INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
Remote Sensing of Regional Crop
Evapotranspiration Estimation
Abhishek Danodia
Scientist/Engineer ‘SD’
Agriculture & Soils Department
abhidanodia@iirs.gov.in
05/08/2020
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
Contents :
Crop evapotranspiration: Fundamentals
RS based Crop evapotranspiration: Traditional approach
RSEB Models (One Source & Two Source)
Trapezoidal Feature Space based method (S-SEBI Model, TVT Model)
Physical equation based method (METRIC)
Two Source Energy Balance (TSEB) Model
Seasonal crop evapotranspiration
Global ET Products
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
ET: Measurement & Computation
Measurement Computation
Lysimeter method Temperature based approach
Atmometer Radiation based approach
Pan evaporative method Combination approach
Aerodynamic methods Energy Balance - RS approach
Soil-water balance methods (Scatter based method, Land
Large Aperture Scintillometer Surface Temperature based
method)
Eddy flux tower
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
Crop evapotranspiration (ETc)
The crop evapotranspiration
under standard conditions,
denoted as ETc, is the
evapotranspiration from disease-
free, well-fertilized crops, grown in
large fields, under optimum soil
water conditions, and achieving full
production under the optimum
climatic conditions.
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
Crop evapotranspiration:
Traditional approach
ETc is governed by weather, crop condition, management
and environmental aspects.
ETc = kc * ET0
Where,
ETc = Actual crop evapotranspiration rate
kc = Crop coefficient
ET0 = Evapotranspiration rate for a grass reference crop /
Reference crop evapotranspiration
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
Reference crop evapotranspiration (ETo)
A consultation of experts &
researchers was organized by
FAO in May 1990, in
collaboration with the
International Commission for
Irrigation & Drainage and with
the World Meteorological
Organization, to review the
FAO methodologies on crop
water requirements.
“A hypothetical reference crop with an assumed crop height of 0.12 m, a
fixed surface resistance of 70 sm-1 and an albedo of 0.23.’’
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
FAO Penman-Monteith Equation
The FAO Penman-Monteith method is selected as the method
by which the evapotranspiration of this reference surface
(ETo) can be unambiguously determined, and as the method
which provides consistent ETo values in all regions and
climates.
Why ?
The method gives most accurate at most of the locations
Physically based
Explicitly incorporates the physiological and aerodynamic
parameters
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
RS based Crop coefficient method
The command area of MRBC in Kheda district of Gujarat state in West-
central India.
MRBC lies between 22°26’N 72°49’E and 22°55’N 73°23’E.
It serves 485 villages of seven talukas covering a cultivable command area
(CCA) of 212.694 thousand hectares.
The main canal has six
branches with 38 distributaries.
Climate- Semi-arid
Annual rainfall 823 mm
Soil- deep, varying from loamy
sand to clay in texture.
Ray & Dadhwal 2001, AWM
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
RS Data:
IRS-1C WiFS data (188m)
One scene from each month from November to February
Used two spectral bands in red (620-680 nm) and near-
infrared (770-860 nm) region
Exercise for crop classification and computation of crop
coefficient.
Meteorological Data:
Monthly temperature, daytime wind speed, sunshine hours,
relative humidity
To compute ET0 using Blaney-Criddle method
Ray & Dadhwal 2001, AWM
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
Methodology:
Ray & Dadhwal 2001, AWM
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
Decision Rules:
Ray & Dadhwal 2001, AWM
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
Ray & Dadhwal 2001, AWM
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
Ray & Dadhwal 2001, AWM
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
The relative
deviation between
these two estimates
were between 7.2 to
12.8%, with RS-
based methodology
giving a lower
estimate. However,
the strength of the
RS-based estimate
lies in giving a
spatial information.
Seasonal crop evapotranspiration map of MRBC command area
Ray & Dadhwal 2001, AWM
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
Remote Sensing Energy Balance (RSEB) Model
The surface energy balance equation is written as:
Rn = H + LE + G
Where,
Rn is the net radiation flux; H is the sensible heat flux, LE is the
latent heat flux and G is the soil/ground heat flux.
Thus, ET is generally determined from satellite imagery by
applying an energy balance at the surface, where energy
consumed by the ET process is calculated as a residual of the
surface energy equation
LE = Rn - H - G
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
REMOTE SENSING ENERGY BALANCE (RSEB) MODEL
S-SEBI (Simplified-Surface Energy Balance Index)
SEBS (Surface Energy Balance System)
TVT (Temperature Vegetation Index)
SEBAL (Surface Energy Balance Algorithm System)
METRIC (Mapping ET at high Resolution with Internalized
Calibration)
SSEBop (Operational Simplified- Surface Energy Balance)
ALEXI (Atmosphere Land Exchange Inverse)
Sim-ReSET (Simple Remote Sensing Evapo Transpiration)
TSEB (Two Source Energy Balance)
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
S-SEBI (Simplified-Surface Energy Balance Index) Model
S-SEBI has two major advantages over other RS-EB
models:
(I) No additional meteorological data is needed for energy
flux estimation if the surface extremes of vegetation cover
and soil moisture are available,
(II) This model concerns about the extreme temperature
of the wet and dry conditions which varies with changing
reflectance values, where other models try to calculate a
static value of temperature for wet and dry conditions,
both for the whole image and/or for each land use class
(Roerink et al. 2000).
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
S-SEBI computes ETc using the evaporative fraction (Λ) theory
Methodology: Flow chart
Danodia et al. 2017, GI
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
RS derived Energy flux components:
Net radiation
Rn = (1- ).RS↓ + RL↓ - RL↑ - (1- ε0).RL↓
Soil heat flux
G = Rn * Ts / (0.0038+0.00742) (1-0.98NDVI4)
Evaporative fraction
λE λE
𝛬= =
λE + H Rn − G
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
In the case of a wet pixel, H can
be assumed as zero, so λEmax can
be estimated by subtracting G
from Rn (i.e. λEmax = Rn - G).
While at the dry pixel H will be the
highest (Hmax), which can be
estimated by subtracting the G
from Rn (i.e. Hmax = Rn - G). In
such case the Λ can be expressed
as:
TH − T0 T0 is LST of the individual pixel,
𝛬= TH and TλE are the LST of dry pixels and
TH − TλE
wet pixels
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
𝑻𝑯 = 𝒂𝑯 + 𝒃𝑯 𝛂 𝐚𝐇 + 𝐛𝐇 𝛂 − 𝐓𝐒
𝚲=
𝐚𝐇 −𝐚𝛌𝐄 +ሺ𝐛𝐇 − 𝐛𝛌𝐄 )𝛂
𝑻𝝀𝑬 = 𝒂𝝀𝑬 + 𝒃𝝀𝑬 𝛂
Once evaporative fraction is determined, Latent Heat Flux (λE)
and Sensible heat flux (H) are calculated as;
λE = Ʌ. (Rn -G)
H = (1- Ʌ) (Rn -G)
Here it can be assumed that the instantaneous evaporative
fraction is equal to the daily evaporative fraction on the basis of
S-SEBI theory:
d =i = λEi / (Rni – Gi) = λEd / (Rnd – Gd)
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
Assuming that the integrated daily soil heat flux is
approximately equal to zero (Allen et al., 1998), thus the daily
ET estimation by S-SEBI can be calculated as:
ET = λEi * ( Rnd / λRni )
where, Rni is instantaneous net radiation at the time of satellite
overpass and Rnd is the daily net radiation.
Seguin and Itier (1983) proposed a procedure for calculation
of Rnd in S-SEBI. They showed that the ratio between
instantaneous and daily values of net radiation is constant
during the daytime.
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
Large Aperture Scintillometer (LAS): Measurement
Record the logarithm of the structure
parameter of the refractive index of air
LAS measures sensible heat flux (H)
Wavelength (840–880 nm) over a
known path length to the receiver
EVATION processing software
Danodia et al. 2017, JESS
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
LAS:
EVATION
Danodia et al. 2017 (IIRS-IARI collaborative project)
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
Large Aperture Scintillometer: Measurement
Positive correlation LAI and LE (R2= 0.80) &
Negative correlation H and LAI (R2= -0.79)
Danodia et al. 2017, JESS
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
Case study: S-SEBI Model
Sugarcane-Wheat Rice-Wheat
Danodia et al. 2017, GI
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
S-SEBI pixels
Evaluation of validation 3*3 5*5 10*10 15*15 20*20 30*30 50*50
procedure obtained for the CC 0.86 0.86 0.85 0.85 0.85 0.84 0.84
S-SEBI model for LAS
RRMSE 0.031 0.029 0.026 0.027 0.029 0.030 0.030
ME 0.53 0.55 0.60 0.59 0.55 0.54 0.54
footprint: Statistical analysis AI 0.84 0.84 0.86 0.86 0.86 0.86 0.86
Danodia et al. 2017, GI
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
Temperature Vegetation Triangle (TVT) Model
Ts–VI Triangle method-
Jiang and Islam (1999)
Priestley–Taylor formulation
with fully RS data
To estimate regional ET
and EF by interpreting the
scatter plot constructed
from remotely sensed LST
and fractional vegetation
cover
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
Methodology: Flow chart
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
Latent Heat Flux
Δ
LE = Φ [(Rn-G) ]
Δ−Υ
Evaporative Fraction
Δ
EF = Φ
Δ−Υ
Combined coefficient
ϕ = {(Tmax,i – Ts,i) / (Tmax,i – Tmin,i)} * (ϕ max,i – ϕmin,i) + ϕmin,i
Slope of sat. vap.
pressure curve
Tang et al. 2010, RSE
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
TVT Model:
Tang et al. 2010, RSE
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
Mapping Evapotranspiration at High Resolution
using Internalized Calibration (METRIC) Model
METRIC originated from versions of the SEBAL model (Surface
Energy Balance Algorithm for Land) developed in 1995
(Bastiaanssen, et al. 1995).
METRIC is designed to produce high quality and accurate
maps of ET for focused regions smaller than a few hundred
kilometers in scale and at high resolution (≤30m).
The design of SEBAL and METRIC cause the final ET estimate
to be relatively insensitive to the parameterization of
aerodynamics and near surface temperature difference (dT)
vs. surface temperature function originated in SEBAL provides
internal and essentially automatic calibration.
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
Assumptions:
Variation in land surface temperature is linearly related to
difference between land surface and air temperatures;
Sensible heat flux varies linearly between hot and cold
reference pixels;
Actual ET for hot reference pixel is 0;
Ratio of actual ET for cold reference pixel to reference ET
is 1.05;
Reference ET fraction (ETrF) is constant throughout day;
Actual ET for study area varies in proportion to changes
in reference ET at weather stations.
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
METRIC Model: Data requirement
Clear sky Satellite data
(High resolution ≤30m)
Crop inventory map
Weather data: Hourly or
fine scale (wind sped,
rainfall, temperature,
Hourly and daily
reference ET
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
Energy flux components:
Net radiation Rn = (1- ).RS↓ + RL↓ - RL↑ - (1- ε0).RL↓
Soil heat flux G/Rn = Ts (0.0038 + 0.0074) (1 - 0.98NDVI4)
G/Rn = 0.1 + 0.17 exp( −0.55 * LAI)
Sensible heat flux
H = (ρ * Cp * dT)/rah
where, ρ is the air density (kg/m3), Cp is the specific heat at a given
pressure (1004 J.kg-1.K-1), dT, the air temperature difference (K)
between the two heights of 0.1 and 2m and rah is the aerodynamic
resistance to heat transfer (S.m-1).
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
Cold and Hot pixels selection procedure
Javadian et al. 2019, WATER
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
dT = Tz1 – Tz2
u* is the friction velocity (m/s) which
quantifies the turbulent velocity fluctuations
in the air, k is von Karman’s constant (0.41),
z1 & z2 are 0.1 and 2m, zom is the
momentum roughness length (m), Tz1 & Tz2
are the air temperature at heights z1 and z2
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
Advantages:
Minimum ground-based measurements required;
Solves for all terms of energy balance model;
Land surface slope and aspect can be applied on more complicated
terrain.
Disadvantage:
Uncertainty from user selection of hot and cold reference pixels;
Time intensive to apply at basin scale
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
Two Source Energy balance (TSEB) Model
LEC = (RN,C + HC )
LES = (RN,S +G + HS )
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
Seasonal ET:
Instantaneous ET
(mm/hr)
ETinst = 3600 * λET/ λ λET = Instantaneous ET
Reference ET fraction
ETr = Reference ET at the
ETrF = ETinst/ ETr time of the image
(ETrF ≈ Kc)
ET24 ETr-24 = Cumulative 24 hour
ET24 = ETrF/ ETr-24
ETr
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
Seasonal ET:
S-SEBI Model TVT Model
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
ET Products:
https://www.mosdac.gov.in/
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
ET Products:
https://modis.gsfc.nasa.gov/data/dataprod/mod16.php
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
ET Products:
https://earlywarning.usgs.gov/fews/product/460
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
ET Products:
https://www.gleam.eu/
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
ET Products:
https://eeflux-level1.appspot.com/
INDIAN INSTITUTE OF REMOTE SENSING, DEHRADUN
Email- abhidanodia@iirs.gov.in
Tel- 0135-2524141
