ArticlePDF Available

Mapping Dynamic Water Fraction under the Tropical Rain Forests of the Amazonian Basin from SMOS Brightness Temperatures

Authors:

Abstract and Figures

Inland surface waters in tropical environments play a major role in the water and carbon cycle. Remote sensing techniques based on passive, active microwave or optical wavelengths are commonly used to provide quantitative estimates of surface water extent from regional to global scales. However, some of these estimates are unable to detect water under dense vegetation and/or in the presence of cloud coverage. To overcome these limitations, the brightness temperature data at L-band frequency from the Soil Moisture and Ocean Salinity (SMOS) mission are used here to estimate flood extent in a contextual radiative transfer model over the Amazon Basin. At this frequency, the signal is highly sensitive to the standing water above the ground, and the signal provides information from deeper vegetation density than higher-frequencies. Three-day and (25 km × 25 km) resolution maps of water fraction extent are produced from 2010 to 2015. The dynamic water surface extent estimates are compared to altimeter data (Jason-2), land cover classification maps (IGBP, GlobeCover and ESA CCI) and the dynamic water surface product (GIEMS). The relationships between the water surfaces, precipitation and in situ discharge data are examined. The results show a high correlation between water fraction estimated by SMOS and water levels from Jason-2 (R > 0.98). Good spatial agreements for the land cover classifications and the water cycle are obtained.
Content may be subject to copyright.
water
Article
Mapping Dynamic Water Fraction under the Tropical
Rain Forests of the Amazonian Basin from SMOS
Brightness Temperatures
Marie Parrens 1,*, Ahmad Al Bitar 1, Frédéric Frappart 2,3 , Fabrice Papa 2,4, Stephane Calmant 2,
Jean-François Crétaux 2, Jean-Pierre Wigneron 5and Yann Kerr 1
1Centre d’Etudes Spatiales de la BIOsphère (CESBIO—Université de Toulouse, CNES, CNRS, IRD),
UMR5126, BPI 2801, 31401 Toulouse CEDEX 9, France; ahmad.albitar@cesbio.cnes.fr (A.A.B.);
yann.kerr@cesbio.cnes.fr (Y.K.)
2Laboratoire d’Etudes en Géophysique et Océanographie Spatiales (LEGOS), UMR5566, Université de
Toulouse, CNES, CNRS, IRD, Observatoire Midi-Pyrénées (OMP), 14 Avenue Edouard Belin, 31400 Toulouse,
France; frederic.frappart@legos.obs-mip.fr (F.F.); fabrice.papa@ird.fr (F.P.); stephane.calmant@ird.fr (S.C.);
jean-francois.cretaux@legos.obs-mip.fr (J.-F.C.)
3Géosciences Environnement Toulouse (GET), UMR5563, Université de Toulouse, CNES, CNRS, IRD,
Observatoire Midi-Pyrénées (OMP),14 Avenue Edouard Belin, 31400 Toulouse, France
4Indo-French Cell for Water Sciences (IFCWS), IRD-IISc-NIO-IITM Joint International Laboratory,
Bangalore 560012, India
5INRA, UMR 1391 ISPA, F-33140 Villenave d’Ornon, Bordeaux, France; wigneron@bordeaux.inra.fr
*Correspondence: marie.parrens@cesbio.cnes.fr; Tel.: +33-5-6155-8531
Academic Editor: Y. Jun Xu
Received: 23 February 2017; Accepted: 11 May 2017; Published: 17 May 2017
Abstract:
Inland surface waters in tropical environments play a major role in the water and carbon
cycle. Remote sensing techniques based on passive, active microwave or optical wavelengths are
commonly used to provide quantitative estimates of surface water extent from regional to global
scales. However, some of these estimates are unable to detect water under dense vegetation and/or
in the presence of cloud coverage. To overcome these limitations, the brightness temperature data at
L-band frequency from the Soil Moisture and Ocean Salinity (SMOS) mission are used here to estimate
flood extent in a contextual radiative transfer model over the Amazon Basin. At this frequency, the
signal is highly sensitive to the standing water above the ground, and the signal provides information
from deeper vegetation density than higher-frequencies. Three-day and (25 km
×
25 km) resolution
maps of water fraction extent are produced from 2010 to 2015. The dynamic water surface extent
estimates are compared to altimeter data (Jason-2), land cover classification maps (IGBP, GlobeCover
and ESA CCI) and the dynamic water surface product (GIEMS). The relationships between the water
surfaces, precipitation and in situ discharge data are examined. The results show a high correlation
between water fraction estimated by SMOS and water levels from Jason-2 (R > 0.98). Good spatial
agreements for the land cover classifications and the water cycle are obtained.
Keywords: water fraction extent; L-band; Amazon Basin
1. Introduction
Terrestrial surface water covers only about 5% of the Earth’s ice-free land surface [
1
,
2
], but plays
a key role in global biogeochemistry, hydrology and wildlife diversity [
3
,
4
]. Consequently, it is critical
to monitor the distribution of terrestrial water at large spatial and high temporal scales [
5
8
]. The work
in [
9
] estimates that nearly two-thirds of all terrestrial freshwater wetlands disappeared between 1997
and 2011. A more recent study used three million Landsat images to provide a high resolution map of
Water 2017,9, 350; doi:10.3390/w9050350 www.mdpi.com/journal/water
Water 2017,9, 350 2 of 26
surface water extent [
10
]. The authors of this study estimated that 90,000 km
2
of permanent surface
water had disappeared between 1984 and 2015.
Several different methods based on mapping water bodies from remote sensing datasets were used
since the development of the Earth’s space observations: (1) visible; (2) infrared; (3) active microwave;
(2) passive microwave and; (4) hybrid approach (passive and active microwave). Each method offers
varying degrees of success in providing quantitative estimates of wetlands and inundation extents.
Water surface can be sensed by optical remote sensing methods. These methods typically exploit
the absorption of longer wavelengths of light in water, especially the near and shortwave infrared
parts of the electromagnetic spectrum [
11
,
12
]. Optical remote sensing provides very accurate mapping
of water bodies. For example, [
13
] senses lakes with a spatial resolution of 15 m, whereas [
14
] sensed
global water bodies at 30-m resolution using the Landsat data. The majority of the studies using optical
remote sensing for water bodies’ detection provided only one snapshot of the hydrology stage. Due to
the low revisit time of the optical sensors, few maps of a large area are available, and the minimum
and/or the maximum of the flooded area are not always observed. The detection of sudden changes
impacting the hydrologic cycle [
10
] is also not sensed with accuracy. These limitations are crucial
issues for hydrology application. However, some studies [
15
,
16
] managed to follow the temporal
dynamics of the water surface in specific places. The most important limitation of the optical sensors is
their inability to penetrate clouds and dense vegetation cover, which is essential during tropical wet
seasons over the Amazon Basin.
Active microwave (scatterometers and Synthetic Aperture Radar (SAR)) is also sensitive to the
water surface and has the ability to penetrate clouds and, to a certain extent, vegetation. Open water
surfaces are generally characterized by low backscattering coefficients. Contrary to passive microwave,
the signal is more contaminated by the vegetation. The spatial resolution of scatterometers is about
25–50 km, whereas the SAR provides higher resolution, typically around 10–150 m. Several studies
have shown the ability of active microwave to map surface water at regional scales, such as over the
Amazon region [
17
] and over the Arctic region [
18
]. Satellite altimeters are radars that observe at nadir
to measure surface topography. They provide accurate measurements of water heights in rivers, lakes
and wetlands [
19
21
]. Due to their high spatial resolution, altimeters do not provide sufficient spatial
coverage to analyze the water bodies’ temporal dynamics, except in polar regions [
22
]. The future
Surface Water Ocean Topography (SWOT) mission [
23
] intended to be launched in 2021 is expected to
provide K-band SAR interferometry, enabling continental altimetry.
Passive microwaves are sensitive to the distribution of liquid water in the landscape; they can
operate day and night for all weather conditions. However, they are limited by a low spatial resolution
(approximately 30 km). They can sense only large wetlands or regions where the cumulative area
of small wetlands comprises a significant portion of the field of view. Consequently, they provide
the capability to map the temporal evolution of surface water over the land surface due to their high
temporal resolution. In previous studies, passive microwave measurements have shown the capability
to sense the dynamics of terrestrial surface water at coarse resolution [
24
31
]. The basic principle of
the surface water measurements based on passive microwave is explained by the difference of the
emissivity between the water and the soil. Flooding surfaces decrease the emissivity in both vertical (V)
and horizontal (H) polarization and increase the difference between the two polarizations, especially
at low frequencies. This approach produces ambiguous estimation of surface water over regions
with mixtures of open water and other complex surfaces (topography effects). The work in [
26
,
27
]
has extensively studied the inundation area over the Amazon Basin with the Scanning Multichannel
Microwave Radiometer (SMMR). However, their studies focus essentially on a restricted area close to
Manaus town from 1979 to 1987.
Hybrid approaches combine the strengths of different types of sensors. For example, altimetry
data are characterized by a high spatial resolution and a low temporal resolution and can be combined
with passive microwave data having low spatial resolution and a high temporal resolution to
obtain a product with both high temporal and spatial resolution. The Global Inundation Extent
Water 2017,9, 350 3 of 26
from Multiple-Satellites (GIEMS) products are based on merged data from passive (Special Sensor
Microwave/Imager (SMM/I)), active microwave (European Remote Sensing satellite (ERS)) and data
from an optical sensor (Advanced Very High Resolution Radiometer (AVHRR)) [32].
Table 1presents the major studies related to the observation and detection of water bodies from
space by using the techniques presented above. Visible and infrared remote sensing methods were
extensively used, but provided static maps of water bodies at the global scale or a dynamic map at
the regional scale [
15
,
33
]. A lack of studies concerning dynamic water surface extent from 2013 to the
present is clearly identified in this table.
The floodplains and wetlands of the Amazon River are important in terms of water volume and
in terms of fluxes between the land and the atmosphere. Mapping water fraction under the Amazon
tropical dense forest is challenging, but sensing water under dense vegetation remains a key issue in
the remote sensing scientific community.
In this study, we developed a method to map the temporal evolution of the water bodies at coarse
spatial resolution and weekly temporal resolution by using a microwave sensor at L-band (1.4 GHz)
called Soil Moisture Ocean Salinity (SMOS) over the Amazon Basin. The SMOS satellite operates at
L-band, and it was shown that this frequency is the most suitable, being less impacted by vegetation
than higher frequencies [
34
36
]. Originally, the SMOS satellite was dedicated to sense soil moisture
over land surfaces and the ocean salinity. The SMOS physical signal (brightness temperature) is highly
impacted by the presence of standing water over the ground.
Our motivation is to use a contextual radiative transfer model and a single dataset to estimate
the water fraction over the tropical basin. The area of study and the datasets used in this work are
presented in the Section 2and 3, respectively. Section 4presents the algorithm permitting retrieving
the water fraction extent from the SMOS data, and Section 5contains the results and the validation.
The discussion and conclusions are presented in Sections 6and 7.
Table 1.
List of selected scientific papers on the observation of the water surface over the continents
from space. References, area of study and sensors are shown.
Remote Sensing Approach Reference Area of Study Sensors Frequency
Passive microwave
[27] Amazon Basin SSMR Q-band
[29] Boreal regions SSM/I-SSMR K- and Ka-band
[37] North Eurasian AMSR-E-QSCAT C-band
Active microwave
See [38] for a review of the SAR technique
[17] Amazon Basin ENVISAT SAR L-band
[39] High latitude regions ENVISAT ASAR C-band
[40] Mekong basin ENVISAT ASAR C-band
[41] Global scale ENVISAT ASAR C-band
[22] Boreal regions Topex-Poseidon C-band
Hybrid approaches
[42] Global scale SSM/I, ERS, AVHRR Ka- and C-band
[32] Global scale SSM/I, ERS, AVHRR Ka- and C-band
[43] Global scale SSM/I, ERS, AVHRR Ka- and C-band
[37] Global scale SSM/I, SSMI/S, ERS, QSAT, ASCAT Ku- and C-band
Optical and infrared
[44] Okavango Delta AVHRR -
[45] Brahmaputra AVHRR -
[46] Inner Niger Delta MODIS -
[47] China MODIS -
[33] Mekong Delta MODIS -
[10] Global scale Landsat -
2. Study Areas
This study focuses on the Amazon Basin, which is the largest tropical basin with an area of
approximately 6,000,000 km
2
and contributes up to 15% of the global river discharge to the ocean
(approximately 200,000 m
3
s
1
discharge). With a sediment load of three million tons near its mouth [
48
]
and drainage area covers about 6,200,000 km
2
, almost 5% of all of the continental masses, the
Water 2017,9, 350 4 of 26
Amazon Basin is one of the most impressive hydrological basins of the world. The Amazon is
highly interconnected by floodplain channels, resulting in complex flow patterns. Figure 1presents the
Amazon Basin with the main rivers and floodplains. Covering more than 300,000 km
2
, the Amazon
extensive floodplains play a crucial role for global climate and biodiversity, but they are still poorly
monitored at a large scale, limiting our understanding of their role in flood hazard, carbon production,
sediment transport, nutriment exchange and air-land interactions. Surface water stored in floodplains
represents about half of the terrestrial water storage and 15–20% of the water that flowed out of
the Amazon
floodplains [4953].
Because it extends over two hemispheres, the Amazon region is
characterized by several rainfall regimes. Rainfall shows opposing phases between the Northern
and the Southern Hemisphere with a rainy season in austral winter in the Northern Hemisphere and
summer in the Southern Hemisphere. The rainfall shows a gradient from northwest to southeast with
decreasing rainfall amount and increasing length in the dry season. For the eastern part of the basin,
the rainy season occurs from March–May, and the dry season prevails from September–November.
For the northern regions, low rainfall seasonality is observed with wet conditions throughout the year.
For more information on the Amazon hydrological regime, see [5456].
Figure 1. Map of the Amazon Basin with the main rivers and floodplains.
3. Data
This section describes the data used to compute the water fraction extent from passive microwave
at L-band (SMOS data, topography data and skin temperature) and the data used to compare and
validate this product (precipitation data, static land cover maps, other dynamic water fraction products,
water level from altimetry and in situ river discharge data).
3.1. L-Band Brightness Temperatures from SMOS
The SMOS mission is a joint program of the European Space Agency (ESA), the Centre National
d’Etudes Spatiales (CNES) and the Centro para el Desarrollo Teccnologico Industrial (CDTI) in the
framework of the Earth Explorer Opportunity Mission initiative. It is the first satellite specifically
dedicated to soil moisture retrievals with a passive microwave radiometer at 1.4 GHz (L-band).
The physical signal of SMOS is the Brightness Temperature (TB). This signal is highly sensitive to
the water under the ground [
34
]. Clouds and rain have a negligible effect [
57
], and the atmospheric
contribution is limited and known [
34
]. The microwave signal is to a lesser extent sensitive to the
vegetation, but actually, the L-band signal is less impacted by the vegetation than higher frequencies.
Water 2017,9, 350 5 of 26
SMOS has a Sun-synchronous orbit at a 757-km altitude with a 06:00 LST ascending Equator
crossing time and an 18:00 LST descending Equator crossing time. The globe is fully imaged twice
every three days. The main innovative feature of SMOS is the capability for multi-incidence-angle
observations at full polarization across a 900-km swath. In this study, the SMOS Level (L) 3 TB
(RE04v300)
products [58]
produced by the Centre Aval de Traitement des Données SMOS (CATDS)
are used. These data are projected on the Equal-Area Scalable Earth (EASE) Grid 2 [
59
] with a spatial
resolution of
25 km ×25 km
. The main differences between the SMOS L3 TB and the other lower
levels of data are: (i) the L3 TB products are expressed at the top of the atmosphere over the terrestrial
reference frame (H and V); and (ii) they are bin averaged every 5
from 2.5
–62.5
. In the present
study, SMOS L3 TB were used from 2010–2015 over the Amazon Basin. Angles of 32
±
5
, 37
±
5
,
42
±
5
and 47
±
5
in both H and V polarization were considered to retrieve the water fraction over
the tropical basin. The SMOS data were downloaded from the CATDS servers (www.catds.fr).
3.2. Topography
The digital elevation model obtained by the Shuttle Radar Topography Mission (SRTM) [
60
] with
a spatial resolution of 30 arc sec(approximately 1km) was used over the Amazon Basin. These data
result from the Global 30 Arc-Second Elevation (GTOPO30) computed at the U.S. Geological Survey’s
EROS Data Center (USGS) and were available at https://Ita.cr.usgs.gov/GOTO30. The elevation
and topographic index maps over the Amazon watersheds were computed by averaging all of the
SRTM elevation values present in an SMOS pixel (Figure 2). These data are used to flag areas with an
elevation higher than 500 and/or a topographic index indicated as moderate in the SMOS flag.
3.3. Skin Temperature
The surface skin temperature produced by the European Centre for Medium-range Weather
Forecasting (ECMWF) was used in this study. This product was obtained by the SMOS L3 preprocessor,
which computed the spatiotemporal average of the ECMWF reanalysis products on the EASE 2.0 grid.
3.4. Precipitation Data
Precipitation measured by the Tropical Rainfall Measuring Mission (TRMM) were used over
the entire Amazon Basin from 2010–2015. TRMM is a joint mission between NASA and the Japan
Aerospace Exploration (JAXA) Agency and provides rainfall estimates at 0.25
×
0.25
spatial resolution.
The TRMM-3B42 product [
61
] uses microwave data to calibrate the infrared-derived estimates and
creates estimates that contain microwave-derived rainfall estimates when and where microwave data
are available and the calibrated infrared estimates where microwave data are not available.
3.5. Static Land Cover Maps
The International Geosphere-Biosphere Programme (IGBP), the GlobeCover land cover classification
and the ESA CCI maps were used. The IGBP land cover map was obtained using images from
the Moderate Resolution Imaging Spectroradiometer (MODIS) with a spatial resolution of 0.005
during 2001–2012 [
62
]. The GlobeCover land surface map is based on data from the Advanced Very
High Resolution Radiometer (AVHRR) data and is at a 1-km spatial resolution. Data were acquired
during 1992–1993 [
63
]. Recently, a new release of the ESA Climate Change Initiative (CCI) Land
Cover map was made available [
64
]. The new water/no water global mask at 150 m was built on
previous achievements using SAR systems and further improved thanks to a combination with recent
Landsat-derived products. This dataset was based on acquisitions from the years 200–2012. Data can
be downloaded at: http://www.esa-landcover-cci.org/?q=node/162. For the three products, the
water classes were aggregated and re-sampled on the EASE v2.0 grid to obtain the water fraction (%)
present in each cell of the EASE v2.0 grid. The three products are static and are presented in Figure 2.
Water 2017,9, 350 6 of 26
Figure 2.
(
a
) Elevation (in meters) of the Amazon Basin from the SRTM data rescaled in the EASE v2.0
grid; spatial distribution of the water surface from: (
b
) ESA CCI; (
c
) IGBP; (
d
) Globe Cover; (
e
) average
inundation extent from GIEMS from 1993–2007 over the Amazon Basin; and (
f
) average inundation
extent from Surface Water Microwave Product Series (SWAMPS) from 2010–20over the Amazon Basin.
3.6. Dynamics and Climatoloy of Water Fraction Data
The GIEMS products provided a long-term global map of inundation at coarse resolution by
merging passive and active data (SSM/I, ERS, AVHRR) from 1993–2007 at monthly time steps.
The GIEMS product is gridded on an equal area grid of 0.25
×
0.25
at the Equator. Over highly
vegetated areas, the GIEMS product has some limitations. For example, over the Amazon Basin,
the GIEMS product has a tendency to overestimate higher inundation fractions [
32
]. This product is
fully described in [
42
]. These data were subsequently employed in estimating surface water storage
variations in large river basins [
50
,
65
,
66
]. To be compared with our data, the GIEMS product was
Water 2017,9, 350 7 of 26
averaged during the full period and considered as a static climatological product. Figure 2shows the
temporal GIEMS average over the Amazon Basin from 1993–2007.
The recent Surface Water Microwave Product Series (SWAMPS) data provided daily surface
water globally at 25-km resolution from 1992–2013. This product was based on the combination of
passive and active microwave sensors (SSM/I, SSMIS, ERS, QuikSCAT, ASCAT) and visible sensors
(MODIS). The data are described in detail and validated in [
37
]. In our study, only data from 2010–2013,
coinciding with SMOS data availability, over the Amazon Basin were considered. Figure 2shows the
temporal SWAMPS average over the Amazon Basin from 2010–2013.
3.7. Water Level from Satellite Altimetry
In the Amazon Basin, water level (in meters) time series for virtual stations calculated from the
Jason-2 altimeter satellite over the period 2008–2012 were downloaded from the Hydroweb database
(http://hydroweb.theia-land.fr/). The Jason-2 satellite was launched on 20 June 2008 in the follow-on
mission to the Jason-1 satellite (2002–2008, CNES/NASA). It operates at Ku-band (13.575 GHz) and
C-band (5.3 GHz) and has a time period of
9.9 days [
67
]. The water level computation method and
the location of the virtual stations are presented in [
68
]. In the present study, water level time series
over 83 virtual stations from 2010–2012 inclusive were used.
3.8. In Situ River Discharge Data
Monthly in situ discharge (m
3
/s) observations for the Amazon River were obtained from the
Obidos gauge station (1
00’ S, 55
00’ W) for the period of January 2010–March 2014. Data are available
on Hidroweb (http://www.hidroweb.ana.gov.fr) from the Brazilian water agency.
4. Methods
This section describes the approach used to derive surface water extent (expressed in %) contained
in an SMOS pixel from the SMOS L3 TB. The contextual radiative transfer model and the selection
of two reference points are described below, followed by the statistical method used to compare and
validate the SWAF product.
4.1. Contextual Radiative Transfer Model to Retrieve the Water Fraction under Dense Forests
4.1.1. Description of the SWAF Algorithm
It is well known that the microwave emission is highly impacted by the dielectric constant [
57
].
In the present algorithm, we assumed that over the tropical basins, the pixels are only composed of
two contributions: the water and the forest, such as represented in Figure 3.
Figure 3.
(
a
) Pixel representation with the two contributions: forest and water; (
b
) location of the
“water”, the “forest” and the mixed pixels.
Water 2017,9, 350 8 of 26
Therefore, for a given pixel, the TB for the entire pixel is the sum of the water contribution and
the forest contribution. The water contribution is the TB of the water weighted by the fraction of the
pixel flooded, called the Surface WAterFraction (SWAF). In the same way, the contribution of the forest
is the TB of the forest weighted by the part of the pixel not flooded (1-SWAF). The TB of the total pixel
can be expressed as:
TB(θ,p)tot =SWAF(θ,p)T B(θ,p)water + (1SWAF(θ,p))TB(θ,p)f ores t (1)
where
θ
is the incidence angle,
p
the polarization (
p
= H or V),
TB(θ
,
p)tot
is the TB of the total pixel,
TB(θ
,
p)water
the TB of the water and
TB(θ
,
p)f orest
the TB of the forest. The fraction of the water present
in an SMOS pixel depends on both the incidence angle and the polarization and can be expressed as:
SWAF(θ,p) = TB(θ,p)tot TB(θ,p)f ore st
TB(θ,p)water T B(θ,p)f o rest
(2)
TB(θ
,
p)tot
is the brightness temperature observed by the SMOS satellite. This observation is done
over each pixel over the Amazon Basin. However, the
TB(θ
,
p)f orest
values are extracted over a selected
pixel located at 2.137
S, 60.803
W (Figure 3) and composed exclusively of forest. The
TB(θ
,
p)f orest
are interpolated in time from the ascending SMOS overpass. In the same way, the
TB(θ
,
p)water
time
series are computed over a selected pixel located close to Obidos (2.142
S, 55.449
W) and composed
of more than 80% water (Figure 3). Indeed, an SMOS pixel composed only of freshwater cannot been
found (water from the ocean is salty, which modifies the emissivity). Therefore, the
TB(θ
,
p)water
has
been computed as the product of the emissivity and the skin temperature provided from the ECMWF.
The Klein and Swift model [
69
] has been used to calculate the water emissivity. An average of the
TB(θ
,
p)water
over the full period is computed to add more stability to the model. The
TB(θ
,
p)water
is
the only contribution that is constant in time in the model. Details about the value and the time series
of both the TB(θ,p)f ores t and TB(θ,p)water are presented in the next section.
At this stage, an illustration is needed for a better comprehension of the algorithm. Figure 4presents
the time series of the
TB(θ
,
p)water
,
TB(θ
,
p)f orest
and a pixel annually flooded considered as “mixed”
i.e., composed of both forest and water (TB(θ,p)mixed) located in Figure 3.
Figure 4.
Time series of TB at H polarization (left) and V polarization (right) at two incidence angles
(32±5and 47±5): the “water”, “forest” and “mixed” pixels.
As shown in Figure 4, the
TB(θ
,
p)
values over the “forest”, the “water” and the “mixed” pixels
differ slightly following the angles and the polarization. For all of the polarization and the incidence
angles, the
TB(θ
,
p)
values of the “water” pixel are the lowest, whereas the values of
TB(θ
,
p)
over the
“forest” pixel are highest. The values of
TB(θ
,
p)
over the “mixed” pixel are included between the two
contributions. Over the six years, the values of TB(θ,p)water and TB(θ,p)f orest are very constant with
time. This is not the case for the time series of
TB(θ
,
p)mixed
, which shows annual cycles due to the
Water 2017,9, 350 9 of 26
annual inundation of the Amazon River. For the “mixed” pixel, at each date, Equation (2) is computed
to obtain the SWAF value. This method is generalized over all of the pixels of the Amazon Basin (the
localization of the “mixed” pixel moved, but the water and forest reference pixels are always the same).
The
SWAF(θ
,
p)
data are computed each day with the SMOS ascending overpass and smoothed with
a sliding window of 17 days.
The passive interferometric technique has some limitations over the areas with complex
topography [
70
]. To avoid artifacts, SWAF data are not computed over areas with moderate topography
according to the SMOS flag. Moreover, pixels with a topography index estimated as moderate or
higher in the SMOS flag over the Amazon Basin were not considered.
4.1.2. Forest and Water TB Reference
As explained in the previous section, the time series of TB values over the “water” pixel are
averaged in time to be sure of the TB(θ,p)water stability. Results for each angle and polarization used
in the algorithm are presented in Table 2. Lower values of TB are obtained in H polarization than in V
polarization. In H polarization, the mean TB over the “water ” pixel decreases with the increase of the
incidence angles. The reverse is observed in V polarization. The signal over the “water” pixel is really
stable during the full period as shown by the standard deviation, which does not exceed 1 K.
To compare the value obtained over the “water” pixel, the TB over the “forest” pixel are also
averaged. However, note that these values are not used in the SWAF algorithm. For the “forest”
pixel, the mean TB values increase for decreasing incidence angles, in both H and V polarization.
This behavior is more marked in H polarization. The standard deviation of the TB over the “forest”
pixel is higher than that observed over the “water” pixel, but does not exceed 4 K. Lower standard
deviations are obtained in V polarization than in H polarization, except at 32±5.
Table 2.
Average and standard deviation (
σ
) of the TB over the “water” and “forest” pixel in both H
and V polarization and at four incidence angles (32±5, 37±5, 42±5, 47±5).
“Water” Pixel “Forest” Pixel
Incidence Angle H-pol V-pol H-pol V-pol
Mean (K) σ(K) Mean (K) σ(K) Mean (K) σ(K) Mean (K) σ(K)
32±594.52 0.51 122.58 0.64 274.43 2.71 276.61 2.72
37±589.96 0.49 128.25 0.67 272.44 2.94 276.12 2.61
42±584.72 0.46 135.27 0.70 271.88 3.57 275.72 2.71
47±578.78 0.43 143.93 0.74 269.71 3.22 274.26 2.51
4.2. Statistic Scores’ Computation
In this study, we use a common set of skill scores: (i) the coefficient of correlation (
r
); (ii) the
p
-value; (iii) the cross-correlation; (iv) the bias; and (v) the Root Mean Square Error (RMSE) value.
The Pearson correlation coefficient (
r
) is used to compare the dynamic behavior of the SWAF data (
x
)
with the dynamic evolution of other variables (y):
r=n
i=1(xix)(yiy)
q(n
i=1(xix)2q(n
i=1(yiy)2with x=1
n
n
i=1
xiand y=1
n
n
i=1
yi(3)
with
n
the number of elements in the
x
and
y
series. Associated with the
r
, the
p
-value is also computed
for the null hypothesis. The authors consider that for a
p
-value higher than 0.05, correlation values are
not significant.
The cross-correlation measures the similarity of two time series (
x
and
y
) as a function of the
displacement of one relative to the other. In this study, the displacement corresponds to the time
(in months). Therefore, the cross-correlation value is the higher correlation value obtained if the
x
time
Water 2017,9, 350 10 of 26
series is moved by nmonths with respect to the ytime series.
The bias between two series (xand y) is defined as:
bias =
n
i=1
(xiyi)
n(4)
and the RMSE value is usually used to define the accuracy of the data. It is computed as:
RMSE =rn
i=1(xiyi)2
n(5)
5. Results over the Amazon Basin
This section provides the SMOS water fraction results and analysis with a focus on the comparison
and the validation of the product using a set of multi-source datasets described in Section 3.
5.1. Spatial Patterns and Temporal Dynamics of the SWAF Maps
This section described the spatial and temporal behavior of the water surface extent estimated by
SMOS for the four angles and the two polarizations presented in Section 3.
Figure 5shows the SMOS water fraction (SWAF) averaged over the 2010–2015 period for the entire
Amazon Basin. Results are presented for four angle bins: (32
±
5
, 37
±
5
, 42
±
5
, 47
±
5
) and the
two polarizations (H and V pol). From Figure 5, it can be seen that independent of the incidence angles
and polarizations, the major spatial patterns of the Amazon Basin are observed: the Amazon River and
its tributaries, the Mamore floodplain in the south of the basin, the Branco floodplain in the north of the
basin and the Balbina lake located in the north of Manaus. For a given angle, the spatial distribution
of the SWAF is close in H and V polarization. However, the percentage of water fraction estimate
is slightly higher at H polarization than V polarization. The major difference concerns the spatial
distribution of the SWAF sensed at low incidence angle (32
±
5
) and at high incidence angle (47
±
5
).
Low incidence angles reveal small patterns of SWAF and, in particular, the smaller Amazon River west
affluent. Conversely, SWAF sensed with a higher incidence angle shows only the major structure of
the flooded areas (Amazon River, Rio Negro River, Mamore plain, etc.).
Figure 6shows the temporal dynamic of the SWAF for the full basin for the eight SMOS
configurations. For all of the angles and the polarizations, water surface extent exhibits a clear
seasonal cycle. The minimum of the inundation is observed in March, whereas the maximum of the
flooding is reached during October. This observation is valid for all of the angles and polarization.
Both in H and V polarization, for all of the incidence angles, the temporal dynamics of the SWAF
are in good agreement, except for the incidence angle of 47
±
5
. SWAF sensed with the highest
incidence angle underestimates the water fraction in the Amazon Basin with respect to lower angles.
In V polarization, on average, the Amazon Basin is less flooded than in H polarization. These results
are in accordance with Figure 5. During the wet season, almost 1% of the Amazon Basin is flooded,
whereas during the dry season, only 0.2% of the basin is flooded.
Figure 7shows the monthly spatial variability of the SWAF product over the Amazon Basin.
The monthly average has been computed during the full period (2010–2015) at V-polarization and
at 32
±
5
of incidence angle. The Amazon River, its main tributaries and the major floodplains
are well represented for all of the months. The maximum of the inundation of the Amazon River
is observed between March and July. The spatial and temporal variation of the Mamore floodplain
is well described by the SWAF data. This floodplain is inundated from January–June, and the least
flooding is observed during September and October. The major Amazon tributaries are more flooded
between January and May. Similar results are observed at H polarization and higher incidence angles
(not shown).
Water 2017,9, 350 11 of 26
Figure 5.
Average in time of the SMOS water fraction during 2010–2015 over the full Amazon Basin.
Both H and V polarization and the four incidence angles (32
±
5
, 37
±
5
, 42
±
5
, 47
±
5
)
are considered.
Figure 6.
Spatial average of the SWAF over the full Amazon Basin in H polarization (lines) and V
polarization (dashed lines) for the four incidence angles: 32
±
5
(blue), 37
±
5
(red), 42
±
5
(green)
and 47±5(black).
Water 2017,9, 350 12 of 26
Figure 7.
Monthly average of the SWAF product from 2010–2015 at V-polarization and at 32
±
5
incidence angle.
5.2. Comparison and Validation
In the following section, the SWAF product is compared to other data sources and variables: static
and dynamic water extent maps, water level measured by altimetry satellite, in situ river discharge
at the outlet of the basin and precipitation data. The static land surface maps obtained by optical
sensors and the GIEMS product were also used to analyze the spatial patterns. The SWAF data were
also compared to the dynamic water fraction product available over the Amazon Basin, the SWAMPS
product. Note that all of these datasets are completely independent from the SMOS water surface
extent maps.
Water 2017,9, 350 13 of 26
5.2.1. Comparison to Static and Climatological Water Extent Maps
The SWAF temporal average over the entire Amazon Basin had been compared to three static
land cover maps (IGBP [
62
], GlobeCover [
63
] and ESA CCI [
64
]) and the mean of the GIEMS water
fraction maps. These maps are detailed in Section 3.5 and presented in Figure 2. Main similar patterns
can be observed from the three different static land cover maps. The number of pixels partially
flooded over the Amazon Basin is higher for the GlobeCover map than for both the IGBP and ESA
CCI maps. Conversely, the ESA CCI map produces less pixels totally flooded over the Amazon
Basin than the others maps. IGBP, GlobeCover and ESA CCI maps present inundation surfaces of
220,000 km
2
, 360,000 km
2
and 210,000 km
2
, respectively. Note that the static maps are based on data
from different time periods (see Section 3.5), and the inland water occupation can change with time
and anthropogenic activities.
The average GIEMS product over the 1993–2007 period is presented in Figure 2. The Amazon
River, its major tributaries, southern and northern floodplains are well depicted. On average, over the
1993–2007 period, 440,000 km2of the surface were flooded.
Figure 8presents the mean distribution (2010–2015) of the water fraction extent for the IGBP
map, GlobeCover map, ESA CCI map, GIEMS averaging and the temporal average of all of the SWAF
products for the eight SMOS configurations over the Amazon Basin. For all of the products, the
majority of the pixels are partly flooded, and a few of them are totally flooded. The distributions of
both IGBP and ESA CCI are very close. The GlobeCover map and the mean GIEMS product (1993–2007)
provide larger estimates of pixels flooded or partially flooded with respect to the other products, in
particular for water fraction higher than 0.4. For all of the SMOS configurations, the distributions of
the SWAF products are comprised between both the GlobeCover and GIEMS distributions and both
the IGBP and ESA CCI distributions. Between the eight SMOS configurations, only a few differences
in their spatial distributions can be noticed. Both in H and V polarizations, a decrease of the incidence
angle leads to a decrease of the detection of pixels partially flooded. In V polarization, fewer pixels are
flooded than in H polarization. This behavior is particularly marked for water fraction ranges between
10% and 40%. This trend makes the SWAF computed in V polarization closer to the IGBP and ESA CCI
maps than the SWAF calculated in H polarization for the moderately flooded pixels. Table 3presents an
average of the number of square kilometers flooded in the Amazon Basin. Figures 5and 8and Table 3
confirm that, on average, the number of square kilometers flooded decreases for increasing incidence
angles, and more flooded areas are detected in H polarization than in V polarization. Independent of
the selected SMOS configuration, the number of square kilometers flooded is in the range between the
IGBP and GlobeCover estimates.
Figure 9presents the bias (reference static maps, SWAF configurations) and the RMSE between
water surface extent for each reference maps used in Figure 8and the water surface extent for the
eight SMOS configurations. For all of the SWAF configurations, lower bias is obtained by comparing
the SWAF data with IGBP (mean bias = 0.6%) and ESA CCI (mean bias =
0.4%). Higher bias is
obtained by comparing SWAF data with the GlobeCover map (mean bias = 5.9%). Bias between the
static maps and the SWAF data is always higher at V-polarization than at H-polarization for all of
the static maps. Moreover, the bias values increase with the growth of the incidence angles. A lower
RMSE value is obtained by comparing the SWAF data with the ESA CCI map (mean RMSE = 5.8%),
and a higher RMSE value is obtained with the GlobeCover map (mean RMSE = 17.8%). Following
the static map considered, the behavior of the SWAF data with respect to angles and polarization
differs. For IGBP and ESA CCI maps, lower values of RMSE are obtained at V-polarization than at
H-polarization, and a slight decrease of the RMSE values is observed with the increase of the incidence
angles. The contrary is noticed for the comparison with the GIEMS data. No trend concerning the
RMSE behavior is observed for the comparison between the SWAF data and the GlobeCover data.
Water 2017,9, 350 14 of 26
Table 3.
Average of the number of square kilometers flooded in the Amazon Basin for the eight
SMOS configurations.
Incidence Angle () H-pol V-pol
32±5290,000 270,000
37±5280,000 260,000
42±5280,000 250,000
47±5280,000 250,000
Figure 8.
Histogram of water fraction for the IGBP map (red), GlobeCover map (black), the ESA
CCI (blue) and SWAF (yellow columns) for eight SMOS TB configurations (32–47 angle bins and
H/V configurations).
Figure 9.
Bias (reference static maps, SWAF) and RMSE values computed between each reference maps
(GlobeCover, IGBP, ESA CCI, GIEMS) and SWAF for the eight configurations.
5.2.2. Comparison with Water Height Measured by Altimetry
For low topography slopes, the water surface extent can be related to the water height.
Other studies [
32
,
66
] had already shown that the seasonal and inter-annual variation patterns of
Water 2017,9, 350 15 of 26
the surface water extent and the water level agree well. In this present study, the water levels
measured by the Jason-2 satellite were compared with the SWAF product over 83 virtual stations
during the 2010–2012 periods. Results are presented in Figure 10. Only stations with significant
results (p-value
<
0.01) are represented. The color dot indicates the correlation value for each virtual
station. The gray color dots show the virtual station with non-significant results. For all of the SMOS
configurations, the correlation value between the water fraction and the water level measured by
Jason-2 is very high (r
>
0.8) throughout the Amazon River and the major tributaries. The lower
correlation values are located over areas where the relation between the water surface’s extent and the
water level is not direct. For a given angle, slight differences in terms of correlation values are observed
with respect to the polarization choice. The number of not significant stations varies following angles
and polarization. At H polarization, the number of no significant stations is equal to 36, 33, 33 and 44
for angles from 32
±
5
–47
±
5
, respectively. At V-polarization, the number of no significant stations
is slightly lower and equal to: 35, 30, 31 and 44, respectively.
To formalize this information, the sum of the correlation values for each SMOS configuration is
presented in Figure 11. Only stations with significant correlations for the eight SMOS configuration
are considered to compute this figure. The sum of the correlation values between the SWAF and the
water level estimated by altimetry is always higher at V-polarization than at H-polarization, except at
32
±
5
. The higher sum of correlation is obtained at V polarization and at 47
±
5
incidence angle.
At high incidence angles, higher correlation values are obtained, but the number of significant stations
is lower. The contrary is observed at low incidence angles.
5.2.3. Comparison with SWAMPS Dynamic Surface Extent
The recent SWAMPS product provides a daily estimation of the surface water extent. Note that the
SWAMPS products are obtained by more complex algorithm merging active and passive microwaves
than the SWAF data. Figure 2shows the average of the SWAMPS water fraction over the Amazon
Basin from January 2010–March 2013. Spatially, the mean SWAF and SWAMPS products are in good
agreement. Both in the SWAF and SWAMPS products, the Amazon River and its tributaries are well
represented, and the major floodplains are present. In the SWAF data, no data are provided over the
southeast part of the Amazon Basin due to high topographic index where as the SWAMPS product
shows some patterns of water surface over the same region. An important difference between the two
products is the spread of the rivers. In the SWAMPS product, all of the rivers have a larger floodplain
area than in the SWAF data. Over the two major floodplains of the Amazon Basin, different patterns
are observed in the SWAF data. This behavior is not observed for the SWAMPS data. The water surface
extents estimated by the SWAMPS data are lower than those estimated by the SWAF data.
Figure 12 shows the temporal correlation values between the SWAMPS and SWAF data from
January 2010–March 2013 for all of the SMOS configurations. The correlation value is computed
only over pixels where both SWAMPS and SWAF water fraction are present. For all of the SMOS
configurations, a good agreement (r
>
0.8) between the SWAF and SWAMPS products is observed
over the Amazon River and the two largest floodplains of the basin. Over these locations, the temporal
dynamics of the surface water are well described by the two products. Concerning the Amazon
tributaries, results are more contrasting, and the SWAF and the SWAMPS seem to have a different
temporal dynamics. In terms of SMOS configurations, results are very similar whatever the incidence
angle and polarization chosen. For example, the number of pixels that obtained a high correlation
value (r
>
0.8) between the SWAF and the SWAMPS products ranges between 147 (for 32
±
5
in
H-pol) and 172 (for 47
±
5
in H-pol). No trend is noticed between high and low angles or V and
H polarization. For the accepted correlation value (r
>
0.6), the number of pixels that satisfied these
criteria ranges between 1097 (for 47
±
5
in V-pol) and 1247 (for 37
±
5
in V-pol). In this case, a clear
trend is observed: increasing the incidence angle leads to decreasing the number of pixels with an
accepted correlation.
Water 2017,9, 350 16 of 26
Figure 10.
For each SMOS configuration, correlation values against the SWAF water surface extent and
the water level measured by Jason-2 during 2010–2012. The color dot represents the correlation value.
Gray color dots show no significant results (p-value >0.05).
Water 2017,9, 350 17 of 26
Figure 11.
For each SMOS configuration, the sum of the correlation value (
r
) obtained in Figure 8.
Only significant stations for all of the SMOS configuration are used for the computation.
Figure 12.
Temporal correlation values between the SWAMPS and SWAF products from January
2010–March 2013 for each SMOS configuration.
Water 2017,9, 350 18 of 26
5.2.4. Link between SWAF and the Hydrological Components
Strong seasonal and interannual variations can be observed in both precipitation and surface
water extent in the Amazon Basin. Figure 13 presents the standardized anomalies (i.e., the time series
of a hydrological parameter minus the average of the time series divided by its standard deviation of
the observation period) of both precipitation from TRMM and SWAF products for the eight SMOS
configurations over the entire Amazon Basin. Note that inundation is highly linked to the precipitation
events, but can also occur in response to snow melt or heavy precipitation at upstream locations. In this
case, flooded areas and precipitation are separated in both time and space. A good agreement between
all of the SWAF products whatever the polarization or the incidence angles can be noticed, except
at
47±5
in H polarization. For the entire Amazon Basin, the cross-correlation values between the
TRMM precipitation and the SWAF products are shown in Table 4. The best correlation values are
obtained at 42±5and 32±5. In H and V polarization, similar correlation values were obtained.
The correlation value is highly impacted by the choice of the incidence angle, whereas the polarization
plays a negligible role. Note that using the angle 47
±
5
strongly degrades the correlation value
between the precipitation and the water surface extent. By computing the time lag correlation values,
a time lag of two months was found between the precipitations and the water surface for all of the
SMOS configurations, except for the angle of 47±5(four months).
Figure 13.
Monthly normalized anomalies of precipitation (TRMM data), in situ discharge at Obidos
and the SWAF at H-pol (top) and V-pol (bottom) with the four incidence angles considered in this
study. The precipitation and SWAF anomalies were computed over the entire Amazon Basin.
The time series of the Amazon River discharge is closely linked to the total amount of the surface
water extent in the whole basin [
43
]. Figure 13 also shows the monthly normalized anomalies of the in
situ river discharge measured at Obidos. By computing the time lag correlation values, it was found
that the maximum water surface extent often precedes the maximum Amazon discharge. For the entire
Amazon Basin, the cross-correlation values between the in situ discharge at Obidos and the SWAF
products are reported in the Table 4. High cross-correlation values are obtained between the discharge
and all of the SWAF products varying from 0.78 (47
±
5
at H-pol) to 0.88 (42
±
5
at H-pol). For all
of the SWAF configurations, the maximum water surface extent precedes the discharge by one month.
The normalized anomalies of SWAF whatever the SMOS configuration are better correlated with the
normalized anomalies of discharge than those of precipitation. The cross-correlation value between
the normalized anomalies of precipitation and discharge is equal to 0.84 with a time lag of four months
for the whole Amazon Basin.
Water 2017,9, 350 19 of 26
Table 4.
Cross-correlation (
r
) values between TRMM precipitation, discharge at the outlet and the
SWAF products for the eight configurations over the 2010–2015 period.
Precipitation Discharge
Incidence Angle () H-pol V-pol H-pol V-pol
32±50.91 0.90 0.83 0.84
37±50.88 0.89 0.86 0.85
42±50.92 0.88 0.88 0.82
47±50.83 0.85 0.78 0.83
6. Discussion
6.1. Water Surface Validation
The SWAF products and the aggregated high spatial resolution of the IGBP and ESA CCI maps
over the Amazon Basin showed a good agreement (see Figures 8and 9). The mean bias and RMSE
between the SWAF for all of the configurations and the IGBP map is equal to 0.6% and 10%, respectively.
Better results are obtained for the comparison of the SWAF and the ESA CCI maps (mean bias =
0.4%
and mean RMSE = 5.8%). The SWAF sensitivity to seasonal and annual water fraction extent was
also demonstrated in the comparison against precipitation and discharge dynamics. The comparison
between the anomaly of water surface dynamics estimated from the SWAF products over the entire
Amazon Basin and the anomaly of discharge at the mouth of the Amazon showed a shift of one month.
This result is in good agreement with previous research papers. For example, results presenting the
same time-lag for surface water extent [
43
], surface water storage [
49
] and terrestrial water storage [
71
].
Our findings showed that precipitation often preceded the water surface dynamic by two months over
the Amazon Basin. These results also agree well with previous studies [43,71].
6.2. Impact of the Angles and Polarization on the Water Surface Retrievals
The SMOS mission provides data in multi-angular and full polarization modes. A specific analysis
is presented to determine the best acquisition configuration to retrieve the water fraction from SMOS.
The use of combined polarization and angles was used in this study as the single angle and channel
gave satisfactory results. As shown in the previous section, the spatio-temporal evolution of the SWAF
estimated using different angles and polarizations was very close in the Amazon Basin. Differences
are mainly present in areas far from the main Amazon river stream characterized by high vegetation
density. Table 5summaries the best SMOS configurations (low or high incidence angles and H or V
polarizations) in order to obtain the higher agreement between the SWAF product and each variable
used for validation (land cover classification, water level, dynamic water fraction, precipitation and
discharge). “NS” stands for Non-Significant, and it is used when no trend in the results (see previous
section) was observed.
Table 5.
Summary of the best SMOS configurations permitting a good agreement between SWAF and
the other variables. Green color means that good agreement was found, whereas the red color means
the opposite. See Section 5for the results. NS, Non-Significant.
Variables H-pol V-pol Low angles High angles
Land cover classification NS NS
Water level NS NS
Dynamic water fraction NS NS
Precipitation NS NS NS NS
Discharge NS NS NS NS
Table 5shows that the low incidence angles were more suitable to detect the dynamic water
surface extent over the tropical regions. This result was expected based on the microwave signal
theory [
72
]. The H-polarization tends to overestimate the water fraction extent for a low fraction
Water 2017,9, 350 20 of 26
(
<
50%). Over the Amazon Basin, low incidence angles at V polarization are the best configuration to
compute the water fraction extent with the SMOS data. For future studies, the authors advise to use
the SWAF products at low incidence angles and at V polarization over tropical areas. In the near future,
SWAF estimated in both H and V polarization at low incidence angle will be combined to extend the
domain of applicability of the algorithm to other environments to increase the water sensibility to the
SWAF product.
6.3. Impact of Vegetation Cover
Estimating water surface extent under dense vegetation with passive microwave is challenging.
The work in [
26
,
27
] provided estimates of the water surface extent over the Amazon Basin by using
the passive microwave SMMR sensor at 37 GHz (Q-band). They found good agreement in the seasonal
changes in inundation area and good correlations values with the water level in Manaus. However,
they noticed the effects of the vegetation on their results, in particular for small patches of open water
intermixed with vegetation canopies or from an attenuating effect of homogeneous canopies overlying
water surfaces. To overcome the vegetation attenuation, the GIEMS product mixed passive and active
microwave products at coarse spatial resolution with the optical dataset at finer resolution, which
enhances the capability to detect the small water fraction. However, this capability could be hampering
over dense vegetated area and frequently cloud-covered regions, such as the tropical ones, due to
the limitation of the optical sensors. The frequencies of the passive microwave data from SSM/I are
at 19 GHz (K band) and 85 GHz (E band), and the active one is at 5.25 GHz (C band). The work
in [
34
] showed that the low frequencies were less sensitive to the vegetation effects than the higher
frequency. The results presented in this study are the first to demonstrate the potential of L-band
(1.4 GHz) brightness temperature to estimate surface water extent under dense vegetation.
6.4. Limitations and Prospects of the SWAF Dataset
The most important limitation of the SWAF product is its coarse resolution inherent to the
passive microwave sensor. This limitation was already noticed for the GIEMS [
42
] products using
microwave sensors. This limitation implies that the water surface lower than 4% for all of the angles
and polarization could not be mapped by the SWAF products.
The SMOS sensor is based on dual polarizations and multi-angular measurements. Mountainous
areas modify local incidence angles and multi-scattering, which impacts the TB values [
70
] and,
consequently, the water surface estimation. The effects led to an overestimation of the water fraction.
To overcome this effect, areas with moderate to strong topographic slopes were not considered in
this study.
The snow is also an important component of the hydrological cycle. Due to the impact of the
topography slopes on the TB and the presence of the snow only over the mountainous areas in the
tropical basins, the temporal evolution of the snow coverage from the Andes Mountains has not been
investigated in this study.
The method developed in this study is based on the impact of surface water at the L-band
signal and based on the stability over time of the TB measured over dense forests. Some additional
computation was performed to measure the sensitivity of the SWAF product to the “forest” reference
point. It was found that instead of choosing only one reference “forest” pixel, but a set of pixels
composed only of forest, the mean TB increases by 3 K for all of the incidence angles and polarization.
This value is included in the incertitude range. However, this growth on the TB “forest” reference
value tends to increase the SWAF by 8%. Concerning the “water ” reference point, no change was
observed by choosing a set of pure water pixels. Moreover, the method applied in this study could
not be applied in areas where the TB over vegetation is not stable over time or not dense enough.
Future work will concentrate on the extension of the current algorithm to other environments by using
multi-angular information. By solving this limitation, the water fraction would be estimated at the
global scale by using only one dataset and a simple approach.
Water 2017,9, 350 21 of 26
7. Conclusions and Prospect
This study presents the validation and the link to other hydrological components of regional
(Amazon Basin) daily and multi-year (2010–2015) water surface extent maps from the SMOS mission
at coarse resolution (25 km
×
25 km). The SWAF product is based on L-band acquisitions. At such a
frequency, the signal is highly sensitive to the standing water above the ground, and it is expected to
penetrate deeper in the vegetation than at higher frequencies, such as visible and infrared or microwave
at higher frequencies. As the L-band signal is more sensitive to open water under dense vegetation,
the SWAF product provides surface water extent estimates (percentage of inundation in a pixel of
25
×
25 km) with a high temporal resolution (
<
3 days) based on the accumulation of daily surface
water extent in the Amazon Basin between 2010 and 2015. The SWAF product is computed from the
L-band, and it can be computed easily and quickly without any ancillary data. Over this basin, the
water surface extent showed a strong seasonal and interannual variability with two marked droughts
in 2010 and 2015.
The SWAF data were compared to three sets of static land cover maps provided from visible
sensors (IGBP, GlobeCover and ESA CCI) and the average inundation extent from GIEMS over
1993–2007. It was found that the SWAF products are close to the IGBP and ESA CCI maps. On average
and during the 2010–2015 period, 270,000 km
2
were inundated over the Amazon Basin. A slight
overestimation of the flooded areas could be noticed. Over the Amazon Basin, the SWAF products
were highly correlated with water levels measured by Jason-2 (
r>
0.8) for the significant stations.
The temporal dynamics of the SWAF products were also validated against precipitation (TRMM
data) and in situ discharge at the mouth of each river. It was found that over the Amazon Basin,
the precipitations often precede the inundation by three months, and the water surface extent impacts
the discharge at the mouth of the Amazon after one month. As expected by the microwave theory,
the mall water fraction could not be detected by the large footprint of SMOS. This implied that
low water fraction extent (
<
4%) could not be mapped by the SWAF products. The mountainous
areas were also a limitation of the SWAF products. The topography-modified local incidence angles
implied significant impact on the microwave signal and, consequently, on the water surface estimation.
The effects led to overestimation of the water fraction. To avoid this effect, the areas with high
topography slopes were flagged in the SWAF products.
Based on the SMOS product, the SWAF products declined with several incidence angles at two
polarizations (H and V). It was clear that high incidence angles (
>
47
±
5
) were not suitable to sense
the water surface from the L-band microwave signal. The H-polarization tended to increase the lower
value of the water fraction extent with respect to the V-polarization. The SWAF products computed
with different angles and polarizations led to similar results with very slight differences over the
Amazon Basin. For future use, the authors advise the use of SWAF computed with low incidence
angles (32±5and/or 37±5) at V for the Amazon Basin.
The methodology permitting retrieval of the water fraction applied in this study does not require
much computation time and can be easily be applied to another L- band microwave dataset, such as
the new Soil Moisture Active and Passive (SMAP) data or an older dataset (SSM/I. . .). The method
had been validated over the Amazon Basin by taking advantage of the numerous data and research
performed over this area.
In the near future, this recent water surface fraction product can be easily extended with the
future SMOS data and the Soil Moisture Active and Passive (SMAP) data to obtain a long record of
inundation products under dense vegetation. These data will be useful to better understand the water,
carbon and methane cycles over the tropical areas. By adding a third component (saturated soil) on
the first-order radiative transfer, this method is likely to be applied in other regions in the world.
Water 2017,9, 350 22 of 26
Acknowledgments:
This work was funded by the program Terre Océan Surfaces Continentales et Atmosphère
(TOSCA, France) and the CNES under the project TOSCA-SOLE and Marie Parrens was funded by the CNES
PostDoc program.
Author Contributions:
Ahmad Al Bitar and Marie Parrens conceived of and designed the algorithms.
Marie Parrens performed the analysis. Frédéric Frappart has contributed to the evaluation of the products.
Frédéric Frappart, Fabrice Papa, Jean-François Crétaux and Stephane Calmant, Jean-Pierre Wigneron and Yann Kerr
provided scientific expertise, datasets and corrections to the manuscript. Marie Parrens and Ahmad Al Bitar wrote
the manuscript with contributions from all of the co-authors.
Conflicts of Interest:
The authors declare no conflict of interest. The founding sponsors had no role in the design
of the study; in the collection, analyses or interpretation of data; in the writing of the manuscript; nor in the
decision to publish the results.
Abbreviations
The following abbreviations are used in this manuscript:
SAR Synthetic Aperture Radar
SWOT Surface Water Ocean Topography
V Vertical
H Horizontal
SMMR Scanning Multichannel Microwave Radiometer
GIEMS Global Inundation Extent from Multi-Satellites
SSMI/I Special Sensor Microwave/Imager
ERS European Remote Sensing
QSAT QuickSCAT
ASCAT Advanced Scatterometer
SSM/S Special Sensor Microwave/Sounder
SMOS Soil Moisture and Ocean Salinity
SMAP Soil Moisture Active and Passive
ESA European Space Agency
CNES Centre National d’Etude Spatiale
CDTI Centro para el Desarrollo Teccnologico Industrial
L Level
EASE Equal-Area Scalable Earth
SRTM Shuttle Radar Topography Mission
USGS U.S. Geological Survey
IGBP International Geosphere Biosphere Programme
AVHRR Advanced Very High Resolution Radiometer
MODIS Moderate Resolution Imaging Spectroradiometer
CCI Climate Change Initiative
NASA North America Space Agency
TRMM Tropical Rainfall Measuring Mission
ECMWF European Center for Medium range Weather Forecasting
SWAF SMOS WAter Fraction
ENVISAT ENVironment SATellite
TB Brightness Temperature
References
1.
Matthews, E.; Fung, I. Methane emission from natural wetlands: Global distribution, area, and
environmental characteristics of sources. Glob. Biogeochem. Cycles 1987,1, 61–86.
2. Mitsch, W.; Gosselink, J. Wetlands, 3rd ed.; John Wiley & Sons, Inc.: Hoboken, NJ, USA, 2000.
3.
Cole, J.J.; Prairie, Y.T.; Caraco, N.F.; McDowell, W.H.; Tranvik, L.J.; Striegl, R.G.; Duarte, C.M.; Kortelainen, P.;
Downing, J.A.; Middelburg, J.J.; et al. Plumbing the global carbon cycle: Integrating inland waters into the
terrestrial carbon budget. Ecosystems 2007,10, 172–185.
4.
Sjögersten, S.; Black, C.R.; Evers, S.; Hoyos-Santillan, J.; Wright, E.L.; Turner, B.L. Tropical wetlands:
A missing link in the global carbon cycle? Glob. Biogeochem. Cycles 2014,28, 1371–1386.
5.
Alsdorf, D.E.; Rodriguez, E.; Lettenmaier, D.P. Measuring surface water from space. Rev. Geophys.
2007
,45,
doi:10.1029/2006RG000197.
Water 2017,9, 350 23 of 26
6. Bakker, K. Water security: Research challenges and opportunities. Science 2012,337, 914–915.
7.
Finlayson, C.; Davidson, N. Global Review of Wetland Resources and Priorities for Wetland Inventory; Preface iv
Summary Report; CM Finlayson & AG Spiers: Canberra, CT, USA, 1999; p. 15.
8.
Vörösmarty, C.; Hoekstra, A.; Bunn, S.; Conway, D.; Gupta, J. Fresh water goes global. Science
2015
,
349, 478–479.
9.
Costanza, R.; de Groot, R.; Sutton, P.; van der Ploeg, S.; Anderson, S.J.; Kubiszewski, I.; Farber, S.;
Turner, R.K. Changes in the global value of ecosystem services. Glob. Environ. Chang. 2014,26, 152–158.
10.
Pekel, J.F.; Cottam, A.; Gorelick, N.; Belward, A.S. High-resolution mapping of global surface water and its
long-term changes. Nature 2016,540, 418–422.
11.
Smith, L.C. Satellite remote sensing of river inundation area, stage, and discharge: A review. Hydrol. Process.
1997,11, 1427–1439.
12.
Frazier, P.S.; Page, K.J. Water body detection and delineation with Landsat TM data. Photogramm. Eng.
Remote Sens. 2000,66, 1461–1468.
13.
Verpoorter, C.; Kutser, T.; Seekell, D.A.; Tranvik, L.J. A global inventory of lakes based on high-resolution
satellite imagery. Geophys. Res. Lett. 2014,41, 6396–6402.
14.
Feng, M.; Sexton, J.O.; Channan, S.; Townshend, J.R. A global, high-resolution (30-m) inland water body
dataset for 2000: First results of a topographic–spectral classification algorithm. Int. J. Digit. Earth
2016
,
9, 113–133.
15.
Crétaux, J.F.; Jelinski, W.; Calmant, S.; Kouraev, A.; Vuglinski, V.; Bergé-Nguyen, M.; Gennero, M.C.;
Nino, F.; Del Rio, R.A.; Cazenave, A.; et al. SOLS: A lake database to monitor in the Near Real Time water
level and storage variations from remote sensing data. Adv. Space Res. 2011,47, 1497–1507.
16.
Frappart, F.; Do Minh, K.; L’Hermitte, J.; Cazenave, A.; Ramillien, G.; Le Toan, T.; Mognard-Campbell, N.
Water volume change in the lower Mekong from satellite altimetry and imagery data. Geophys. J. Int.
2006
,
167, 570–584.
17.
Hess, L.L.; Melack, J.M.; Novo, E.M.; Barbosa, C.C.; Gastil, M. Dual-season mapping of wetland inundation
and vegetation for the central Amazon Basin. Remote Sens. Environ. 2003,87, 404–428.
18.
Morrissey, L.A.; Durden, S.L.; Livingston, G.P.; Steam, J.A.; Guild, L.S. Differentiating methane source
areas in arctic environments with multitemporal ERS-1 SAR data. IEEE Trans. Geosci. Remote Sens.
1996
,
34, 667–673.
19.
Birkett, C.M. Contribution of the TOPEX NASA radar altimeter to the global monitoring of large rivers
and wetlands. Water Resour. Res. 1998,34, 1223–1239.
20. Fung, L.; Cazenave, A. Satellite Altimetry and Earth Science. A Handbook of Techniques and Applications,
International Geophysical Series; Academic Press: San Diego, CA, USA, 2001; Volume 69.
21.
Riˇcko, M.; Birkett, C.M.; Carton, J.A.; Crétaux, J.F. Intercomparison and validation of continental water
level products derived from satellite radar altimetry. J. Appl. Remote Sens. 2012,6, 061710.
22.
Papa, F.; Prigent, C.; Rossow, W.; Legresy, B.; Remy, F. Inundated wetland dynamics over boreal regions
from remote sensing: The use of Topex-Poseidon dual-frequency radar altimeter observations. Int. J.
Remote Sens. 2006,27, 4847–4866.
23.
Biancamaria, S.; Lettenmaier, D.P.; Pavelsky, T.M. The SWOT mission and its capabilities for land hydrology.
Surv. Geophys. 2016,37, 307–337.
24.
Giddings, L.; Choudhury, B. Observation of hydrological features with Nimbus-7 37 GHz data, applied to
South America. Int. J. Remote Sens. 1989,10, 1673–1686.
25.
Choudhury, B.J. Passive microwave remote sensing contribution to hydrological variables. In Land
Surface—Atmosphere Interactions for Climate Modeling; Springer: Berlin, Germany, 1991; pp. 63–84.
26.
Sippel, S.; Hamilton, S.; Melack, J.; Choudhury, B. Passive microwave satellite observations of seasonal
variations of inundation area in the Amazon River floodplain. Brazil. Remote Sens. Environ.
1994
,4, 70–76.
27.
Sippel, S.; Hamilton, S.; Melack, J.; Novo, E. Passive microwave observations of inundation area and the
area/stage relation in the Amazon River floodplain. Int. J. Remote Sens. 1998,19, 3055–3074.
28.
Fily, M.; Royer, A.; Goıta, K.; Prigent, C. A simple retrieval method for land surface temperature and
fraction of water surface determination from satellite microwave brightness temperatures in sub-arctic
areas. Remote Sens. Environ. 2003,85, 328–338.
Water 2017,9, 350 24 of 26
29.
Mialon, A.; Royer, A.; Fily, M. Wetland seasonal dynamics and interannual variability over northern
high latitudes, derived from microwave satellite data. J. Geophys. Res. Atmos.
2005
,110,
doi:10.1029/2004JD005697.
30.
Temimi, M.; Leconte, R.; Brissette, F.; Chaouch, N. Flood monitoring over the Mackenzie River Basin using
passive microwave data. Remote Sens. Environ. 2005,98, 344–355.
31.
Grippa, M.; Mognard, N.; Le Toan, T.; Biancamaria, S. Observations of changes in surface water over the
western Siberia lowland. Geophys. Res. Lett. 2007,34, doi:10.1029/2007GL030165.
32.
Prigent, C.; Papa, F.; Aires, F.; Rossow, W.; Matthews, E. Global inundation dynamics inferred from
multiple satellite observations, 1993–2000. J. Geophys. Res. Atmos. 2007,112, doi:10.1029/2006JD007847.
33.
Sakamoto, T.; Van Nguyen, N.; Kotera, A.; Ohno, H.; Ishitsuka, N.; Yokozawa, M. Detecting temporal
changes in the extent of annual flooding within the Cambodia and the Vietnamese Mekong Delta from
MODIS time-series imagery. Remote Sens. Environ. 2007,109, 295–313.
34.
Kerr, Y.; Waldteufel, P.; Wigneron, J.; Martinuzzi, J.; Font, J.; Berger, M. Soil moisture retrieval from
space: The Soil Moisture and Ocean Salinity (SMOS) mission. IEEE Trans. Geosci. Remote Sens.
2001
,
39, 1729–1735.
35.
Parrens, M.; Wigneron, J.P.; Richaume, P.; Al Bitar, A.; Mialon, A.; Wang, S.; Fernandez-Moran, R.; Al-Yaari, A.;
O’Neill, P.; Kerr, Y. Considering Combined or Separated Roughness and Vegetation Effects in Soil Moisture
Retrievals. Int. J. Appl. Earth Obs. Geoinform.2017,55, 73–86.
36.
Parrens, M.; Wigneron, J.P.; Richaume, P.; Mialon, A.; Al Bitar, A.; Fernandez-Moran, R.; Al-Yaari, A.;
Kerr, Y.H. Global-scale surface roughness effects at L-band as estimated from SMOS observations.
Remote Sens. Environ. 2016,181, 122–136.
37.
Schroeder, R.; McDonald, K.C.; Chapman, B.D.; Jensen, K.; Podest, E.; Tessler, Z.D.; Bohn, T.J.;
Zimmermann, R. Development and Evaluation of a Multi-Year Fractional Surface Water Data Set Derived
from Active/Passive Microwave Remote Sensing Data. Remote Sens. 2015,7, 16688–16732.
38.
Henderson, F.M.; Lewis, A.J. Radar detection of wetland ecosystems: A review. Int. J. Remote Sens.
2008
,
29, 5809–5835.
39.
Bartsch, A.; Trofaier, A.; Hayman, G.; Sabel, D.; Schlaffer, S.; Clark, D.; Blyth, E. Detection of open water
dynamics with ENVISAT ASAR in support of land surface modelling at high latitudes. Biogeosciences
2012
,
9, 703–714.
40. Kuenzer, C.; Guo, H.; Huth, J.; Leinenkugel, P.; Li, X.; Dech, S. Flood mapping and flood dynamics of the
Mekong Delta: ENVISAT-ASAR-WSM based time series analyses. Remote Sens. 2013,5, 687–715.
41.
Santoro, M.; Wegmüller, U. Multi-temporal synthetic aperture radar metrics applied to map open water
bodies. IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens. 2014,7, 3225–3238.
42.
Prigent, C.; Matthews, E.; Aires, F.; Rossow, W.B. Remote sensing of global wetland dynamics with multiple
satellite data sets. Geophys. Res. Lett. 2001,28, 4631–4634.
43.
Papa, F.; Prigent, C.; Aires, F.; Jimenez, C.; Rossow, W.; Matthews, E. Interannual variability of surface
water extent at the global scale, 1993–2004. J. Geophys. Res. Atmos. 2010,115, doi:10.1029/2009JD012674.
44.
McCarthy, J.; Gumbricht, T.; McCarthy, T. Ecoregion classification in the Okavango Delta, Botswana from
multitemporal remote sensing. Int. J. Remote Sens. 2005,26, 4339–4357.
45.
Jain, S.K.; Saraf, A.K.; Goswami, A.; Ahmad, T. Flood inundation mapping using NOAA AVHRR data.
Water Resour. Manag. 2006,20, 949–959.
46.
Bergé-Nguyen, M.; Crétaux, J.F. Inundations in the inner Niger delta: monitoring and analysis using
MODIS and global precipitation datasets. Remote Sens. 2015,7, 2127–2151.
47.
Xiao, X.; Boles, S.; Liu, J.; Zhuang, D.; Frolking, S.; Li, C.; Salas, W.; Moore, B. Mapping paddy rice
agriculture in southern China using multi-temporal MODIS images. Remote Sens. Environ.
2005
,95, 480–492.
48.
Molinier, M.; Guyot, J.L.; De Oliveira, E.; Guimarães, V. Les Regimes Hydroiogiques de L’Amazone et de Ses
Affluents; IAHS Publication: Paris, France, 1996; pp. 209–222.
49.
Frappart, F.; Papa, F.; da Silva, J.S.; Ramillien, G.; Prigent, C.; Seyler, F.; Calmant, S. Surface freshwater
storage and dynamics in the Amazon Basin during the 2005 exceptional drought. Environ. Res. Lett.
2012
,
7, 044010.
50.
Papa, F.; Frappart, F.; Güntner, A.; Prigent, C.; Aires, F.; Getirana, A.C.; Maurer, R. Surface freshwater
storage and variability in the Amazon Basin from multi-satellite observations, 1993–2007. J. Geophys. Res.
Atmos. 2013,118, doi:10.1002/2013JD020500.
Water 2017,9, 350 25 of 26
51. Diegues, A.C.S. An Inventory of Brazilian Wetlands; Number 15; IUCN: Gland, Switzerland, 1994.
52.
Frappart, F.; Seyler, F.; Martinez, J.M.; Leon, J.G.; Cazenave, A. Floodplain water storage in the Negro
River basin estimated from microwave remote sensing of inundation area and water levels. Remote Sens.
Environ. 2005,99, 387–399.
53.
Junk, W.J. General aspects of floodplain ecology with special reference to Amazonian floodplains. In The
Central Amazon Floodplain; Springer: Berlin, Germany, 1997; pp. 3–20.
54.
Salati, E.; Marques, J. Climatology of the Amazon region. In The Amazon; Springer: Berlin, Germany, 1984;
pp. 85–126.
55.
Marengo, J.A. Interannual variability of surface climate in the Amazon Basin. Int. J. Climatol.
1992
,
12, 853–863.
56.
Espinoza Villar, J.C.; Ronchail, J.; Guyot, J.L.; Cochonneau, G.; Naziano, F.; Lavado, W.; De Oliveira, E.;
Pombosa, R.; Vauchel, P. Spatio-temporal rainfall variability in the Amazon Basin countries (Brazil, Peru,
Bolivia, Colombia, and Ecuador). Int. J. Climatol. 2009,29, 1574–1594.
57.
Ulaby, F.T. Microwave Remote Sensing: Active and Passive, Radar Remote Sensing and Surface Scattering and
Emission Theory; Addison-Wesley: Reading, MA, USA, 1982; Volume 2.
58.
Bitar, A.; Mialon, A.; Kerr, Y.; Cabot, F.; Richaume, P.; Jacquette, E.; Quesney, A.; Mahmoodi, A.; Tarrot, S.;
Parrens, M.; et al. The global SMOS Level 3 daily soil moisture and brightness temperature maps. ESSD
2017, doi:10.5194/essd-2017-1.
59. Brodzik, M.J.; Billingsley, B.; Haran, T.; Raup, B.; Savoie, M.H. EASE-grid 2.0: Incremental but significant
improvements for Earth-gridded data sets. ISPRS Int. J. Geo-Inform. 2012,1, 32–45.
60.
Jarvis, A.; Reuter, H.I.; Nelson, A.; Guevara, E. the CGIAR-CSI SRTM 90 m Database. Hole-Filled SRTM
for the Globe Version 4. 2008. Available online: http://srtm.csi.cgiar.org (accessed on 16 May 2017).
61.
Huffman, G.J.; Bolvin, D.T.; Nelkin, E.J.; Wolff, D.B.; Adler, R.F.; Gu, G.; Hong, Y.; Bowman, K.P.; Stocker, E.F.
The TRMM multisatellite precipitation analysis (TMPA): Quasi-global, multiyear, combined-sensor
precipitation estimates at fine scales. J. Hydrometeorol. 2007,8, 38–55.
62.
Friedl, M.A.; Sulla-Menashe, D.; Tan, B.; Schneider, A.; Ramankutty, N.; Sibley, A.; Huang, X. MODIS
Collection 5 global land cover: Algorithm refinements and characterization of new datasets. Remote Sens.
Environ. 2010,114, 168–182.
63.
Loveland, T.; Reed, B.; Brown, J.; Ohlen, D.; Zhu, Z.; Yang, L.; Merchant, J. Development of a global land
cover characteristics database and IGBP DISCover from 1 km AVHRR data. Int. J. Remote Sens.
2000
,
21, 1303–1330.
64.
Bontemps, S.; Defourny, P.; Radoux, J.; Van Bogaert, E.; Lamarche, C.; Achard, F.; Mayaux, P.; Boettcher, M.;
Brockmann, C.; Kirches, G.; et al. Consistent global land cover maps for climate modelling communities:
current achievements of the ESA’s land cover CCI. In Proceedings of the ESA Living Planet Symposium,
Edimburgh, UK, 9–13 September 2013; pp. 9–13.
65.
Papa, F.; Güntner, A.; Frappart, F.; Prigent, C.; Rossow, W.B. Variations of surface water extent and water
storage in large river basins: A comparison of different global data sources. Geophys. Res. Lett.
2008
,35,
doi:10.1029/2008GL033857.
66.
Frappart, F.; Papa, F.; Famiglietti, J.S.; Prigent, C.; Rossow, W.B.; Seyler, F. Interannual variations of river
water storage from a multiple satellite approach: A case study for the Rio Negro River basin. J. Geophys.
Res. Atmos. 2008,113, doi:10.1029/2007JD009438.
67.
Birkett, C.M.; Beckley, B. Investigating the performance of the Jason-2/OSTM radar altimeter over lakes
and reservoirs. Mar. Geod. 2010,33, 204–238.
68.
Da Silva, J.S.; Calmant, S.; Seyler, F.; Rotunno Filho, O.C.; Cochonneau, G.; Mansur, W.J. Water levels in the
Amazon Basin derived from the ERS 2 and ENVISAT radar altimetry missions. Remote Sens. Environ.
2010
,
114, 2160–2181.
69.
Klein, L.; Swift, C. An improved model for the dielectric constant of sea water at microwave frequencies.
IEEE Trans. Antennas Propag. 1977,25, 104–111.
70.
Mialon, A.; Coret, L.; Kerr, Y.H.; Sécherre, F.; Wigneron, J.P. Flagging the topographic impact on the SMOS
signal. IEEE Trans. Geosci. Remote Sens. 2008,46, 689–694.
Water 2017,9, 350 26 of 26
71.
Frappart, F.; Ramillien, G.; Ronchail, J. Changes in terrestrial water storage versus rainfall and discharges
in the Amazon Basin. Int. J. Climatol. 2013,33, 3029–3046.
72.
Ulaby, F.; Moore, R.; Fung, A. Microwave Remote Sensing: Active and Passive; Advanced Systems and
Applications Inc.: Dedham, MA, USA, 1986; pp. 1797–1848.
c
2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).
... More specifically, the high-resolution, dualseason classification of the Japanese Earth Resources Satellite-1 (JERS-1) L-band SAR data for the entire lowland Amazon basin by Hess et al. (2015), validated with airborne videography images, has been used as a benchmark for the inundation extent of Amazon wetlands. Since these initial studies, and with the availability of other imagery (e.g., Advanced Land Observing Satellite (ALOS) 1 and 2 missions), the remote sensing community seeking to map and characterize inundation employed various combinations of active and passive microwave data to benefit from the higher spatial resolution of the former and the higher temporal resolution of the latter (Aires et al., 2013;Jensen and McDonald, 2019;Papa et al., 2010;Parrens et al., 2019Parrens et al., , 2017Prigent et al., 2007Prigent et al., , 2020Schroeder et al., 2015). ...
... Passive microwave (PM) data are the basis of SWAF-HR, GIEMS family (GIEMS-D15, GIEMS-D3, GIEMS-2), and SWAMPS, while ancillary data (i.e., optical imagery and microwave scatterometry) are used to complement the PM signal. SWAF-HR data result from the disaggregation of water surface fraction in a dataset at coarser spatial resolution (SWAF), based on L-band passive microwave observations from the Soil Moisture and Ocean Salinity (SMOS) satellite (Parrens et al., 2017). The disaggregation of SWAF relies on water occurrence maps from GSWO and the Digital Elevation Model (DEM) Multi-Error-Removed-Improved-Terrain (MERIT) (Parrens et al., 2019). ...
... WAD2M is the only dataset to exclude open water areas (removal based on GSWO) due to its goal of estimating wetland methane emissions. SWAF-HR (Parrens et al., 2019) and GIEMS-D3 use additional data and methodologies to downscale the original 25-km passive microwavebased SWAF (Parrens et al., 2017) and GIEMS (Papa et al., 2010;Prigent et al., 2007) datasets to 1 km and 90 m, respectively. While GIEMS-D3 has a different inundation magnitude than the original GIEMS due to merging with ancillary data, SWAF-HR conserves the same inundation magnitude across scales. ...
Article
The Amazon River basin harbors some of the world's largest wetland complexes, which are of major importance for biodiversity, the water cycle and climate, and human activities. Accurate estimates of inundation extent and its variations across spatial and temporal scales are therefore fundamental to understand and manage the basin's resources. More than fifty inundation estimates have been generated for this region, yet major differences exist among the datasets, and a comprehensive assessment of them is lacking. Here we present an intercomparison of 29 inundation datasets for the Amazon basin, based on remote sensing only, hydrological modeling, or multi-source datasets, with 18 covering the lowland Amazon basin (elevation <500 m, which includes most Amazon wetlands), and 11 covering individual wetland complexes (subregional datasets). Spatial resolutions range from 12.5 m to 25 km, and temporal resolution from static to monthly, spanning up to a few decades. Overall, 31% of the lowland basin is estimated as subject to inundation by at least one dataset. The long-term maximum inundated area across the lowland basin is estimated at 599,700 ± 81,800 km² if considering the three higher quality SAR-based datasets, and 490,300 ± 204,800 km² if considering all 18 datasets. However, even the highest resolution SAR-based dataset underestimates the maximum values for individual wetland complexes, suggesting a basin-scale underestimation of ~10%. The minimum inundation extent shows greater disagreements among datasets than the maximum extent: 139,300 ± 127,800 km² for SAR-based ones and 112,392 ± 79,300 km² for all datasets. Discrepancies arise from differences among sensors, time periods, dates of acquisition, spatial resolution, and data processing algorithms. The median total area subject to inundation in medium to large river floodplains (drainage area > 1000 km²) is 323,700 km². The highest spatial agreement is observed for floodplains dominated by open water such as along the lower Amazon River, whereas intermediate agreement is found along major vegetated floodplains fringing larger rivers (e.g., Amazon mainstem floodplain). Especially large disagreements exist among estimates for interfluvial wetlands (Llanos de Moxos, Pacaya-Samiria, Negro, Roraima), where inundation tends to be shallower and more variable in time. Our data intercomparison helps identify the current major knowledge gaps regarding inundation mapping in the Amazon and their implications for multiple applications. In the context of forthcoming hydrology-oriented satellite missions, we make recommendations for future developments of inundation estimates in the Amazon and present a WebGIS application (https://amazon-inundation.herokuapp.com/) we developed to provide user-friendly visualization and data acquisition of current Amazon inundation datasets.
... Synthetic aperture radar (SAR) is often used to map water providing good results but they often over estimate the presence of water and have a low time revisiting surfaces (Hess et al., 2003(Hess et al., , 2015Martinez and Le Toan, 2007). A passive microwave sensor can map water under cloud and dense vegetation with a high revisit time but at low spatial resolution (Parrens et al., 2017). The potential of radar altimetry for monitoring water levels at a local scale has already been demonstrated (de Oliveira Campos et al., 2001;Birkett et al., 2002) but measurements vary from 10 days up to 35 days, depending on the mission, which is a low revisit time for some hydrologic issues. ...
... It provides brightness temperature (TB) emitted from the Earth over a range of incidence angles (0 • to 55 • ) with a spatial resolution between 35 and 50 km. Parrens et al. (2017) and Fatras et al. (2021) conditions (even during heavy precipitation and cloudy sky). SWAF data provide water surface fraction over the Amazon Basin from 2010 to date each week over lowlands but it is not available over mountainous areas. ...
Article
Full-text available
The Madeira is one of the major tributaries of the Amazon River and is characterized by a large alluvial floodplain throughout the stream continuum. This study aims to better assess the hydrological functioning of the Madeira Basin over its alluvial floodplains at both local and global scales. We used the semi-distributed hydrological Soil and Water Assessment Tool (SWAT) model to simulate water discharge at a daily time step and water resources for each hydrological compartment. A new hydraulic module for water routing was implemented in the SWAT model considering the floodplain either as a simple reservoir or as a continuum where the water can flow along with the floodplain network. Both water surface estimated by L-band passive microwaves (SWAF data) and digital elevation model—shuttle radar topography mission (DEM–SRTM data) were used to delineate the floodplain, as inputs for the model. On the global scale, the amount of water stored in the Madeira floodplain is between 810 ± 230 km ³ per year when the floodplains are delimited with SWAF and 1,300 ± 350 km ³ per year with the DEM floodplain delineation between 2008 and 2018. Spatial altimetry (Jason 2-3) data were also applied to alluvial areas to validate the water height dynamic in floodplains at a local scale. Results show that more than 60% of the alluvial validation points display a correlation above 0.40 ± 0.02 regardless of the floodplain delineation. This study permits us to better characterize the spatio-temporal storage dynamics of the Madeira floodplains at both local and global scales, and it underlines the importance of a precise floodplain delineation, before computing biogeochemical fluxes and sediment yield.
... Most studies located in South America concentrate on the Amazon River Basin. Sixtyseven percent of studies focus on the hydrosphere, particularly surface water area dynamics [88,104,149,264,267] and surface water storage dynamics [138,185,186]. Due to the high cloud cover and dense vegetation in the research area, most studies work with multi-sensor approaches. ...
... An alternative approach based on Soil Moisture and Ocean Salinity (SMOS) passive microwave satellite data has been developed in recent years. It offers higher temporal resolution (3 days) but is limited in its spatial resolution (25 km) [88,267]. The findings from the Guayas watershed [263], the Orinoco River [138], and the Amazon basin [22] are in agreement that the intra-and inter-annual water dynamics are correlated by variations in precipitation. ...
Article
Full-text available
Inland surface water is often the most accessible freshwater source. As opposed to groundwater, surface water is replenished in a comparatively quick cycle, which makes this vital resource—if not overexploited—sustainable. From a global perspective, freshwater is plentiful. Still, depending on the region, surface water availability is severely limited. Additionally, climate change and human interventions act as large-scale drivers and cause dramatic changes in established surface water dynamics. Actions have to be taken to secure sustainable water availability and usage. This requires informed decision making based on reliable environmental data. Monitoring inland surface water dynamics is therefore more important than ever. Remote sensing is able to delineate surface water in a number of ways by using optical as well as active and passive microwave sensors. In this review, we look at the proceedings within this discipline by reviewing 233 scientific works. We provide an extensive overview of used sensors, the spatial and temporal resolution of studies, their thematic foci, and their spatial distribution. We observe that a wide array of available sensors and datasets, along with increasing computing capacities, have shaped the field over the last years. Multiple global analysis-ready products are available for investigating surface water area dynamics, but so far none offer high spatial and temporal resolution.
... We believe that a CYGNSS-derived global flood product would be of high interest due to the characteristics of the mission. In effect, actual wetland monitoring derives either from optical data (Pekel et al., 2016), which is problematic especially in tropical areas -temporal averaging due to clouds, no detection of the inundations below the canopy -, or from active and passive microwave data (Bartsch et al., 2009, Parrens et al., 2017, Prigent et al., 2020. Active radar measurements such as Sythetic Aperture Radar, suffer from a lower temporal repeat (ALOS-1 and 2) and double bounce effect (EN-VISAT, RADARSAT-1 and 2, SENTINEL 1) or weak penetration depth (TerraSAR-X) in vegetated areas, while passive microwave measurements operated by radiometers (e.g., SSM/I, AMSR-E, SMAP, SMOS) have spatial resolutions coarser than 20 km. ...
... This is due to the mixing of reflections over open water and riverbanks in the same CYGNSS pixels, producing a high value of Γ std . However, the dense tree canopy out of the water bodies in these regions drastically reduces the penetration of L-band signals (Parrens et al., 2017). This affects a potential mapping of floods in equatorial forests using CYGNSS, as it is confirmed with the example of the Cuvette Centrale of Congo. ...
Article
Full-text available
This study uses the observations from the Cyclone GNSS (CYGNSS) mission to analyze their potential for a global mapping of the floods dynamics in the pan-tropical area using Global Navigation Satellite System (GNSS) Reflectometry (GNSS-R). We base our analysis on the coherent reflectivity derived from CYGNSS observations. We show that the CYGNSS mission configuration allows a gridding at a spatial resolution of 0.1° (∼11 km at the equator), with a time sampling of 1 week. We calculate the average and standard deviation values of reflectivity in the grid pixels at each time step. A Gaussian weighted window of one month is used to fill the gaps which appear in the time series due to the pseudo-random sampling of CYGNSS observations. The maps of these two parameters are then compared to elevation data from the Shuttle Radar Topography Mission (SRTM) digital elevation model (DEM), to Land Cover information from the European Space Agency’s (ESA) Climate Change Initiative (CCI), and to a reference set of static inundation maps. We observe a strong correspondence between CYGNSS reflectivity-based parameters, and the percentage of flooded areas established in the literature. The detection of the major floodplains, irrigated crops, open water areas, and the hydrological network using CYGNSS data is clear. We observe some limitations over the areas with high elevation – due to the CYGNSS mission specificities – and over the most densely vegetated areas. At some point it could prevent the correct extraction of flood patterns. For a future complete CYGNSS-based flood product, the integration of ancillary data describing the major role of land cover, biomass and topography on the GNSS-R returned signals should be necessary to extract the correct features of water cycle.
... Products making use of passive microwave-derived observations have been also specifically developed to monitor surface water extent over large African basins, such as the Surface Water Fraction (SWAF) product based on multi-angular and dual polarization passive L-band (1.4 GHz) microwave signal from Soil Moisture Ocean Salinity (SMOS) (Fatras et al. 2021). This product has been recently used to map the spatiotemporal variability of water bodies in the Congo River basin at ~ 50 km spatial resolution and weekly temporal resolution from 2010 to 2017, with the ability of the L-Band frequency to retrieve water under dense canopy (Parrens et al. 2017). They reported that the mean flooded area of the Congo extent was between 2 and 3% of the entire basin. ...
... (Schroeder et al., 2015) derived the Surface WAter Microwave Product Series (SWAMPS) product from the ERS/SSMI and ASCAT brightness temperatures. Parrens et al., (2017) demonstrated the ability of L-Band microwave brightness temperatures (Al from SMOS to provide the SMOS WAter Fraction (SWAF) under the vegetation in the Amazonian basin. An enhanced algorithm using multi-angular and dual-polarization brightness temperatures has been applied to obtain a global dataset G-SWAF (Al . ...
Thesis
Full-text available
The biosphere is undergoing an environmental crisis caused by our energy-intensive economic model. Understanding the regulation and exchange processes of global ecosystems is the key challenge to improve our resilience and reduce our impact on the Earth. The aim of this thesis is to assess the dynamic role of wetland ecosystems in the global nitrogen cycle. The Soil Denitrification Model (SDM) was developed using soil moisture and temperature satellite Earth Observations. The model was validated at local scale with laboratory measurements, then upscale global scale. Identification of hot moments and hot spots in natural wetlands. Southeast Asia and Oceania was identified as the main hot spot. The hot moments varied by region and by typology. May was the hot moment for freshwater marshes and complex wetlands. While brackish wetlands are active all year long and flooded forest, have their peak in December and January. Quantitative estimation of denitrification was estimated at 169.32± 18.31 TgN (N2O-N +N2-N).yr-1 and evolution of denitrification in the last 8 year was analysed and linked to global climate anomalies.
... (Schroeder et al., 2015) derived the Surface WAter Microwave Product Series (SWAMPS) product from the ERS/SSMI and ASCAT brightness temperatures. Parrens et al., (2017) demonstrated the ability of L-Band microwave brightness temperatures (Al from SMOS to provide the SMOS WAter Fraction (SWAF) under the vegetation in the Amazonian basin. An enhanced algorithm using multi-angular and dual-polarization brightness temperatures has been applied to obtain a global dataset G-SWAF (Al . ...
Thesis
The biosphere is undergoing an environmental crisis caused by our energy-intensive economic model. Understanding the regulation and exchange processes of global ecosystems is the key challenge to improve our resilience and reduce our impact on the Earth. The aim of this thesis is to assess the dynamic role of wetland ecosystems in the global nitrogen cycle. The Soil Denitrification Model (SDM) was developed using soil moisture and temperature satellite Earth Observations. The model was validated at local scale with laboratory measurements, then upscale global scale. Identification of hot moments and hot spots in natural wetlands. Southeast Asia and Oceania was identified as the main hot spot. The hot moments varied by region and by typology. May was the hot moment for freshwater marshes and complex wetlands. While brackish wetlands are active all year long and flooded forest, have their peak in December and January. Quantitative estimation of denitrification was estimated at 169.32± 18.31 TgN (N2O-N +N2-N).yr-1 and evolution of denitrification in the last 8 year was analysed and linked to global climate anomalies.
... Products making use of passive microwave-derived observations have been also specifically developed to monitor surface water extent over large African basins, such as the Surface Water Fraction (SWAF) product based on multi-angular and dual polarization passive L-band (1.4 GHz) microwave signal from Soil Moisture Ocean Salinity (SMOS) (Fatras et al. 2021). This product has been recently used to map the spatiotemporal variability of water bodies in the Congo River basin at ~ 50 km spatial resolution and weekly temporal resolution from 2010 to 2017, with the ability of the L-Band frequency to retrieve water under dense canopy (Parrens et al. 2017). They reported that the mean flooded area of the Congo extent was between 2 and 3% of the entire basin. ...
Article
Full-text available
The African continent hosts some of the largest freshwater systems worldwide, characterized by a large distribution and variability of surface waters that play a key role in the water, energy and carbon cycles and are of major importance to the global climate and water resources. Freshwater availability in Africa has now become of major concern under the combined effect of climate change, environmental alterations and anthropogenic pressure. However, the hydrology of the African river basins remains one of the least studied worldwide and a better monitoring and understanding of the hydrological processes across the continent become fundamental. Earth Observation, that offers a cost-effective means for monitoring the terrestrial water cycle, plays a major role in supporting surface hydrology investigations. Remote sensing advances are therefore a game changer to develop comprehensive observing systems to monitor Africa’s land water and manage its water resources. Here, we review the achievements of more than three decades of advances using remote sensing to study surface waters in Africa, highlighting the current benefits and difficulties. We show how the availability of a large number of sensors and observations, coupled with models, offers new possibilities to monitor a continent with scarce gauged stations. In the context of upcoming satellite missions dedicated to surface hydrology, such as the Surface Water and Ocean Topography (SWOT), we discuss future opportunities and how the use of remote sensing could benefit scientific and societal applications, such as water resource management, flood risk prevention and environment monitoring under current global change.
Article
The monitoring of flood and wetland dynamics at global scale is hampered by several limitations, including a reduced data availability in tropical areas due to the presence of clouds affecting visible and infrared imagery, or low spatial and/or temporal resolutions affecting passive and active microwave Earth Observation (EO) data. As a consequence, surface water extent estimates and their temporal variations remain challenging especially in equatorial river basins. Global Navigation Satellite System Reflectometry (GNSS-R) L-band signals recorded onboard Cyclone GNSS (CYGNSS) mission, composed of 8 Low Elevation Orbit (LEO) satellites, provide information on surface properties at high temporal resolution from 2017 up to now. CYGNSS bistatic observations were analyzed for detecting permanent water and seasonal floodplains over the full coverage of the mission, from 40°S to 40°N. We computed CYGNSS reflectivity associated to the coherent component of the received power, that was gridded at 0.1° spatial resolution with a 7-day time sampling afterwards. Several statistical metrics were derived from CYGNSS reflectivity, including the weighted mean and standard deviation, the median and the 90th percentile (respectively Γmean,Γstd,Γmedian and Γ90%) in each pixel. These parameters were clustered using the K-means algorithm with an implementation of the Dynamic Time Warping (DTW) similarity measure. They were compared to static inundation maps, and to dynamic estimations of surface water extent both at the global and regional scales, using the Global Inundation Extent from Multi-Satellites (GIEMS) and MODIS-based products. The difference between Γ90% and Γmedian shows the best sensitivity to the presence of water. The river streams and lakes are correctly detected, and a strong seasonality is identified in CYGNSS reflectivity over the largest floodplains, with the exception of the Cuvette Centrale of Congo which is covered by dense vegetation. This seasonal reflectivity signal correlates well with inundation maps: Pearson’s correlation coefficient between Γmedian and surface water extent from both GIEMS and MODIS is over 0.8 in the largest floodplains. The spatial patterns of reflectivity are consistent with static inundation maps: at the time of maximum flooding extent, a spatial correlation coefficient around 0.75 with Γmedian is obtained for several basins. We also evaluated the dependence of CYGNSS-derived clusters and reflectivity on the dominant land cover type and on the density of Above Groud Biomass (AGB) in the pixel. On the one hand, misclassifications of flooded pixels were observed over vegetated regions, probably due to uncertainties related to the attenuation by the vegetation in both CYGNSS and reference datasets. On the other hand, flooded pixels with a mean AGB up to ∼300 Mg/ha were correctly detected with the clustering. High reflectivity values are also observed over rocky soils in arid regions and create false alarms. Finally, strong winds on large lakes cause surface roughness, and lower reflectivity values are observed in this case which weaken the detection of open water. While these constraints are to be taken in account and corrected in a future model, a pan-tropical mapping of surface water extent dynamics using CYGNSS can be envisaged.
Article
Full-text available
The Congo River Basin (CRB) is the second largest river system in the world, but its hydroclimatic characteristics remain relatively poorly known. Here, we jointly analyze a large record of in situ and satellite-derived observations, including long-term time series of Surface Water Height (SWH) from radar altimetry (a total of 2,311 virtual stations) and Surface Water Extent (SWE) from a multi-satellite technique, to characterize the CRB surface hydrology and its variability. Firstly, we show that SWH from altimetry multi-missions agree well with in situ water stage at various locations, with root mean square deviation varying from 10 cm (with Sentinel-3A) to 75 cm (with European Remote Sensing-2). SWE variability from multi-satellite observations also shows a plausible behavior over a ~25-year period when evaluated against in situ observations from sub-basin to basin scale. Both datasets help to better characterize the large spatial and temporal variability of hydrological patterns across the basin, with SWH exhibiting annual amplitude of more than 5 m in the northern sub-basins while Congo main-stream and Cuvette Centrale tributaries vary in smaller proportions (1.5 m to 4.5 m). Furthermore, SWH and SWE help illustrate the spatial distribution and different timings of the CRB annual flood dynamic and how each sub-basin and tributary contribute to the hydrological regime at the outlet of the basin (the Brazzaville/Kinshasa station), including its peculiar bi-modal pattern. Across the basin, we estimate time lag and water travel time to reach the Brazzaville/Kinshasa station, ranging from 0-1 month in its vicinity in downstream parts of the basin and up to 3 months in remote areas and small tributaries. Northern sub-basins and the central Congo region highly contribute to the large peak in December-January while the southern part of the basin supplies water to both hydrological peaks, in particular to the moderate one in April-May. The results are supported using in situ observations at several locations in the basin. Our results contribute to a better characterization of the hydrological variability in the CRB and represent an unprecedented source of information for hydrological modeling and to study hydrological processes over the region.
Article
Full-text available
The objective of this paper is to present the multi-orbit (MO) surface Soil Moisture (SM) and angle binned Brightness Temperature (TB) products for the SMOS (Soil Moisture and Ocean Salinity) mission based on the a new multi-orbit algorithm. The Level 3 algorithm at CATDS (Centre de Traitement Aval des Données SMOS) makes use of multi-orbit (multi-revisits) retrieval to enhance the robustness and quality of SM retrievals. The motivation of the approach is to make use of the temporal auto-correlation of the vegetation optical depth (VOD) to enhance the retrievals when an acquisition occurs at the border of the swath. The retrieval algorithm is implemented in a unique operational processor delivering multiple parameters (e.g. SM and VOD) using angular signatures, dual polarization and multiple revisits. A subsidiary angle binned TB product is provided. In this study the L3 TB V300 product is showcased and compared to SMAP (Soil Moisture Active Passive) TB. The L3 SM V300 product is compared to the single-orbit (SO) retrievals from Level 2 SM processor from ESA (European Space Agency) with aligned configuration. The advantages and drawbacks of the Level 3 SM product (L3SM) product are discussed. The comparison is done at global scale between the two datasets and at local scale with respect to in situ data from AMMA-CATCH and USDA-ARS WATERSHEDS networks. The results obtained from the global analysis show that the MO implementation enhances the number of retrievals up to 9 % over certain areas. The comparison with the in situ data shows that the increase of the number of retrievals does not come with a decrease of quality. But rather at the expense of an increased lag of product availability from 6 hours to 3.5 days which can be a limiting factor for forecast applications like flood forecast but reasonable for drought monitoring and climate change studies. The SMOS L3 soil moisture and L3 brightness temperature products are delivered using an open licence and free of charge by CATDS (http://www.catds.fr).
Article
Full-text available
The location and persistence of surface water (inland and coastal) is both affected by climate and human activity and affects climate, biological diversity and human wellbeing. Global data sets documenting surface water location and seasonality have been produced from inventories and national descriptions, statistical extrapolation of regional data and satellite imagery, but measuring long-term changes at high resolution remains a challenge. Here, using three million Landsat satellite images, we quantify changes in global surface water over the past 32 years at 30-metre resolution. We record the months and years when water was present, where occurrence changed and what form changes took in terms of seasonality and persistence. Between 1984 and 2015 permanent surface water has disappeared from an area of almost 90,000 square kilometres, roughly equivalent to that of Lake Superior, though new permanent bodies of surface water covering 184,000 square kilometres have formed elsewhere. All continental regions show a net increase in permanent water, except Oceania, which has a fractional (one per cent) net loss. Much of the increase is from reservoir filling, although climate change is also implicated. Loss is more geographically concentrated than gain. Over 70 per cent of global net permanent water loss occurred in the Middle East and Central Asia, linked to drought and human actions including river diversion or damming and unregulated withdrawal. Losses in Australia and the USA linked to long-term droughts are also evident. This globally consistent, validated data set shows that impacts of climate change and climate oscillations on surface water occurrence can be measured and that evidence can be gathered to show how surface water is altered by human activities. We anticipate that this freely available data will improve the modelling of surface forcing, provide evidence of state and change in wetland ecotones (the transition areas between biomes), and inform water-management decision-making.
Conference Paper
Full-text available
Led by the European Space Agency, the Climate Change Initiative land cover project focuses on the land cover observed as an Essential Climate Variable. Consultation mechanisms were established with the climate modelling community in order to identify its specific needs in terms of satellite-based global land cover products. Key finding was the needs for stable land cover data and a dynamic component in form of time-series. An innovative land cover concept is proposed, along with an new global land cover mapping approach, based on multi-year earth observation datasets. The corresponding products are presented, which consist of three successive and consistent global LC maps centred to the epochs 2000, 2005 and 2010.
Article
Full-text available
The sensitivity of Earth’s wetlands to observed shifts in global precipitation and temperature patterns and their ability to produce large quantities of methane gas are key global change questions. We present a microwave satellite-based approach for mapping fractional surface water (FW) globally at 25-km resolution. The approach employs a land cover-supported, atmospherically-corrected dynamic mixture model applied to 20+ years (1992–2013) of combined, daily, passive/active microwave remote sensing data. The resulting product, known as Surface WAter Microwave Product Series (SWAMPS), shows strong microwave sensitivity to sub-grid scale open water and inundated wetlands comprising open plant canopies. SWAMPS’ FW compares favorably (R2 = 91%–94%) with higher-resolution, global-scale maps of open water from MODIS and SRTM-MOD44W. Correspondence of SWAMPS with open water and wetland products from satellite SAR in Alaska and the Amazon deteriorates when exposed wetlands or inundated forests captured by the SAR products were added to the open water fraction reflecting SWAMPS’ inability to detect water underneath the soil surface or beneath closed forest canopies. Except for a brief period of drying during the first 4 years of observation, the inundation extent for the global domain excluding the coast was largely stable. Regionally, inundation in North America is advancing while inundation is on the retreat in Tropical Africa and North Eurasia. SWAMPS provides a consistent and long-term global record of daily FW dynamics, with documented accuracies suitable for hydrologic assessment and global change-related investigations.
Article
Full-text available
Surface water storage and fluxes in rivers, lakes, reservoirs and wetlands are currently poorly observed at the global scale, even though they represent major components of the water cycle and deeply impact human societies. In situ networks are heterogeneously distributed in space, and many river basins and most lakes—especially in the developing world and in sparsely populated regions—remain unmonitored. Satellite remote sensing has provided useful complementary observations, but no past or current satellite mission has yet been specifically designed to observe, at the global scale, surface water storage change and fluxes. This is the purpose of the planned Surface Water and Ocean Topography (SWOT) satellite mission. SWOT is a collaboration between the (US) National Aeronautics and Space Administration, Centre National d’Études Spatiales (the French Spatial Agency), the Canadian Space Agency and the United Kingdom Space Agency, with launch planned in late 2020. SWOT is both a continental hydrology and oceanography mission. However, only the hydrology capabilities of SWOT are discussed here. After a description of the SWOT mission requirements and measurement capabilities, we review the SWOT-related studies concerning land hydrology published to date. Beginning in 2007, studies demonstrated the benefits of SWOT data for river hydrology, both through discharge estimation directly from SWOT measurements and through assimilation of SWOT data into hydrodynamic and hydrology models. A smaller number of studies have also addressed methods for computation of lake and reservoir storage change or have quantified improvements expected from SWOT compared with current knowledge of lake water storage variability. We also briefly review other land hydrology capabilities of SWOT, including those related to transboundary river basins, human water withdrawals and wetland environments. Finally, we discuss additional studies needed before and after the launch of the mission, along with perspectives on a potential successor to SWOT.
Article
For more than six years, the Soil Moisture and Ocean Salinity (SMOS) mission has provided multi angular and full-polarization brightness temperature (TB) measurements at L-band. Geophysical products such as soil moisture (SM) and vegetation optical depth at nadir (τnad) are retrieved by an operational algorithm using TB observations at different angles of incidence and polarizations. However, the quality of the retrievals depends on several surface effects, such as vegetation, soil roughness and texture, etc. In the microwave forward emission model used in the retrievals (L-band Microwave Emission Model, L-MEB), soil roughness is modelled with a semi-empirical equation using four main parameters (Qr, Hr, Nrp, with p = H or V polarizations). At present, these parameters are calibrated with data provided by airborne studies and in situ measurements made at a local scale that is not necessarily representative of the large SMOS footprints (43 km on average) at global scale. In this study, we evaluate the impact of the calibrated values of Nrp and Hr on the SM and τnad retrievals based on SMOS TB measurements (SMOS Level 3 product) over the Soil Climate Analysis Network (SCAN) network located in North America over five years (2011–2015). In this study, Qr was set equal to zero and we assumed that NrH = NrV. The retrievals were performed by varying Nrp from −1 to 2 by steps of 1 and Hr from 0 to 0.6 by steps of 0.1. At satellite scale, the results show that combining vegetation and roughness effects in a single parameter provides the best results in terms of soil moisture retrievals, as evaluated against the in situ SM data. Even though our retrieval approach was very simplified, as we did not account for pixel heterogeneity, the accuracy we obtained in the SM retrievals was almost systematically better than those of the Level 3 product. Improved results were also obtained in terms of optical depth retrievals. These new results may have key consequences in terms of calibration of roughness effects within the algorithms of the SMOS (ESA) and the SMAP (NASA) space missions.
Article
The Soil Moisture and Ocean Salinity (SMOS) mission is the first satellite dedicated to providing global surface soil moisture products. SMOS operates at L-band (1.4 GHz) and, at this frequency, the signal not only depends on soil moisture and vegetation optical depth but is also significantly affected by surface effects and, in particular, by the soil roughness. However, when dense vegetation is present, the L-band signal is poorly sensitive to the surface effects. First, by using multiple regressions between soil moisture (SM) and brightness temperature (TB) at different incidence angles and polarizations, the SMOS sensitivity to the surface effects was evaluated. A global-scale map of SMOS sensitivity to the surface effects was computed and showed that, for 87% of the land surface, the SMOS observations were sensitive to these effects, while very low sensitivity to the surface effects was estimated over 13% of the land surfaces. For instance, over broadleaf evergreen forest (mainly the Amazon and Congo forests), SMOS was sensitive to the surface effects over only half of the pixels considered. In a second step, in L-MEB (L-band Microwave Emission of the Biosphere), the forward emission model of the SMOS algorithm, the vegetation and roughness effects were combined in a single parameter, referred to as TR in this study. By inverting L-MEB, SM and TR were retrieved at global scale from the SMOS Level 3 (L3) TB observations during 2011. Assuming a linear relationship between TR and the Leaf Area Index (LAI) obtained from MODIS data, the effects of roughness (Hr) and vegetation were decoupled and a global map of soil roughness effects was estimated. It was found that the spatial pattern of the Hr values could be related to the main vegetation types. Higher values of roughness (Hr = 0.32-0.39) were obtained for forests (broadleaf evergreen, deciduous and mixed coniferous) while lower values (Hr = 0.14-0.16) were obtained for deserts, shrubs and bare soils. Intermediate values (Hr = 0.20-0. 23) were obtained over grasslands, tundra and cultivated land. Over vegetation biomes composed of forests and wooded grasslands, the Hr values were mainly correlated to the vegetation density (r ~ 0.55). For deserts, shrubs and bare soils, the Hr values were mainly correlated to the topography slopes (r ~ 0.53). The global maps presented in this study could lead to improved retrievals of soil moisture and vegetation optical depth for present and future microwave remote sensing missions such as SMOS and Soil Moisture Active Passive (SMAP).
Article
The advent of precision microwave radiometry has placed a stringent requirement on the accuracy with which the dielectric constant of sea water must be known. To this end, measurements of the dielectric constant have been conducted at S-band and L-band with a quoted uncertainty of tenths of a percent. These and earlier results are critically examined, and expressions are developed which will yield computations of brightness temperature having an error of no more than 0.3 K for an undisturbed sea at frequencies lower than X-band. At the higher microwave and millimeter wave frequencies, the accuracy is in question because of uncertainties in the relaxation time and the dielectric constant at infinite frequency.
Article
The aim of this project was to determine the accuracy of using simple digital image processing techniques to map riverine water bodies with Landsat 5 TM data. This paper quantifies the classification accuracy of single band density slicing of Landsat 5 TM data to delineate water bodies on riverine floodplains. The results of these analyses are then compared to a 6-band maximum likelihood classification over the same area. The water boundaries delineated by each of these digital classification procedures were compared to water boundaries delineated from colour aerial photography acquired on the same day as the TM data. These comparisons show that Landsat TM data can be used to map water bodies accurately. Density slicing of the single mid-infrared band 5 proved as successful as multispectral classification achieving an overall accuracy of 96.9%, a producer's accuracy for water bodies of 81.7% and a user's accuracy for water bodies of 64.5%.
Article
Water management is a central responsibility of civil society. Major questions persist regarding practice, policy, and the underlying evidence and methods to inform both. Over the next 3 weeks, Science presents essays invited to debate key issues in freshwater research and management. This week: local versus global. When, and to what extent, should a global viewpoint replace, or work in tandem with, enduring localized perspectives?