| | Background
Map of the broader area around Lake Volvi
On 8 October 2006, extensive flash floods occurred in the Prefecture of Thessaloniki, Central Macedonia, Greece, damaging agricultural fields, properties and infrastructure. Flash floods are caused by heavy rainfall in a short period of time (i.e. a few hours), and are usually characterised by raging torrents that rip through rural and urban areas, sweeping everything before them. In order to study the effects of this flood, we will look at the area north of Lake Volvi.
The area belongs to a semi–mountainous region that also includes the tectonically active basin of Mygdonia, located about 60km east of the city of Thessaloniki. Several small villages exist in the area, which is principally covered by agricultural fields.
In this case study you will see how remote sensing can contribute to flood detection and management. You will learn how to use radar satellite images acquired before and after the flood, in order to delineate the affected areas and extract information on the flood.
Earth Observation Synthetic Aperture Radars (SAR) are active, side-looking sensors onboard satellites that can provide images day and night. Their wavelengths also allow them to penetrate clouds, smoke and some vegetation, contrary to optical systems which are, in most cases, useful only during the day and under favourable weather conditions. For more information on radar technology, go to the respective “Remote Sensing Principles”/”Remote sensing in depth” section on Eduspace.
With respect to water monitoring, the strength of the backscattering recorded on a radar image is influenced by several parameters, including the average surface roughness and dielectric properties of the target.
- For smooth surfaces, specular reflection takes place and the recorded backscattering is weak. This results in dark tones on radar images, making calm water areas relatively easy to recognise. However, the presence of wind and/or currents increases the roughness of the water surface and produces a high backscatter (brighter tones).
- Under dry conditions, most natural materials have a dielectric constant between 3 and 8. Water has a high dielectric constant (80) - at least 10 times higher than that of dry soil. Therefore, increased soil moisture affects the dielectric constant of the soil, which results in higher backscattering.
Different features are seen on this SAR image
In a SAR image water bodies are generally dark, while very wet (but not covered by water) areas appear bright. Other smooth surfaces such as roads, highways and bare fields appear darker, whereas some human infrastructure, such as buildings, villages and cities, appear very bright. Hills and mountains appear distorted, due to the side-looking characteristics of radar systems.
The images we will use in the exercises come from the ASAR radar sensor onboard the Envisat satellite.
The data that will be used in this exercise are three Envisat/ASAR images with the following characteristics:
1. Date of acquisition: 10/10/2006, Descending, track 7, swath IS2, Incidence angle: 19.20° – 26.70°, Spatial resolution: about 28m, Pixel size: 12.5m
2. Date of acquisition: 25/10/2005, Descending, track 7, swath IS2, Incidence angle: 19.20°– 26.70°, Spatial resolution: about 28m, Pixel size: 12.5m
3. Date of acquisition: 5/10/2004, Descending, track 7, swath IS2, Incidence angle: 19.20° – 26.70°, Spatial resolution: about 28m, Pixel size: 12.5m
|  | Indicative damages from the October 2006 floods | |
The first image (10/10/2006) was acquired two days after the flood event, while the other two images were acquired under dry conditions, one year (25/10/2005) and two years (5/10/2004) before the flood respectively. As can be seen, all the selected images have the exact same technical characteristics, in order to ensure maximum consistency in terms of acquisition conditions. They were also taken during the same time of the year (October), avoiding vegetation differences that would have been seen if taken during different seasons.
A zipped file (Thessaloniki.zip) containing all the images can be downloaded from the right menu.
| |
