ArticlePDF Available

Abstract and Figures

Surface soil moisture has a wide application in climate change, agronomy, water resources, and in many other domain of science and engineering. Measurement of soil moisture at high spatial and temporal resolution at regional and global scale is needed for the prediction of flood, drought, planning and management of agricultural productivity to ensure food security. Recent advancement in microwave remote sensing, especially after the launch of Sentinel operational satellites has enabled the scientific community to estimate soil moisture at higher spatial and temporal resolution with greater accuracy. This study evaluates the potential of Sentinel-1A satellite images to estimate soil moisture in a semi-arid region. Exactly at the time when satellite passes over the study area, we have collected soil samples at 37 different locations and measured the soil moisture from 5 cm below the ground surface using ML3 theta probe. We processed the soil samples in laboratory to obtain volumetric soil moisture using the oven dry method. We found soil moisture measured from calibrated theta probe and oven dry method are in good agreement with Root Mean Square Error (RMSE) 0.025 m 3 /m 3 and coefficient of determination (R 2 ) 0.85. We then processed Sentinel-1A images and applied modified Dubois model to calculate relative permittivity of the soil from the backscatter values ( σ∘ ). The volumetric soil moisture at each pixel is then calculated by applying the universal Topp’s model. Finally, we masked the pixels whose Normalised Difference Vegetation Index (NDVI) value is greater than 0.4 to generate soil moisture map as per the Dubois NDVI criterion. Our modelled soil moisture accord with the measured values with RMSE = 0.035 and R 2 = 0.75. We found a small bias in the modelled soil moisture ( 0.02m3/m3 ). However, this has reduced significantly ( 0.001m3/m3 ) after applying a bias correction based on Cumulative Distribution Function (CDF) matching. Our approach provides a first-order estimate of soil moisture from Sentinel-1A images in sparsely vegetated agricultural land.
Content may be subject to copyright.
remote sensing
Article
Estimation of Soil Moisture Applying Modified
Dubois Model to Sentinel-1; A Regional Study from
Central India
Abhilash Singh 1, Kumar Gaurav 1,*, Ganesh Kumar Meena 1and Shashi Kumar 2
1Department of Earth and Environment Sciences, Indian Institute of Science Education and Research,
Bhopal 462066, Madhya Pradesh, India; sabhilash@iiserb.ac.in (A.S.); gkmeena@iiserb.ac.in (G.K.M.)
2Department of Photogrammetry and Remote Sensing, Indian Institute of Remote Sensing, ISRO,
Dehradun 248001, Uttarakhand, India; shashi@iirs.gov.in
*Correspondence: kgaurav@iiserb.ac.in; Tel.: +91-755-669-1383
Received: 11 May 2020; Accepted: 3 June 2020; Published: 15 July 2020


Abstract:
Surface soil moisture has a wide application in climate change, agronomy, water resources,
and in many other domain of science and engineering. Measurement of soil moisture at high
spatial and temporal resolution at regional and global scale is needed for the prediction of
flood,
drought, planning and
management of agricultural productivity to ensure food security.
Recent advancement
in microwave remote sensing, especially after the launch of Sentinel operational
satellites has enabled the scientific community to estimate soil moisture at higher spatial and temporal
resolution with greater accuracy. This study evaluates the potential of Sentinel-1A satellite images
to estimate soil moisture in a semi-arid region. Exactly at the time when satellite passes over the
study area, we have collected soil samples at 37 different locations and measured the soil moisture
from
5 cm
below the ground surface using ML3 theta probe. We processed the soil samples in
laboratory to obtain volumetric soil moisture using the oven dry method. We found soil moisture
measured from calibrated theta probe and oven dry method are in good agreement with Root Mean
Square Error (RMSE) 0.025 m
3
/m
3
and coefficient of determination (R
2
) 0.85. We then processed
Sentinel-1A images and applied modified Dubois model to calculate relative permittivity of the soil
from the backscatter values (
σ
). The volumetric soil moisture at each pixel is then calculated by
applying the universal Topp’s model. Finally, we masked the pixels whose Normalised Difference
Vegetation Index (NDVI) value is greater than 0.4 to generate soil moisture map as per the Dubois
NDVI criterion.
Our modelled
soil moisture accord with the measured values with RMSE = 0.035
and R
2
= 0.75. We found a small bias in the modelled soil moisture (
0.02 m3/m3
). However, this has
reduced significantly (
0.001 m3/m3
) after applying a bias correction based on Cumulative Distribution
Function (CDF) matching. Our approach provides a first-order estimate of soil moisture from
Sentinel-1A images in sparsely vegetated agricultural land.
Keywords: soil moisture; theta probe; Sentinel-1A; NDVI; modified Dubois model
1. Introduction
Soil moisture is a temporary storage of water in soil pores that controls various processes occurring
at the air–soil interface [
1
5
]. Quantification of soil moisture is required on a regular basis for predicting
flood, drought, agricultural productivity, hydrological modelling and climate studies [
6
11
].
Based on
the specific application, soil moisture is needed at different spatial and temporal scales. At a local
scale, it can be measured in the field using Time Domain Reflectometry (TDR) or gravimetric methods.
These measurement techniques provide a more accurate estimate of soil moisture, but they are tedious
Remote Sens. 2020,12, 2266; doi:10.3390/rs12142266 www.mdpi.com/journal/remotesensing
Remote Sens. 2020,12, 2266 2 of 19
and time-consuming. This limits the use of in-situ measurement techniques to measure soil moisture
on global or regional scales.
An alternative to in-situ measurements, surface soil moisture, can be modelled from remote
sensing images. Active microwave remote sensing, specifically Synthetic Aperture Radar (SAR)
imaging has emerged as an effective tool to estimate surface soil moisture. The SAR sensors transmit
microwave electromagnetic pulses and record the backscattered energy from the earth’s surface.
The microwave pulses have high sensitivity towards the dielectric properties of the target and surface
roughness [
12
]. At a given incidence angle, when a SAR signal interacts with the soil-water mixture,
a permittivity (
e
) gradient exists between the dry soil (
e
= 2) and water (
e
= 80), that reflects in the
intensity of radar backscatter [1316].
To retrieve the relative soil permittivity and surface roughness component from the SAR
backscatter, various empirical, semi-empirical and theoretical models have been proposed [
17
24
].
These models are developed for the quad polarised SAR images and are mainly applicable on barren
land. However, the Water Cloud and Dubois models have been successfully used to estimate soil
moisture over barren and sparsely vegetated land [2528].
In a study, Zribi et al.
[29]
have applied Water Cloud Model (WCM) on L-band PALSAR/ALOS-2
satellite data to estimate soil moisture in a tropical agricultural area under dense vegetation cover
conditions. Yang et al.
[30]
have used the fully polarimetric C-band Radarsat-2 SAR data for
soil moisture mapping in Juyanze Basin, China. They concluded that increasing the number of
polarimetric parameters at C-band can provide a more robust estimate of surface soil moisture.
El Hajj et al. [31]
and Bousbih et al.
[32]
have shown that the synergic use of radar (Sentinel-1) and
optical (Sentinel-2) data can be utilised to estimate soil moisture at higher spatial resolution at field
scale. They have used
a neural network
(NN) model to estimate soil moisture by the inversion of
radar signals. In doing so,
they have
generated a synthetic database of the backscattering coefficient
in the VV and HH polarisation for a range of soil moisture, surface roughness and NDVI values by
the parameterisation of coupled WCM and modified Integral Equation Model (IEM). They concluded
that their approach can be used to estimate soil moisture in agricultural plots having NDVI less than
0.75. Further, Qiu et al. [33] used the similar WCM and IEM coupled model to evaluate the impact of
different vegetation indices (NDVI, Enhanced Vegetation Index, and Leaf Area Index) in the estimation
of soil moisture. They reported the accuracy of estimated soil moisture is independent of the choice of
specific vegetation indices.
Hachani et al. [34]
used sentinel-1 images to estimate soil moisture in an
arid climate in Tunisia. They have developed an artificial neural network (ANN) using the training
samples obtained by combining satellite measurements and the simulated backscatter values using
the Integral Equation Model (IEM). They claimed their model is almost site independent and able
to simulate soil moisture content with limited or no ancillary information (i.e., DEM, local Incidence
angle, NDVI). More recently,
Ezzahar et al. [35]
have used Support Vector Machine (SVM), IEM and
Oh models to estimate soil moisture over bare agricultural soil in the Tensfit basin of Morocco from
sentinel-1 satellite images.
Hosseini and McNairn
[36]
applied WCM-Ulaby model on C-band (Radarsat-2) and L-band
(UAVSAR) SAR data to estimate soil moisture and biomass in wheat fields in western Canada. Some
authors [
37
,
38
] have used modified Dubois model and dual polarised (HH and HV) RISAT-1, C-band
data to estimate soil moisture of the Bhal region in Gujrat, India. They found promising results with
good correlation during the initial period of the crop. However, the accuracy decreases greatly at the
locations of dense canopy cover and higher NDVI values in the study area.
After the launch of Sentinel-1A satellite mission in April 2014, the global coverage of SAR data is
easily available at higher spatial and temporal resolution, and is being widely used for soil moisture
estimation [
39
]. Sentinel-1A satellite sensor operates in C-band at frequency 5.405 GHz and acquire
information about the earth’s surface in selectable single (HH or VV) and dual polarisation (HH + HV,
VV + VH) [
40
]. This data is freely available and is widely used in various applications, including soil
moisture estimation [34,38,4143].
Remote Sens. 2020,12, 2266 3 of 19
This study uses dual polarised Sentinel-1A satellite images to estimate soil moisture in a semi-arid
region in central India. We used modified Dubois model to calculate the relative soil permittivity from
SAR backscatter values. Eventually, we input the relative soil permittivity in universal Topp’s model
to obtain volumetric soil moisture in barren or sparsely vegetated farmlands.
2. Study Area
This study is conducted in Bhopal district of Madhya Pradesh in central India (Figure 1).
Bhopal is
divided into two administrative blocks, Berasia in the north and Phanda in the south having
a surface
area of about 1424 km
2
and 1348 km
2
respectively (Figure 1). Climatically, Bhopal lies in
a semiarid
zone and is typically covered by agriculture land (64.5%), barren land (7.3%), forest (13%), and water
bodies (4.6%) [
44
]. The average elevation in the study area varies between 450 and 550 m from the
mean sea level with the gently undulated landscape. The average air temperature ranges between 6
C
and 41 C. About 75% of the study area is covered by black cotton soil formed due to the weathering
of basaltic rocks. The remaining 25% is covered with yellowish-red, mixed soils [45,46].
!!!!
!
!
!
!
!
!
!
!!
!
!
!
!
!
#
###
#
#
#
#
#
#
#
#
#
#
#
#
#
##
77°30'E77°10'E
23°40'N23°20'N
±
0 20
km
Ü
3 km
3 km
18 km
18 km
20 m
#
#
#
#
#
#
#
#
!
17/01/2019
#
29/01/2019
Not in scale
In-situ locations (N=37)
Phanda
Berasia
Figure 1.
Landsat-8 image in False Color Composite (FCC) shows the administrative blocks (Phanda
and Berasia) of Bhopal district, Madhya Pradesh. Circles in Yellow and triangles in black are the
locations of soil moisture measurement in the field. Grid on the top right illustrates the random
sampling strategy to collect soil moisture.
In this study, we have considered Phanda block as a test site to estimate soil moisture. About 44%
of the area of Phanda is cultivable and used for agricultural purposes. The agricultural practice in the
region largely depends on the Indian summer monsoon in June, July, August, and September (JJAS),
where it receives about 92% of total rainfall.
In recent years, the frequency of droughts in central India has increased, which has adversely
impacted the agricultural productivity [
47
]. To obtain the frequency of drought events in the study
Remote Sens. 2020,12, 2266 4 of 19
area, we have calculated Standardized Precipitation Index (SPI) for the monsoon period (JJAS) from
1990 to 2018 using gridded (0.25
0×
0.25
0
) rainfall data obtained from the Indian Meteorological
Department [
48
,
49
]. SPI is used to characterise meteorological drought. For example, SPI values in
a range
between (
0.99 to 0.99) is considered normal, whereas SPI values less than
1 and greater
than 1 are considered to be dry and wet period respectively [
50
]. The SPI values of Phanda block
(Figure 2), clearly suggests that the frequency of drought events has increased in last two decades.
In total,
six drought
events (2002, 2004, 2010, 2014, 2015, and 2017) have occurred between the years
2000 to 2018 (Figure 2).
In this scenario, monitoring soil moisture at higher spatial and temporal resolution has become
important in planning and management of agricultural productivity, water resources and ensuring
food security. The landuse, geology, and soil types of the Phanda block makes it an ideal field site to
study the soil moisture.
Figure 2.
Areal averaged Standardized precipitation Index (SPI) for Indian summer monsoon (June, July,
August, September) from 1990 to 2018. Shaded region (SPI
0.99 to 0.99) shows the normal precipitation
condition. SPI in range (1≤ −2 ) and (1 2) suggests drought and wet condition respectively.
3. Material and Methods
3.1. Satellite Data
This study uses microwave and optical satellite images (Table 1) to estimate soil moisture.
We have
downloaded publicly available Sentinel-1A images of two consecutive pass, i.e., 17 and 29 January
2019 from European Space Agency (https://scihub.copernicus.eu/). Sentinel-1A, C-band SAR records
the backscatter signals day and night, independent of the illumination and weather conditions.
It acquires
microwave images in four exclusive modes; Stripmap (SM), Interferometric Wide swath
(IW),
Extra Wide swath (EW)
, Wave (WV). It can capture images (in terms of polarization) using same
set of transmitted pulses by using its antenna. Depending upon the acquisition mode, Sentinel-1
can acquires images in dual polarisation modes, Vertical–Vertical (VV) and Vertical–Horizontal (VH),
or in single polarisation (HH or VV) at 10 m
×
10 m cell size with a swath 250 km. At this swath,
the incidence angle ranges between 29
to 46
for near and far range respectively. It has a temporal
resolution of 12 days (in combination with Sentinel-1B the temporal resolution is 6 days). Microwave
signals at C-band can penetrate up to 5 cm deep below the soil surface [
51
,
52
]. The Sentinel-1A level-1
data is categorised into two product types: Ground Range Detected (GRD) and Single Look Complex
(SLC). Wide range of applications requires Sentinel-1A GRD product with standard corrections.
Remote Sens. 2020,12, 2266 5 of 19
After applying standard corrections, the Sentinel-1A GRD images will have square pixels (consisting
of amplitude) with reduced speckle [
53
57
]. We have also downloaded Landsat 8 images from the
United States Geological Survey (https://earthexplorer.usgs.gov/) pertaining to the date closest to
the Sentinel-1A images. The Landsat 8 satellite mission carries a two-sensor payload, the Operational
Land Imager (OLI) and the Thermal Infrared Sensor (TIRS). The OLI sensor consists of nine spectral
bands (0.43–1.38
µm
) and acquire images at 16 days revisit time with the spatial resolution 15 m for the
panchromatic band (0.50–0.68µm) and 30 m for the remaining bands [58].
Table 1. Details of the Sentinel-1 and Landsat 8 images used in this study.
Sentinel-1A
Date Polarisations Incidence angle
()
Pixel size
(m ×m) Direction
17/01/2019 VH + VV 38.4 10 ×10 NE
29/01/2019 VH + VV 38.5 10 ×10 NE
Landsat 8
Row/Path Band Wavelength
(µm)
Resolution
(m)
15/01/2019 44/145 4, 5 0.64–0.67, 0.85–0.88 30
22/01/2019 43/146 4, 5 0.64–0.67, 0.85–0.88 30
3.2. Field Measurement
We have used ML3 theta probe sensor and gravimetric method to measure soil moisture in the
field. Theta probe works on the principle of TDR, and it measures the bulk soil relative permittivity
(
e
) at a frequency 100 MHz. This bulk relative permittivity is then converted into volumetric soil
moisture by applying a soil specific calibration of TDR. We have calibrated the theta probe for three
different dominant soil types of our study area. A detailed procedure of calibration is mentioned in the
Appendix A. To measure soil moisture, we have conducted a field campaign on 17 and
29 January 2019
in Bhopal, Madhya Pradesh. These dates coincide with the Sentinel-1A pass over Bhopal at 5:50 a.m.
(IST). At the time of satellite passes, we have measured the soil moisture at 37 locations.
To conduct measurements, we overlay a square grid of
3 km ×3 km
on our study area [
59
].
We then
randomly select few grids to measure the soil moisture. At the centre of each grid,
we measured
the soil moisture by inserting the metal rod of the theta probe at 5 cm below the ground
surface and record the soil moisture value (
m3/m3
) in the data logger and acquired their locations
using
a Garmin-64S
handheld GPS (Figure 3a). We repeated this procedure at least in 8 to 10 different
locations within the grid and finally averaged them to get soil moisture (Figure 1). Simultaneously
at each location, we have collected about 100 g of soil samples at 5 cm below the ground surface
using tubular samplers (Figure 3b). We used these samples to measure soil moisture by oven drying
method (
Figure 3c
). At each of the measurement location, we have also taken observations on land
use,
soil type
, vegetation height and weather parameters (temperature, precipitation, etc.). Table 2
reports the detailed characteristics and location of our sampling sites.
Remote Sens. 2020,12, 2266 6 of 19
Figure 3.
Measurement of soil moisture using (
a
) ML3 sensor and (
b
) Collection of soil samples in
the field. Photograph (
c
) Shows the sample processing in the laboratory for oven drying to calculate
soil moisture.
Remote Sens. 2020,12, 2266 7 of 19
Table 2. Details of the ground condition and soil moisture measured in the field using ML3 theta probe.
Date Site ID Latitute Longitude Elevation
(msl) Land Use Sampling Depth
(cm)
Vegetation Height
(cm)
SM (m3/m3)
(TDR)
17-01-2019 1 23.22913 77.26613 516 Agriculture land (Chickpeas) 5 10–15 0.291 ±0.053
17-01-2019 2 23.22978 77.28679 517 Agriculture land (Chickpeas) 5 15–20 0.170 ±0.028
17-01-2019 3 23.22081 77.31552 515 Agriculture land (Chickpeas) 5 0–5 0.167 ±0.020
17-01-2019 4 23.20669 77.34076 538 Barren land 5 0 0.043 ±0.017
17-01-2019 5 23.18931 77.38358 507 Barren land 5 0 0.109 ±0.032
17-01-2019 6 23.16962 77.33730 532 Agriculture land (Mustard) 5 20–30 0.185 ±0.025
17-01-2019 7 23.16144 77.31003 546 Barren land 5 0 0.111 ±0.014
17-01-2019 8 23.31624 77.41531 495 Agriculture land (Wheat) 5 15–20 0.266 ±0.052
17-01-2019 9 23.33788 77.39777 497 Agriculture land (Wheat) 5 30–35 0.175 ±0.022
17-01-2019 10 23.33280 77.35484 494 Agriculture land (Coriander) 5 10–15 0.275 ±0.046
17-01-2019 11 23.31749 77.32511 503 Agriculture land (Mixed) 5 20–30 0.198 ±0.032
17-01-2019 12 23.32754 77.25596 529 Agriculture land (Wheat) 5 20–30 0.255 ±0.051
17-01-2019 13 23.32683 77.29851 510 Agriculture land (Chickpeas) 5 10–15 0.254 ±0.037
17-01-2019 14 23.34099 77.36872 502 Agriculture land (Wheat) 5 30–40 0.098 ±0.015
17-01-2019 15 23.38717 77.39961 488 Agriculture land (Daikon) 5 0 0.270 ±0.042
17-01-2019 16 23.43098 77.37777 487 Agriculture land (Chickpeas) 5 15–20 0.218 ±0.038
17-01-2019 17 23.50044 77.40357 476 Agriculture land (Wheat) 5 20–25 0.285 ±0.052
17-01-2019 18 23.53360 77.40742 478 Agriculture land (Chickpeas) 5 10-15 0.186 ±0.021
29-01-2019 19 23.41329 77.39910 484 Agriculture land (Wheat) 5 30–40 0.189 ±0.025
29-01-2019 20 23.35487 77.40400 483 Agriculture land (Wheat) 5 40–50 0.174 ±0.024
29-01-2019 21 23.34471 77.43551 490 Agriculture land (Wheat) 5 20–25 0.175 ±0.024
29-01-2019 22 23.35516 77.47569 512 Agriculture land (Wheat) 5 10–15 0.270 ±0.060
29-01-2019 23 23.37469 77.45610 492 Agriculture land (Wheat) 5 40–50 0.129 ±0.018
29-01-2019 24 23.39272 77.44480 478 Agriculture land (Wheat) 5 20–25 0.069 ±0.013
29-01-2019 25 23.36028 77.45131 501 Agriculture land (Mixed) 5 10–15 0.262 ±0.033
29-01-2019 26 23.32369 77.48873 518 Agriculture land (Wheat) 5 30–40 0.099 ±0.024
29-01-2019 27 23.28805 77.51691 491 Agriculture land (Chickpeas) 5 15–20 0.267 ±0.041
29-01-2019 28 23.29685 77.27975 516 Barren land 5 0 0.221 ±0.042
29-01-2019 29 23.29730 77.99768 519 Agriculture land (Chickpeas) 5 20–25 0.180 ±0.027
29-01-2019 30 23.29316 77.26544 528 Agriculture land (Wheat) 5 15–20 0.267 ±0.042
29-01-2019 31 23.29380 77.28096 520 Agriculture land (Wheat) 5 10–15 0.159 ±0.027
29-01-2019 32 23.28579 77.28620 523 Agriculture land (Mixed) 5 5–10 0.234 ±0.046
29-01-2019 33 23.28255 77.28209 519 Agriculture land (Wheat) 5 40–50 0.117 ±0.026
29-01-2019 34 23.28405 77.28794 520 Agriculture land (Chickpeas) 5 10–15 0.153 ±0.028
29-01-2019 35 23.29396 77.26690 526 Agriculture land (Wheat) 5 40–50 0.276 ±0.032
29-01-2019 36 23.25948 77.27253 511 Barren land 5 0 0.102 ±0.034
29-01-2019 37 23.26410 77.29583 515 Agriculture land (Chickpeas) 5 10–15 0.124 ±0.029
Note: Temperature range (at 2 m): (18.4–19.1) C; Earth surface temperature: (13.7–18.1) C; Precipitation: Nil (Source: Weather station Bhopal).
Remote Sens. 2020,12, 2266 8 of 19
3.3. Data Processing
3.3.1. Soil Samples
We processed the soil samples in laboratory to measure their moisture content. We took about
100 g
each of soil samples from a grid and placed them in a separate beaker (Figure 3c). We placed the
samples in an electric oven at 105
C for about 24 h [
60
,
61
]. Once the samples are completely dried,
we weighed
them. The difference between initial and dry weight provides the amount of moisture
present in the soil samples. Eventually, we divide the weight of moisture content with the dry weight
of soil samples. This quantity is the gravimetric soil moisture (mg) and expressed in [kg/kg].
Since theta probe measures the volumetric soil moisture (
mv
), we need to convert our laboratory
measurement into comparable metric for any meaningful comparison. We multiply the gravimetric
soil moisture (
mg
) to the density ratio of soil (
ρsoil
) and water (
ρwater
) to compute the volumetric soil
moisture (mv) in [m3/m3] according to;
mv=mg·ρsoil
ρwater (1)
3.3.2. Satellite Images
We have used Sentinel Application Platform (SNAP) v6.0 to process the Sentinel-1A images.
The processing involves four major steps, radiometric calibration, multilook, speckle noise reduction
using refined Lee filter, and terrain correction. Terrain or geometric calibration uses the Shuttle Radar
Topography Mission (SRTM) digital elevation model of spatial resolution 30 m. The resulting image
pixels contain the true backscatter (
σ
) values on a linear scale. Finally, we convert the backscatter
values into decibel scale (
σdB
) according to,
σ
dB =
10
log10(σ)
. We have also processed Landsat-8
images to compute NDVI. It helps to measure the intensity of vegetation cover in terms of vegetation
density and vegetation height. We take the ratio of the difference between band 5 (Near Infrared) and
band 4 (Red) to the sum of band 5 and band 4 of Landsat-8 images. The NDVI is required to specify
the validity range (NDVI 0.4) of modified Dubois model for soil moisture.
3.4. Soil Moisture Modelling
We have used the backscatter values of Sentinel-1A images and incidence angle in the modified
Dubois model to calculate the relative soil permittivity. Finally, we used the relative permittivity in
universal Topp’s model to compute volumetric soil moisture.
3.4.1. Radar Backscattering Model
Dubois et al.
[19]
developed an empirical model to calculate the relative soil permittivity from
quad polarised SAR images. Initially, this model was developed for L-, C- and X-band data obtained
from scatterometer and later applied on airborne images as well. The model structure is developed on
a strong physical reasoning; however, some of the unknown coefficients are obtained by fitting to the
experimental data. The backscattering coefficient for HH and VV polarisation is given by Equations (2)
and (3).
σ
HH =102.75 cos1.5 θ
sin5θ!100.028etan θ(k.s. sin θ)1.4λ0.7 (2)
σ
VV =102.35 cos3θ
sin3θ!100.046etan θ(k.s. sin θ)1.1λ0.7 (3)
where
θ
is the incidence angle,
e
is the relative soil permittivity, sis the surface roughness (cm),
k=(2π/λ)
is the wavenumber, and
λ
is the SAR wavelength. These parameters can be grouped
into two; sensor parameters (
θ
and
λ
) and target parameters (
e
and s). In Equations (2) and (3),
Remote Sens. 2020,12, 2266 9 of 19
target parameters
are unknown. These equations can be inverted to compute the relative soil
permittivity and surface roughness parameters.
Dubois model is applicable for the SAR images acquired at incidence angle
θ
between 30
to 65
and in the frequency range between 1.5–11 GHz. Further, this model is valid on the region having
sparsely vegetation or barren land. The performance of the Dubois model is maximum where NDVI of
the image pixels is less than 0.4 (Dubois NDVI criterion) or in the region on SAR images where the
ratio of cross-polarised σ
HV /σ
VV is less than 11 dB [19].
Dubois model was initially developed for quad polarised SAR images; it can not be directly
applied to dual polarised data. In a study Rao et al.
[41]
modified the Dubois model by incorporating
in-situ measurements and field conditions and used the Equation (2) to calculate the soil moisture
from the SAR images acquired in HH polarisation.
To derive the unknown parameter (surface roughness) of Equation (3), we obtain the parameter
from regression model proposed by Srivastava et al.
[62]
. Once surface roughness is known,
Equation (2) or (3) can be solved to compute the other unknown, the relative soil permittivity.
A flowchart in Figure 4illustrates the detailed methodology used for in-situ data acquisition,
image processing and modelling to estimate soil moisture.
Sentinel 1A
(VV + VH)
Landsat 8
Input
Pre-processing
Model Input
Backscattering
(VV + VH)
Incidence
Angle
Unsupervised
Classification
(Iso-data)
NDVI
Modified
Dubois model
Regression model Identification of sparse
vegetation and bare land
patches
Soil moisture
estimation
over bare land
Subset, Calibration,
Multi-looking, Georeferencing
& Geometrical correction
Radiometric, Atmospheric
& Geometrical correction,
Georeferencing
Surface Roughness
PCA on
Landsat-8
images
In-situ Measurement
Gravemetric
soil moisture
Voltage of the
dry & wet soil
using TDR
TDR
calibration
Soil moisture
Soil
permittivity
Gravemetric to
volumetric
soil moisture
mvmg*ρsoil
Bulk density
of the soil
Topp's
model
σVH σVV
[dB
Vs
s
Validation
[
Figure 4.
Flow chart illustrates the methodology used for the estimation of soil moisture from
Sentinel-1 images.
3.4.2. Estimation of Soil Moisture
To estimate soil moisture, we used Topp’s model [
63
]. It takes the relative soil permittivity
derived from Sentinel-1A image as an input to estimate volumetric soil moisture content according to
Equation (4)
. This model does not require a priori knowledge of soil properties (texture, grain size)
and is proven to be a robust approach for the estimation of soil moisture [64].
mv=5.3 ×102+2.92 ×102e5.5 ×104e2+4.3 ×106e3(4)
Remote Sens. 2020,12, 2266 10 of 19
4. Results
4.1. Performance of Calibrated Theta Probe
We compared the soil moisture measured from the theta probe and oven dry method.
We observed
no obvious difference, despite a mild scatter, all data points seem to gather around a single line
(
Figure 5
). This suggests a good agreement between (R
2
= 0.84 with RMSE = 0.025 m
3
/m
3
) both the
methods, provided theta probe is correctly calibrated for the specific soil types in the study area.
Hereafter we will use the measurements of theta probe for further analysis.
0.1 0.15 0.2 0.25 0.3
Oven dry SM [m3/m3]
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
TDR SM [m3/m3]
N = 37, RMSE = 0.025, R2= 0.85
Figure 5.
Soil moisture measured using theta probe as a functions of gravimetric method. Shaded gray
region is the 95% confidence level of the regression curve.
4.2. Soil Moisture from Sentinel-1A
Once we calculated the soil moisture from Sentinel-1A images, we evaluated their accuracy with
the in-situ measurement. We first identified the valid region for modified Dubois model using the
Dubois NDVI criterion. For the locations where ground measurement is valid, we extracted the
corresponding pixels from the modelled soil moisture maps. For both the dates, we observed
a good
agreement (R
2
) between the measured and modelled soil moisture (Figure 6).
However, at some
locations, the difference between the modelled and measured values of soil moisture is comparatively
large. This is probably due to the spatial scale mismatch, heterogeneous field conditions, measurement
uncertainty, and model bias. When in-situ measurement is compared with the satellite derived soil
moisture, a representative error arises due to the difference in the spatial scale of in-situ measurement
and satellite observation. The spatial scale mismatch becomes more prominent when the land surface
is heterogeneous. This limits the competency to compare the point measurement with the satellite [
65
].
The error in the retrieval of soil moisture increases with the heterogeneity in land surface
[66]
.
Further, the measurement uncertainty is mainly due to the error associated with in-situ measurement.
For example, error resulting from the conversion of dielectric constant to soil moisture in TDR, presence
Remote Sens. 2020,12, 2266 11 of 19
of organic matter in the soil, overestimation of drying condition in oven and difference in the sampling
depth of the in-situ and satellite measurements [67].
To reduce the measurement uncertainty, we have calibrated the TDR, removed the organic matter
from soil the soil samples, and collected soil samples in the field according to the simulated penetration
depth of the C-band SAR signal in the ground [
51
]. Model error (or bias) is mainly due to the
assumptions followed by its use.
0.05 0.1 0.15 0.2 0.25
TDR SM [m3/m3]
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
Satellite SM [m3/m3]
N = 37, Bias = 0.018 m3/m3, R2= 0.75
Figure 6.
Regression curve between satellite derived and in-situ (TDR) measured soil moisture.
Shaded gray region is the 95% confidence level of the regression curve.
To overcome the systematic errors (instrument calibration and model error), we performed a bias
correction by applying the CDF matching approach [
68
]. This is one of the widely used statistical
methods to minimise the bias in satellite-derived soil moisture [
69
71
]. In principle, we adjusted the
satellite-derived soil moisture according to the CDF of in-situ (TDR) soil moisture to minimise their
difference (Figure 7).
Applying the CDF correction on our data, biases in the soil moisture derived from Sentinel-1A
has reduced from 0.02 m
3
/m
3
to 0.001 m
3
/m
3
. Finally, we used the bias corrected values to generate
soil moisture maps for 17 and 29 January 2019 (Figure 8) of the study area.
Remote Sens. 2020,12, 2266 12 of 19
0.05 0.1 0.15 0.2 0.25 0.3 0.35
Soil moisture [m3/m3]
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
CDF
TDR SM (Reference)
Sentinel 1A SM (Biased)
Sentinel 1A SM (Corrected)
Figure 7.
Bias correction using the CDF matching technique. Solid black line is the CDF of in-situ
measured soil moisture using TDR, dashed and dotted lines in black are the CDF of biased and
corrected soil moisture estimated from sentinel-1A respectively.
00.05 0.1 0.15 0.2 0.25 0.3 0.35
Soil moisture [m3/m3]
17/01/2019 29/01/2019
Figure 8.
Spatial distribution of soil moisture estimated from Sentinel-1. Pixels shown in white
correspond to the region where modified Dubois model is not valid.
Remote Sens. 2020,12, 2266 13 of 19
5. Discussion
This study uses modified Dubois model to estimate soil moisture from dual polarised (VV and VH)
Sentinel-1A images. Our result suggests that the model derived soil moisture accord well (
R2= 0.75
and RMSE = 0.035 m
3
/m
3
) with the in-situ measurement (Figure 6). This is consistent with the range
of RMSE (0.03–0.04) m3/m3reported by other researchers [72].
The modelled soil moisture is subject to bias due to various geometric, atmospheric and modelling
errors. The magnitude of these uncertainties depends on various factors such as choice of a backscatter
model, frequency of SAR images, ground condition and vegetation types. Several methods such as
mean based (linear, local intensity and variance scaling) and the distribution-based (quantile and CDF
matching) have been developed to correct the bias from the modelled soil moisture values [
68
,
73
77
].
We have used the CDF matching approach to minimize the bias from our modelled soil moisture.
In doing so, we have adjusted the CDF of biased values according to the CDF of reference data.
This has
significantly reduced the bias from our modelled soil moisture. Figure 9, shows the relative
soil moisture values estimated from Sentinel-1A images before and after the CDF correction.
5 10 15 20 25 30 35
Sample ID
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
Relative soil moisture [m3/m3]
TDR SM (Reference) Sentinel 1A SM (Biased) Sentinel 1A SM (Corrected)
Figure 9.
Relative soil moisture plotted at their corresponding sample location ID. Difference between
the measured (solid line) and modelled (dotted lines) soil moisture is shown before and after the
bias correction.
We observed, at some locations the modelled soil moisture is overestimated or underestimated
with respect to the measured values. The underestimation is probably related to low backscattering
coefficients from relatively smooth surfaces. Similarly, the overestimation of soil moisture could be
associated with the high surface roughness resulting in high backscattering coefficients.
Though Sentinel-1A images provide a first-order estimation of soil moisture, modelling soil
moisture from satellite images has substantial limitations and challenges. Most of the backscatter
models used for the estimation of the relative soil permittivity from SAR backscatter are only valid for
a specific range of soil moisture. For example, the modified Dubois model is valid for soil moisture in
a range between 0 and 0.35
m3/m3
. Microwave SAR images with appropriate backscatter models can
estimate soil moisture only upto a few cms below the soil surface. In our case, we have estimated the soil
moisture from the top about 5 cm below the ground surface. Further, many of the existing backscatter
models are applicable for specific landuse and landcover classes. For example, Oh and water cloud
models are applicable for barren and vegetated lands, respectively. The modified Dubois model can
be applied on both barren and sparsely vegetated (NDVI
0.4) land. In summary, modified Dubois
model performs well over the semi-arid region on agriculture and barren land. The performance is
further improved when the bias correction method is used.
Moreover, estimation of soil moisture from C-band SAR is sensitive to various parameters such
as; the incidence angle, polarisation, vegetation height and vegetation indices, i.e., Leaf Area Index
Remote Sens. 2020,12, 2266 14 of 19
(LAI). Ulaby
[13]
observed that in a typical agricultural field (LAI
>
0.5), the effect of vegetation is
more prominent in the radar backscatter compared to other target properties.
As a consequence, when the vegetation effect starts dominating, it complicates the soil moisture
retrieval. The backscatter values results from the combined signature of vegetation and underlying
soil water [
78
]. In such condition we can not ignore the attenuation due to vegetation. To minimise the
effect of vegetation, scattering model such as the WCM can be implemented [7983].
6. Conclusions
We have shown that the modified Dubois model provides a good estimate (R
2
= 0.75) of soil
moisture in a region of heterogeneous land cover. The modelled soil moisture is subject to error due to
the model (systematic) and sampling (random) errors. We have shown that the systematic error (also
called systematic bias) can be largely minimised by the CDF matching technique. Whereas random
errors can be reduced by increasing the number of sample size.
We found that the VV polarisation of Sentinel-1A is suitable for soil moisture monitoring.
This is
mainly because the VV polarisation is more sensitive to the soil contribution. In contrast,
VH polarisation
is more sensitive to the volume scattering, and it describes the vegetation contribution
more effectively. Since our model is only for VV or HH, we have used the VV polarisation based on
the polarisation of the Sentinel-1A data.
Further, the potential of VV and HH polarisation of Sentinel-1A may be evaluated, especially for
the models (i.e., WCM) that can estimate soil moisture in diverse vegetation condition. Finally, our
first-order analysis calls for a more detailed study on soil moisture modelling from Sentinel-1A images
in diverse soil, landuse and landcover conditions.
This study is a step towards monitoring the surface soil moisture at higher spatial and temporal
resolution from remote sensing data. Our methodology can be used to predict and monitor
meteorological droughts, agricultural productivity and managing water resources in the region.
Author Contributions:
Conceptualization, A.S. and K.G.; methodology, A.S., G.K.M. and K.G.; software,
A.S. and K.G.; validation, G.K.M., A.S. and K.G.; formal analysis, A.S. and G.K.M.; investigation, K.G. and
A.S.; resources, K.G. and S.K.; data curation, G.K.M., A.S. and K.G.; writing—original draft preparation, A.S.
and K.G.; writing—review and editing, K.G. and A.S.; visualization, K.G. and S.K.; supervision and project
administration, K.G.; funding acquisition, K.G. and S.K. All authors have read and agreed to the published version
of the manuscript.
Funding: This research was funded by the Space Applications Centre (SAC-ISRO) under NASA-ISRO Synthetic
Aperture Radar (NISAR) mission through grant Hyd-01.
Acknowledgments:
We would like to acknowledge IISER Bhopal for providing institutional support.
Abhilash Singh
PhD is supported by the Department of Science and Technology (DST), Government of India
through DST-INSPIRE fellowship. We gratefully acknowledge S.K. Tandon for fruitful suggestions. We thanks to
the editor and all the three anonymous reviewers for providing helpful comments and suggestions.
Conflicts of Interest: The authors declare no conflict of interest.
Appendix A
Theta probe (ML3 sensor) requires an intensive soil specific and a sensor specific calibration
before being used for data acquisition in the field. The soil specific calibration explores a functional
relationship between the relative soil permittivity and soil moisture. In contrast, the sensor specific
calibration explores the functional relationship between the relative soil permittivity and ML3 output
(volts). Both the soil and sensor specific calibration equations are combined to finally convert the
sensor output directly into soil moisture. The soil specific calibration allows to find the two constants
(a0,a1) of Equation (A1). e=a0+a1·mv(A1)
Remote Sens. 2020,12, 2266 15 of 19
For sensor calibration, ML3 sensor measures the bulk relative soil permittivity using the empirical
relation given by Equation (A2);
e=1.0 +6.175 ·V+6.303 ·V273.578 ·V3+183.44 ·V4184.78 ·V5+68.017 ·V6(A2)
where, Vis the voltage in Volt. Combining soil and sensor specific calibration equation reads
(Equation (A3));
mv=ea
a1(A3)
Before the field campaign, we have calibrated the theta probe at three different locations in the
study area (Figure A1).
S3
S2
S1
Figure A1.
Soil samples used from three different fields for the calibration of MLT-3 theta probe.
Samples from S1 and S3 corresponds to agricultural lands composed of black cotton soil and rich
in organic matters. S2 consists of black soil with granules rubble (about 20–25%) formed due to the
subsequent weathering of Deccan basalt.
The corresponding values of the calibration constant is given Table A1.
Table A1. Field derived calibration constants for MLT-3 theta probe.
Measuring Unit Site 1 (S1) Site 2 (S2) Site 3 (S3)
Ls(cm3) 563 563 563
Ww(gm) 491 505 504
Vw(mV) 342 333 350
W0(gm) 410 415 421
V0(mV) 101 72 90
a01.6 1.5 1.6
a17.1 7.4 7.6
To calculate
a0
and
a1
, we have measured two voltage from Equation (A2).
V0
and
Vw
correspond
to the dry and wet samples respectively. We have also measured wet weight (
Ww
) and dry weight (
W0
)
of the samples.
Calculation of a0
: For dry soil (i.e.,
mv
=0), Equation (A1), is reduced to
e0
=
a0
.
Substituting the
value of
V0
in Equation (A2), we have calculated the value of
e0
. Eventually we calculated
a0
according to;
a0=e0(A4)
Calculation of a1: For wet soil, we calculated the soil moisture as;
mv=WwW0
Ls(A5)
Remote Sens. 2020,12, 2266 16 of 19
where,
Ls
is the volume occupied by the sample in the beaker. By substituting the value of
Vm
in
Equation (A2), we have calculated the values of em. Finally a1is calculated according to;
a1=ewe0
mv(A6)
References
1.
Kite, G.; Pietroniro, A. Remote sensing applications in hydrological modelling. Hydrol. Sci. J.
1996
,41,
563–591.
2.
Jackson, T.J. Measuring surface soil moisture using passive microwave remote sensing. Hydrol. Process.
1993
,
7, 139–152.
3.
Entekhabi, D. Retrieval of Soil Moisture Profile by Combined Remote Sensing and Modeling. 1993. Available
online: https://trs.jpl.nasa.gov/handle/2014/34543 (accessed on 11 May 2020).
4.
Jackson, T.; Hawley, M.; O’neill, P. Preplanting soil moisture using passive microwave sensor. JAWRA J. Am.
Water Resour. Assoc. 1987,23, 11–19.
5.
Martin, M.E.; Aber, J.D. High spectral resolution remote sensing of forest canopy lignin, nitrogen, and
ecosystem processes. Ecol. Appl. 1997,7, 431–443.
6.
Sivapalan, M.; Takeuchi, K.; Franks, S.; Gupta, V.; Karambiri, H.; Lakshmi, V.; Liang, X.; McDonnell, J.;
Mendiondo, E.; O’connell, P. IAHS Decade on Predictions in Ungauged Basins (PUB), 2003–2012: Shaping
an exciting future for the hydrological sciences. Hydrol. Sci. J. 2003,48, 857–880.
7.
Field, C.B.; Randerson, J.T.; Malmström, C.M. Global net primary production: Combining ecology and
remote sensing. Remote Sens. Environ. 1995,51, 74–88.
8.
Seneviratne, S.I.; Corti, T.; Davin, E.L.; Hirschi, M.; Jaeger, E.B.; Lehner, I.; Orlowsky, B.; Teuling, A.J.
Investigating soil moisture–climate interactions in a changing climate: A review. Earth-Sci. Rev.
2010
,
99, 125–161.
9.
Kerr, J.T.; Ostrovsky, M. From space to species: Ecological applications for remote sensing. Trends Ecol. Evol.
2003,18, 299–305.
10.
Mulla, D.J. Twenty five years of remote sensing in precision agriculture: Key advances and remaining
knowledge gaps. Biosyst. Eng. 2013,114, 358–371.
11.
Walker, J.P. Estimating Soil Moisture Profile Dynamics from Near-Surface Soil Moisture Measurements and
Standard Meteorological Data. Ph.D. Thesis, University of Newcastle, Callaghan, Australia, 1999.
12.
Narvekar, P.S.; Entekhabi, D.; Kim, S.; Njoku, E.G. Soil Moisture Retrieval Using L-Band Radar Observations.
IEEE Trans. Geosci. Remote Sens. 2015,53, 3492–3506.
13.
Ulaby, F.T. Microwave remote sensing active and passive. In Rader Remote Sensing and Surface Scattering and
Emission Theory; NASA Technical Report; Kansas University: Lawrence, KS, USA, 1982; pp. 848–902.
14.
Altese, E.; Bolognani, O.; Mancini, M.; Troch, P.A. Retrieving soil moisture over bare soil from ERS 1 synthetic
aperture radar data: Sensitivity analysis based on a theoretical surface scattering model and field data.
Water Resour. Res. 1996,32, 653–661.
15.
Verhoest, N.E.; Lievens, H.; Wagner, W.; Álvarez-Mozos, J.; Moran, M.S.; Mattia, F. On the soil roughness
parameterization problem in soil moisture retrieval of bare surfaces from synthetic aperture radar. Sensors
2008,8, 4213–4248.
16.
Barrett, B.W.; Dwyer, E.; Whelan, P. Soil moisture retrieval from active spaceborne microwave observations:
An evaluation of current techniques. Remote Sens. 2009,1, 210–242.
17.
Oh, Y.; Sarabandi, K.; Ulaby, F.T. An empirical model and an inversion technique for radar scattering from
bare soil surfaces. IEEE Trans. Geosci. Remote Sens. 1992,30, 370–381.
18.
Oh, Y.; Sarabandi, K.; Ulaby, F.T. An inversion algorithm for retrieving soil moisture and surface roughness
from polarimetric radar observation. In Proceedings of the IEEE International Geoscience and Remote
Sensing Symposium, Pasadena, CA, USA, 8–12 August 1994; Volume 3, pp. 1582–1584.
19.
Dubois, P.C.; Van Zyl, J.; Engman, T. Measuring soil moisture with imaging radars. IEEE Trans. Geosci.
Remote Sens. 1995,33, 915–926.
20.
Oh, Y.; Sarabandi, K.; Ulaby, F.T. Semi-empirical model of the ensemble-averaged differential Mueller matrix
for microwave backscattering from bare soil surfaces. IEEE Trans. Geosci. Remote Sens. 2002,40, 1348–1355.
Remote Sens. 2020,12, 2266 17 of 19
21.
Oh, Y. Quantitative retrieval of soil moisture content and surface roughness from multipolarized radar
observations of bare soil surfaces. IEEE Trans. Geosci. Remote Sens. 2004,42, 596–601.
22. Attema, E.; Ulaby, F.T. Vegetation modeled as a water cloud. Radio Sci. 1978,13, 357–364.
23.
Fung, A.K.; Li, Z.; Chen, K.S. Backscattering from a randomly rough dielectric surface. IEEE Trans. Geosci.
Remote Sens. 1992,30, 356–369.
24.
Fung, A.K.; Cheng, K.S. Microwave Scattering and Emission Models and Their Applications; Artech House Inc.:
Boston, MA, USA, 1994.
25.
Baghdadi, N.; Cresson, R.; El Hajj, M.; Ludwig, R.; La Jeunesse, I. Soil parameters estimation over bare
agriculture areas from C-band polarimetric SAR data using neural networks. Hydrol. Earth Syst. Sci. Discuss.
2012,9, 2897–2933.
26.
El Hajj, M.; Baghdadi, N.; Zribi, M.; Belaud, G.; Cheviron, B.; Courault, D.; Charron, F. Soil moisture retrieval
over irrigated grassland using X-band SAR data. Remote Sens. Environ. 2016,176, 202–218.
27.
Dabrowska-Zielinska, K.; Musial, J.; Malinska, A.; Budzynska, M.; Gurdak, R.; Kiryla, W.; Bartold, M.;
Grzybowski, P. Soil moisture in the Biebrza Wetlands retrieved from Sentinel-1 imagery. Remote Sens.
2018
,
10, 1979.
28.
Wagner, W.; Pathe, C.; Doubkova, M.; Sabel, D.; Bartsch, A.; Hasenauer, S.; Blöschl, G.; Scipal, K.;
Martínez-Fernández, J.; Löw, A. Temporal stability of soil moisture and radar backscatter observed by the
Advanced Synthetic Aperture Radar (ASAR). Sensors 2008,8, 1174–1197.
29.
Zribi, M.; Muddu, S.; Bousbih, S.; Al Bitar, A.; Tomer, S.K.; Baghdadi, N.; Bandyopadhyay, S. Analysis of
L-band SAR data for soil moisture estimations over agricultural areas in the tropics. Remote Sens.
2019
,
11, 1122.
30.
Yang, L.; Feng, X.; Liu, F.; Liu, J.; Sun, X. Potential of soil moisture estimation using C-band polarimetric
SAR data in arid regions. Int. J. Remote Sens. 2019,40, 2138–2150.
31.
El Hajj, M.; Baghdadi, N.; Zribi, M.; Bazzi, H. Synergic use of Sentinel-1 and Sentinel-2 images for operational
soil moisture mapping at high spatial resolution over agricultural areas. Remote Sens. 2017,9, 1292.
32.
Bousbih, S.; Zribi, M.; El Hajj, M.; Baghdadi, N.; Lili-Chabaane, Z.; Gao, Q.; Fanise, P. Soil moisture and
irrigation mapping in A semi-arid region, based on the synergetic use of Sentinel-1 and Sentinel-2 data.
Remote Sens. 2018,10, 1953.
33.
Qiu, J.; Crow, W.T.; Wagner, W.; Zhao, T. Effect of vegetation index choice on soil moisture retrievals via the
synergistic use of synthetic aperture radar and optical remote sensing. Int. J. Appl. Earth Obs. Geoinf.
2019
,
80, 47–57.
34.
Hachani, A.; Ouessar, M.; Paloscia, S.; Santi, E.; Pettinato, S. Soil moisture retrieval from Sentinel-1
acquisitions in an arid environment in Tunisia: Application of Artificial Neural Networks techniques. Int. J.
Remote Sens. 2019,40, 9159–9180.
35.
Ezzahar, J.; Ouaadi, N.; Zribi, M.; Elfarkh, J.; Aouade, G.; Khabba, S.; Er-Raki, S.; Chehbouni, A.; Jarlan, L.
Evaluation of Backscattering Models and Support Vector Machine for the Retrieval of Bare Soil Moisture
from Sentinel-1 Data. Remote Sens. 2020,12, 72.
36.
Hosseini, M.; McNairn, H. Using multi-polarization C-and L-band synthetic aperture radar to estimate
biomass and soil moisture of wheat fields. Int. J. Appl. Earth Obs. Geoinf. 2017,58, 50–64.
37.
Dave, R.; Kumar, G.; Kr. Pandey, D.; Khan, A.; Bhattacharya, B. Evaluation of modified Dubois model for
estimating surface soil moisture using dual polarization RISAT-1 C-band SAR data. Geocarto Int.
2019
, 1–11.
doi:10.1080/10106049.2019.1655801.
38.
Alexakis, D.D.; Mexis, F.D.K.; Vozinaki, A.E.K.; Daliakopoulos, I.N.; Tsanis, I.K. Soil moisture content
estimation based on Sentinel-1 and auxiliary earth observation products. A hydrological approach. Sensors
2017,17, 1455.
39.
Wagner, W.; Sabel, D.; Doubkova, M.; Bartsch, A.; Pathe, C. The potential of Sentinel-1 for monitoring soil
moisture with a high spatial resolution at global scale. In Proceedings of the Earth Observation and Water
Cycle Science, Frascati, Italy, 18–20 November 2009.
40. Bauer-Marschallinger, B.; Freeman, V.; Cao, S.; Paulik, C.; Schaufler, S.; Stachl, T.; Modanesi, S.; Massari, C.;
Ciabatta, L.; Brocca, L.; et al. Toward global soil moisture monitoring with Sentinel-1: Harnessing assets and
overcoming obstacles. IEEE Trans. Geosci. Remote Sens. 2018,57, 520–539.
Remote Sens. 2020,12, 2266 18 of 19
41.
Rao, S.S.; Das, S.; Nagaraju, M.; Venugopal, M.; Rajankar, P.; Laghate, P.; Reddy, M.S.; Joshi, A.; Sharma, J.
Modified Dubois model for estimating soil moisture with dual polarized SAR data. J. Indian Soc. Remote
Sens. 2013,41, 865–872.
42.
Thanabalan, P.; Vidhya, R. A synergistic approach for soil moisture estimation using modified Dubois model
with dual-polarized SAR and optical satellite data. Proc. SPIE 2016,9877, doi:10.1117/12.2223751.
43.
Mirsoleimani, H.R.; Sahebi, M.R.; Baghdadi, N.; El Hajj, M. Bare soil surface moisture retrieval from
sentinel-1 SAR data based on the calibrated IEM and Dubois models using neural networks. Sensors
2019
,
19, 3209.
44.
NRSC. District and Category Wise Distribution of Land Use and Land Cover in Madhya Pradesh
(2015-16). Available online: https://bhuvan-app1.nrsc.gov.in/2dresources/thematic/LULC503/MAP/MP.
pdf (accessed on 25 May 2020).
45.
Ministry of Water Resources, Central Ground Water Board, North Central Region, Bhopal. District Ground
Water Information Booklet. Available online: http://cgwb.gov.in/District_Profile/MP/Bhopal.pdf (accessed
on 25 May 2020).
46.
Central Ground Water Board, Ministry of Water Resources, River Development and Ganga Rejuvenation,
Government of India. Aquifer Mapping Report Phanda Block, Bhopal District, Madhya Pradesh.
Available online: http://cgwb.gov.in/AQM/NAQUIM_REPORT/MP/PhandaBlockBhopal.pdf (accessed
on 25 May 2020).
47.
Mishra, V.; Shah, R.; Garg, A. Climate Change in Madhya Pradesh: Indicators, Impacts and Adaptation; Indian
Institute of Management Ahmadabad: Ahmedabad, India, 2016.
48.
Pai, D.; Sridhar, L.; Rajeevan, M.; Sreejith, O.; Satbhai, N.; Mukhopadhyay, B. Development of a new high
spatial resolution (0.25
×
0.25) long period (1901–2010) daily gridded rainfall data set over India and its
comparison with existing data sets over the region. Mausam 2014,65, 1–18.
49.
Indian Meteorological Department. Rainfall Data. Available online: http://www.imdpune.gov.in/Clim_
Pred_LRF_New/Grided_Data_Download.html (accessed on 25 May 2020).
50.
Hayes, M.J.; Svoboda, M.D.; Wiihite, D.A.; Vanyarkho, O.V. Monitoring the 1996 drought using the
standardized precipitation index. Bull. Am. Meteorol. Soc. 1999,80, 429–438.
51.
Singh, A.; Meena, G.K.; Kumar, S.; Gaurav, K. Analysis of the effect of incidence angle and moisture content
on the penetration depth of L-and S-band SAR signals into the ground surface. ISPRS Ann. Photogramm.
Remote Sens. Spat. Inf. Sci. 2018,4, 197–202.
52.
Singh, A.; Meena, G.K.; Kumar, S.; Gaurav, K. Evaluation of the Penetration Depth of L-and S-Band (NISAR
mission) Microwave SAR Signals into Ground. In Proceedings of the 2019 URSI Asia-Pacific Radio Science
Conference (AP-RASC), New Delhi, India, 9–15 March 2019.
53.
Potin, P.; Rosich, B.; Grimont, P.; Miranda, N.; Shurmer, I.; O’Connell, A.; Torres, R.; Krassenburg, M.
Sentinel-1 mission status. In Proceedings of the EUSAR 2016: 11th European Conference on Synthetic
Aperture Radar, Hamburg, Germany, 6–9 June 2016; pp. 1–6.
54.
Filipponi, F. Sentinel-1 GRD Preprocessing Workflow. 3rd International Electronic Conference on Remote
Sensing Proceedings 2019,18, 11.
55.
Frison, P.L.; Fruneau, B.; Kmiha, S.; Soudani, K.; Dufrêne, E.; Le Toan, T.; Koleck, T.; Villard, L.; Mougin, E.;
Rudant, J.P. Potential of Sentinel-1 data for monitoring temperate mixed forest phenology. Remote Sens.
2018
,
10, 2049.
56.
Gao, Q.; Zribi, M.; Escorihuela, M.J.; Baghdadi, N. Synergetic use of Sentinel-1 and Sentinel-2 data for soil
moisture mapping at 100 m resolution. Sensors 2017,17, 1966.
57.
Martinis, S.; Plank, S.; ´
Cwik, K. The use of Sentinel-1 time-series data to improve flood monitoring in arid
areas. Remote Sens. 2018,10, 583.
58.
Roy, D.P.; Wulder, M.A.; Loveland, T.R.; Woodcock, C.; Allen, R.G.; Anderson, M.C.; Helder, D.; Irons, J.R.;
Johnson, D.M.; Kennedy, R.; et al. Landsat-8: Science and product vision for terrestrial global change
research. Remote Sens. Environ. 2014,145, 154–172.
59. Guo, M. Soil sampling and methods of analysis. J. Environ. Qual. 2009,38, 375.
60.
Gardner, C.M.; Robinson, D.A.; Blyth, K.; Cooper, J.D. Soil water content. In Soil and Environmental Analysis:
Physical Methods; Smith, K.A., Mullins, C.E., Eds.; Marcel Dekker: New York, NY, USA, 2001; pp. 1–64.
61.
Reynolds, S. The gravimetric method of soil moisture determination Part I. A study of equipment, and
methodological problems. J. Hydrol. 1970,11, 258–273.
Remote Sens. 2020,12, 2266 19 of 19
62.
Srivastava, H.; Patel, P.; Navalgund, R.; Sharma, Y. Retrieval of surface roughness using multi-polarized
Envisat-1 ASAR data. Geocarto Int. 2008,23, 67–77.
63.
Topp, G.C.; Davis, J.; Annan, A.P. Electromagnetic determination of soil water content: Measurements in
coaxial transmission lines. Water Resour. Res. 1980,16, 574–582.
64.
Song, K.; Zhou, X.; Fan, Y. Empirically adopted IEM for retrieval of soil moisture from radar backscattering
coefficients. IEEE Trans. Geosci. Remote Sens. 2009,47, 1662–1672.
65.
Talone, M.; Camps, A.; Monerris, A.; Vall-llossera, M.; Ferrazzoli, P.; Piles, M. Surface topography and
mixed-pixel effects on the simulated L-band brightness temperatures. IEEE Trans. Geosci. Remote Sens.
2007
,
45, 1996–2003.
66.
Zhang, T.; Zhang, L.; Jiang, L.; Zhao, S.; Zhao, T.; Li, Y. Effects of spatial distribution of soil parameters on
soil moisture retrieval from passive microwave remote sensing. Sci. China Earth Sci. 2012,55, 1313–1322.
67.
Lee, J.H.; Zhao, C.; Kerr, Y. Stochastic bias correction and uncertainty estimation of satellite-retrieved soil
moisture products. Remote Sens. 2017,9, 847.
68.
Michelangeli, P.A.; Vrac, M.; Loukos, H. Probabilistic downscaling approaches: Application to wind
cumulative distribution functions. Geophys. Res. Lett. 2009,36, 1–6.
69. Reichle, R.H. Data assimilation methods in the Earth sciences. Adv. Water Resour. 2008,31, 1411–1418.
70.
Reichle, R.H.; Koster, R.D. Bias reduction in short records of satellite soil moisture. Geophys. Res. Lett.
2004
,
31, 1–4.
71.
Draper, C.; Mahfouf, J.F.; Walker, J.P. An EKF assimilation of AMSR-E soil moisture into the ISBA land
surface scheme. J. Geophys. Res. Atmos. 2009,114, 1–13.
72.
Hosseini, R.; Newlands, N.K.; Dean, C.B.; Takemura, A. Statistical modeling of soil moisture, integrating
satellite remote-sensing (SAR) and ground-based data. Remote Sens. 2015,7, 2752–2780.
73.
Kumar, S.; Peters-Lidard, C.; Santanello, J.; Reichle, R.; Draper, C.; Koster, R.; Nearing, G.; Jasinski, M.
Evaluating the utility of satellite soil moisture retrievals over irrigated areas and the ability of land data
assimilation methods to correct for unmodeled processes. Hydrol. Earth Syst. Sci. 2015,19, 4463–4478.
74.
Leander, R.; Buishand, T.A. Resampling of regional climate model output for the simulation of extreme river
flows. J. Hydrol. 2007,332, 487–496.
75.
Schmidli, J.; Frei, C.; Vidale, P.L. Downscaling from GCM precipitation: A benchmark for dynamical and
statistical downscaling methods. Int. J. Climatol. 2006,26, 679–689.
76.
Lenderink, G.; Buishand, A.; Deursen, W.V. Estimates of future discharges of the river Rhine using two
scenario methodologies: Direct versus delta approach. Hydrol. Earth Syst. Sci. 2007,11, 1145–1159.
77.
Panofsky, H.A.; Brier, G.W. Some Applications of Statistics to Meteorology; Pennsylvania State University:
Pennsylvania, PA, USA, 1958.
78.
He, B.; Xing, M.; Bai, X. A synergistic methodology for soil moisture estimation in an alpine prairie using
radar and optical satellite data. Remote Sens. 2014,6, 10966–10985.
79.
Bindlish, R.; Barros, A.P. Parameterization of vegetation backscatter in radar-based, soil moisture estimation.
Remote Sens. Environ. 2001,76, 130–137.
80.
Gherboudj, I.; Magagi, R.; Berg, A.A.; Toth, B. Soil moisture retrieval over agricultural fields from
multi-polarized and multi-angular RADARSAT-2 SAR data. Remote Sens. Environ. 2011,115, 33–43.
81.
Du, J.; Shi, J.; Sun, R. The development of HJ SAR soil moisture retrieval algorithm. Int. J. Remote Sens.
2010
,
31, 3691–3705.
82.
Lievens, H.; Verhoest, N.E. On the retrieval of soil moisture in wheat fields from L-band SAR based on
water cloud modeling, the IEM, and effective roughness parameters. IEEE Geosci. Remote Sens. Lett.
2011
,
8, 740–744.
83.
De Roo, R.D.; Du, Y.; Ulaby, F.T.; Dobson, M.C. A semi-empirical backscattering model at L-band and C-band
for a soybean canopy with soil moisture inversion. IEEE Trans. Geosci. Remote Sens. 2001,39, 864–872.
c
2020 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/).
... Sentinel-1A and 1B satellites together acquire images at a temporal resolution of 6-days (12-days individually) into four different modes; wave, extra-wide swath, interferometric wide swath, and stripmap (Singh et al., 2020;DeVries et al., 2020). The images are available in three different levels; level-0, level-1, and level-2. ...
... Before measuring the soil moisture, we need to calibrate the TDR. This we have done by following the procedures explained in Singh et al. (2020). The penetration depth of Sentinel-1 SAR pulses ranges between 1-5 cm depending upon the target and sensor properties. ...
... It provides reliable estimates if the soil moisture is within a range between 0-0.35 m 3 /m 3 . Wherever quad polarised SAR images are not available, a modified version of the Dubois model is used (Sahebi and Angles, 2010;Rao et al., 2013;Dave et al., 2019;Singh et al., 2020;Thanabalan et al., 2021). Sahebi and Angles (2010) have eliminated the surface roughness through field measurements and reduced the computation to one equation (either σ HH or σ VV depending on the availability and satellite mission) with one unknown (i.e., ∊). ...
Article
Full-text available
We use surface soil moisture content as a proxy to assess the effect of drainage congestion due to structural barriers on the alluvial Fan of the Kosi River on the Himalayan Foreland. We used Sentinel-1 satellite images to evaluate the spatial distribution of soil moisture in the proximity of structural barriers (i.e., road network). We applied modified Dubois and a fully connected feed-forward artificial neural network (FC-FF-ANN) models to estimate soil moisture. We observed that the FC-FF-ANN predicts soil moisture more accurately (R = 0.85, RMSE = 0.05 m3/m3, and bias = 0) as compared to the modified Dubois model. Therefore, we have used the soil moisture from the FC-FF-ANN model for further analysis. We identified the road network that traverses on the Kosi Fan horizontally, vertically, and with inclination. We create a buffer of 1 km along either side of the road. Within this, we assessed the spatial distribution of soil moisture. We observed a high concentration of soil moisture near the structural barrier, and decreases gradually as we move farther in either direction across the orientation of the road. The impact of structural barriers on the spatial distribution of soil moisture is prominent in a range between 300 to 750 m within the road buffer. This study is a step towards assessing the effect of structural interventions on drainage congestion and flood inundation.
... A modified Dubois model was also used to estimate soil moisture based on Sentinel-1 SAR data in central India. Those results showed that SSM varied from 0.05 to 0.35 m 3 /m 3 , and the modeled SSM was found to be either overestimated or underestimated compared to the measured values [61]. ...
... A modified Dubois model was also used to estimate soil moisture based on Sentinel-1 SAR data in central India. Those results showed that SSM varied from 0.05 to 0.35 m 3 /m 3 , and the modeled SSM was found to be either overestimated or underestimated compared to the measured values [61]. The scatter plot was made between satellite-based SSM and in-situ measurements of 54 samples ( Figure 5). ...
... The results exhibited a good agreement with R 2 of 0.77 (p-value < 0.001), RMSE of 0.06 m 3 /m 3 , and percentage bias of 2% between the measured and satellitebased soil moisture. A similar range of R 2 (0.75) and RMSE (0.035 m 3 /m 3 ) was also reported by Singh et al. [61] over central India. Moreover, this statistical performance was consistent with the RMSE reported from 0.03-0.06 ...
Article
Full-text available
Surface Soil Moisture (SSM) is a key factor for understanding the physical process between the land surface and atmosphere. With the advancement of Synthetic Aperture Radar (SAR) technology and backscattering models, retrieval of SSM over the land surface at higher spatial resolution became effective and accurate. This study examines the potential of C-band Sentinel-1 SAR data to derive SSM in a dry season (February 2020) over bare soil and vegetated agricultural fields in the Kosi River Basin (KRB) in North Bihar. Field campaigns were conducted simultaneously with Sentinel–1A acquisition date, and measurements comprised 54 in-situ sampling plots for the top of the soil (0–7.6 cm depth) using time-domain reflectometry (TDR–300). The modified Dubois model was employed to estimate relative soil permittivity from the backscatter values (σ°) of VV polarization. With the help of Topp’s model, volumetric SSM (m3/m3) was derived for all areas with normalized difference vegetation index (NDVI) less than 0.4 that majorly covered bare land or sparse vegetation. The key findings demonstrated that model-derived SSM was well correlated with the in-situ SSM with the coefficient of determination (R2) of 0.77 and root mean square error (RMSE) of 0.06 m3/m3. The spatial distribution of SSM ranged from 0.05 to 0.5 m3/m3 over the KRB, and the highest moisture was found in the Kosi Megafan. The modified Dubois model was effective in providing SSM from Sentinel–1A data in bare soil and agricultural fields and, thus, supporting use in hydrological, meteorological and crop planning applications.
... The gravimetric method is used to calculate in-situ volumetric SM in barren or sparsely vegetated farmland (Sekertekin et al., 2020), and we proposed a model using Sentinel-1 derived Stokes parameters (Raney, 2006;Stokes, 1851) and partial least square regression (PLSR) to estimate volumetric SSM in barren or sparsely vegetated farmland with NDVI less than 0.4. For comparison of our results, we worked on a database published by (Singh et al., 2020) as a second experiment. ...
... The second experiment was realized to verify the consistency of our results, thus a total of 36 surface soil samples (0-5 cm) were collected and measured over a semi-arid area in central India by Singh et al. (2020). More details about this experiment can be accessed in their study (Singh et al., 2020). ...
... The second experiment was realized to verify the consistency of our results, thus a total of 36 surface soil samples (0-5 cm) were collected and measured over a semi-arid area in central India by Singh et al. (2020). More details about this experiment can be accessed in their study (Singh et al., 2020). ...
Article
Surface soil moisture (SSM) is an important key aspect highly applied in several fields as water resources, climate change, and agronomy. The soil moisture estimation at high spatial and temporal resolution is considered as an important aspect for a sustainable environment. This study seeks to examine the potential of Stokes parameters and backscattering coefficients of C-band Sentinel-1 to estimate SSM over bare and sparsely vegetated agricultural fields. In this work, a total of 33 surface soil samples (0–5 cm) were collected from 2 provinces in a semi-arid region of northwestern Morocco, the study is conducted between December 24-2020 and February 23-2021, then analyzed in the laboratory to determine the percentage of soil moisture volume. As a second experiment to verify the consistency of our results, a total of 36 surface soil samples (0–5 cm) were collected and measured over a semi-arid area in central India. In addition to the soil samples, the Sentinel-1 data are acquired. In this study, two models are proposed for the estimation of the SSM, namely a model based on partial least squares regression (PLSR) and radar backscatter, and a model based on partial least squares regression (PLSR) and Stokes parameters. The sensitivity of the backscattering coefficients and Stokes parameters of Sentinel-1 was investigated. In the first experiment, the highest sensitivity was obtained by the model based on Stokes parameters and PLSR (R² = 0.679 and RMSE = 1.96%), for the model that uses the backscattering coefficient and PLSR, we found (R² = 0.60 and RMSE = 2.17). In the second experiment, the greatest sensitivity was also obtained by the model based on Stokes parameters and PLSR (R² = 0.49 and RMSE = 4.94%), for the model based on backscattering coefficient and PLSR we found (R² = 0.39 and RMSE = 5.41). The results showed that the potential of the Stokes and PLSR parameters for estimating the SSM was higher than that of that using the backscattering coefficient and PLSR. Both models showed a significant relationship (p < 0.001) with soil moisture. The results showed that PLSR is a very effective technique for modeling soil moisture.
... Depending on the soil type and moisture conditions, at this frequency, the SAR signals can penetrate up to 5 cm of the topsoil surface [54,55]. Sentinel-1 satellites have a temporal resolution of 12 days, that jointly (1A and 1B) result in a 6-day repeat pass over the equator [56,57]. Sentinel-1 acquires images in four different modes: Stripmap (SM), Interferometric Wide swath (IW), Extra-Wide swath (EW), and Wave (WV). ...
... Based on the acquisition mode, they record the signals in co-polarisation (i.e., VV) or crosspolarisation (i.e., VH) at 10 m × 10 m cell size with 250 km swath. The incidence angle ranges between 29 • and 46 • in near-and far-range, respectively [56]. For our purpose, we have downloaded the dual polarised (VV & VH) Ground Range Detected (GRD) product (Table 1). ...
... At each sampling location, we have also measured the surface soil moisture using TDR-Probe (Theta-Probe). Before taking the measurements, we have calibrated the Theta-Probe (ML3 sensor) for our field using the procedure described in Singh et al. [56]. ...
Article
Full-text available
We apply the Support Vector Regression (SVR) machine learning model to estimate surface roughness on a large alluvial fan of the Kosi River in the Himalayan Foreland from satellite images. To train the model, we used input features such as radar backscatter values in Vertical–Vertical (VV) and Vertical–Horizontal (VH) polarisation, incidence angle from Sentinel-1, Normalised Difference Vegetation Index (NDVI) from Sentinel-2, and surface elevation from Shuttle Radar Topographic Mission (SRTM). We generated additional features (VH/VV and VH–VV) through a linear data fusion of the existing features. For the training and validation of our model, we conducted a field campaign during 11–20 December 2019. We measured surface roughness at 78 different locations over the entire fan surface using an in-house-developed mechanical pin-profiler. We used the regression tree ensemble approach to assess the relative importance of individual input feature to predict the surface soil roughness from SVR model. We eliminated the irrelevant input features using an iterative backward elimination approach. We then performed feature sensitivity to evaluate the riskiness of the selected features. Finally, we applied the dimension reduction and scaling to minimise the data redundancy and bring them to a similar level. Based on these, we proposed five SVR methods (PCA-NS-SVR, PCA-CM-SVR, PCA-ZM-SVR, PCA-MM-SVR, and PCA-S-SVR). We trained and evaluated the performance of all variants of SVR with a 60:40 ratio using the input features and the in-situ surface roughness. We compared the performance of SVR models with six different benchmark machine learning models (i.e., Gaussian Process Regression (GPR), Generalised Regression Neural Network (GRNN), Binary Decision Tree (BDT), Bragging Ensemble Learning, Boosting Ensemble Learning, and Automated Machine Learning (AutoML)). We observed that the PCA-MM-SVR perform better with a coefficient of correlation (R = 0.74), Root Mean Square Error (RMSE = 0.16 cm), and Mean Square Error (MSE = 0.025 cm2). To ensure a fair selection of the machine learning model, we evaluated the Akaike’s Information Criterion (AIC), corrected AIC (AICc), and Bayesian Information Criterion (BIC). We observed that SVR exhibits the lowest values of AIC, corrected AIC, and BIC of all the other methods; this indicates the best goodness-of-fit. Eventually, we also compared the result of PCA-MM-SVR with the surface roughness estimated from different empirical and semi-empirical radar backscatter models. The accuracy of the PCA-MM-SVR model is better than the backscatter models. This study provides a robust approach to measure surface roughness at high spatial and temporal resolutions solely from the satellite data. Keywords: surface roughness; Sentinel-1; Sentinel-2; machine learning models; AutoML; backscatter models
... Moreover, they do not require any pre-installed infrastructure for their operations [4]. As a result, they have a large number of civilians as well as defence applications like precision agriculture through soil moisture, landslide prediction, vehicle traffic monitoring, internet of things (IoT), healthcare, telecommunication, enemy tacking, battlefield surveillance, reconnaissance [5][6][7][8][9][10][11][12][13][14][15][16][17][18], and so on. ...
... Ignoring BEs, the average probability of event detection can be calculated by putting Eq. (3) (for binary sensing range model) and Eq. (6) (for log-normal sensing range model) in Eq. (10), which in turn is put in Eq. (12) to achieve network k-coverage. Further, the average probability of an event detection incorporating BEs can be obtained by putting Eq. (4) and Eq. ...
Article
Coverage is a crucial quality of service (QoS) parameter for wireless sensor networks (WSNs), which tells how effectively the deployed network monitors a given region. Analytical models available for the coverage analysis of finite WSNs are not scalable for large networks due to boundary effects (BEs).The effective coverage area (ECA) of a sensor node lying near the boundary regions is less than the ECA of a sensor node lying in the region's middle. Also, with the presence of obstacles in the transmission path and the radio irregularities, there are frequent changes in the wireless channel characteristics known as shadowing effects (SEs). Therefore, it becomes crucial to include BEs and SEs while investigating the coverage performance of WSNs. In this work, we analyze the-coverage performance of a WSN spread in a circular region of interest (RoI) by considering BEs and using a binary and a log-normal sensing range model.Furthermore, we also assess the effect of various network parameters viz.,the number of sensor nodes, maximum sensing range, and standard deviation of SEs on the-coverage of the WSN. Also, we compare the-coverage outcomes obtained by considering BEs with the results obtained by ignoring BEs. It is found that both BEs and SEs have a significant effect on the-coverage performance of the WSNs. The simulation outcomes substantiate analytical results and match upto a great extent.
... The distribution mapping-based bias correction is capable of removing the bias present in the projected data effectively. We applied the cumulative distribution function (CDF) matching for bias correction from the projected data (rainfall and temperature) by taking the observed values as a reference before feeding them into the hydrological model (Reichle and Koster 2004;Singh et al. 2020;Zhang et al. 2018;Teutschbein and Seibert 2012;Luo et al. 2018). It transfers the CDFs of projected data based on past observations. ...
Article
Full-text available
We use Soil and Water Assessment Tool (SWAT) to simulate the combined effects of land use/land cover (LU/LC) and climate change on the hydrological response of the Upper Betwa River Catchment (UBRC), a semi-arid region in Central India. We execute this model for two different time periods, 1982-2000 and 2001-2018, using the LU/LC data of 1990 and 2018, respectively. We classified the Landsat satellite images of 1990 and 2018 to obtain the dominant LU/LC classes (water body, built-up, forest, agriculture, and open land) in the catchment. The water body, built-up areas, and cropland have increased by 63%, 65%, and 3%, respectively, whereas forest cover and open land decreased by 16% and 23% in the UBRC from 1990 to 2018. The observed climate data in UBRC shows an increase in the average temperature and decrease in the total rainfall during the period between 1980 to 2018. Once the model is set up, we perform the calibration and validation by using the SWAT Calibration Uncertainty Program (SWAT-CUP). We considered two time periods (1991-1994 and 2001-2007) for the calibration and (1995-1998 and 2008-2014) for the validation. For both these time periods, the calibration and validation result of our model is satisfactory. The output of our calibrated model shows a relative decrease in rainfall (12%), surface runoff (21%), and percolation (9%) in the catchment during the period between 2001-2018 as compared to 1982-2000. Finally, we simulate the surface runoff and percolation in the UBRC using the future climate change scenario. We used the bias-corrected multi-model ensemble of CMIP6 GCMs for four different climate scenarios (2023-2100) by assuming no change in the existing LU/LC. We do this for two different time slices: one from 2023-2060 and the other from 2061-2100. For all the climate scenarios , rainfall and surface runoff in the catchment are expected to decrease by 15-40% and 50-79% as compared to the baseline period of 1982-2018. Percolation in the catchment will have a mixed response. It is expected to decrease by 18% in the middle part of the catchment and increase about 25% in the remaining parts of the catchment. Keywords Climate change · CMIP6 · ERA5 · Land use and land cover change · SWAT · Semi-arid catchment
... We found that VV polarization of Sentinel-1A is suitable for soil moisture mapping, because the VV is sensitive to soil contribution. In contrast, the VH polarization is more sensitive to the volume scattering and it describes the vegetation contribution more effectively (Singh et al. 2020). The over estimation of predicted soil moisture values is due to the different landuse classes particularly the settlement/builtup area gives high backscattering values, which is reflect in comparison with barren field. ...
Article
In past studies, several researchers took potential use of multi-temporal optical data and dual-polarized SAR data to assess drought by estimating soil moisture. In this study, Modified Dubois Model (MDM) semi-empirical model with Topp's model is used for retrieval of soil moisture.It involves retrieving the backscattering coefficient from RISAT-1 and SENTINEL-1 datasets to derive the surface roughness and soil moisture conditions. The estimated soil moisture retrieved from microwave SAR parameters is validated with field measurements provides soil moisture spatial variability over different land use classes and bare soil condition. The RISAT-1 derived soil moisture has R² =0.53, whereas SENTINEL-1 shows R² =0.84. It also confirms the possibility of two different polarization σ°HH and σ°VV backscatter involving MDM. It observes that SENTINEL-1 was found well correlated with ground-measured soil moisture. Also, the averaged NDVI sounds reliable with soil moisture ratio, which helps to understand the impact of agricultural drought monitoring.
... Optical data is mainly sensitive to the color of the top surface of the object and the SAR data has sensitivity to the electrical and structural properties of an object. Due to these advantages, the spaceborne SAR data has been widely used for soil characterization (Singh et al., 2020), object detection (Grover et al., 2018), volcanic mapping (Babu and Kumar, 2019), vegetation characterization (Behera et al., 2016;Kumar et al., 2012;Tomar et al., 2019) and several other applications. Due to these advantages, SAR-based remote sensing can give better performance compared to optical sensors (Angiulli et al., 2005). ...
Article
This study compares the utility of multifrequency SAR and Optical multispectral data for land-cover classification of Mumbai city and its nearby regions with a special focus on water body mapping. The L-band ALOS-2 PALSAR-2, X-band TerraSAR-X, C-band RISAT-1, and Sentinel-2 datasets have been used in this work. This work is done as a retrospective study for the dual-frequency L and S-band NASA-ISRO Synthetic Aperture Radar (NISAR) mission. The ALOS-2 PALSAR-2 data has been pre-processed before the implementation of machine learning algorithms for image segmentation. Multi-looking is performed on ALOS-2 PALSAR-2 data to generate square pixels of size 5.78 m and then target decomposition is applied to generate a false-color composite RGB image. While in the case of TerraSAR-X and RISAT-1 datasets, no multi-looking was performed and direct target decomposition was applied to generate false-color composite RGB images. Similarly, for the optical dataset that has a resolution of 10 m, a true color composite, and a false color composite RGB image are generated. For the comparative study between ALOS-2 PALSAR-2 and Sentinel-2 dataset, the RGB images are divided into smaller chunks of size 500*500 pixels each to create a training and testing dataset. Ten image patches were taken from the large dataset, out of which eight patches were used to train the machine learning models Random Forest (RF), K Nearest Neighbor (KNN), and Support Vector Machine (SVM), and two patches were kept for testing and validation purpose. For training the machine learning models, feature vectors are generated using the Gabor filter, Scharr filter, Gaussian filter, and Median filter. For patch 1, the mIOU for true-color composite based Optical image varies from 0.2323 to 0.2866 with the RF classifier performing the best and the mIOU for false-color composite based Optical image varies from 0.4130 to 0.4941 with the RF classifier performing the best while for ALOS-2 PALSAR-2 data, the mIOU varies from 0.4033 to 0.4663 with the RF classifier outperforming the KNN and the SVM classifiers. For patch 2, the mIOU for true-color composite based Optical data varies from 0.3451 to 0.4517 with KNN performing the best and the mIOU for false-color composite based Optical image varies from 0.5156 to 0.5832 with the RF classifier performing the best while for ALOS-2 PALSAR-2 data, the mIOU varies from 0.4600 to 0.5178 with the RF classifier outperforming the KNN and the SVM classifiers. The gap between the performance of ALOS-2 PALSAR-2 data and Sentinel-2 optical data is observed when the IOU of water class is compared, with IOUw for the true-color composite based optical image at a maximum of 0.2525 and for false-color composite based optical image at a maximum of 0.7366 while for ALOS-2 PALSAR-2 data a maximum IOUw of 0.7948 is achieved. The better performance of SAR data as compared to true-color composite based optical image data is due to the misclassification of ground and water classes into urban and forest in the case of the true-color composite based optical dataset which can be attributed to the high similarity between water and forest classes in the case of true-color composite based optical data whereas both these classes are easily separable in case of SAR data. This issue is however resolved by using the false color composite based optical image dataset for the classification task which performs slightly better than ALOS-2 PALSAR-2 data in the overall classification task. However, the SAR data works best in water body detection as notable from the high IOU for water class in the case of SAR data. In addition to the comparative analysis between Sentinel-2 optical and ALOS-2 PALSAR-2 data, land-cover classification has been performed on X-band TerraSAR-X and C-band RISAT-1 data on a single patch and it has been found that the RF classifier performs the best, recording the mIOU 0.5815 for X-band TerraSAR-X data, mIOU of 0.4031 for the C-band RISAT-1 data, and mIOU of 0.6153 for the L-band ALOS-2 data.
Chapter
Globally, floods are attributed to be one of the leading natural hazards responsible for recurrent major economic losses, population affected, and mortality. The rapid assessment of flood hazard dynamics at regional scale during flood crisis is one of the few elements which is required by the agencies involved on ground for relief and rescue operations. Due to the weather-independent and day and night acquisition capability offered through microwave sensors, space-borne remote sensing for flood hazard management has undergone a paradigm change. Today, globally data from synthetic aperture radar (SAR) has emerged as invaluable source for monitoring flood hazard. From demonstrating the proof of concept in its initial launch campaigns, the SAR technology has matured to be competent enough to provide operational support for major flood disasters. In recent times, the continuously streaming of free SAR datasets from Sentinel-1 mission and together with emergence of advanced cloud-based computing and processing technologies like the Google Earth Engine (GEE), automated, and quasi-real-time flood mapping services have evolved. The future missions like the NISAR in conjunction with Sentinel-1 C-band data will help in providing more accurate and faster response during flood crisis and see application of SAR data grow multi-fold in coming years for flood hazard mitigation. This chapter attempts to provide a broad overview of the active microwave remote sensing for flood hazard studies. The first part of the chapter discusses about the interaction of the SAR signal for flooding in open, vegetated, and urbanized areas, followed by the role of the sensor parameters like the wavelength, polarization, and incidence angle on the backscattering of SAR signal. The latter half of the chapter discusses about the flood mapping techniques, SAR satellite mission contributing to flood hazard mapping, various applications of SAR derived flood hazard information, the Indian nationwide near-real-time (NRT) flood hazard mapping under ISRO DMS program, and the emergence of Web-based cloud computing techniques and open-source data policies revolutionizing the flood hazard mitigation activities.
Article
In this research work, models for the scattering from rough surfaces and semi-empirical soil moisture models are presented. Models using the “Bidirectional reflectance distribution function (BRDF)” are focussed in this study. Therefore the effect of the “surface roughness” on the moisture content estimation can be analyzed. “Kirchhoff Approximation Model,” “Small Perturbation Model,” “Integral Equation Model,” and “Small Slope Approximation Method” are mathematically investigated in this research work. Integral Equation Model and Small Slope Approximation Method are having the most extensive range of validity. The main disadvantage of these methods is that they follow a complex methodology and are pretty tricky to implement. Two semi-empirical soil moisture models, popularly known as the “Dubois model” and “Oh model,” are also presented and analyzed in this research work. These models assess the moisture content present in the soil depending upon several factors, i.e., incidence angle (θ), wavelength (λ), frequency (ν), scattering coefficients, etc. Scattering coefficients provide information in terms of polarization in “horizontal-horizontal (HH),” “vertical-vertical (VV),” “horizontal-vertical (HV),” or “vertical-horizontal (VH)” directions. Finally, a novel IoT-based resistive soil moisture sensor is developed and presented in this research work which provides voltage values corresponding to different moistured soil surfaces. Thus in this work, complex mathematics behind the scattering from rough surfaces is presented. Popular semi-empirical soil moisture models for moisture content estimation are presented. Finally, a prototype of the soil moisture sensor is developed to predict the moisture conditions for the different soil surfaces.
Article
Full-text available
The main objective of this work was to retrieve surface soil moisture (SSM) by using scattering models and a support vector machine (SVM) technique driven by backscattering coefficients obtained from Sentinel-1 satellite images acquired over bare agricultural soil in the Tensfit basin of Morocco. Two backscattering models were selected in this study due to their wide use in inversion procedures: the theoretical integral equation model (IEM) and the semi-empirical model (Oh). To this end, the sensitivity of the SAR backscattering coefficients at V V ( σ v v ∘ ) and V H ( σ v h ∘ ) polarizations to in situ soil moisture data were analyzed first. As expected, the results showed that over bare soil the σ v v ∘ was well correlated with SSM compared to the σ v h ∘ , which showed more dispersion with correlation coefficients values (r) of about 0.84 and 0.61 for the V V and V H polarizations, respectively. Afterwards, these values of σ v v ∘ were compared to those simulated by the backscatter models. It was found that IEM driven by the measured length correlation L slightly underestimated SAR backscatter coefficients compared to the Oh model with a bias of about − 0.7 dB and − 1.2 dB and a root mean square (RMSE) of about 1.1 dB and 1.5 dB for Oh and IEM models, respectively. However, the use of an optimal value of L significantly improved the bias of IEM, which became near to zero, and the RMSE decreased to 0.9 dB. Then, a classical inversion approach of σ v v ∘ observations based on backscattering model is compared to a data driven retrieval technic (SVM). By comparing the retrieved soil moisture against ground truth measurements, it was found that results of SVM were very encouraging and were close to those obtained by IEM model. The bias and RMSE were about 0.28 vol.% and 2.77 vol.% and − 0.13 vol.% and 2.71 vol.% for SVM and IEM, respectively. However, by taking into account the difficultly of obtaining roughness parameter at large scale, it was concluded that SVM is still a useful tool to retrieve soil moisture, and therefore, can be fairly used to generate maps at such scales.
Article
Full-text available
The present study focuses on the soil moisture estimation using dual polarimetric RISAT-1, C-band SAR data in HH and HV polarization. The semi empirical approach derived by Dubois and the same modified and proposed as Modified Dubois Model (MDM) is worked for winter wheat crop from initial vegetative to maturity stage. Surface roughness model derived roughness is replaced with ‘s’ parameter for retrieval of dielectric constant which is further used in Topp’s model to derive soil moisture. Model derived soil moisture and ground measured soil moisture demonstrates pragmatic correlation after performing linear regression. The behavior of inversion (MDM) model varies at different crop stages depending on the soil exposure and crop cover. Normalized Difference Vegetation Index (NDVI) from Moderate Resolution Imaging Spectroradiometer (MODIS) has also been evaluated to monitor crop status. Performance of MDM is promising for early and maturity stages of crop for soil moisture estimation.
Article
Full-text available
The main purpose of this study is to investigate the performance of two radar backscattering models; the calibrated integral equation model (CIEM) and the modified Dubois model (MDB) over an agricultural area in Karaj, Iran. In the first part, the performance of the models is evaluated based on the field measurement and the mentioned backscattering models, CIEM and MDB performed with root mean square error (RMSE) of 0.78 dB and 1.45 dB, respectively. In the second step, based on the neural networks (NNS), soil surface moisture is estimated using the two backscattering models, based on neural networks (NNs), from single polarization Sentinel-1 images over bare soils. The inversion results show the efficiency of the single polarized data for retrieving soil surface moisture, especially for VV polarization.
Conference Paper
Full-text available
Microwave remote sensing has emerged as an efficient tool for the estimation of soil moisture due to its higher sen-sitivity to the dielectric properties of soil. Synthetic Aperture Radar (SAR) sensors have been used to estimate soil moisture at large scale. The penetration depth of microwave signals into the ground soil vary significantly with the available moisture content. It has been found that the longer wavelengths have a higher capability to penetrate soil,however, their penetration capability decreases with increasing dielectric behaviour of the soil. Moisture contentin the soil increases the dielectric behaviour of soil.Here we study the effect of soil texture, incidence angle and soil moisture content on the penetration depth of microwave pulses into ground. In doing so, we have compared the penetration depth of SAR signals in soil usingthe Dobson empirical [1], Dobson semi-empirical [2] and Hallikainen empirical model [3] over samples collectedat three different locations in Bhopal, Madhya Pradesh, India. We observed all these models results into differentpenetration depth for the same set of soil parameters at same frequency. Further, we explored the potential of theproposed L- and S-band sensor of the upcoming NISAR mission for the estimation penetration depth in soil atdifferent incident angles. According to the NISAR mission science user handbook, the frequency used for L- andS-band SAR systems are 1.25 GHz and 3.22 GHz respectively with variation in the incident angle of 330to 470respectively [4].We observed the penetration depth of microwave signals into the ground decreases as moisture content in the soilsample, incident angle and frequency of SAR sensors increases. The maximum penetration depth is observedat nadir. The decrease in the soil penetration depth is significant for the first 10% increase in Volumetric WaterContent (VWC). However, any further increase in moisture content has a reduced effect on soil penetration depth.The penetration depth is more in L-band SAR signal as compared with the S-band SAR signal. Finally, we conclude that the Dobson empirical model results into higher penetration depth whereas Dobson's semi-empiricalmodel into the lowest penetration depth.
Conference Paper
Full-text available
The Copernicus Programme has become the world’s largest space data provider, providing complete, free and open access to satellite data, mainly acquired by Sentinel satellites. Sentinel-1 Synthetic Aperture Radar (SAR) data have improved spatial resolution and high revisit frequency, making them useful for a wide range of applications. While few research applications need Sentinel-1 Ground Range Detected (GRD) data with few corrections applied, a wider range of users needs products with a standard set of corrections applied. In order to facilitate the exploitation of Sentinel-1 GRD products, there is the need to standardise procedures to preprocess SAR data to a higher processing level. A standard generic workflow to preprocess Copernicus Sentinel-1 GRD data is presented here. The workflow aims to apply a series of standard corrections, and to apply a precise orbit of acquisition, remove thermal and image border noise, perform radiometric calibration, and apply range Doppler and terrain correction. Additionally, the workflow allows spatially snapping of Sentinel-1 GRD products to Sentinel-2 MSI data grids, in order to promote the use of satellite virtual constellations by means of data fusion techniques. The presented workflow allows the production of a set of preprocessed Sentinel-1 GRD data, offering a benchmark for the development of new products and operational down-streaming services based on consistent Copernicus Sentinel-1 GRD datasets, with the aim of providing reliable information of interest to a wide range of communities.
Article
Full-text available
The main objective of this study is to analyze the potential use of L-band radar data for the estimation of soil moisture over tropical agricultural areas under dense vegetation cover conditions. Ten radar images were acquired using the Phased Array Synthetic Aperture Radar/Advanced Land Observing Satellite (PALSAR/ALOS)-2 sensor over the Berambadi watershed (south India), between June and October of 2018. Simultaneous ground measurements of soil moisture, soil roughness, and leaf area index (LAI) were also recorded. The sensitivity of PALSAR observations to variations in soil moisture has been reported by several authors, and is confirmed in the present study, even for the case of very dense crops. The radar signals are simulated using five different radar backscattering models (physical and semi-empirical), over bare soil, and over areas with various types of crop cover (turmeric, marigold, and sorghum). When the semi-empirical water cloud model (WCM) is parameterized as a function of the LAI, to account for the vegetation’s contribution to the backscattered signal, it can provide relatively accurate estimations of soil moisture in turmeric and marigold fields, but has certain limitations when applied to sorghum fields. Observed limitations highlight the need to expand the analysis beyond the LAI by including additional vegetation parameters in order to take into account volume scattering in the L-band backscattered radar signal for accurate soil moisture estimation.
Article
Full-text available
In this study, the potential of Sentinel-1 data to seasonally monitor temperate forests was investigated by analyzing radar signatures observed from plots in the Fontainebleau Forest of the Ile de France region, France, for the period extending from March 2015 to January 2016. Radar backscattering coefficients, σ0 and the amplitude of temporal interferometric coherence profiles in relation to environmental variables are shown, such as in situ precipitation and air temperature. The high temporal frequency of Sentinel-1 acquisitions (i.e., twelve days, or six, if both Sentinel-1A and B are combined over Europe) and the dual polarization configuration (VV and VH over most land surfaces) made a significant contribution. In particular, the radar backscattering coefficient ratio of VV to VH polarization, σVV0/σVH0, showed a well-pronounced seasonality that was correlated with vegetation phenology, as confirmed in comparison to NDVI profiles derived from Landsat-8 (r=0.77) over stands of deciduous trees. These results illustrate the high potential of Sentinel-1 data for monitoring vegetation, and as these data are not sensitive to the atmosphere, the phenology could be estimated with more accuracy than optical data. These observations will be quantitatively analyzed with the use of electromagnetic models in the near future.
Article
Full-text available
The objective of the study was to estimate soil moisture (SM) from Sentinel-1 (S-1) satellite images acquired over wetlands. The study was carried out during the years 2015–2017 in the Biebrza Wetlands, situated in north-eastern Poland. At the Biebrza Wetlands, two Sentinel-1 validation sites were established, covering grassland and marshland biomes, where a network of 18 stations for soil moisture measurement was deployed. The sites were funded by the European Space Agency (ESA), and the collected measurements are available through the International Soil Moisture Network (ISMN). The SAR data of the Sentinel-1 satellite with VH (vertical transmit and horizontal receive) and VV (vertical transmit and vertical receive) polarization were applied to SM retrieval for a broad range of vegetation and soil moisture conditions. The methodology is based on research into the effect of vegetation on backscatter (σ°) changes under different soil moisture and Normalized Difference Vegetation Index (NDVI) values. The NDVI was derived from the optical imagery of a MODIS (Moderate Resolution Imaging Spectroradiometer) sensor onboard the Terra satellite. It was found that the state of the vegetation expressed by NDVI can be described by the indices such as the difference between σ° VH and VV, or the ratio of σ° VV/VH, as calculated from the Sentinel-1 images in the logarithmic domain. The most significant correlation coefficient for soil moisture was found for data that was acquired from the ascending tracks of the Sentinel-1 satellite, characterized by the lowest incidence angle, and SM at a depth of 5 cm. The study demonstrated that the use of the inversion approach, which was applied to the newly developed models using Water Cloud Model (WCM) that includes the derived indices based on S-1, allowed the estimation of SM for wetlands with reasonable accuracy (10 vol. %). The developed soil moisture retrieval algorithms based on S-1 data are suited for wetland ecosystems, where soil moisture values are several times higher than in agricultural areas.
Article
In this paper, an approach for estimating the soil moisture content (SMC) in arid environment in Tunisia is presented. In countries characterized by arid and semi-arid climate, it is very important to obtain reliable estimates of soil moisture evolution for water management purposes, in order to reduce water wastes and properly schedule agricultural practices. On the other hand, the retrieval of SMC is often hampered by the small humidity range (below 10%). A retrieval algorithm aiming at estimating the soil moisture and based on artificial neural networks (ANN) has therefore been implemented, using the data collected by the Synthetic Aperture Radar (SAR) sensor of Sentinel-1. By taking advantage of the fast computation and high retrieval accuracy, ANN are able to generate reliable output maps of SMC starting from the complex SAR images and using little auxiliary information, as Digital Elevation Models, Local Incidence angle, Normalized Difference Vegetation Index (NDVI), and so on. The peculiar strategy adopted for the training, which has been obtained by combining satellite measurements with data simulated by electromagnetic model (based on the Integral Equation Model, IEM), made this algorithm robust and almost site independent. The obtained results demonstrated that ANN represent a powerful tool for estimating SMC, provided that they have been trained with consistent datasets, made up by both experimental and theoretical data. The relationship of the algorithm validation between the estimated and measured SMC showed Pearson’s correlation coefficient, r = 0.77, and RMSE = 1.84%, in spite of the very low SMC values found on the area.
Article
The recent launch of the Sentinel-1 A and Sentinel-1B synthetic aperture radar (SAR) satellite constellation has provided high-quality SAR data with fine spatial and temporal sampling characterizations (6˜12 revisit days at 10m spatial resolution). When combined with high-resolution optical remote sensing, this data can potentially be used for high-resolution soil moisture retrieval over vegetated areas. However, the suitability of different vegetation index (VI) types for the parameterization of vegetation water content in SAR vegetation scattering models requires further investigation. In this study, the widely-used physical-based Advanced Integral Equation Model (AIEM) is coupled with the Water Cloud Model (WCM) for the retrieval of field-scale soil moisture. Three different VIs (NDVI, EVI, and LAI) produced by two different satellite sensors (Moderate Resolution Imaging Spectroradiometer (MODIS) and Landsat) are selected to examine their impact on the parameterization of vegetation water content, and subsequently, on soil moisture retrieval accuracy. Results indicate that, despite the different sensitivity of estimated surface roughness parameters to various VIs (i.e., this sensitivity is highest when utilizing MODIS EVI and lowest in the LAI-based model), the optimum roughness parameters derived from each VI exhibit no discernible difference. Consequently, the soil moisture retrieval accuracies show no noticeable sensitivity to the choice of a particular VI. Generally, meadow and grassland sites with small differences in VI-derived roughness parameters exhibit good performance in soil moisture estimation. With respect to the relative components in the coupled model, the vegetative contribution to the scattering signal exceeds that of soil at VI about 0.6∼0.8 [-] in NDVI-based models and 0.4∼0.6 [-] in EVI-based models. This study provides insight into the proper selection of vegetation indices during the use of SAR and optical imagery for the retrieval of high resolution surface soil moisture.