The Use of Radar Imagery in Riverine Flood Inundation Studies - Fluvial Remote Sensing for Science and Management - page 123

Table 6.4 Reported water level retrieval techniques based on remotely sensed flood extent and DEM fusion and their accuracies
(After Di Baldassarre et al., 2011).

Method Error (RMSE) Validation Data Source

Landsat TM-derived flood extent
superimposed on topographic
contours for volume estimation

± 21 percent Field data Gupta and

Banerji (1985)

ERS SAR flood extent overlain on
topographic contours

0.5 m-2 m Field data Oberstadler et al.

(1997)

ERS SAR flood extent overlain on
topographic contours

up to 2 m Model outputs Brakenridge et al.

(1998)

Inter-tidal area water from multiple
ERS images superimposed onto
simulated water heights

mean error of 0.2-0.3 m Field data Mason et al.

(2001)

Flooded vegetation maps from
combined airborne Land C-SAR
integrated with LiDAR vegetation
height map

around 0.1 m Field data Horritt et al.

(2003)

Integration of high-resolution
elevation data with event wrack
lines

< 0.2 m Model outputs Lane et al. (2003)

Fusion of RADARSAT-1 SAR flood
edges with LiDAR

correlation coefficient of 0.9 TELEMAC-2D model

outputs

Mason et al.
(2003)

Complex fusion of flood aerial
photography and field based
water stages from various
floodplain structures

0.23 m Mean between

maximum and
minimum estimation

Raclot (2006)

Fusion of ENVISAT ASAR flood
edges with LiDAR and
interpolation modelling

0.4-0.7 m Field data Matgen et al.

(2007)

Fusion of ENVISAT ASAR flood
edges with LiDAR and regression
modelling

< 0.2 m Field data Schumann et al.

(2007)

Fusion of ENVISAT ASAR flood
edges with LiDAR/topographic
contours/SRTM and regression
modelling

< 0.35 m/0.7 m/1.07m 1D model outputs Schumann et al.

(2008)

Fusion of hydraulically sensitive
flood zones from ENVISAT
ASAR imagery and LiDAR

0.3 m; 0.5 m uncertainty Mean between

maximum and
minimum estimation

Hostache et al.
(2009)

Fusion of ERS SAR flood edges from
active contour modelling with
LiDAR

mean error of up to 0.5 m Aerial photography Mason et al.

(2009)

Fusion of TerraSAR-X flood edge
with LiDAR (rural area)

error of 1.17 m (note that
relative changes in levels from
TerraSAR-X and aerial
photography were mapped
with an accuracy of 0.35 m
compared to gauge data)

One field gauge Zwenzner and

Voigt (2009)

Fusion of uncertain ENVISAT
ASAR WSM flood edges with
SRTM heights

error in median estimate of
0.8 m

LiDAR derived water
levels

Schumann et al.
(2010a)