INTRODUCTION Kelud volcanic lake is one of the most active and most dangerous stratovolcano in Indonesia. Many lives were claimed by Kelud eruptions in the past six centuries mostly due to pyroclastic flows, surges and especially lahars, as in 1919 eruption (5160 victims). Before the last eruption (1990) a system of drainage was done to maintain the volume of the lake at low level (2,000,000 m3) and to control the lahars. Hydro-acoustic monitoring is the first system of surveillance that has been performed at the crater lake and has revealed intense gas bubbling a year before seismic precursor records of the 1990 eruption. | |
Location map of Kelud volcano |
Image ASTER courtesy NASA (USA), http://asterweb.jpl.nasa.gov/ |
CARBON DIOXIDE FLUX: METHOD OF ACCUMULATION CHAMBER The CO2 flux emitted by the surface of the lake was measured during 2001-2006 by IR spectrophotometry. We modified the technique initially developed for soil gas flux monitoring in order to work on a crater lake by using a floating accumulation chamber. During each field campaign, around 260 measurements were performed to cover the entire lake surface (104,000 m2). Results show a decrease in the CO2 flux from 36,000 T/year in 2001 to 13,000 T/year in 2006. Most of the CO2 degas from the lake surface and escape to the atmosphere. Only a small fraction (<15%) is trapped and converted as HCO3- in the lake waters. |
|
CO2 system on soil and floating accumulation chamber
|
|
|
CO2 MAPPING CO2 flux measurements surveys were carried out in periods of dry seasons (from July to September) between 2001 and 2006. The technique used to measure the CO2 flux degassing from the lake is based on the method of the accumulation chamber (Chiodini et al, 1998). This system has been modified for CO2 flux measurements by using a floating accumulation chamber equipped with a Dräger Polytron Infrared spectrometer (West Systems, Italy). It is the first time that CO2 flux measurements have been applied for studying CO2 degassing on a volcanic lake surface. The estimation of the CO2 discharge and the uncertainty of the estimation were computed with the application of the sequential Gaussian simulation (sGs). This approach has already been used for soil CO2 degassing at other volcanic systems (Cardellini et al, 2003; Chiodini et al, 2004; Frondini et al, 2004). The sGs method produces the nearest realistic images of the distribution of the CO2 fluxes from the histogram and semi-variogram of the original data. The results of the simulations are also used to define the uncertainty of the total CO2 output estimates. On the base of semi-variogram models 500 sGs were performed. Except for 2001 data (67,669 m2) the areas covered are closed to the real surface of the lake (103,600 m2). The results of the 500 simulations were used to compute the total CO2 output The estimated averages of CO2 flux, for the surface of the lake, were about 99 t/d, 90 t/d, 78 t/d, 36 t/d, 33 t/d , 36 t/d from 2001 to 2006. |
DISTRIBUTION OF CO2 FLUX
The computation of the CO2 flux data is based on the graphical statistical approach (GSA) procedure (Chiodini et al., 2001; Brombach et al, 2001). The data were partitioned into log-normal populations using graphical statistical technique of Sinclair (1974).
Three log-normal populations of FCO2 were determined and for each population, the proportion and the statistical parameters were computed . The study of the distribution of CO2 flux permits to distinguish three populations that contribute to the transfer of CO2 from water to air. The population A and B correspond to the CO2 flux resulting from rising bubbles at shallow depth (population A) and where the bubbles are visible on the entire surface of the lake (population B). Population C may result of a transfer by diffusion of CO2 from water to air.
Exemple of histogram and probablility graph to differenciate populations