Remote Sensing

Remote Sensing

Contact: Paul Bates
Funding: NERC, PhD studentships

In collaboration with:

Bristol Researchers:

Bristol PhD Students:

Remotely sensed data is used in various areas of hydrology within the Hydrology Group, often closely linked to our modelling research. The main flooding and flood risk applications researched at Bristol include: monitoring water levels, identifying inundation extent and in the parameterisation and validation of hydrodynamic models. Details of the remote sensing systems being actively used in our research can be found below.

Water levels

The ability to track water level changes of remote rivers and floodplains, such as the River Amazon, is extremely valuable for many areas of research. With recent advances in remote sensing systems, the ability to monitor catchments from space has greatly improved. Amanda Hall is currently carrying out research into Amazon water levels, under the supervision of Dr Guy Schumann, using the Ice, Cloud and land Elevation Satellite (ICESat). This sensor was designed primarily to monitor ice sheets elevations; however, the satellite’s ~70 m footprint with observations every 172 m along-track make it more suitable for observing river water levels than the previously used radar altimeters.

Floodplain water levels obtained with ICESat will then be compared to predicted elevations from the LISFLOOD-FP Amazon model.  ICESat water elevations have also been investigated in relation to modelling work being carried out for the Niger and Zambezi Rivers by Jeff Neal and Guy Schumann.

Inundation extent

Synthetic Aperture Radar (SAR) imagery has also been used to provide synoptic views of flood extent for inundation model validation and calibration and to compare performance of different models for various study areas, from the River Amazon, Brazil to the River Dee, UK. This has required the development of image processing algorithms, such as statistical snakes.  As well as using satellite sensors such as ERS-1/2 and RADARSAT, the use of airborne sensors producing high resolution and multi-band imagery has been investigated. These can produce higher quality validation data sets and can detect flooding beneath vegetation canopies, for example using HH-VV phase difference or enhanced backscatter.  Research using these techniques has been carried out by Liz Stephens.

Contact

Paul Bates

Funded by: EPSRC Flood Risk Management Research Consortium (GR/S76304/01)
NERC Urgency award, Mapping and modelling water elevations for the 2005 Carlisle flood event (NE/D521222/1)

In collaboration with:

Bristol researchers:

Remotely sensed data can be used for both parameterisation and validation of LISFLOOD-FP and this is an area of active research. Airborne laser altimetry (or LiDAR) data have been used to provide both topography and roughness parameters for inundation models. This allows the effects of small scale topographic features, such as dykes and ditches, to be represented in hydraulic models, and the effects of spatially variable roughnesses to be investigated.

LiDAR image of Avonmouth, nr Bristol.

Synthetic Aperture Radar (SAR) imagery has also been used to provide synoptic views of flood extent for inundation model validation and calibration and to compare performance of different models. This has required the development of image processing algorithms such as statistical snakes. As well as using satellite borne sensors such as ERS-1/2 and RADARSAT, the use of airborne sensors producing high resolution and multi-band imagery has been investigated. These can produce higher quality validation data sets and can detect flooding beneath vegetation canopies, for example using HH-VV phase difference or enhanced backscatter.

Airborne Synthetic Aperture Radar classified at a spatial resolution of 1m, using a statistically active contour (Snake)

Edit this page