Use of ENVISAT ASAR Global Monitoring Mode to complement optical data in the mapping of rapid broad-scale flooding in Pakistan

Abstract

Envisat ASAR Global Monitoring Mode (GM) data are used to produce maps of the extent of the flooding in Pakistan which are made available to the rapid response effort within 24 h of acquisition. The high temporal frequency and independence of the data from cloud-free skies makes GM data a viable tool for mapping flood waters during those periods where optical satellite data is unavailable, which may be crucial to rapid response disaster planning, where thousands of lives are affected. Image differencing techniques are used, with pre-flood baseline image backscatter values being deducted from target values to eliminate regions with a permanent flood-like radar response due to volume scattering and attenuation, and to highlight the low response caused by specular reflection by open flood water. The effect of local incidence angle on the received signal is mitigated by ensuring that the deducted image is acquired from the same orbit track as the target image. Poor separability of the water class with land in areas beyond the river channels is tackled using a region-growing algorithm which seeks threshold-conformance from seed pixels at the center of the river channels. The resultant mapped extents are tested against MODIS SWIR data where available, with encouraging results

Operational Surface Water Detection and Monitoring Using Radarsat 2

“…In our experiments, using a single threshold resulted in missing water bodies in the far range, because there were no seeds captured: the threshold value excluded more pixels on open water in the far range, which usually has lower signal to noise ratios and higher calibration uncertainty. Work by O'Grady et al [21,46] also concluded that there are variations in the backscatter based on the incidence angle, given the varying noise levels of the scene. Another drawback of threshold-based methods for water mapping from SAR is that the threshold used for water mapping is severely affected by the noise (and noise floor) of the scene.…”

“…Thresholding techniques on SAR image for water mapping had been covered at length in the literature [4][5][6][7]20,21]. In its most simplistic form, thresholding consists of establishing a value below or equal to which pixel values are selected.…”

“…Due to its all-weather capabilities, and its image acquisition capacity during day or night or in cloudy conditions, SAR imagery offers better alternatives for water mapping than optical imagery [5]. As a result of the radar's unique response to water, water mapping using intensity thresholding methods on SAR image had been extensively used [4][5][6][7]20,21]. In this method, water is separated from land in intensity images, but the accuracy of the results relies on the ability to differentiate land vs. water pixels in the intensity domain, which becomes especially difficult when the water backscatter is affected by wind-induced roughness.…”

“…Since the values of backscatter vary depending on the incidence angle, image quality and wind-induced roughness, the threshold needs to be modified on a scene per scene basis [4][5][6]. In some other instances, image thresholding is combined with texture information [7] or region-growing segmentation algorithms used to determine the water pixels [14,16,21,22]. Segmentation algorithms are quite successful but computationally expensive.…”

Automated global water mapping based on wide-swath orbital synthetic-aperture radar

“…Mason et al (2010) mention the problem of misclassification due to topography, vegetation or canopy. O'Grady et al (2011) conclude that misclassification due to low backscatter values from non-flooded areas can be reduced via image differencing approaches. Matgen et al (2011) and Giustarini et al (2012) present a method relying on the calibration of a statistical distribution of "open water" backscatter values inferred from SAR images of floods.…”