INFLUENCES OF CHANGING LUNAR CYCLE AND BAROMETRIC PRESSURE ON THE ERUPTION OF NATROCARBONATITE LAVA, OL DOINYO LENGAI VOLCANO, TANZANIA

Volcanologists have long been interested in what effect, if any, external factors may have on the onset or intensity of volcanic eruptions. Research typically consists of correlating historical eruption frequencies to recorded seasonal changes and weather patterns over periods of many years, or over relatively short periods of eruption at active volcanoes. Our research involved a comparison of eruption frequencies and strengths to corresponding barometric pressure, theoretical Earth tides, and the lunar cycle during a 30-day period. From June 29 to July 29, 2004, a continuous presence was maintained in the active crater of Tanzania's Ol Doinyo Lengai volcano, within the African Rift Valley. Ol Doinyo Lengai was chosen for study because it erupts extremely low viscosity natrocarbonatite lava, and because the volcano has been in a near-continuous state of eruption since 1983. It is hypothesized that due to the unusually low viscosity of natrocarbonatite lava, possible influences of external factors are more easily detected at Ol Doinyo Lengai than at most other volcanoes. Barometric pressure was recorded using a NovaLynx barograph. Theoretical earth tides were calculated using TSOFT software, developed by the International Center for Earth Tides in Brussels. An hourly rating of the volcano's eruptive activity and strength was recorded using a discrete numerical scale, developed on the basis of prior observations during eight crater visits from 1997 through 2003. Time series analysis reveals that 68% of observed increases in activity occur while barometric pressure decreases or is at a minimum - a result statistically significant to 99%. Spectral analysis using TSOFT indicates no correlation between activity and the diurnal tidal cycle. However, it does reveal activity cycles with periods of 8.5 days and 30 hours, respectively. The 2004 observation period is not long enough to determine if these cycles are significant. Additional eruption observations and barometric pressure data were collected at fifteen minute intervals over a two week period during Summer, 2005. These are being analyzed and compiled with that of 2004 to improve the statistical significance of results.