References: Morton, R.R, Taylor, G., F.R.S, Turner, J., 1956. Turbulent gravitational convection from maintained and instantaneous sources. Prodeedings of the Royal Society of London. Series A, Mathematical and Physical Sciences A234, 1-23. Sparks, R., Bursik, M., Carey, S., Gilbert, J., Glaze, L., Sigurðsson, H., Woods, A., 1997. Volcanic Plumes. * Björn Oddsson, email: bjornod@hi.is,Institute of Earth Sciences, Sturlugata 7, 101 Reykjavík Comparison of results from estimation of mass with models relating discharge rate and top height of volcanic plumes with measured mass. All models overestimate total discharge but constant plume height results are slightly more closer than an oscillating plume. The empirical relation between discharge rate and column height. The cross marks indicates the original dataset used by Sparks et al 1997 to find the formula H =1.67Q 0.259 where H is the height in km and Q is the DRE discharge rate of magma in m 3 /s. The red diamond is the discharge rate in the Grímsvötn eruption, calculated from measured mass. The maximum plume height recorded by the weather radar in Keflavík, the times of visual inspection, altitude measurements done by aircraft, and calibrated plume height. All heights are relative to the eruption site (1400 m a.s.l) The red line indicates the maximum height of the eruption plume. The blue line is the maximum height of the plume, calibrated by using the altitude recorded by a monitoring aircraft altimeter. The altitude records of the weather radar show 1-2 km higher altitude than the altimeter in the aircraft. The path of the radar beam is a function of the curvature of the Earth and the elevation angle. The beam width increases with distance from the radar, and these two properties control the minimum height of the recorded object. Grímsvötn is 260 km away from the radar in Keflavík, and the beam width is around 4 km above Grímsvötn. During the eruption there were inversions in the vertical temperature profile of the atmosphere. Under these surconstances super-refraction can occur which can cause bending of the radar beam, about 1.6 km. Results Measured total mass In the summers of 2005 and 2006, systematic sampling of tephra was carried out on Vatnajökull, resulting in a comprehensive isomass map for the deposit. The map was intergrated to get the total mass of erupted material. The results was 5.4 · 10 10 kg. Model calculations Theoretical Morton (1956) proposed an equation to express the relationship between the input rate of thermal energy to the plume, and the top height of a maintained plume from a steady source: H = 31(1 + n) −3/8 ˙ Q 1/4 , (1) where H is the height (m), ˙ Q is the rate of production of thermal energy at source (kW), and n is the ratio of the vertical gradient of the absolute temperature to the lapse rate. For basaltic magma the relationship between ˙ Q V in m 3 /s, and H in km is: H =1.85 ˙ Q 1/4 V (2) Empirical Conclusions • Radar records from the weather radar in Keflavík should be used with caution when estimating volcanic plume heights over long distances. The radar beam width increases with distance and is about 4 km over Grímsvötn resulting in ±2 km error. Atmospheric super-refraction can cause bending of the radar beam towards the Earth, leading to overestmation of plume height of a few kilometers. Finally the discrete stepping in elevation angles result in an apparent fluctuation in the plume height and a corresponding uncertainity of 2 km plume height over Grímsvötn. Visual inspections during the eruption, show that the plume did not oscillate as much as the radar records are indicate. The apparent oscillation in the data is probably due to the long distance effects mentioned above. • Estimation of total mass by theoretical and empirical models introduced by Morton 1956 and Sparks et al 1997 resulted in 120% - 160% more mass than the measured values, when calculated for constant plume (8 km), and for oscillating plume (jumps in height from 6-10 km in radar records), resulted in values 220% - 280% higher than the measured. The models give the order of magnitude of the mass of the erupted material, but the results suggest that resonably accurate determination of deposit mass or volume can only be obtained from detailed mapping of deposit mass or thickness. • Due to the long distance error effects in the radar records, it is important for safety, economical and scientific reasons to install an weather radar on North-East Iceland to provide cover adequate for eruptions within Vatnajökull and elsewhere with North and East Volcnic Zones. Stepping in radar records The stepping of the maximum height recorded, can be explained by the error limits due to the width of the radar beam (figure 1). The maximum height recorded is depending on whether the plume is detected as the highest point in one elevation angle or the lowest in the next elevation angle above. Weather radar in Keflavík The Icelandic Meteorological office remotely operates a weather radar located about 3 km from Ke- flavík International Airport on the Reykjanes Peninsula, SW Iceland. Currently, scanned images are routinely acquired every 15 minutes for normal weather monitoring and every 5 minutes during vol- canic eruptions. The radar beam circles with an initial elevation angle of 0.5 ◦ . Its revolution is gradually raised 12 times up to an angle of 15 ◦ during a full scanning cycle of 2 minutes. During the volcanic eruption in Grímsvötn, images were generated every 5 minutes representing a 2D map graphically showing the maximum reflectivity present in the vertical column over each surface point. The eruption Small to medium sized phreatomagmatic, basaltic eruptions are the dominant style of activity in Grímsvötn. The last eruption occurred 1-6 November 2004. The crater was located in the south- west corner of the Grímsvötn caldera, where the eruption broke through 150-200 m thick ice in about an hour. This brief subglacial phase was followed by the main phase of the eruption, which lasted about 33 hours. The main phase was monitored by weather radar as well as by visual observations from aircraft. During this period the eruption plume reached a height of 6-10 km (relative to vent), and formed a well-defined tephra layer towards the north and northeast Björn Oddsson 1 ∗ , Magnús Tumi Gudmundsson 1 , Guðrún Larsen 1 , Sigrún Karlsdóttir 2 1 University of Iceland, 2 Icelandic Meteorological Office Grímsvötn 2004: Weather radar records and plume transport models applied to a phreatomagmatic basaltic eruption