Using Sentinel-2 multispectral imagery to describe wet channel extension in temporary rivers

Abstract:

The knowledge of the hydrologic regime of temporary rivers in Mediterranean region, is very important for integrated river management and biodiversity conservation. However, almost all of small streams are ungauged and there is a lack of field data and observations. Public domain multispectral data, such as Sentinel-2 and Landsat, provide cost-effective tools for river monitoring at multidecadal time scales, over large spatial domains and with high temporal frequency. Literature works showed that these data are effective to quantify planform changes, meanders migration, and sand bar evolution of large rivers such as those in the tropical Amazon, and also medium sized rivers (e.g. Fiume Po in Italy). The main idea of this work is to demonstrate the potential of Sentinel-2 images for the characterization of the hydrological regime in small Mediterranean streams.

The Sentinel-2 satellite (launched in 2015) acquires thirteen channels in the visible, near infrared (VNIR), and short-wave infrared (SWIR), with a spatial resolution of 10 m and temporal frequency of about 5 days. The actual revisit time of a specific target increases in case of cloud cover during the acquisitions. The analysis of the reflectivity of surface covered by water, in different spectral bands, allowed to identify the specific bands in which it is possible to distinguish these elements from the surrounding scene. An automatic unsupervised classification algorithm was then developed to discriminate the wet channel. The comparison with field observations, carried out on the same days of images acquisition, showed a good ability to discriminate the water presence when the wet channel width or the pool length are bigger than the image pixels. It was also possible to detect the presence of water when the water surface was slightly below the image resolution. The developed tool can be used to assess the effect of peak flows on channel dynamics and to monitor the plant colonization of river bed. It is also possible to use the derived information for the classification of flow regime (e.g., perennial, ephemeral or intermittent) by using the intensity of intermittency as a metric to group river reaches.