year 2 progress report and plans for year 3pmm/ids_team_report_yr2.pdfyear 2 progress report:...

Year 2 Progress Report: NAG5-10640

Effects of Volcanoes on the Natural EnvironmentGrant NAG5-10640

Year 2 Progress Report and Plans for Year 3

ii. TABLE OF CONTENTS

PAGE No.

i. Cover Pages i

ii. Table of Contents 1

iii Team Members 2

1. Overview 3

2. Publications and Presentations in Year 2 5

3. Field Activities 6

4. Summary of Science Investigationsa. Thermal Studies 7b. Gravity 7c. Gas Studies 8d. Thermal IR Plume Studies 9e. Aerosol Studies 11f. Vegetation Studies 12g. Plume Models 13h. Soil Science 14

5. References 15

6. Figures 16

7. Plans for Year 3 29


Effects of Volcanoes on the Natural Environment

iii. Team Members

Peter Mouginis-Mark (Univ. Hawaii): Team Leader

Steve Businger (University of Hawaii)Luke Flynn (University of Hawaii)Nancy Dise (Open University, England)Lori Glaze (Proxemy Research, Inc.)Andrew Harris (University of Hawaii)Keith Horton (University of Hawaii)Clive Oppenheimer (Cambridge University, England)John Porter (Univ. Hawaii)Fred Prata (CSIRO, Australia)Vince Realmuto (Jet Propulsion Laboratory)David Rothery (Open University, England): United Kingdom LeaderHazel Rymer (Open University, England)Glyn Williams-Jones (University of Hawaii)

Collaborators:Tamar Elias (Hawaii Volcano Observatory, USGS)Jim Kauahikaua (Hawaii Volcano Observatory, USGS)Emanuele Marchetti (Universita’ di Camerino, Italy)Jeff Sutton (Hawaii Volcano Observatory, USGS)


Effects of Volcanoes on the Natural EnvironmentYear 2 Progress Report and Plans for Year 3 (April 2003 - March 2004)

1. OverviewThis Progress Report covers Year Two of our 3-year investigation, which nominally

extends from April 1, 2002 to March 31, 2003. By virtue of the reporting period, theresults described here actually reflect our activities from January 1, 2002 to December31, 2002.

Through their persistent activity, basaltic volcanoes such as Kilauea (Hawaii) andMasaya (Nicaragua) contribute significant amounts of sulfur dioxide and other gases tothe lower troposphere. Although primarily of local rather than regional impact, thecontinuous nature of these eruptions means that they can have a major impact on thetropospheric atmosphere for years or decades. Since mid-1986, Kilauea has emittedabout 2,000 tonnes per day, while between 1995 and 2000 Masaya has emitted about1,000 - 1,500 tonnes per day (Duffell et al., 2001; Delmelle et al., 2002; Sutton and Elias,2002; Williams-Jones et al., 2002). These emissions have a significant effect on the localenvironment. The volcanic smog (“vog”) that is produced affects the health of localresidents, impacts the local ecology via acid rain deposition and the generation of acidicsoils, and is a concern to local air traffic due to reduced visibility.

Until recently, most measurements of sulfur dioxide and the vog plumes have beenmade from the ground or from small aircraft. Our NASA interdisciplinary science (IDS)investigation therefore has as its goal the development of space-based methods forstudying these inputs to the atmosphere, the correlation of these gas emissions with thelevel of activity at the volcano, the dispersal of the gas and aerosol plume, and finally theimpact of the eruption on the surrounding environment. These satellite-based techniquesare developments of our earlier (1989 – 2000) IDS investigation, but require significantadditions both in the field validation of observations of eruptions, and in theincorporation of atmospheric chemistry, meteorology, and ecology studies. Innovativefield measurements using light-weight UV-spectrometers and lidars, as well ashyperspectral imaging and continuous thermal radiometry, are being developed by ourTeam to provide a regional view of gas emission and dispersion. One of our ultimategoals is to have predictive atmospheric dispersion models using real-time MODIS data tocalculate aerosol optical depths that will then provide advance warning (a day or two) forvog plumes that may impact down-wind centers of population such as Honolulu andKona (Kilauea) and El Crucero (Masaya).

Much of our work in Year 2 has focused on the development of field validationtechniques for the monitoring of the thermal properties and gas flux of volcanoes,coupled with the development of remote sensing techniques to study the conversion rateof sulfur dioxide to aerosols. We have also worked on the development of a highresolution 4-D atmospheric transport model to predict the dispersal pattern of a volcanicplume. We have completed three major field experiments during the last 12 months,working on Masaya (Nicaragua) and Poás (Costa Rica) volcanoes in Central America,Stromboli (Italy), and Kilauea (Hawaii). We have collected a year’s worth of MODISdata for Kilauea to study the utility of satellite data to determine the temporal variationsin thermal output. ASTER data have been used to derive total column abundance ofsulfur dioxide at Kilauea. Vegetation spectra were also collected in the field to enable us


to evaluate the effects of volcanic gases on neighboring vegetation. A total of 6 papersand 10 conference presentations, including one at the EOS IWG meeting in Maryland inNovember 2002, were derived from this Year 2 funding (see Section 2).

While our studies are relevant to any persistently active volcano, under directionfrom NASA Headquarters, the volcanoes selected for this effort are:

Kilauea (Hawaii; 19.425°N, 155.292°W): The on-going activity at this basaltic shieldvolcano has enabled many of the algorithms to be developed under our previous NASAEOS IDS activity. Ease of access to the volcano and extensive knowledge of the history,petrology and meteorology of the volcano permits us to use Kilauea as our “fieldlaboratory” where we can develop techniques that can then be taken to the othervolcanoes of interest

Masaya (Nicaragua; 11.984°N, 86.161°W) is a basaltic shield volcano that rises to just635 m above sea level. Several million people live within a 30-km radius of Masaya’scaldera rim and are potentially at risk if a large eruption were to occur. In addition, alarge region downwind from the volcano is adversely affected by the gas plume, makingit an ideal test case for our modeling investigation. Much of the adjacent forest and cropssuch as coffee and citrus plants are killed or severely damaged yearly by the plume andpeople and livestock living in this zone are subject to respiratory problems. Plumeacidity is strong enough to affect buildings and introduce severe lung ailments in childrenwho have to go to school in the plant kill-off zone.

Poás (Costa Rica; 10.20°N, 84.233°W) is a complex andesitic strato-cone in Costa Ricarising to 2700 m above sea level. The volcano has been continuously active throughouthistorical times and for much of this time there has been a warm, acidic crater lake,fumaroles and a ubiquitous shallow hydrothermal system within the summit area. Poássupports a tropical ecosystem; and one of our goals will be to search for spectral changesin foliage as one moves perpendicular to the downwind axis of the plume.

We have created a web site that provides a summary of our recent field and remotesensing activities, including many of our field photos. You can view this material at:

http://www.higp.hawaii.edu/~pmm/IDS.html

Our original project goals included an integrated field, air and space-based analysis offour key aspects of the degassing of these volcanoes:

1. Thermal energy and eruption rates: Here we seek to better understand how tointerpret the thermal output of a volcano, and how this correlates with the gas flux.Thus we have paired our development of a continuously monitoring thermal systemwith the development of a new correlation spectrometer (FLYSPEC; Horton et al.,2002) that we have built in-house. We are studying high-temporal (at about onemeasurement per second) thermal data on the ground (Harris et al., 2002), while atthe same time observing magma production rates using other field techniques (e.g.,infrasonics, gravity and very low frequency radio waves). We have also obtained


satellite data (GOES, Terra MODIS/ASTER, Landsat 7 and EO-1) to study thethermal output of Kilauea and Masaya.

2. Determination of volcanic gas concentration and spatial distribution: Sulfur dioxideis the main volcanic gas of importance here, but we are also trying to developmethods for studying other gases (HCl, HF, CO etc.) using UV and FTIR techniquesin order to better characterize the emissions from a volcano (Porter et al., 2002). Akey objective here is also to develop predictive atmospheric models for the dispersionof the eruption plume using the HY-SPLIT atmospheric model.

3. Aerosol optical depths and particle size distributions: We seek to understand the rateof conversion of the measured sulfur dioxide into sulfate aerosols as they aretransported downwind, and to explore how airborne and spaceborne observations(particularly from sensors such as MODIS) may be used to monitor the plumedispersion.

4. Impact of the plume on the land surface: Concurrent with modeling of plumedispersal, we are studying the impact of plumes on the local environment. This studyis based on the analysis of spectral properties of vegetation located down-wind of theplume and the analysis of soil samples taken from down-wind locations on the BigIsland at distances that range up to ~100 km from the volcano. We have alsocollected spectral data at Masaya for correlation with sulfur deposition plates(collected by Pierre Delmelle) and FLYSPEC measurements. At these test locations,we measured the spectra of grass, bananas, coffee, and ti leaves.

2. Publications and Presentations in Year 2Dise, N.B. and Howell, G.H. 2002. Soil acidification from persistently active volcanoes. Proceedings of the

annual meeting of the U.K. Mineralogical Society Volcanic and Magmatic Studies Group, Universityof Edinburgh 16-18 December 2002.

Flynn L.P., A J Harris, P J Mouginis-Mark, L R Geschwind, S K Rowland, and K A Horton (2002). ThePu’u O’o eruption: Space-borne remote sensing of the evolving lava flow field. EOS Trans. AGU 83(47). Fall 2002 meeting abstract V62C-04.

Geschwind L.R., L P Flynn, A J Harris, S Sahetapy-Engel (2002). Satellite-borne and field-basedhyperspectral measurements of active lava flows at Kilauea volcano. EOS Trans. AGU 83 (47). Fall2002 meeting abstract V71A-1260.

Glaze, L. S. and S. M. Baloga (2002). Volcanic plume heights on Mars: Limits of validity for convectivemodels. J Geophys Res, 107:10.1029/2001JE001830.

Harris, A. J., D J Pirie, K Horton, L P Flynn, H Garbeil, J B Johnson, H Ramm, E Pilger (2002). DUCKS:A continuous thermal presence on the rim of Pu’u O’o. EOS Trans. AGU 83 (47). Fall 2002 meetingabstract V71A-1256.

Horton, K. A., P. J. Mouginis-Mark, H. Garbeil, J. N. Porter, A. J. Sutton, T. Elias, and C. Oppenheimer(2002). FLYSPEC: Expanding options for low-cost instruments to remotely measure volcanic gasemissions. AGU Chapman Conference on Volcanism and the Earth’s Atmosphere, Santorini, Greece.

Horton, K., G. Williams-Jones, H. Garbeil, A. J. Sutton, and T. Elias (2002). FLYSPEC: A new ultravioletcorrelation spectrometer for the detection of volcanic SO2. EOS Trans. AGU 83 (47). Fall 2002meeting abstract V12A-1418.

McConigle, A. J. S. and C. Oppenheimer. Optical sensing of volcanic gas and aerosol emissions. In:Volcanic degassing, Geol. Soc. Sp. Pub., in press.


Mouginis-Mark, P.J. and L.S. Glaze (2002). Effects of Volcanism on the Environment. Presentation to theEOS IWG, Ellicott City, MD, November 20th, 2002.

Mouginis-Mark, P.J., K. Horton, J. N. Porter, A. Harris, L. Flynn, H. Garbeil, D.A. Rothery, H. Rymer, N.Dise, G. Williams-Jones, C. Oppenheimer, V. Realmuto, F. Prata, L. S. Glaze, and S. Businger (2002).Satellite and Field Remote Sensing of Low-Level Degassing of Volcanoes: Preliminary Results fromKilauea and Masaya Volcanoes. AGU Chapman Conference on the Effects of Volcanism on theAtmosphere, Santorini, Greece.

Oppenheimer, C. Volcanic degassing. Treatise on Geochemistry, vol. 3, Chap. 6, in press.Porter, J.N., K. A. Horton, P. J. Mouginis-Mark, B. Lienert, S. K. Sharma, E. Lau, A. J. Sutton, T. Elias,

and C. Oppenheimer (2002). Sun photometer and lidar measurements of the plume from the HawaiiKilauea volcano Pu’u O’o vent: Aerosol flux and SO2 lifetime. Geophys. Res. Lttrs. 29:10.1029/2002GL014744.

Porter, J.N., Obtaining Quantitative Aerosol Extinction Values From An Aircraft Forward Looking Lidar,(submitted to Applied Optics, Dec. 2002).

Williams-Jones, G., L Flynn, A J Harris, B Gibson, P J Mouginis-Mark. The Effects of PersistentlyDegassing Volcanoes on the Natural Environment as Exemplified by Kilauea, Masaya and PoásVolcanoes. EOS Trans. AGU 83 (47). Fall 2002 meeting abstract V21B-1194.

Williams-Jones, G., Delmelle, P., Rymer, H., Rothery, D.A., Stix, J., Harris, A.J.L., Fournier, N.E., St-Amand, K., Strauch, W., Navarro, M., Alvarez, J., Williams, S.N., & Stoiber, R.E. (2002).Mechanisms of magma degassing at an open-system, active basaltic volcano, Masaya, Nicaragua.Bulletin Volcanology, Submitted.

Williams-Jones, G., Rymer, H., & Rothery, D.A. (2002). Gravity changes and passive degassing at theMasaya caldera complex, Nicaragua. J. Volcanol. Geotherm. Res., In press.

3. Field Activities

1) Masaya and Poás, February 2002. Thermal, gas, gravity, and vegetation data werecollected. This was also the first opportunity for some of the U.K. Team to interactwith the US team members.

2) Kona, Hawaii. Volcanic aerosol experiment downwind of the Pu’u O’o plume inorder to investigate aerosol scatter. Lidar, Sun photometer and filter massmeasurements were made.

3) Field experiments on Stromboli volcano, June 2002. Thermal, gas, and infrasonicsmeasurements were made for the five active vents.

4) Kilauea, July, 2002. This was an instrument development experiment to inter-compare methods for measuring gas flux (lidar, COSPEC, Sun photometer, and ourin-house built FLYSPEC instrument). We also measured the spectra of grass,bananas, ti leaves, and coffee plants.

5) Kilauea, November 2002. We conducted an intensive field investigation incollaboration with the U.S. Geological Survey’s Hawaii Volcano Observatory.Thermal, gas, and infrasonics data from the Pu’u O’o vent, open skylights and (forgas studies) downwind of the vent were collected.


4. Summary of Science Investigations:a. Thermal (led by Andy Harris, UH)

The focus of this thermal work to date has been to devise and explore means to recordand track variations in the eruption rate. Continuously monitoring ground-based methodshave been developed (Harris et al., 2002) in order to more fully understand the thermalobservations that are made every 15 minutes by GOES, MODIS, or the “snap shots”obtained once every 16 days by Landsat 7. In this context, we have started a time-seriesanalysis of the MODIS thermal alerts that have been generated over the last year (Figure1), in an attempt to determine how these cycles of activity correlate with the ground data.

A major field experiment was undertaken in early November 2002 that involvedarranging and deployment of thermal, infrasonic, gas, and microgravity experiments onKilauea. The aim of this experiment was to determine whether any cyclic variation, overthe time scale of minutes to hours, could be detected at the active vent (Pu‘u ‘O‘o) andalong the tube system that carries lava to any surface and ocean entry activity. Inpreparation for this experiment, we designed and built three environmentally robust,portable thermal infrared thermometers. These thermometers consist of an Omega 8-14µm thermal infrared thermometer, linked to a data logger housed in a pelican case. Thethermometers view through a selenium-germanium-arsenic window, a design that allowsthe temperature of the targeted vent to be measured while protecting the system fromrain, water vapor and corrosive, acid gases. Preliminary results from this experimentwere reported at the Fall 2002 AGU meeting, but the detailed correlative studies will be amajor effort during Year 3.

b. Gravity (led by David Rothery at the Open University):Our gravity measurements are part of our effort to understand the shallow sub-surface

structure of a volcano, and has the potential to identify when new batches of magma areintruded at depths of a few hundred meters. As such, these gravity data may correlatewell with changes in the thermal output of a volcano, as measured either by MODIS orour field instruments (Investigation #1, above). Under our IDS effort, the microgravitynetwork at Masaya was re-measured during our field trip in February 2002. Gravitychanges have been previously detected there and are believed to be related to changes indegassing rate of the volcano. During our surveys, a continuation of the trend of stable(flat) gravity since 1997 was seen in the immediate vicinity of the active craters (Figure2). Specifically, the gravity anomaly appears to be located over the Nindiri lava lake.This is “flat” gravity coincides with a decreasing/flat trend of SO2 degassing since 1997.Recent activity at Masaya (1993-present) is characterised by these repeated fluctuationsin microgravity and gas flux and is likely to be due the formation and oscillation of a gas-rich vesiculated zone immediately beneath the San Pedro, Nindirí and Santiago craters.Four stages of activity within the most current degassing episode have been identified,with the most recent data suggesting that the system appears to now be in another periodof reduced degassing.

During field work on Masaya in 2002, using a manually-read LaCoste-Romberg G-meter, we observed a 200 microgal gravity excursion over a period of hours whichappears to have been correlated with gas and thermal variations. A gravity change of thismagnitude cannot be explained by shallow changes of magma depth in any likely conduit


below Masaya’s summit, and the change was far too rapid and ephemeral to reflectdeeper changes. We conclude that this was probably an instrumental response to groundvibration. During the 2002 Masaya experiment the instrument proved to be poorlyunsuited for measuring micro-gravity changes over period of hours and days such ascould be caused by magma migration in the conduit because there are too manyuncontrollable drift effects.

We nevertheless remain confident that this is a valuable field technique because usingcontinuously recording gravity meters (LaCoste-Romberg D-meters) worked well atStromboli in June and October 2002 and showed excellent detection of explosions withwaveforms matching the ground displacement from broadband seismometers. We stillhave to analyze the thermal changes that we observed at Masaya at the same time to seeif we can specifically correlate these two dynamic events.

The microgravity network at Poás was re-measured in February 2002. Gravitychanges have been detected there in the past, associated with changes in the style oferuptive activity. A continuation of the trend seen in recent years of increased gravity inthe northern part of the crater bottom and a decrease in the southern part continued. Thisis interpreted in terms of a hydrothermal effect in the north, where the lake has risen andso presumably has the subsurface extension of the system. In the south, close to the domeand active fumarole area, the decrease in gravity is interpreted in terms of a migration ofthe steam-water interface within the hydrothermal system, such that the porous region isunderlain increasingly by superheated steam rather than fluid brine.

A new network of stations was established in 2002 designed to monitor the behaviourof this interface over short periods (days to weeks). The data are still to be processed, butare expected to shed light on the shallow processes of degassing and hydrothermalbuffering at Poás, which will be invaluable to the devlopment of physical models incollaboration with other IDS workers using different observations.

c. Gas Studies (led by Keith Horton, University Hawaii)A key component of our Year 2 work has been the development of a new instrument

(FLYSPEC) for measuring volcanic gases (Horton et al., 2002). The reason why thisdevelopment work has been under-taken is two-fold: first, we are trying to produce smallinstruments that can either be flown on a small airplane or carried easily into the field.The most widely used instrument at the present time is the Correlation Spectrometer (orCOSPEC) which is bulky and uses late 1960’s technology. Key to this effort has beenthe development of FLYSPEC, which we have tested on several occasions at Kilauea inconjunction with a field experiment carried out by Jeff Sutton and Tamar Elias usingHVO’s COSPEC (Figure 3). FLYSPEC was also taken to Masaya, Poás, Vulcano,Villarica and Stromboli in 2002.

Our FLYSPEC instrument is designed to measure the downwelling scattered solarspectral radiance in the UV window between 300-330 nm in which SO2 exhibitscharacteristic spectral absorption features, specifically at 302.1 nm, 304.1 nm, 306.5 nm,308.6 nm, and 310.6 nm, and where atmospheric ozone absorption is sufficiently low andSO2 absorption is sufficiently high to provide adequate signal-to-noise. By use of SO2calibration cells of known concentrations and application of Lambert-Beer’s law, thecolumn abundance of SO2 gas can be measured in ppmm. Scattered solar radiation from


the sky acts as the UV radiance source. Initial field tests indicate that an integration timeof ~1 second provides adequate signal without saturation of the any pixels on the detectorarray. The current version of the acquisition software now incorporates on-the-flycalibration and real-time data reduction and integration of GPS position and time-stampin the data stream. In addition to display and storage of calibrated data in real-time, allraw spectra are simultaneously archived to disk.

During Year 2, we collected a sufficient number of FLYSPEC data sets that werepaired to COSPEC observations that we can demonstrate ~98% correlation with theaccepted community standard. We have tested FLYSPEC in very low emissionconditions (a few ppmm) around Halemaumau Crater, Hawaii, and industrial smokestacks on Oahu. Also, we had a paired experiment with COSPEC on the rim of Pu’u O’owhere the concentration was a few hundred ppmm. In each case we obtained verycomparable results to COSPEC.

By virtue of this strong correlation with COSPEC, and the much lighter weight andlower cost of FLYSPEC, we are now able to plan different types of experiments that will(a) enable us to better refine the flux rate; and (b) help measure the conversion rate tosulfate aerosols. In particular, in Year 3 we plan to deploy at least two (possibly 3)FLYSPEC instruments at different locations downwind of the vent. Lori Glaze is helpingus to refine pattern-recognition software that would enable us to use data from multipleFLYSPECs to determine drift-rate of the plume (see Section 4.g on plume models).

d. Thermal IR Studies of Plumes (led by Vince Realmuto, JPL)

d1. Much progress was made on the SO2 estimation algorithm and the application thatruns this algorithm, as summarized below.

1. Spatially-Variable Optimal Water Vapor and Ozone CorrectionAtmospheric absorption and emission due to water vapor and ozone are among the

biggest obstacles to the detection of volcanic SO2 from space. Water vapor and ozoneabsorption and emission are controlled by atmospheric temperature as well as the gasconcentrations, and the strength of the absorption/emission can vary from pixel to pixelwithin a scene. The effects of water vapor are especially acute in tropical regions, suchas Hawaii and Central America, where the atmosphere is humid and warm.

The new technique is based on the principal that a “spectrum” of ground temperaturesshould show no variation with wavelength. If the emissivity of the ground is known, thenvariations in the ground temperature spectrum can be attributed to atmosphericabsorption and emission. Using radiative transfer modeling, we iterate with differentamounts of ozone and water vapor until we achieve the flattest possible groundtemperature spectrum. The correction scheme is automatically applied to every pixelinvolved in emissivity, ground temperature, or SO2 retrievals. Either the ozone or watervapor corrections can be turned off if such corrections are not necessary. For example,ASTER does not have a channel near the broad ozone absorption feature and the ozonecorrection should be turned off to speed up processing time.

The atmospherically-corrected emissivity and ground temperature spectra are nowautomatically saved to disk files. These data products can be applied to other


investigations, such as the retrieval of ash and aerosol estimates, that are not based on theMODTRAN radiative transfer model.

2. Georeference DataASTER and MODIS data radiance are geo-referenced – the approximate latitude and

longitude of each pixel are determined from spacecraft (Terra and Aqua) ephemeris data.We make use of these data to generate DEM’s for ASTER or MODIS scenes (the DEM’sare required to determine the length of the optical path through the atmosphere) fromstandard topographic data sets, such as GTOPO30, ETOPO5, or GLOBE. The scene-specific DEM’s can be saved to disk, saving the re-generation of a DEM each time anASTER or MODIS scene is analyzed. In addition, pointing a cursor at the pixel andclicking a mouse button can access the geo-reference data for any pixel in a scene. Thiscapability facilitates the comparison of satellite- and ground-based retrievals.

3. ASTER Data ImportThe plume mapping tool has been modified to accommodate ASTER data. The TIR

subsystem of ASTER can be pointed up to 8.6 degrees from nadir, and this pointing anglemust be considered in the calculation of optical paths through the atmosphere. TheASTER-enabled mapping tool was used to produce the first satellite-based map of thePu’u O’o SO2 plume (Figure 4).

4. Backwards CompatibilityThroughout the modification of the mapping tool, every effort has been made to

maintain backwards-compatibility. For example, the mapping tool can still be applied toaircraft data sets, which typically lack geo-reference information, and ground elevationscan still be entered manually, eliminating the need for a DEM. To date the mapping toolhas been used to analyze TIMS, MAS, and MASTER data, all airborne instruments, aswell as ASTER and MODIS data.

d2. MISR and AIRS Data

In his role as Supervisor of the Visualization and Scientific Animation Group (VSA),Vince Realmuto has been working with the MISR and AIRS Projects. The MISR workhas been focused on the retrieval of wind speeds, cloud morphology, and cloud heightfrom the MISR Multi-angle data. Such work has obvious applications to remote sensingvolcanology, since plume height is a critical factor in TIR-based retrievals of SO2 or ashconcentrations, and the wind speed controls the emission rates of volcanic products. TheMISR_Shift tool was developed for interactive estimation of wind speed and cloudheight. Figure 5 demonstrates the application of MISR_Shift to the analysis of theOctober 29, 2002 Mount Etna Eruption plume.

AIRS is an atmospheric sounding instrument currently in orbit aboard the Aquaspacecraft. Last October and November the AIRS Science Team demonstrated that theycould detect SO2 in the Mount Etna eruption plume. The AIRS Team will not estimateSO2 on an operational basis (outside the scope of the project), but we have the


opportunity to use our mapping tool for this purpose. AIRS has over 2300 spectralchannels, and after resampling the AIRS spectra to the highest resolution output byMODTRAN we will obtain 200 channels in the 8 to 12 um spectral region currently usedto map SO2 from space. This unprecedented spectral resolution will help us separate thecontributions of SO2, sulfate aerosol, silicate ash, and ice particles to the radiance spectrameasured over volcanic plumes and clouds.

e. Aerosol Studies (led by John Porter, Univ. Hawaii)In Year 2 of this project, we collected a series of field measurements of the Pu’u

O’o plume and published our preliminary Sun photometer and lidar results (Porter et al.,2002). During this time period we have also been working on satellite algorithms whichwe will use to derive volcanic aerosol flux rates. As part of the volcanic efforts we havealso developed a new way to derive aerosol extinction, which can be used to measurevolcanic emissions. Each of efforts is discussed below.

e.1 Measurements in the Hawaii Volcano PlumeThe Hawaii Department of Health (Clean Air Branch) has been providing a Kona

vog index to the public since 1997. This index is based on aerosol nephelometermeasurements (METONE model ES-640) collected at the Konawaena High School (nearCaptain Cook town) as well as along with visual observations. Values range from 1 to 10with 10 being the largest vog concentrations. These vog index alerts are published in theKona newspaper, radio broadcast, and are available from a telephone recording. Thepublic has shown a strong interest in these vog alerts. Several limitations exist in theaccuracy of the current vog alerts. First the point measurement (at Captain Cook town)may not be representative of the larger Kona region. Secondly, no information isavailable on the vertical distribution of vog and people near beach level may havesignificantly different concentrations from those further up the mountain. Third, thediurnal cycle has never been studied and comparisons with meso-scale models areneeded. Finally, the vog alert is currently based on aerosol optical nephelometermeasurements and human visual observations. In order to convert the nephelometermeasurements into vog aerosol mass concentrations (which is important for dosagecalculations), values provided by the nephelometer manufacturer are currently being usedwhich are based on different aerosol types, which may not be valid.

During April 16-18, 2002 we carried out a field experiment at the DOH site onKona Hawaii. One of the goals was to collect filter mass measurements at this site andcompare them to the aerosol scattering measurements obtained by the Hawaii DOH.Figure 6 shows the results of Moudi impactor filter measurements analyzed by ionchromatography (collected and processed by Barry Huebert’s group at UH). The aerosolcollected during this period was composed of an accumulation mode sulfate aerosol andcoarse mode sea salt aerosol. This was not surprising but we also found that the aerosolswere very neutralized being between ammonium sulfate and ammonium bisulfate. Othermeasurement directly in the volcano plume (and downwind of the plume) have shownthat the aerosols there are much more acidic (sulfuric acid). The ammonia present in theaerosol is likely from ground activities and therefore raises further questions of the spatialvariability of the aerosol acidity.


Figure 7 shows the aerosol mass inferred from the aerosol scattering measurements(nephelometer) collected by the Hawaii Department of Health. The aerosol massconcentration derived from the filter measurements is also shown showing that theHawaii Department of Health may be slightly overestimating the aerosol massconcentrations. As these results are still preliminary and require further work.

Another main goal of this experiment was to investigate the spatial distribution ofvog around the Kona region. Lidar and sun photometer measurements were collectedduring a drive from Kona to the volcano region. Although these measurements have notbeen fully analyzed, an example of the sun photometer measurements is shown in Figure8. As we moved from the Kona side to the windward side the aerosol optical depthsdecreased and had less spectral dependence. We expect to process the lidar data andcarry out aerosol size distribution inversions and compare these data sets to satellitemeasurements.

e.2 Satellite Aerosol StudiesSatellites offer an ideal approach to derive volcanic emissions if they can be made to

quantitatively derive the aerosol optical depths. Various problems exist includingestimating the ocean surface reflection and accounting for variability in the aerosol phasefunction and aerosol single scatter albedo. Estimating the ocean surface reflectiondepends on the wind speeds which can be quite variable around islands. In order toaddress this problem we have begun using meso-scale wind fields in deriving surfacereflections. This provides a spatial resolution of 10 km as opposed to the 100 kmavailable from synoptic models (which are used in all other aerosol products). Figure 9gives an example of this effort where we have processed a MODIS image using anaerosol algorithm we have developed specific for volcanic aerosols. In the future we planto validate this product and use it to derive volcanic aerosol flux rates from the satellite.

f. Vegetation Studies (led by Luke Flynn, Univ. Hawaii)An important link in our investigation of the impact of active volcanism on the

environment is the assessment of the damage done to the ecosystem by the volcanicplumes. The steady trade winds are responsible for very restricted areas of intensefumigation and vegetation kill-off. An important factor at these three volcanoes is theinfluence of topography. At Masaya, the localized kill-offs are most notable in areaswhere the plume is in direct contact with the vegetation. Where the plume is able todecouple from the surface, (e.g., in the valley between the 2 ridges, and past the LlanoPacaya ridge towards the Pacific), the effect is reduced. A similar relationship is visibleat Poás and Kilauea. However, further work is required to fully constrain these effects aswell as the differences between the “wet” (Poás, Kilauea) and “dry” (Masaya) volcanoes.We have limited our spectral studies to a series of plants available at Masaya andKilauea.

Generally, we have used ti leaves, coffee plants, bananas, and grass as our test caseplants. One of the difficulties that we envisioned with this project would be the degree towhich human intervention, or care, in keeping these plants healthy would affect ourresults. However, especially in the case of Masaya, kill-off zones are extremely sharp.As little as 30 – 100 m exists between the fairly robust vegetation (cultivated coffee


plantations) and the grass-only kill-off zones. This correlates with observations of thedeterioration of structures within and outside the kill-off zone. Sheet metal roofs andhinges corrode in less than a year within the kill-off zone leaving the area very sparselypopulated, while metal roofs a few kilometers away (outside the kill-off zone) areunaffected. More importantly, there is a marked higher incidence of asthma and otherlung-related problems for students attending schools within, as opposed to outside, thekill-off zone.

In order to investigate the environmental effects of persistent degassing, we useremote sensing data (Landsat ETM+, IKONOS) with NDVI and Tassled Cap band ratioalgorithms to delineate poorly vegetated areas downwind of each volcano. These data areincorporated, through a GIS, with Digital Elevation Models as well as various groundtruth data (soil pH, dry deposition rates, precipitation acidity, etc.). Extremely distinctzones of vegetation “kill off” are noted. These appear to correlate with changes intopography. It appears that sharp topographic changes allow the gas plume to decouple orcouple with the ground, hence lessening or increasing its impact at any down windlocation. This integrated study of degassing at persistently active volcanoes may aide inlimiting the effects on human populations and agriculture downwind of such systemsthrough improved land use management.

While at Masaya and Poás, we also collected several hundred spectra of healthyplants outside of the plume, and sick plants within the plume (Figure 10). These spectraare being used to better understand the NDVI images of these volcanoes derived fromLandsat 7 data, as well as Kilauea. A correlative study, being undertaken by GlynWilliams-Jones, is the analysis of dry deposition and sulfation rate data from Masaya(Figure 11) that were reported by Delmelle et al. (2001).

g. Plume Models (led by Steve Businger, UH, and Lori Glaze, Proxemy Research)Making the connection between the ground-base observations, single satellite

images and the known meteorological conditions is the primary focus of our plumestudies. Some of this work is focused on simulating plume dispersal patterns on dayswhen we were in the field. For instance, Steve Businger has conducted a dispersionmodel run for Nicaragua for the same day that we were collecting gas measurements. Ananimation of this simulation can be seen at:

http://imina.soest.hawaii.edu/MET/Faculty/businger/poster/vog/masaya_animation.gif

The primary goal of this research is to create a prognostic tool to aid in theprediction of vog plume concentration and dispersion. The wind fields and thermal datafrom the Meso-scale Spectral Model (MSM) were used as input into the Hybrid SingleParticle Lagrangian Integrated Trajectory Model (HY-SPLIT) in order to produce vogsimulations. Validation of model results was conducted using aerosol concentrationsderived from aircraft and ground-based data, as well as satellite images. Satellite imageswere also used to validate plume size, shape, and location.

Another aspect of our plume rise modelling activities resulted in a 2002 publication(Glaze et al., 2002). Much of Lori’s recent attention has been directed toward thedownwind diffusion of SO2 and simulations that can be used to help correlate FLYSPEC


measurements taken at two or more locations. These correlations can then be used toconstrain wind speeds, and hence used to estimate volumetric emission rates. Untilrecently, the standard method for measuring volcanic SO2 emissions in the field was touse a COSPEC (correlation spectrometer). The COSPEC can be fixed to a tripod on theground, mounted inside a vehicle and driven under a plume, or mounted in an airplaneand flown underneath a plume. While the COSPEC does an excellent job of measuringvertical SO2 concentrations, it is bulky and costly. One of the most common uses ofvolcanic COSPEC data has been to estimate the volumetric emission rate of SO2.However, this calculation is always limited by the uncertainty on estimated wind speeds.Because the FLYSPEC is so small and inexpensive to operate, it is possible to deploy twoor more FLYSPECs in order to better constrain the wind speed parameter.

In order to constrain wind speed with multiple FLYSPEC measurements, it iscritical to be able to correlate data collected at one location, to another locationdownwind. In theory, one can identify “puffs” of higher concentration and then see howlong these puffs take to reach the next FLYSPEC. Of course, in practice, this is not aseasy as it might seem. This is partly because as the plume is transported downwind, it isalso diffusing in the atmosphere. Not only does this affect our estimates of wind speed,but it also affects our ability to identify specific puffs at two separate locations.

To address these issues, we have begun to conduct some numerical simulationsdirected toward identifying correlations between the wind speed and gas data sets.Currently, this effort is focused on smoothing an initial time series through translationand decay. Working with FLYSPEC data collected near the summit of Stromboli on 29September 2002, we have an outstanding opportunity for quantitatively characterizing thedecay of each individual SO2 pulse. Figure 12 shows over an hour’s worth of data on9/29/02. These data were collected from a fixed position, with the FLYSPEC pointedupward through the plume. A clear exponential decay is seen for several “puffs” ofhigher SO2 concentrations. To date, a crude simulation has been done of a single peak inthe SO2 concentration. One peak has been taken and fitted to an exponential curve of thedecay, then used to quantify the variation from that curve with the standard deviation ofthe residuals. The diffusion of the pulse as it moves downwind is simulated bylengthening the elapse time of the pulse (the pulse spreads out longitudinally), loweringthe peak value, and adding the noise back in.

h. Soil Science (led by Nancy Dise, Open University)In this year (between the first and second field seasons) we analyzed the soils from

our first trip to volcanoes Poás (Costa Rica), Masaya (Nicaragua) and Kilauea (Hawaii).All soils (150 samples) were analyzed for pH (1:2.5 soil:water slurry), and 60 sampleswere selected for major exchangeable cations (drip column exchange with NH4Cl), majorwater-soluble anions (saturation paste), loss-on-ignition (estimate of organic carboncontent), and trace metals (ICP-AES, not yet complete). Allophane content, cationexchange capacity and anion exchange capacity will be measured on 20 samples (not yetcomplete).

The results show clear differences in the acidification levels of the soils in the threelocations, with the rank order Poás >Masaya>Kilauea (Figure 13). The pH of Poás soils(subsoil pH 4.0-5.7) suggests that soil acid-base reactions are buffered by the ‘non-


renewable’ (until the next eruption deposits fresh material) exchange of basic cations,and, in more acidified soils, by the dissolution of aluminum from the clay lattice. Thisreleases monomeric aluminum, Al3+, a known toxic to plants, microbes and aquatic life.Most of the subsurface soils measured at Poás show significant proportions of theexchangeable surface (up to 100%) saturated with Al3+. Along the mean direction of theplume, acidification was clearly still evident in the most distant sampling site about 12km from the source. Samples away from the main direction of the plume show much lessdamage. Other chemical measurements support this pattern (although the full analysesare not yet complete).

Soils in Masaya (pH 5.1-6.7) and Kilauea (5.9-6.8) are much less acidified, and showcorrespondingly low levels of dissolved Al3+ (Figure 13). The apparent discrepancybetween these sites and Poás may be in part due to the ages of the volcanic deposits, withPoás soils much older, and thus with more time for basic cations to leach out of the soilfrom the intense acid deposition.

Work in our second field season will concentrate on determining the full ‘footprint’of soil acidification to Poás, determining if finer-scale soil acidification can be detectedon Masaya and Kilauea (e.g. close to the plume), collecting plant samples from all sites todetermine metal/nutrient contents, and potentially visiting a third site in Central Americato increase the amount of data for developing empirical transfer functions. We will alsoliase closely with other team members to link the soil data to measurements of gasemission and deposition, remote sensing of plant damage, etc.

5. ReferencesDelmelle, P., J. Stix, P.J. Baxter, J. Garcia-Alvarez, and J. Barquero (2002). Atmospheric dispersion,

environmental effects and potential health hazard associated with the low-altitude gas plume ofMasaya volcano, Nicaragua. Bulletin of Volcanology 64: 423 – 434.

Duffell, H., C. Oppenheimer and M. Burton (2001). Volcanic gas emission rates measured by solaroccultation spectroscopy. Geophys. Res Lttrs. 28: 3131-3134.

Glaze, L. S. and S. M. Baloga (2002). Volcanic plume heights on Mars: Limits of validity for convectivemodels. J Geophys Res, 107:10.1029/2001JE001830.

Harris, A. J., D J Pirie, K Horton, L P Flynn, H Garbeil, J B Johnson, H Ramm, E Pilger (2002). DUCKS:A continuous thermal presence on the rim of Pu’u O’o. EOS Trans. AGU 83 (47). Fall 2002 meetingabstract V71A-1256.

Horton, K., G. Williams-Jones, H. Garbeil, A. J. Sutton, and T. Elias (2002). FLYSPEC: A new ultravioletcorrelation spectrometer for the detection of volcanic SO2. EOS Trans. AGU 83 (47). Fall 2002meeting abstract V12A-1418.

Porter, J.N., K. A. Horton, P. J. Mouginis-Mark, B. Lienert, S. K. Sharma, E. Lau, A. J. Sutton, T. Elias,and C. Oppenheimer (2002). Sun photometer and lidar measurements of the plume from the HawaiiKilauea volcano Pu’u O’o vent: Aerosol flux and SO2 lifetime. Geophys. Res. Lttrs. 29:10.1029/2002GL014744.

Sutton, A.J. and T Elias (2002). Twenty Years of Continuous gas Release at Kilauea: Effusive Lessons in aVolatile Time. . EOS Trans. AGU 83 (47). Fall 2002 meeting abstract V62C-06.

Williams-Jones, G., Delmelle, P., Rymer, H., Rothery, D.A., Stix, J., Harris, A.J.L., Fournier, N.E., St-Amand, K., Strauch, W., Navarro, M., Alvarez, J., Williams, S.N., & Stoiber, R.E. (2002).Mechanisms of magma degassing at an open-system, active basaltic volcano, Masaya, Nicaragua.Bulletin Volcanology, Submitted.


6. Figures

Figure 1. Time series set of observations of the thermal anomaly associated with thePu’u O’o vent, as derived from the Terra MODIS during 2001.


Figure 2: Average monthly gravity change (black diamonds) at a representative craterrim station on Masaya compared with the average monthly sulfur dioxide flux (greycircles). Grey shaded region represents one standard deviation of SO2 flux values(~30%). Dashed lines denote the five stages of activity between 1992 and 2002. Theschematic cross-section illustrates the possible stages of thinning and thickening of thevesiculated gas-rich layer due to changes in gas flux.


Figure 3: We have done a number of field comparisons between our new FLYSPECinstrument and COSPEC, which has been the community’s standard technique for morethan 20 years. Here we see our field set-up, along with a direct comparison of how thetwo instruments measure SO2 in both high and low emission conditions. Theseinstruments produce almost identical results, but the lower cost of FLYSPEC (~10% ofCOSPEC) and smaller size make it possible for us to plan deployments of multipleinstruments in a semi-permanent array to fully study the sulfur dioxide content of aplume.


Figure 4: Color-coded sulfur dioxide concentrations superposed onto a false-colorcomposite of ASTER’s visible and near infrared channels. High SO2 concentrations arecolored white, lower concentrations are red, orange, yellow, green and blue. ASTER isthe only instrument in orbit that can detect passive venting of SO2 in plumes as small asthe Pu’u O’o plume (typically <1.5 km wide over land). Scan line noise in the SO2retrievals over the ocean results from the lower temperature contrast between the plumeand ocean surface (relative to the plume over land). Despite this lower contrast, themisfits (differences between the predicted and actual radiance at a pixel) were lowestover the ocean, due to the thermal and spectral uniformity of the ocean surface. Thiscapability is essential as we try to monitor SO2 emissions from space. The image wasacquired on October 30, 2001 and covers an area of 42 x 44 km.


Figure 5: Screen shot demonstrating the application of our MISR_Shift tool for theanalysis of the October 29, 2002 Mount Etna Eruption plume.


Figure 6. Aerosol size distribution derived from cascade impactor (Moudi Impactor)measurements taken in Kona Hawaii at the Hawaii Department of Health site in Kona(Captain Cook area). Molar ratios showed that the sulfate aerosol was somewherebetween ammonium sulfate and ammonium bi-sulfate meaning that it was nearly fullyneutralized. This is possibly due to the DOH site, which is near a baseball field whereammonium nitrate is used for fertilizer. Although the aerosol near the volcano has beenshown to be very acidic, the degree of aerosol neutralization in other parts of Kona is stillnot well known.


Figure 7. These lines show the aerosol mass concentrations as a function of time of dayderived from the aerosol scattering measurements obtained by the Hawaii Department ofHealth (at the Kona, Captain Cook site). The green line is the average for April (2002)and the blue line is the average for March (2002). The thin red line shows the values forApril 18, 2002 when filter samples were also collected. The solid red line gives theaerosol mass concentration from the filter measurements suggesting that the conversionfrom aerosol scatter to mass may be slightly overestimating the aerosol mass.


Figure 8: On April 17, 2002 aerosol optical depths measurements were made whiledriving from Captain Cook (on Kona Hawaii) to the volcano area. These measurementsare shown above going from A through J. On the Kona side the aerosol optical depthsare larger and have a stronger spectral dependence. On this day the plume near the venthad strong easterly component so that south of the volcano area clean marine air wasencountered as opposed to volcanic aerosol, which can occur under more typical tradewind conditions.


Figure 9. Aerosol optical depths are shown here which were derived from a MODISsatellite image collected from the University of Hawaii direct broadcast ground station(run by Torben Nielsen). The image was processed by an algorithm developed by JohnPorter and uses meso-scale wind fields (~10 km resolution) to derive the surfacereflection around the complex island topography. The plume from Kilauea volcano isclearly visible.


Figure 10: Spectral signatures for coffee plants and grasses in and outside of areasaffected by intense degassing at Masaya volcano. Note the drop in the chlorophyll peakfor the coffee leaves and its complete absence in the upper of the grass spectra. Bycomparing variations in the spectral intensities, and index is being developed forcomparison with the Landsat 7 NDVI.


Figure 11: Map of the network of dry deposition stations and sulfatation rates (mg sulfurdioxide m2 per day) in the vicinity of Masaya volcano. Note that the highestconcentrations correlate well with the areas of strong topography and are stronglycontrolled by the dominant trade winds. Modified from Delmelle et al. (2002).


Figure 12: Time-series of the sulfur dioxide measurements that we obtained atStromboli volcano, Italy, using FLYSPEC.

0

500

1000

1500

2000

2500

3000

3500

15:36:00 16:04:48 16:33:36Time (HH:MM:SS)

Con

cent

ratio

n (p

pm-m

)


Figure 13: Soil data collected at Poás and Masaya show a wide variation in pH, but ingeneral it is clear that much more Ca-rich soils due to the lower rainfall.


7. Plans for Year 3a. At the time of writing, we are preparing for another visit to Masaya volcano in March

2003, to collect gas and thermal data, but also to participate in the Workshop onVolcanic Gases being organized by John Stix (McGill University). Team memberswho are expected to be present will be Harris, Williams-Jones, Horton and Busingerfrom UH, Oppenheimer (Cambridge) and Realmuto (JPL). Businger will attendprimarily to help study the airflow to the west of the volcano as input data for his HY-SPLIT simulations. We will conduct a HY-SPLIT simulation for the entire period ofthe workshop and compare this simulation to our field observations.

b. We have just taken a very crude first step in the area of measuring gas fluxes usingmultiple FLYSPEC instruments. Keith Horton plans to deploy at least twoFLYSPECs on Kilauea in order to derive mass flux estimates. Co-incident lidar andSun photometer data will also be collected to maintain the linkage between gas andaerosols. As we collect additional field data from multiple instruments it should bepossible to develop statistical tools for finding significant correlations between twodata sets. To first order, this tool will be ready to constrain wind speeds. In parallel,Lori Glaze will work on the diffusion modeling to add real physics into our estimatesof how diffusion affects the shape and magnitude of a “puff” as it is transporteddownwind.

c. We believe that there is a wealth of information contained with our thermal andinfrasonics data already collected for Kilauea, Masaya and Stromboli. We expect thata considerable part of our effort in year 3 will be devoted towards trying tounderstand what these data are telling us about the near-surface conduit structure, aswell as understand how the micro-gravity observations may relate to theseobservations.

d. In terms of satellite observations of plumes, and our efforts to determine the SO2concentration of the plume, we will concentrate on the collection and analysis ofMODIS time series data set for Masaya Volcano, concentrating on images coincidentwith the February 2002 field campaign as well as the upcoming (March 2003)IAVCEI Volcanic Gas Workshop. This work will be in collaboration with John Stixof McGill University, as well as several of the IDS Team members. We also plan tocontinue the collection and analysis of ASTER and MODIS data coincident with fieldwork at Kilauea Volcano in 2001, 2002, and scheduled for 2003. This work will beclosely coordinated with John Porter and Keith Horton. Development of tools for there-sampling of AIRS data to MODTRAN resolution, thus enabling us to use theplume-mapping tool to analyze AIRS data.

e. We also recognize the need for documenting all of the field/satellite correlations thatwe have derived in this project to date. Many of our techniques are either new (e.g.,time-series MODIS thermal flux [Figure 1], FLYSPEC, lidar/sun photometer data,and the ASTER-derived fluxes), or data are obtained at unprecedented rates (e.g., ourHY-SPLIT models and the thermal radiometer data). Producing a synthesis paperthat describes all of our methods in a single journal article is expected to be of greatvalue to volcanology community. This paper will most likely be submitted to aninternational volcanology journal (e.g., Bulletin of Volcanology), but we may alsowrite a paper with a remote sensing focus for IEEE or Remote Sensing of the


Environment. We will also explore options for comparable summaries beingsubmitted to atmospheric chemistry and ecology journals.

f. We intend to link our spectral observations of plants with those derived from Landsat7 and Hyperion observations. Our point spectral studies on Masaya and Kilauea willprovide sub-pixel resolution for this study. We will publish these results. We willalso compare our observations of plant health with local topographic changes and inthe case of Masaya the dry deposition of sulfur on sulfur plates.

year 2 progress report and plans for year 3pmm/ids_team_report_yr2.pdfyear 2 progress report:...

Documents