Applicability of InSAR to tropical volcanoes: insights from Central America

Measuring volcano deformation is key to understanding the behaviour of erupting volcanoes and detecting those in periods of unrest. Satellite techniques provide the opportunity to do so on a global scale but, with some notable exceptions, the deformation of volcanoes in the tropics has been understudied relative to those at higher latitudes, largely due to technical difficulties in applying Interferometric Synthetic Aperture Radar (InSAR). We perform a systematic survey of the Central American Volcanic Arc to investigate the applicability of Interferometric Synthetic Aperture Radar (InSAR) to volcanoes in the tropics. Volcano characteristics that may prevent InSAR measurement include: (1) dense vegetation cover; (2) persistent activity; and (3) steep slopes. Measurements of deformation are further inhibited by atmospheric artefacts associated with: (4) large changes in topographical relief. We present a systematic method for distinguishing between water vapour artefacts and true deformation. Our data show a linear relationship (c. 2 cm/km) between the magnitudes of water vapour artefacts and volcano edifice height. For high relief volcanoes (e.g. Fuego, Guatemala, 3763 m a.s.l. (above sea level)) errors are of the order of 4–5 cm across the volcano’s edifice but are less than 2 cm for lower relief (e.g. Masaya, Nicaragua, 635 m a.s.l.). Examples such as Arenal, Atitlan and Fuego illustrate that satellite acquisition strategies incorporating ascending and descending tracks are particularly important for studying steep-sided volcanoes. Poor coherence is primarily associated with temporal decorrelation, which is typically more rapid in southern Central America where Evergreen broadleaf vegetation dominates. Land-use classification is a better predictor of decorrelation rate than vegetation index. Comparison of coherence for different radar wavelengths match expectations; high resolution X-band radar is best suited to local studies where high-resolution digital elevation models (DEMs) exist, while L-band wavelengths are necessary for regional surveys. However, this is the first time that relationships between phase coherence and time, perpendicular baseline, radar wavelength, and land use have been quantified on the scale of a whole volcanic arc. Interferometric synthetic aperture radar has been used to detect and measure deformation at over 70 volcanoes worldwide since the 1990s (Fournier et al. 2010). A variety of different types of deformation have been observed using InSAR surveys, including the movement of magma during co-eruptive deformation (e.g. Lu & Dzurisin 2010) or intrusive processes (e.g. dyke and sill intrusion: Hamling et al. 2009; Biggs et al. 2010), as well as a variety of shallower surface processes. These include hydrothermal activity (Pritchard & Simons 2004), slow edifice subsidence (Ebmeier et al. 2010) and lava flow contraction (Stevens et al. 2001). Regional-scale surveys have detected magma movement at volcanoes previously thought to be quiescent (e.g. on the East African Rift: Biggs et al. 2009) and at locations not obviously associated with a particular volcanic edifice (e.g. the central Andes: Pritchard & Simons 2004). However, the global distribution of InSAR measurements of volcano deformation is currently uneven, due, in part, to the differences in radar returns from different types of land surface. The majority of radar satellites have operated at C-band wavelengths, l 1⁄4 5.6 cm, which are known to be affected by vegetation cover. This is a particular problem in the tropics where dense, rapidly growing evergreen vegetation is especially prevalent and causes a high rate of change in surface scatterer properties and therefore rapid decorrelation. This has presented significant obstacles for some C-band studies of volcano deformation in the From: Pyle, D. M., Mather, T. A. & Biggs, J. (eds) Remote Sensing of Volcanoes and Volcanic Processes: Integrating Observation and Modelling. Geological Society, London, Special Publications, 380, http://dx.doi.org/10.1144/SP380.2 # The Geological Society of London 2013. Publishing disclaimer: www.geolsoc.org.uk/pub_ethics 10.1144/SP380.2 Geological Society, London, Special Publications published online March 4, 2013 as doi: