SURFACE DEGASSING AND MODIFICATIONS TO VESICLE SIZE DISTRIBUTIONS IN ACTIVE BASALT FLOWS

The character of the vesicle population in lava flows includes several measurable parameters that may provide important constraints on lava flow dynamics and rheology. Interpretation of vesicle size distributio ns (VSDs), however, requires an understanding of vesiculation processe s in feeder conduits, and of post-eruption modifications to VSDs durin g transport and emplacement. To this end we collected samples from act ive basalt flows at Kilauea Volcano: (1) near the effusive Kupaianaha vent; (2) through skylights in the approximately isothermal Wahaula an d Kamoamoa tube systems transporting lava to the coast; (3) from surfa ce breakouts at different locations along the lava tubes; and (4) from different locations in a single breakout from a lava tube 1 km from t he 51 vent at Pu'u 'O' o. Near-vent samples are characterized by VSDs that show exponentially decreasing numbers of vesicles with increasing vesicle size. These size distributions suggest that nucleation and gr owth of bubbles were continuous during ascent in the conduit, with min or associated bubble coalescence resulting from differential bubble ri se. The entire vesicle population can be attributed to shallow exsolut ion of H2O-dominated gases at rates consistent with those predicted by simple diffusion models. Measurements of H2O, CO2 and S in the matrix glass show that the melt equilibrated rapidly at atmospheric pressure . Down-tube samples maintain similar VSD forms but show a progressive decrease in both overall vesicularity and mean vesicle size. We attrib ute this change to open system, ''passive'' rise and escape of larger bubbles to the surface. Such gas loss from the tube system results in the output of 1.2 x 10(6) g/day SO2, an output representing an additio n of approximately 1% to overall volatile budget calculations. A stead y increase in bubble number density with downstream distance is best e xplained by continued bubble nucleation at rates of 7-8/cm's. Rates ar e approximately 25% of those estimated from the vent samples, and thus represent volatile supersaturations considerably less than those of t he conduit. We note also that the small total volume represented by th is new bubble population does not: (1) measureably deplete the melt in volatiles; or (2) make up for the overall vesicularity decrease resul ting from the loss of larger bubbles. Surface breakout samples have di stinctive VSDs characterized by an extreme depletion in the small vesi cle population. This results in samples with much lower number densiti es and larger mean vesicle sizes than corresponding tube samples. Simi lar VSD patterns have been observed in solidified lava flows and are i nterpreted to result from either static (wall rupture) or dynamic (bub ble rise and capture) coalescence. Through comparison with vent and tu be vesicle populations, we suggest that, in addition to coalescence, t he observed vesicle populations in the breakout samples have experienc ed a rapid loss of small vesicles consistent with 'ripening' of the VS D resulting from interbubble diffusion of volatiles. Confinement of ri pening features to surface flows suggests that the thin skin that form s on surface breakouts may play a role in the observed VSD modificatio n.