IEEE Geoscience and Remote Sensing Magazine - March 2020 - 129

Congo floodplain is governed more by local topography
and shows less channel connectivity. In addition, the Congo floodplain does not have well-defined boundaries and
is broadly distributed. Poncos et al. [63] measured 2h/2t
across the Danube delta by using PALSAR InSAR. 2h/2t
maps were used to create water-flow directions based on
the mass-continuity theory, which revealed the spatialtemporal complexity of the flow dynamics.
Yuan et al. [64] mapped multitemporal 2h/2t in the Congo wetlands with ALOS PALSAR interferograms and Envisat
altimetry data. They found that, during the low-water season, 2h/2t with a magnitude of 1 dm occurred mostly at the
wetland-upland boundary and the distal wetlands (that is,
those far from the river), which suggested that there was no
hydraulic connectivity with the river. During the high-water
season, 2h/2t at the proximal wetlands (those close to the
river) became higher than the distal wetlands, with magnitude of up to 2 m, which suggested that water flowed mostly
from the floodplain to the river, as illustrated in Figure 8.
Jaramillo et al. [65] assessed the hydrologic connectivity in
the ungauged Ciénaga Grande de Santa Marta wetlands using
ALOS PALSAR interferograms. They found that the river inflow could explain fewer than 17% of the water-level changes
and that man-made infrastructure, such as channels and
roads, played a barrier role by reducing hydrologic connectivity. Palomino-Ángel et al. [66] analyzed the Atrato River wetlands, again with ALOS PALSAR interferograms. They found
that the river played different roles in flooding the wetlands.
Its level had more influence over flooding in the lower basin
than the middle one, where topography and superficial channels influenced the water flow and connectivity.
InSAR has been used to characterize water-flow dynamics in managed and natural wetlands. For example,
Wdowinski et al. [17] mapped the 2h/2t pattern for the
managed and natural-flow wetlands in the Everglades using JERS-1 L-band SAR images. In the managed wetlands,
the InSAR fringes are organized and follow patterns that
reflect water-level changes caused by gate operations. In
the natural-flow areas, the fringes become irregular and
have a low fringe rate due to the naturally diffuse water
flow from ocean tides.
Coastal wetlands inundated by ocean tides have been
characterized through InSAR measurements. For example,
Wdowinski et al. [67] applied InSAR to map the 2h/2t pattern that results from tide propagation in the coastal wetlands in the southern Everglades. The spatially detailed
2h/2t demonstrates that the tide propagates much faster in
the channel than the vegetation zone. It has been shown
that the tide flow can impact coastal wetlands that expand
2-3 km on both sides of the channel and 3-4 km at the
end of it. Cabrera and Wdowinski mapped 2h/2t in Louisiana coastal wetlands using ALOS PALSAR and Radarsat-1
images [62]. They showed that the horizonal tidal-flow
extent is limited to 5-15 km in the wetlands and that the
vertical 2h/2t reaches up to 30 cm. The high spatial resolution of InSAR-derived 2h/2t enabled the identification of
MARCH 2020

IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE

man-made infrastructure (for example, canals and roads)
that disrupted the ocean-tide flow.
On the other hand, InSAR has been useful for characterizing the flows in wetlands fed by ground water. Gondwe
et al. [68] applied Radarsat-1 SAR images to map 2h/2t in
Mexico's Sian Ka'an wetlands. The fringe pattern corresponded to the water-flow direction and local-scale water
divides. The authors also found that the wetlands' main water input was ground water transported from the catchment
area via faults. Kim et al. [69] analyzed PALSAR InSAR data
over the Great Dismal Swamp located across South Virginia
and North Carolina to measure the relative surface waterlevel changes and estimate groundwater changes.
WATER-STORAGE CHANGES
Water-storage changes can be estimated using InSAR measurements, such as

TS = / A ) 2h/2t , (11)

where A is the area of each InSAR pixel and TS represents
water-storage changes.
Alsdorf et al. [12] mapped 2h/2t in one large Amazon lake
using JERS-1 interferograms and converted 2h/2t to a storage
change of 0.28 km3. Later, Alsdorf used SIR-C interferograms
and found that the one-day-interval 2h/2t in the Amazon
wetlands correlated with its distance to the main channel
[70]. That functional relationship was used to extrapolate
2h/2t beyond the SIR-C swath for the central Amazon wetlands. The TS estimated from the extrapolated 2h/2t was
roughly 30% less than the estimation derived from the floodrouting models.
These studies suggest that the estimation of TS across
wetlands from flood-routing models has great errors compared to TS estimates from InSAR. However, InSAR has
been used to estimate TS only between acquisition timings. It has not been used to estimate seasonal and interannual TS. This is mostly due to SAR images' limited temporal resolutions. To overcome this obstacle, Yuan et al. [13]
developed a method to estimate water storages by modeling the relationships between depths and storages using
2h/2t mapped from InSAR. This depth-storage relationship
was used to estimate water storages across the Congo wetlands, from 2002 to 2011, through Envisat radar-altimetry
measurements. The method mapped 2h/2t at high-water
season and assumed that the water surface changed homogeneously across the wetlands, as illustrated in Figure 9.
The presumption was almost true for the Congo wetlands.
However, its application can be limited if 2h/2t is spatially
heterogeneous, as in the Amazon wetlands [9].
FORESTED-WETLANDS TOPOGRAPHY
Wetland topography is one of the major factors that controls
2h/2t and a dominant element of wetland hydrodynamic
modeling [71]. Yuan et al. [72] developed a method that
combined InSAR and radar altimetry to map the bare-Earth
topography of Congo's forested wetlands, with improved
129

IEEE Geoscience and Remote Sensing Magazine - March 2020

Table of Contents for the Digital Edition of IEEE Geoscience and Remote Sensing Magazine - March 2020