1. Fate and Transport of Fine Volcanic Ash William I Rose Michigan Tech University HOUGHTON, MI 49931 USA raman@mtu.edu www.geo.mtu.edu/~ramandu/~raman 26 May 201026 May 2010 ESRIN Frascati MODIS Aqua 17 April 2010

2. fine ash proportions erupted radar and---ascending ash ice or water in volcanic cloud? remote sensing dynamics----early fallout stage 2 ash size and distance t-storm analogy---forecast possible? is the aged cloud dangerous? http://www.geo.mtu.edu/~raman/Ashfall/

3. Origin of particles in volcanic clouds Explosive vesiculation-- As pressure drops in ascending magma--overpressured bubbles burst Hydrothermal explosions-- rock fractured by thermal shock from contact between magma and water Milling-- abrasion and grinding of particles can occur in pyroclastic flows and in the vent Chemical and meteorological processes-- condensation, sublimation, surface chemical reactions forming acids, salts, hydrometeors and aggregates of mixed origin

4. Tephra is classified on the basis of pyroclast size: ASH -- Very fine-grained fragments (< 2 mm), generally dominated by broken glass shards, but with variable amounts of broken crystal and lithic (rock) fragments. Courtesy of USGS. LAPILLI -- Pea- to walnut-size pyroclasts (2 to 64 mm). They often look like cinders. In water-rich eruptions, the accretion of wet ash may form rounded spheres known as accretionary lapilli (left). Courtesy of USGS. BLOCKS AND BOMBS -- Fragments >64 mm. Bombs are ejected as incandescent lava fragments which were semi-molten when airborne, thus inheriting streamlined, aerodynamic shapes. Blocks (not shown) are ejected as solid fragments with angular shapes. Courtesy of J.P. Lockwood, USGS. www.geology.sdsu.edu/how_volcanoes_work

5. Φ phi Φ = - log 2 d (mm) Lognormal size distributions are “expected” and we use a “biased” system to define them ASH LAPILLI BOMBS, BLOCKS

6. Total Grain-size distribution weighted by mass and by isopach volume, compared to Carey and Sigurdsson [1982]. 18 May 1980 Mount St Helens Fall deposit 40% of mass is <30 microns in diam

7. 7 Fine and very fine ash Linked to fall and air resistance volcanic ash, <2 mm diameter fine ash, <1 mm : intermediate flow regime very fine ash, < 30 microns : laminar flow regime ash > 1 mm falls in ~30 min as we progress after 30 min, exponential thinning decreases and may reverse, and atmospheric sorting decreases to zero Rose & Durant, 2009, JVGR 186: 31-39

8. Three stages of volcanic clouds Stage 1--near volcano, first 1-2 hrs of ash residence, exponentially thinning fallout of pyroclasts in turbulent flow, radar and webcam Stage 2-- several hours to 2 days, accelerated aggregate fallout of 90% of fine and very fine ash, infrared remote sensing (mie scattering) Stage 3--days to weeks? drifting volcanic clouds, very fine ash present but danger uncertain... trajectory models, particle/SO 2 remote sensing

9. Volcanic Cloud Stages Rose et al, 2001, J Geology, 109: 677-694

10. Keflavik Radar

12. Freezing occurs rapidly over narrow height range Release of latent heat as all droplets freeze; burst of positive buoyancy Influences maximum plume height attained Droplets form as T falls during rise -17<T<-24˚C (not to scale!) freezing during volcanic plume rise freezing during volcanic plume rise freezing level Adam Durant

13. Thunderstorm Formation Ingredients –warm, moist air (often mT) –unstable (or conditionally unstable if lifting mech.) –encouraged by diverging air aloft

14. Rabaul, PNG, 1994 Reventador, Ecuador, 2002 Ice affects plume buoyancy and maximum plume height, and particle fallout Maximum plume height is an essential input parameter for volcanic cloud dispersion modelling used in real-time hazard mitigation convective column ice-rich stratospheric ‘umbrella’ cloud ~16 km max. height ice-rich (>20 MT) stratospheric ‘umbrella’ cloud ~20 km max. height Adam Durant

15. During much of the Eyjafjallajökull eruption, the ash column did not reach the level of ice formation. This affects the aggregation and fallout.

16. Laminar flow; RN = 10 -2 Turbulent flow; RN = 10 6 RN = 20 RN = 40RN = 10 4 Coarse ash Fluid dynamics applies dimensionless analysis of fall of spheres in the atmosphere, which shows that experience with large pyroclasts might not apply to smaller ones which fall much more slowly… RN =dv t  /  Fine and very fine ash 10 m/s D = 1mm D = 1µm.01 cm/s

17. Fall of spherical particles in earth’s atmosphere Schneider et al., 1999, J Geophys Res 104 4037-4050

18. d < 1000 μm Φ > 0 d < 30 μm Φ > 5 Each stage 2 ashfall has a size distribution which does not reflect distance or age, and which includes ash down to submicron diameters.

19. 2 Map from Sarna-Wojcicki et al. [1981]; isomass contours in g/cm 3.

20. 3 Ave. Cumulative Mass Fraction

21. 6 Aggregate Growth From Gilbert and Lane [1994] Binding forces: electrostatic surface-tension from liquid films ice formation mechanical interlocking Collisions: differences in particle terminal fall velocities electrostatic attraction (if separation distance is low)

22. 5 MSH80 Aggregate Fall Map adapted from Sarna-Wojcicki et al. [1981]; isomass contours in g/cm 3

23. Meteorological Cloud Volcanic Cloud Many IN Small ice HM Little Precip Sublimation Few IN Bergeron Large Ice HM Precipitation Durant et al., 2008, JGR 113

24. Mammatus simulation: thunderstorm cirrus outflow anvil Kanak and Straka, Atmos. Sci. Let. 7: 2–8 (2006) ~6000 m Simulation time: 20 minutes! 10 µm snow aggregate diameter contours dry sub-cloud layer snowflake aggregation induced Cloud descent rate: ~6.5 ms -1 10 µm ice crystal descent rate: <10 -2 ms -1

25. 9 Conceptual Model: Distal Fallout

26. FLEXPART forecasts ash cloud motion, but how does it account for stage 2 fallout, and beyond?

27. CIMSS experimental product using SEVIRI and showing ash loading, cloud height and particle size. This data is a potential step forward, and may allow for quite sophisticated interpretation, and fits well after ground based radar and before/with trajectory models. M Pavolonis, NOAA CIMSS.

28. The sizes of ash particles sensed optimally by these methods are about 1-25 µm (~5-9  ).

29. “Aged” volcanic clouds Coarse ash falls out of cloud within ~30 minutes Most remaining ash has fallen out of cloud within 24-36 hours Trackable up to 4 days with IR split window, mostly SO 2 and sulfate after that Aged = over 2 days old Do they still pose a risk to aircraft?

30. GOES - visible Courtesy of Scott Bachmeier, U. Wisconsin, Madison