispersion with a 3-D hemispheric tracer model. Tellus, 41B, 391-412.

Realmuto, V.J., A.J. Sutton and T. Elias (1997) Multispectral thermal infrared mapping of sulfur dioxide plumes: A case study from the East Rift Zone of Kilauea Volcano, Hawaii. Journal of Geophysical Research, 102, 15,057-15,072.

Riley, C.M., W.I. Rose, C. Textor, H-F. Graf, G.J.S. Bluth and R.G. McGimsey (2000) Modelling particle aggregation and microphysical processes in the August 1992 Spurr eruption plume using ATHAM. Fall AGU Meeting.

Rose, W. and G. Mayberry (in press) Use of GOES thermal infrared imagery for eruption scale measurements, Soufriere Hills, Montserrat. Geophysical Research Letters.

Rose, W.I., G.J.S. Bluth and G.G. J. Ernst (2000a) Integrating retrievals of volcanic cloud characteristics from satellite remote sensors: a summary. Philosophical Transactions of the Royal Society of London, v. 358, 1585-1606.

Rose, W.I., G. Bluth, C. Riley, M. Watson, T. Yu, and G.G. Ernst (2000b) Potential mitigation of volcanic cloud hazards using satellite data: a case study of the February 2000 Hekla event and an unexpected NASA DC8 encounter. Fall AGU Meeting.

Schneider, D.J., W.I. Rose, L.R. Coke, G.J.S. Bluth, I.E. Sprod and A.J. Krueger (1999) Early Evolution of a stratospheric volcanic eruption cloud as observed with TOMS and AVHRR, Journal of Geophysical Research, 104, 4037-4050.

Schneider, D.J., W.I. Rose, and G.J.S. Bluth (2000) Satellite observations of the separation of volcanic ash and sulfur dioxide in the atmosphere. Fall AGU Meeting.

Schnetzler, C.C., G.J.S. Bluth, A.J. Krueger, and L.S. Walter (1997) A proposed volcanic sulfur dioxide index (VSI). Journal of Geophysical Research, 102, 20,087-20,092.

Schoeberl, M.R., Lait, L.R., Newman, P.A., and Rosenfield, J.E. (1992) The structure of the polar vortex. Journal of Geophysical Research, 97, 7859-7882.

Schoeberl, M.R., Doiron, S.D., Lait, L.R., Newman, P.A., and Krueger, A.J. (1993) A simulation of the Cerro Hudson SO2 cloud. Journal of Geophysical Research, 98, 2949-2955.

Searcy, C., K. Dean, and W. Stringer (1998) PUFF: A high-resolution volcanic ash tracking model. Journal of Volcanology and Geothermal Research, 80, 1-16.

Seftor, C.J., N.C. Hsu, J.R. Herman, P.K. Bhartia, O. Torres, W.I. Rose, D.J. Schneider, and N. Krotkov (1997) Detection of volcanic ash clouds from Nimbus 7/total ozone mapping spectrometer. Journal of Geophysical Research, v. 102, 16,749-16,759.

Shannon, J.M. (1996) 3-D reconstruction of the Mt. Spurr volcanic clouds using AVHRR, TOMS and wind trajectory data. M.S. Thesis, Michigan Technological University.

Shocker, H.L. (1996) Evolution of volcanic clouds from the 1993 Lascar, Chile eruption. M.S. Thesis, Michigan Technological University.

Wen, S. and Rose (1994) Retrieval of sizes and total masses of particles in volcanic clouds using AVHRR bands 4 and 5.

Yu, T. and Rose (2000) Retrieval of sulfate and silicate ash masses in young (1 to 4 days old) eruption clouds using multiband infrared HIRS/2 data. Remote Sensing of Active Volcanism, Geophysical Monograph 116, 87-100.

Yu, T. (2000) Improved algorithms for volcanic clouds using multispectral infrared remote sensing. Ph.D. Thesis, Michigan Technological University.