Jökull - 01.06.2000, Side 70

Earthquakes in the Mýrdalsjökull area, 1978–1985
occur in one season, the process loading stress in the
area must do it at a rate that is similar to the rate of
the modulating effects of the seasonal changes. In the
absense of magma accumulation, the only other pro-
longed process that comes to mind as a loading pro-
cess is the regional stress accumulation due to plate
movements. Even though the Mýrdalsjökull volca-
noes are not on the plate boundary they are within
the zone where significant strain accumulates (Jóns-
son et al., 1997). We suggest that the stress concen-
tration around the magma chambers of the Mýrdals-
jökull volcanoes is sufficient to sustain the persistent
seismicity there. In view of these arguments the very
existence of the western cluster is a strong indication
of an independent magma chamber west of the Katla
caldera.
How then is the triggering achieved? Assuming
that the earthquakes are caused by brittle failure the
trigger must either change the state of stress or the
failure conditions, or possibly both. Let us consider
the state of stress within the seismically active crustal
volume, characterized by the maximum and minimum
principal compressive stresses, ! and !#" . Because of
the persistent nature of the seismicity the rocks must
be close to failure most of the time, or as shown in
Figure 8, the Mohr circle of the stress field must be
close to the failure envelope. The long term mass
fluctuations of the glacier are modulated by seasonal
changes. Snow accumulates on the top of the glacier
most of the year but is partly melted in the summer
months. The net increase in snow is balanced by ice
flow down to the ablation areas. So it must be as-
sumed that the load of the glacier is slowly increas-
ing during the larger part of the year but decreasing
during the summer months. This load will affect the
stress field in the underlying crust. A load term, $&% ('
is added to both principal stresses, where $&% is the
ice density,

is the acceleration of gravity, and
'
is
the thickness of the additional ice. The Mohr circle
thus moves to the right during increasing load, i. e.
away from the failure envelope, and to the left during
times of decreasing load. The crustal volume can, in
this way, be brought to failure by reducing the load of
the glacier. But this triggering mechanism only works
during the summer months.
The Mýrdalsjökull glacier is a temperate glacier
and a good part of the melt water finds its way into
the glacier and down into the crust below. So the
load of the melted ice is not carried off immediately
and the load reduction is prolonged into the autumn
months. This may be sufficient to explain why earth-
quakes usually continue until the end of the year. But
now a different mechanism is likely to become effec-
tive. The increased water in the crust will lead to el-
evated pore pressure, ) , which in turn will affect the
failure envelope. The failure envelope will move to
the right by the amount ) (Figure 8).
Figure 8. A Mohr-diagram for a hypothetical state of
stress and a failure envelope in the crust beneath a
glacier. The crust is fractured when the Mohr circle
and the failure envelope intersect. The small arrow
shows the displacement of the circle * +'( $ % when
the thickness of the ice is reduced by h. The large
arrow shows the displacement of the failure envelope
when the same amount of melt water is added to the
crust increasing the pore pressure by ) -,/.0'( $1% . –
Mohr-graf sem sýnir ímyndað spennuástand og brot-
þolsferil í jarðskorpunni undir jökli. Skorpan brest-
ur þegar Mohrshringurinn og brotþolsferillinn sker-
ast. Litla örin sýnir hvernig Mohrshringurinn færist
þegar jökullinn þynnist vegna bráðnunar. Stóra ör-
in sýnir hvernig brotþolsferillinn færist ef sama magn
af bræðsluvatni fer niður í skorpuna og hækkar poru-
þrýsting í berginu.
JÖKULL No. 49 69