Evapotranspiration Measurement and Modeling: Shrubs to Basins

1
Evapotranspiration Measurement and Modeling
Shrubs to Basins
U. S. Geological Survey National Ground-Water
Workshop Aug. 2008 David Stannard NRP,
Denver, CO David Sumner FWSC, Orlando,
FL Michael Moreo NWSC, Henderson, NV Kip
Allander NWSC, Carson City, NV Amanda Garcia
NWSC, Carson City, NV

2

• Talk given by David I. Stannard as the
  introductory conceptual lecture to a short course
  on evapotranspiration (ET).
• Subsequent lectures were given by the other
  authors on the first slide, about specific field
  studies methods.
• Together, these talks comprised a ½ day class
  providing a quick primer on USGS ET research, as
  part of the 2008 USGS National Groundwater
  Workshop held in Lakewood, CO, August 4-8.

3
Klemes
4
Florida Budgets
WSC budget
Water budget
ET
P
ET
P
SW
GW
SW
GW
inches per year
dollars per year
5
Global ET table
6
Continental ET Table
7
Evapotranspiration
The transport of water between the earths
surface and the atmosphere accompanied by a
change in phase from liquid or solid
(sublimation) at or below the surface, to vapor
in the atmosphere. Transpiration is evaporation
through the stomates of plants. Stomates are
microscopic holes in the leaves, through which
carbon dioxide diffuses in (for growth) and water
vapor diffuses out. This upward flow of water
also provides nutrients to the leaves for
photosynthesis.
8
Mass-Energy Conversion
9
ET-Link between water and energy
10
Factors
11
POTENTIAL EVAPOTRANSPIRATION (PET)
The rate of evapotranspiration that occurs from a
given land surface under a given set of
meteorological conditions when water is supplied
freely to the evaporating surfaces.
12
Micrometeorological Methods to Measure Actual ET

• Aerodynamic Profile Method
• Bowen Ratio Energy Balance
• Eddy Correlation (Covariance)

13
Aerodynamic Profile Method - historical

• ET depends on profiles of e, T, and u
• Requires multi-level data, iterative calculations
• Small sensor bias creates large errors in ET
• Concepts valid, and led to Bowen ratio method

14
Energy balance
Energy balance
Available Energy Turbulent Flux
15
BOWEN-RATIO METHODMeasures Rn, G, DS, and
partitions thatavailable energy into H and LE
according to the temperature and humidity profiles
16
BR Method
17
BOWEN-RATIO ENERGY BALANCE STATION
18
A PAIR OF NET RADIOMETERS
19
SOIL HEAT FLUX SENSORS
20
TEMPERATURE AND VAPOR-PRESSURE DIFFERENCE SENSORS
Shielded and Aspirated
21
SENSORS EXCHANGE EVERY 15 MINUTES
SENSOR BIAS REMOVAL
SENSORS EXCHANGE EVERY 15 MINUTES
22
Radiation Energy Balance Systems (REBS)
Bowen-ratio System
Exchange Mechanism-3k
Two Temp-Hum Probes--2.4k
23
Meadow ET - spring
24
Forest ET - spring
25
Meadow ET - fall
26
Forest ET - fall
27
Eddies...
Random, swirling, chaotic parcels of air that
transport water vapor, heat, etc. through upward
and downward motions. Idealized as spheres, but
actually complex shapes, best visualized with
tracers.
28
Turbulent flux of water vapor
29
Turbulent flux of heat
30
EDDY-CORRELATION METHODMeasures upward and
downward motions, and water vapor content of
all eddies, very fast (10 hz). High-speed
book-keeping of flux.
31
EC Method
32
EDDY-CORRELATION SENSORS
Krypton Hygrometer (vapor density) Sonic
Anemometer (wind vector, temperature)
33
EC Source Area
34
EDDY-CORRELATION TOWER
At least 1.5 times canopy height
35
NET RADIATIONCOMPLETE THE ENERGY BALANCE EVEN
WITH EDDY CORRELATIONSOIL HEAT FLUX
36
RELATED DATA
Solar Radiation
Rainfall
Temperature Humidity
Leaf Temperature (Infrared)
Soil Moisture
Wind Speed Direction
37
SOLAR POWERED STATIONS12-volt sensors and data
loggersLow power demand supplied by solar
panelsCan operate in remote areas
38
COSTS FOR MICROMET STATIONS

• Bowen-Ratio or Eddy-Correlation
• 15,000 to 20,000 depending on options
• Tower (if needed)
• 1,000 (40 ft.) to 4,500 (200 ft.)
• Tower Installation
• 1,000 to 5,000 depending on location and access

39
Forest Dry-down
40
ET Comparison
41
Infil 1999
P 26.4 in ET 18.7 in
42
Infil 2000
P 18.2 in ET 17.3 in
43
Infil 7 yr
44
OTHER SITE-SCALE MEASUREMENT METHODS

• Chambers
• Lysimeters
• Sap-Flow Sensors
• Water-Table Fluctuations (White, 1932)
• Scintillometers (sensible-heat flux)
• Mass Balance Method
• Soil Moisture Balance

45
Chambers 1
Chambers No Fetch Needed (can use in city) Can
Study Species Differences, E vs. T (Garcia et
al.) Labor Intensive
46
Chambers 2
Fans Inside Keep Air Well Mixed Temperature-Humidi
ty Sensor Measures Vapor Density Leave in Place
About One Minute
47
Chamber time series
48
Lysimeters
49
Sap-flow Sensors
Measure Sap Velocity Using Heat Advection
Principles (Transpiration Only) Difficult to
Define Radial Velocity Profile Must Sample
Multiple Trees Scale Up to Forest Level
50
Whites Method
51
Scintillometers

• 100 m8 km path
• Weighted avg. of H over pathlength
• Need Rn, G and wind profile at same scale to
  compute reliable LE

                                                                                   
52
Scintillometer Costs

• LAS (100 m 4.5 km)
• 41 K
• XLAS (500 m 8 km)
• 54 K
• LAS ET system
• 80 K
• XLAS ET system
• 95 K
• Radioadd 15-20K

53
Mass Balance Method
54
Soil Moisture Balance
55
Basin-Scale Methods

• Extrapolation of Point Measurements
• Water Balance
• Chloride Balance
• Remote Sensing

56
EXTRAPOLATION OF POINT MEASUREMENTS TO BASIN SCALE

• If basin is uniform, use point measurement
  directly
• If basin characteristics vary predictably, select
  site in average setting
• Watershed models (PRMS) and ET models need
  coverages of soil, vegetation, elevation, slope,
  aspect, precip
• Can use models alone, or with weather stations
  and/or point ET measurements

57
EXTRAPOLATION TO BASIN SCALE(continued)

• Mesoscale Meteorological Models - Large numerical
  models to help refine watershed and ET model
  input data. Useful in rugged terrain, and where
  weather stations are scarce.
• Remote Sensing Multiple uses.

58
Basin Water Balance
Whole Basin GIN GOUT ( 0) Long Time DS 0
59
Empirical Basin Models
60
Chloride Balance (Claassen Halm, 1996)

• Works in ungaged basins
• Integration period must be matched to basin
• Chloride assumed to be conservative within basin

61
REMOTE SENSING USES FOR ETDirect or indirect,
but large area

• Energy-Balance Approach LE computed as
  residual, from space. NASA ARS. Sparse
  vegetation problem finally solved (TSM Kustas
  Norman, 2000 ALEXI Anderson et al. 2004)
• Map major vegetation types . Distribute
  ground-based ET values in major types using
  vegetation maps (Moreo, Laczniak, DeMeo)
• Principle Component Analysis Extrapolate
  network of ground-based measurements across
  landscape using coverages.

62
REMOTE SENSING USES FOR ET(continued)

• Microwave bands used to measure soil moisture for
  use in models. Problems seeing through dense
  vegetation.
• Can help characterize landscape types to select
  ET measurement sites
• All remote sensing subject to cloud interference

63
ET MODELING METHODS

• Potential ET Models
• Actual ET Models

64
POTENTIAL EVAPOTRANSPIRATION (PET)
The rate of ET that occurs from a given land
surface under a given set of meteorological
conditions when water is supplied freely to the
evaporating surfaces. Feedbacks can bias PET
estimates high if input data are collected in dry
conditions.
65
PENMAN EQUATION
The most rigorous and accurate of PET equations.
However, also most data intensive, and strongly
affected by feedbacks if input data obtained
under dry conditions (overestimates).
66
Penman Potential ET Equation Combines energy
balance and aerodynamic principles
LEP Potential latent-heat flux (W m-2) s
slope of saturation vapor-pressure curve (kPa
oC-1) Rn Net radiaton (W m-2) G Soil heat
flux (W m-2) g psychrometric constant (kPa
oC-1) VPD Vapor pressure deficit (kPa) f(u)
wind function (W m-2 kPa-1) u wind speed (m
s-1) f(u) C1 C2u Wind function empirical
led to many versions to account for variable
surface roughness, and, unknowingly, surface
resistance.
67
OTHER PET EQUATIONS
A wide range of equations based on correlations
between factors that affect PET. The more complex
and data intensive, the more potential for
accuracy. Ones that dont require humidity are
more robust when data not collected under PET
conditions. Priestley-Taylor works surprisingly
well.
68
Priestley-Taylor Potential ET Equation Abbreviated
version of Penman equation

• LEP Potential latent-heat flux (W m-2)
• a 1.26
• s slope of saturation vapor-pressure curve (kPa
  oC-1)
• Rn Net radiaton (W m-2)
• G Soil heat flux (W m-2)
• psychrometric constant (kPa oC-1)
• Approximates that under potential conditions,
  aerodynamic term is 26 of energy term.
  Eliminates need for humidity and windspeed data.
  More accurate than Penman when data not collected
  under potential conditions.

69
ACTUAL ET MODELS

• Characterize water availability
• Tend to be site specific
• Often calibrated using actual ET measurements

70
Penman-Monteith
71
Shuttleworth-Wallace
72
USGS WATER SCIENCE CENTER ET STUDIES

• Florida
• Nevada
• WashingtonHanford Low-level Site
• TexasHoney Creek Juniper Removal
• NebraskaPlatte River Riparian Zone
• OklahomaNorman Landfill
• OregonKlamath Lake Wildlife Refuge

Many projects
73
Florida Evapotranspiration Network
A collaborative effort between the USGS
and SWFWMD, SJRWMD, SFWMD, SRWMD, and RCID
74
ET network February 2002
Planned
Installed
75
Death Valley Area
ET STUDIES for Quantifying Ground-Water
Discharge From Death Valley Regional Flow
System Randy Laczniak, Guy DeMeo, LaRue Smith
76
DELINEATION BY SPECTRAL CLASSIFICATION
DEATH VALLEY REGIONAL FLOW SYSTEM
77
SPECTRAL GROUPING FOR CLASSIFICATION
78
TYPICAL ET SITE
79
SALT DEPOSITS

BADWATER PLAYA, DEATH VALLEY
80
RESULTS

• Major differences in ground-water discharge
  estimates between these results and earlier
  estimates
• In general, results indicate greater discharge
  from northern discharge areas and less from
  southern discharge areas
• Travel-times from Yucca Mountain significantly
  revised

81
Texas
TEXAS DISTRICT 6-yr study - Asquith Effects of
juniper on water supply near Austin Paired
basins Measure ET before, clear juniper in one
basin, measure change in ET
82
FLORIDA

• Lake evaporation studies 80s - present
• Adams, Swancar, Lee
• Carlton Mem. Reserve Bidlake/Lopez 93
• Disney Studies Sumner 96
• Rainbow Silver Spgs Basins Knowles 96
• Everglades German 00
• Tiger Bay Watershed Sumner 01

83
Principle Component Analysis John
Jones National Mapping Discipline
84
Water-Resources Evaluation of Ruby Valley,
northeastern Nevada

• David L. Berger
• U.S. Geological Survey
• District Program Review
• Nevada District
• March 2002

85

• Bowen-ratio sites
• Bulrush marsh
• Phreatophyte-1
• Open-water
• Desert-shrub upland

Natural-color composite-April 8, 2000 Enhanced
thematic mapperUSGS Landsat 7
86
(No Transcript)
87
(No Transcript)
88
Phase 1 resultsOct. 1999 - Sept. 2000
Bowen-ratio sites Daily Average (inches) Winter daily average (inches) Summer daily average (inches) Winter total Oct.-Apr. (inches) Summer total May-Sept. (inches) Annual total (inches)
Open water 0.174 0.112 0.260 23.85 39.99 63.64
Bulrush marsh 0.137 0.062 0.242 13.18 37.06 50.24
Phreatophyte-1 0.043 0.028 0.065 5.96 9.93 15.89
Desert-shrub uplands 0.033 0.029 0.035 6.17 5.78 11.96
89
Determination of Ground-Water Evapotranspiration
in Riparian Woodlands Along the Platte River,
Nebraska

• M.K. Landon, D.L. Rus,
• B.J. Dietsch, M.R. Johnson,
• U.S. Geological Survey

90
Objectives of Current Study

• Determine total evapotranspiration (ET) and
  estimate ET from ground water (GWET) in
  representative riparian woodlands
• Calibrate methods that can be used to estimate
  riparian ET across the COHYST study area to ET
  measurements

91
WASHINGTON

• Tomlinson 94 96 (4 reports)
• Several sites in south-central and eastern
  Washington related to waste isolation at Hanford
  site
• Bidlake
• Klamath and Tule Lake Wildlife preserves