METRIC / SEBAL Approaches to Regionalized Remote Sensing of

1
METRIC / SEBAL Approaches to Regionalized Remote
Sensing of Evapotranspiration

• Richard G. Allen, Univ. Idaho Kimberly
• Development Partners
• Wim Bastiaanssen, WaterWatch, the Netherlands
• Ayse Irmak, Univ. Nebraska-Lincoln
• Ricardo Trezza, Univ. Idaho
• Jan Hendrickx, New Mexico Tech
• Justin Huntington, Desert Research Institute,
  Nevada
• Jeppe Kjaersgaard, South Dakota State Univ.

2
Presentation Coverage

• SEBAL/METRIC approach(es)
• Operational needs
• Model Calibration via Inversion
• Interpolation in Time
• Applications

3

Why Energy balance?

• ET is calculated as a residual of the energy
  balance

(radiation from sun and sky)
ET
H (heat to air)
Basic Truth Evaporation consumes Energy
G (heat to ground)
4
Energy balance gives us actual ET

• We can see impacts on ET caused by
• water shortage
• disease
• crop variety
• planting density
• cropping dates
• salinity
• management

5
METRICtm and SEBAL

• Mapping EvapoTranspiration with high Resolution
  and Internalized Calibration
• METRIC was derived from Surface Energy Balance
  Algorithm for Land

Allen et al., (2002, 2007)
Bastiaanssen et al., (1995, 1998, 2005)
METRICtm and SEBAL are, in general, complementary
processes
6
Surface Temperature southcentral Idaho August
14, 2000
Large Contrast in LST
Irrigated Ag.
Desert
7
Heat Flux to Air southcentral Idaho August
14, 2000
8
Instantaneous ET southcentral Idaho August
14, 2000
9
24-hour ET southcentral Idaho August 14, 2000
500,000 irrig. acres
Desert
24-hour ET Magic Valley, Idaho August 14,
2000 --by METRIC Satellite-based ET Procedure
Lake Walcott
ASABE June 22, 2010
10
Satellite Energy Balances are Plagued by
Uncertainty, Bias, and Error in EB components

• Surface temperature
• Aerodynamic vs. Radiative Temperature
• Bias in Satellite Sensor Calibration
• Atmospheric Correction
• Air temperature
• Albedo calculation
• Net radiation calculation (incoming long-wave)
• Soil heat flux
• Aerodynamic resistance calculation
• Wind speed field
• Extrapolation of instantaneous ET to 24-hour
  periods

11
Sensible Heat Flux (H) Classical
H r cp (Taero - Tair) / rah
Taero aerodynamic temperature
rah the aerodynamic resistance
u friction velocity k von karmon
constant (0.41)
12
Sensible Heat Flux (H) CIMEC models
H (r cp dT) / rah
dT floating near surface temperature
difference (K)
rah the aerodynamic resistance from
z1 to z2
u friction velocity k von karmon
constant (0.41)
13
METRIC and SEBAL assume dT to float in the
blended EBL above the canopy-soil complex
CGIAR Workshop on Surface Energy Balance Models
of Agricultural Areas from Earth Observation
Data Universidad Nacional Agraria La Molina
(Peru), 13 March 2008
14
Solution to dT function Use Inverse Modeling

• Calibrate against known ET at extreme conditions
  (end points)
• Incorporate biases of all inputs into the
  internal calibration
• Biases then fallout during the final estimation
  process

15
Calibration of SEBAL and METRIC CIMEC models
ASCE Penman-Monteith for 0.5 m (alfalfa)
reference is used for the wet extreme
Inversion of H eqn

Rn G - 1.05 ETref alfalfa (METRIC) or 0 (SEBAL
classical)
Rn - G
16
Near Surface Temperature Difference (dT)

• Tair is unknown and unneeded
• SEBAL and METRICtm assume a linear relationship
  between Ts and dT
• dT b aTs

Bastiaanssen breakthrough

• Ts is used only as an index and can have large
  bias and does not need to represent aerodynamic
  surface temperature

17
Calibration of METRIC/SEBAL
The Sensible Heat (H) Function calibrates around
Biases in many of the Energy balance
components (Biases exist in net radiation,
soil heat flux, aerodynamic stability,
aerodynamic roughness, absolute surface
temperature, atmospheric correction)
unbiased

biases
H Rn G LE (for calibration)
LE Rn G H (during application)
Biases cancel out
18
METRIC analyses suggest that the relationship
between Relative ET and Ts can vary with land use
(/- 100) even within one Landsat Scene and can
be nonlinear due to buoyancy effects
May 22, 2002
August 26, 2002
Landsat P33, R34-35, New Mexico Rio Grande
Therefore, good to consider albedo, soil heat
flux, stability correction, roughness
19
Use ETrF to extrapolate to 24-hours
Sugar beets
Assume simple scaling with the ET reference
20
Comparing METRIC vs. traditional Kc ETref methods
(relatively good agreement among very independent
approaches, with some variation during the
shoulder periods when ground has partial cover)
21
Uses of Landsat-based ET to Cover Regional Areas
Seasonal ET from METRIC and Landsat (100 mile x
200 mile area)
ASABE June 22, 2010
22
Cloud Mitigation
Dr. Ayse Irmak, UNL
23
Adjusting for background evaporation at image
time to derive Correct Monthly ET
ET from August 13 1997 not adjusted
for background soil evaporation
ET from August 13 1997 adjusted for background
soil evaporation
A daily Gridded Evaporation/Land Process model is
run
24
ET Precipitation ------Indication of Recharge /
Depletion
Scottsbluff, NE
Use in high-res hydrologic studies
Irrigated areas
Rangeland / Dryland Ag.
25
Thanks
ET
ASABE June 22, 2010