3. OTHER MONITORING SURVEYS
Several surveys have been undertaken since June 2006 to assess the extent and the effects of the mud eruptions.
GPS and geodetic surveys were carried out between June 06 and June 07 to assess subsidence. These surveys were not repeated.
Other surveys, however, were repeated at irregular periods such as micro-gravity and micro-seismic surveys.
In 2008 shallow ground temperature surveys were added to the programme.
3.1 Assessment of Subsidence
Earlier GPS surveys occupied stations located mainly in the then not-flooded part of the subsiding area but showed significant variation of short term rates between adjacent stations (Istadi et al., in prep; Abidin et al., 2008).
After flooding of the N sector in December 2006, subsidence surveys could only be undertaken on dams where soil compaction disturbed measurements.
At the relief well site RW01 (see location in Figure 3), subsidence was recorded for 123 days between September 2006 and January 2007 (average subsidence c. 1.8 cm/d).
Another 33 day long GPS record of subsidence was taken between November and
An independent assessment of subsidence between May 06 and February 07 was obtained by analysing radar data recorded by the ALOS satellite using the INSAR method.
Results of this study (Fukushima et al., in press; Abidin et al., 2009), converted to subsidence rates in m/yr, are shown in Figure 3.
Other long-term subsidence information comes from observations of sites not affected by soil compaction, such as the cemented, 4 m high standing conductor pipe at well RW01 where a total subsidence of c. 8 m (with reference to the nearby carriage dam) is indicated when visited by us in October 2008.
This estimate points to a long-term subsidence of at least 4.5 m/yr (over a 22 months period).
Railway tracks were lifted along and outside the W boundary dam to compensate for local subsidence of the order of 1 m between 2007 and 2008.
These data indicate that most of the early subsidence data derived from spot GPS surveys contain errors owing to site effects and should not be interpreted in terms of long-term estimates.
The caveat also applies to the previously cited GPS readings at the RW02 site taken over a one month period for which an apparent subsidence rate of c. 14 m/yr was predicted.
Construction and maintenance of all dams at the LUSI site has been a tremendous labour-intensive effort (c. 70 Mill t of fill has been used for the construction of these dams over a two year period) although all dams, especially the c. 15 m high, inner circular dam, suffer additional subsidence as a result of compaction.
3.2 Interpretation of Subsidence
Possible causes of the subsidence and its pattern (shown in Figure 3) have been investigated by Fukushima et al. (in press), namely:
· loading effect of the deposited mud layer,
· subsidence due to the creation of an equivalent vertical mud conduit,
· as well as pore pressure decrease and/or removal of mass at depth.
The effects of the first two are insufficient to explain the magnitude of the observed subsidence.
Limits of the third explanation were investigated by computing subsidence patterns for a set of oblate spheroidal models resulting in best fits for models with compaction centres between c. 0.6 km and 0.7 km depth.
These depths are much shallower than the inferred source depths of the mud discharged at the surface.
Because of the important implication of these findings, we used characteristics of other compaction-subsidence models (Geertsma, 1974) to assess the maximum depth (z max) of an anomalous compaction centre from the characteristic wavelength λ /2, where λ defines the (horizontal) distance of the subsidence anomaly.
For data in Figure 3, along a profile (070o) running through RW01, a range of λ /2 data was obtained which yield z max values between c. 0.5 km and 0.7 km; the early subsidence rate at RW01 becomes an important controlling parameter in the assessment of λ.
Our treatment is less rigorous than that of Fukushima et al. but produces a similar result. It can be inferred that subsidence has been caused by pore pressure reduction of sediments at rather shallow (< 0.7 km) depths (sediments at these depths were probably also over-pressured).
An inferred section of the shallow de-gassed sediments is shown in Figure 2. The triggering mechanism for the discharges is unknown but is probably not related to activities during the last drilling phase of the 2.8 km deep uncased well before the LUSI eruption.
3.3 Micro-Gravity Surveys
Two separate micro-gravity studies were made with the aim of detecting the likely location and magnitude of subsurface mass withdrawal. The first study used three surveys between August and October 2006 and was undertaken by ITB staff.
The second study involved four surveys during May to July 2008 by BMG staff. In each case, the gravity data were not reduced for station height changes between start and end of the survey nor for near-surface mass changes (changes of ground water level and other near-field effects). Such reductions are a fundamental requirement for the interpretation of gravity anomalies caused by underground mass removal (Hunt, 1995).
The computed lapse-time anomalies of both micro-gravity studies contain therefore systematic errors and can not be interpreted in terms of underground mass changes.
3.4 Micro-Seismicity and Ground Temperature Surveys
A second micro-earthquake survey was made between May and July 2008 coinciding with the micro-gravity study (BMG, 2008). Three-component seismometers were installed at seven stations close to the circumference of the flooded area. A total of c. 60 local events were observed.
All, except probably one, were of shallow origin (< 0.5 km depth).
The epicentre of most events occurred in the S and E quadrant at distances of c. 1.5 to 2 km from the eruption centre; a single event occurred in the N sector, also c. 2 km away from the mud discharge centre. The shallow seismic velocity structure of the flooded area is not well constrained and is disturbed by near-surface, gas-rich sediments.
Monitoring of ground temperatures was started in May/June 2008 (80 sites); a repeat survey was made in October 2008 (65 sites). Temperatures were measured at 0, 1 and 2 m depth (occasionally at 1.5 m) along the foot of dams inside the flooded areas and near short road segments outside. D
uring the 2ndsurvey, shallow T-profiles were also measured along the inner dam-hot mud boundary around the semi-circular dam and extended towards the pumping stations.
With reference to climate data from Surabaya airport, anomalous ‘cold’ and ‘hot’ ground temperatures could be defined for data from 1 m depth ( T < 23 deg C was defined as ‘cold’ and T > 33 deg C as ‘hot’). The observed mean minimum and maximum air temperatures do not exceed 25 and 31 deg C respectively during the whole year.
‘Cold’ ground temperatures, sometimes as low as 19 deg C, were measured on the inner N side of the semi-circular dam and near the outer W, NW, and S boundary dams. All ‘cold’ sites were close to nearby gas discharge features. Anomalously ‘hot’ temperatures (up to 47 deg C) were observed over short, outer segments at the foot of the circular dam.
The ‘cold’ temperatures are probably caused by near-surface expansion of rising gas (for rising CO2 a pressure drop of 1 bar can be associated with a cooling effect of c. 0.7 deg C).
Sites with ‘hot’ temperatures at the foot of dams are affected by conductive heating. Hot mud temperatures down to 2 m were also measured during the 2nd survey.
At stations near the gas-charged mud ring (inner margin of circular dam) temperatures were still 90 deg C at 2 m depth but the mud had cooled rapidly to c. 40 deg C at the surface, possibly due to the combined effect of gas expansion and evaporation.
The second survey also detected anomalously ‘cold’ ground in the SE sector outside the dam which had suffered a dam breach in early 2008.