Operational Use of InSAR for Volcano Observatories: Experience from Montserrat

Volcanoes as targets for InSAR vary greatly in the quality of their returned phase signal. There are two end-member volcano types from this perspective. Basaltic, low-relief shield volcanoes with frequent effusive eruptions and shallow magma reservoirs generally give excellent coherence and large magnitude ground deformations (~1m) that can be easily detected and modelled. However, their individual (lava flow) deposits tend to be too thin ( <10m) for accurate InSAR measurement. In contrast, silicic volcanoes with substantial relief, infrequent eruptions and deeper magma reservoirs tend to be more difficult targets because of greater vegetation cover, layover/shadowing and small magnitude deformation signals. New deposits, however, can be thick (~100m) and can , potentially, be measured with sufficient accuracy by InSAR. Sometimes silicic volcanoes continue to erupt for years, usually explosively and dangerously, and present a long-term challenge to the local observatory. Here we report early experiences in learning how useful InSAR can be in such a case – the 1995-99 eruption of Soufriere Hills volcano, Montserrat. We found that the volcano became a more accessible target to ERS InSAR with time as more of its slopes were covered by deposits and became coherent. We were able to map the thickness of these deposits in this way. However, the summit lava dome did not become coherent and hence could not be measured by InSAR. Long-term coherence was also limited, partly because of decorrelation effects due to ash fall/erosion and this also limited the ability to build up a time series of differential movements. The ERS repeat cycle is not optimum for volcano monitoring but could provide some useful operational support if data can be acquired fast enough (local reception/fast transmission) . Local processing is also probably desirable because of the benefits of immediate interaction with other scientists concerned with crisis management.