From an initial state considered as a time-reference, gravity may change on volcanoes because of deformation of the volcano displacement of mass at the surface of within the edifice (water, magma, ...) change in density within the edifice Tha amplitude changes are of the order of centimeters to meters for the deformation, whereas it is of the order of 1-1000 microgal (1 microgal=10-8 m/s2). Very accurate instruments are required and we need to obtain observation or to model other variables at the same time, like the deformation (accuracy for height change at 1 cm level) or the pressure. Figure caption Schematic temporal evolution of an active volcano. Possibilities to produce gravity changes are reported in the figure. From an initial state (state 1), where blacks dots represent the location of measurement points, the gravity variation at the second state (state 2) may change due to surface movement (which involves free-air elevation change and volume change of the volcano), mass redistribution of the topography at surface (extrusions) and internal mass redistribution, the amount of which we are looking for (after Jousset et al., 1999;Jousset et al., 2000;).

Volcanological Survey of Indonesia

Usu Volcano Observatory, Japan

Institut de Physique du Globe de Paris, France

Merapi, one of Indonesia's most active volcanoes, lies in one of the world's most densely populated areas and dominates the landscape immediately N of the major city of Yogyakarta. The steep-sided modern Merapi edifice, its upper part unvegetated due to frequent eruptive activity, was constructed to the SW of an arcuate scarp cutting the eroded older Batulawang volcano. Pyroclastic flows and lahars accompanying growth and collapse of the steep-sided active summit lava dome have devastated cultivated and inhabited lands on the volcano's western-to-southern flanks and caused many fatalities during historical time. The volcano is the object of extensive monitoriong efforts by the Merapi Volcano Observatory of the Volcanological Survey of Indonesia." (from GVP/USGS Weekly Volcanic Activity Report). Within the Arrangement between the Volcanological Survey of Indonesia and the French Goverment, a series of studies have been carried out in the mid-90's, involving the observations of gravity and deformation, during an intense activity of the volcano. The gravity variations were linked to the growth and collapse of the dome (figure), but also allowed to propose a model for the activity of the volcano at the time. The crystallisation and degassing of implaced magma would be responsible of the dome growth. Figure caption Observed gravity variations for 1993-1994 period at Merapi volcano. Despite the small deformation, large gravity variations occured at the summit area, whereas no changes occured far from the volcano summit. This effect is mainly due to the dome activity. The open squares indicate the theoretical attraction of the extruded dome, computed form the estimations of the dome shape (using a density oof 2400 kg.m-3). The error bar associated to these squares refelct the range of the possible theoretical attractions computed from different cases of dome, including shape, volume and density (after Jousset et al., 2000).

Figure caption Modelling of the gravity variations between November 1997 and May 1998. Observations are consistent with a double-source model: (1) a deep increase of mass at a depth of about 4 to 5 km below the surface, which contributes to the large-scale positive residual gravity signal (red curve); and (2) a shallow decrease of mass within the first hundred meters below the surface, which contributes to the crater-scale residual relative negative signal (green curve). The addition of these two computed signals explains the observed data (after Jousset et al., 2000).

Several small phreatic eruption occured in March 1996, October 1998 and November 2000. The Coordinating Committee for Prediction of Volcanic Eruption (CCPVE) noted that repeating small-scale phreatic eruptions may be possible advance signs of a large eruption, such as for the 1929 eruption. Within this context, a two-year project of the Japanese Society for the Promotion of Science, 4 sessions of  gravity and displacement have been obtained by teams of the Usu Volcano Observatory. Our observations are consistent with a model of underground shallow-water evaporation occuring as a result of heat rising from a hot and dense body (density contrast of 200 kg.m-3) of about 10^11 kg at 4-5 kg in depth.

Within the same project of the Japanese Society for the Promotion of Science at Komagatake volcano, several sessions of observations have been carried out at Usu Volcano between September 1996 and August 1998, using both microgravity and GPS observations on a large network covering Results revealed that the dome associated with the 1977-1978 eruption was still contracting due to compaction and cooling. There was also a rise of 1.5 m of the water table, which might be a precursor to the 2000 eruption (see below). Figure caption Gravity Variations as a function of height variations. Top: Normally gravity and elevation changes are inversely corralated. If no mass change is associated with the elevation change, then the gravity follows the free-air gradient. Both theoretical and estimated gradients are indicated on the figure. Similarly, if no density change occurs with elevation changes, then gravity follows the Bouguer-corrected gradient. When observations reveal departure from these predicted gradients, sub-surface and/or density changes are implied. Bottom: Residual gravity after estimated free-air gradient correction as a function of elevation. Normally, residual variations should follow the variations of the attraction of the topography. We clearly see that at Usu-Shinzan dome, these models do not explain the ecess of residual gravity, indicating an increase of density within the crypto-dome. We modelled the excess in gravity as a combination of underground water level rise by 1.5 m and the density Jousset et al., 1999).

At 13:10 (JST), 31 March [2000], an eruption started at about 4 km NW of the summit, and at about 2 km NE of the epicenter of volcanic earthquakes, that occured within the few days before the outburst.
The National Committee of Volcanic Eruption Prediction, chaired by Prof. Yoshiaki Ida (U. Tokyo) showed the official comments on 22 May as follows:

      "At Usu volcano, eruption has continued since March 31, though it became small in the activity. In the initial stage of eruption, many craters appeared and issued volcanic ash. The eruption activity concentrated into a few craters in the mid-April, repeating intermittent explosions. The degree of explosion and amount of eruption cloud gradually decreased with time. Magmatic material (pumice) that was found in the initial stage of eruption were hardly observed in products of successive eruptions.
      Although many felt earthquakes occurred before and after the commencement of eruption (the maximum M=4.6), such the seismicity declined soon. Though earthquakes have still continued several kilometers below the western and southern parts of the volcano, their distribution, frequency and magnitude (M2 to 3) have not been changed since the mid-April.
      During the activity, the West-Nishiyama area was largely uplifted, the highest part reaching about 60 m, and the volume of uplifting is
estimated about 40 million cubic meters. The uplifting rate decreased with time, and is 10 cm a day these days. The time needed for the uplifting rate declining is smaller by one order than those associated with the dome formation in the 1944-45 and 1977-78 eruptions, and the same in order as that in the 1910 eruption.
      As the above, the magma activity had gradually declined and there is a possibility that the activity comes to an end. As eruption, uplifting and seismicity have still continued, however, it is still probably that explosions damaging the areas around the present active craters take place due to new interaction of magma with underground water.
      Although it is impossible to deny the possibility of a new lager eruption due to the increase of magma supplied from beneath, it is likely that such the eruption can be predicted as changes in ground deformation, seismicity, and mode of eruption cloud. Should still pay attention to the eruptive activity near the crater areas. Careful and continuous monitoring of the eruptive activity is critical to detection of its reactivation."

Within this context the Comite d'Evaluation des Risques Volcaniques (France) decided to learm from this experience how to manage a volcanic crisis. New observations of gravity have been obtained and are currently under process. The huge deformation associated with this eruption (figure) is the first reason of the gravity change in this eruption. With gravity variation ranging 0 to more than 3 mGal  at some station, these observations include some of the most large values of gravity change ever recorded on a volcano. This excellent data should be able to help quantifying the amount of new magma that intruded within the edifice, in association with this eruption.