View source for Temporal analysis: Remote sensing

Remote sensing approaches usually use instruments that are not in contact with the ground or water to measure their characteristics (e.g. elevation, spectral signature, etc). They may employ passive sensors that detect the electromagnetic radiation emanating from an object (e.g. photography) or active sensors that emit a signal and measure the properties of the signal after it has reflected off the object (e.g. radar). The sensors may be mounted on satellites, aircraft or at points on the Earth’s surface. Data types that are collected using remote sensing approaches include aerial photographs, multispectral, radar and laser-derived information. For an overview of remote sensing and its use in fluvial geomorphology, see Jensen (2000), Gilvear et al. (2003) and Carbonneau and Piégay (2012).


Remotely-sensed data can be used at all of the spatial scales and to assess temporal changes in most hydromorphological characteristics. Its application is dependent generally on the resolution of the data and the size of the features being identified or the magnitude of change being detected. For example, high-altitude aerial photography and most freely-available satellite data (e.g. Landsat, ASTER) have high spatial coverage and low spatial resolution making them best suited for the identification and monitoring of catchment, landscape unit and segment scale characteristics. These data sources can also be used for exploring some reach characteristics on medium to large rivers (e.g. channel position and width for rivers greater than ca. 100 m in width). Conversely, low-altitude aerial surveys (e.g. photography, multi- and hyperspectral), high resolution satellite imagery (sub-metre), and laser-based techniques (airborne LiDAR and terrestrial laser scanning) have lower spatial coverage but higher spatial resolution making them better suited to segment and reach scale characterisation. Another distinction should be drawn between sources that obtain plan (2D) information (e.g. aerial photographs, multi – and hyperspectral data) that may also be interpreted to estimate heights, and those that directly produce altimetry data (e.g. LiDAR, TLS, radar).


The timescale over which remote sensing can be used to investigate changes in hydromorphological characteristics is highly variable, as is the frequency of measurements that are collected. For example, airborne surveys are relatively expensive to commission, thereby limiting the frequency with which they are conducted, but these surveys often have sufficient temporal resolution to observe broad decadal to annual changes in river geomorphology. Furthermore, aerial photograph archives often date back to the mid 20th century and so predate other types of remotely sensed information. In contrast, satellite datasets offer amazing opportunities to observe changes over very short timescales: annually, seasonally or even weekly, and thus immediately before and after specific events (e.g. floods, earthquakes, etc.). For example, since the launch of NASA’s Landsat 4 satellite in 1982, Landsat Thematic Mapper data has been collecting data across the Earth’s surface every 16 days, providing multispectral information with a spatial resolution of 30 to 120 m. Landsat imagery is freely available from the USGS  within 24 hours of acquisition.
 

Remotely-sensed data products obtained from national governments or commercial sources will typically have well-defined accuracies that are detailed in accompanying manuals and technical documents. When not specified, accuracy / uncertainty can be estimated using standard techniques, such as photogrammetric and GIS-based approaches in the case of aerial photographs.