InfraRed imaging
- Thermal anomalies are easier detectable in MIR (right) than in TIR (left) spectra (MT. Etna peak)
- Clouds often cover the peak of volcanoes (example on Java) and disable anomalies detection
- The difference between background and hot spot signal is the strogest in 3.5–4 µm spectra
State of the art
Monitoring volcanic activity from space using radiometer or other imaging systems has been used for at least 25 years and the suitability of space-borne infrared sensors for the detection and analysis of volcanic thermal surface features and plumes is well known. Early on resolution and repeat cycles were not as good as they are now but it has been shown by several authors that thermal anomalies due to effusive events could be detected, monitored, and measured. During the last decade most observation of volcanic eruptions were carried out using data from AVHRR, GOES, MODIS, ASTER, BIRD, and Landsat-7 / ETM system. All systems differ in spatial resolution ranging from 8×8 km for 6.47–7.02 µm band on GOES all the way down to 15×15 m pixels for the VNIR bands on ASTER. Availability of similar scenes varies also strongly, with geostationary systems sending images very 15–30 min and ASTER/MODIS every 16/2 days.
Most of the detection of high temperature hotspots is based on the so called dual band or bi-spectral method. Given a medium size lava lake (e.g. 1000 m2) at 500–1000 °C an observation of the MIR (pixel size 1 km2) would result in a temperature of 52–115 °C. In the TIR however this would only result in a temperature of 1.5–3.6 °C given that the background is at 0 °C. This temperature difference allows a detection of subpixel high temperature anomalies. Using refined techniques solar heated anomalies can also be eliminated.
Goals and workplan
The problem that remain, are:
- the Mid and Thermal IR channels, MIR at 4 µm and TIR at 10 µm, respectively, saturate from major eruptive volcanic events, and
- their coarse spatial resolution (pixel size > 1 km) does not allow the reliable detection of small volcanic events, especially thermal precursors of events.
We will follow a follow a two way strategy with the following goals:
- include the MODIS observations into the VFRS
- perform an analysis of existing at DLR BIRD data sets obtained in 2002–2004 over small and major volcanic events, in order to
- assess the expected potential of prospective IR sensor systems (ESA’s GMES Infrared Element and FIRES as information sources for volcanic unrest observation).