15.2 Study Area and Data
The study area extends over 70 % of the Province of New Brunswick (NB) in
eastern Canada, from NB’s southern coast along the Bay of Fundy to northcentral
NB (Fig. 15.1). The area is characterized by its temperate evergreen-deciduous mix,
Table 15.1 “Multi-temporal SAR for SWC change detection” approaches used in the past
Source
Approach
a
Shoshany
et al. (2000)
Introduced multi-temporal backscatter ratios, such as the simple ratio
(SR) and normalized radar backscatter soil moisture index (NBMI); used
ERS SAR images over humid to semi-arid regions of Israel. Obtained
strong relations for both ratios (r
2
>85 %). However, the NBMI produced
a stronger relationship with SWC in the 20–40 % range.
Wagner and Scipal
(2000)
Determined a relative measure of SWC from dry, wet, and instantaneous
values of radar backscattering coefficient, σ
0
; used ERS Scatterometer
images over western Africa. Demonstrated promising qualitative results
with soil water index (SWI) over wet-dry climatic zones.
Wickel et al. (2001) Established relations between σ
0 with SWC; used RADARSAT SAR
images over the Southern Great Plains 1997 Hydrology Experiment Sites
in Oklahama, USA. Observed strong correlation for wheat stubble fields
(r
2 ¼ 89 %) and no correlation for pasture fields. Demonstrated potential
of using ScanSAR modes of RADARSAT for multi-temporal estimates of
SWC (e.g., Boisvert et al. 1996).
Lu and Meyer
(2002)
Used a correlation image computed from radar-image pairs for explaining
observed changes in SAR intensity; used ERS SAR over southeast New
Mexico, USA. Correlated SWC increments (within the 5–20 % range)
with increments in SAR intensity.
Thoma et al. (2006) Employed four approaches; i.e., empirical, physical, semi-empirical, and
image difference-based (i.e., delta index; DI) approaches; used ERS SAR
and RADARSAT SAR images over Southern Arizona, USA. DI produced
strong correlations with SWC (r
2
¼91 %) and provided overall better
results in comparison to the other methods considered.
Pathe et al. (2009)
Calculated a relative SWC as a function of dry, wet, and instantaneous
σ
0 from ENVISAT ASAR acquired over Oklahoma, USA. ASAR-derived
SWC were compared against ground-based measurements of SWC and
found that in 75 % of the cases, the standard deviation fell within 13–27 %
of field-based SWC.
Baghdadi
et al. (2011)
Used a change-detection technique between data acquired during rainy
and dry seasons (considered as the reference images) to retrieve SWC.
Derived SWC-values were approximately 2.3 % (RMSE) from groundbased measurements.
Qin-Xue and You
(2013)
Applied approaches similar to those of Wagner and Scipal (2000) and
Pathe et al. (2009) to calculate relative SWC using ENVISAT ASAR-data
over Hubei Province, China. Demonstrated reasonable correlation for
cotton fields (r
2
¼78 %).
a
SR-ratio of σ
0 from two different dates; NBMI-a function of instantaneous σ
0 and σ
0 for dry soil;
SWI-a function of actual SWC, wilting point and field capacity; DI-a function of instantaneous
σ
0 (i.e., wet condition) and σ
0 for the same soil under dry conditions; ASAR-advanced Synthetic
Aperture Radar
15 Development of a New Wetness Index Based on RADARSAT-1 ScanSAR Data
303
The study area extends over 70 % of the Province of New Brunswick (NB) in
eastern Canada, from NB’s southern coast along the Bay of Fundy to northcentral
NB (Fig. 15.1). The area is characterized by its temperate evergreen-deciduous mix,
Table 15.1 “Multi-temporal SAR for SWC change detection” approaches used in the past
Source
Approach
a
Shoshany
et al. (2000)
Introduced multi-temporal backscatter ratios, such as the simple ratio
(SR) and normalized radar backscatter soil moisture index (NBMI); used
ERS SAR images over humid to semi-arid regions of Israel. Obtained
strong relations for both ratios (r
2
>85 %). However, the NBMI produced
a stronger relationship with SWC in the 20–40 % range.
Wagner and Scipal
(2000)
Determined a relative measure of SWC from dry, wet, and instantaneous
values of radar backscattering coefficient, σ
0
; used ERS Scatterometer
images over western Africa. Demonstrated promising qualitative results
with soil water index (SWI) over wet-dry climatic zones.
Wickel et al. (2001) Established relations between σ
0 with SWC; used RADARSAT SAR
images over the Southern Great Plains 1997 Hydrology Experiment Sites
in Oklahama, USA. Observed strong correlation for wheat stubble fields
(r
2 ¼ 89 %) and no correlation for pasture fields. Demonstrated potential
of using ScanSAR modes of RADARSAT for multi-temporal estimates of
SWC (e.g., Boisvert et al. 1996).
Lu and Meyer
(2002)
Used a correlation image computed from radar-image pairs for explaining
observed changes in SAR intensity; used ERS SAR over southeast New
Mexico, USA. Correlated SWC increments (within the 5–20 % range)
with increments in SAR intensity.
Thoma et al. (2006) Employed four approaches; i.e., empirical, physical, semi-empirical, and
image difference-based (i.e., delta index; DI) approaches; used ERS SAR
and RADARSAT SAR images over Southern Arizona, USA. DI produced
strong correlations with SWC (r
2
¼91 %) and provided overall better
results in comparison to the other methods considered.
Pathe et al. (2009)
Calculated a relative SWC as a function of dry, wet, and instantaneous
σ
0 from ENVISAT ASAR acquired over Oklahoma, USA. ASAR-derived
SWC were compared against ground-based measurements of SWC and
found that in 75 % of the cases, the standard deviation fell within 13–27 %
of field-based SWC.
Baghdadi
et al. (2011)
Used a change-detection technique between data acquired during rainy
and dry seasons (considered as the reference images) to retrieve SWC.
Derived SWC-values were approximately 2.3 % (RMSE) from groundbased measurements.
Qin-Xue and You
(2013)
Applied approaches similar to those of Wagner and Scipal (2000) and
Pathe et al. (2009) to calculate relative SWC using ENVISAT ASAR-data
over Hubei Province, China. Demonstrated reasonable correlation for
cotton fields (r
2
¼78 %).
a
SR-ratio of σ
0 from two different dates; NBMI-a function of instantaneous σ
0 and σ
0 for dry soil;
SWI-a function of actual SWC, wilting point and field capacity; DI-a function of instantaneous
σ
0 (i.e., wet condition) and σ
0 for the same soil under dry conditions; ASAR-advanced Synthetic
Aperture Radar
15 Development of a New Wetness Index Based on RADARSAT-1 ScanSAR Data
303