ET Study

Cooperative Study Proposal
Approved by the Inyo/Los Angeles Standing Committee March 23, 2000.

Project Title:
Evapotranspiration from groundwater dependent plant communities: Comparison of micrometeorological and vegetation-based measurements.

Principal Investigators:
Robert Harrington and Aaron Steinwand
Inyo County Water Department
163 May Street
Bishop, CA 93514
(760) 872-1168

Paula J. Hubbard and David Martin
Los Angeles Department of Water and Power
300 Mandich Street
Bishop, CA 93514
(760) 872-1104

Purpose:
Current Green Book methods (III.D.) for estimating plant-water requirements rely on leaf-scale transpiration measurements and plot-scale leaf area index (LAI) measurements to forecast plant-water requirements at permanent vegetation monitoring sites. These methods are central to the calculations that govern well on/off status, but their validity has not been tested. The Green Book (V.B.2.) notes that the methods presented in section III.D. should be verified or improved upon in a future cooperative study.

The purpose of this study is to measure evapotranspiration (ET) using micrometeorological methods, and compare these measurements to current methods of computing plant-water requirements. This will provide an independent and direct measurement of ET by which to evaluate the accuracy of Green Book methods. The proposed ET measurements will be representative of ET integrated over an area of vegetation of the approximate spatial scale of a monitoring site, thus assessing the validity of the Green Book method of forecasting plant-water requirements. Because of the importance of plant-water requirement forecasts in on/off calculations, this is an essential evaluation of Green Book methods.

This study addresses two questions:

  • Is the Green Book method of extrapolating leaf-scale measurements of transpiration to produce monitoring-site-scale forecasts of plant-water needs valid?
  • Do the currently used Green Book forecasting methods provide accurate estimates of plant water requirements at vegetation monitoring sites?

Ancillary products of this study will be refinements to ET computations in the USGS groundwater model (Danskin, in press), additional data regarding the relationship between transpiration and depth to groundwater, and better quantification of the ET component of the regional water balance.

Background:
Transpiration is the key process linking plants with subsurface hydrology and it is one of the largest components of the regional water balance. Yet, it is one of the least well quantified fluxes in the water balance of the Owens Valley. The methods for forecasting plant water requirements in III.D of the Green Book rely on extending leaf-scale measurements up to the plot- and wellfield-scale for the purpose of setting on/off conditions for groundwater extraction.

Extending measurements made at small scales up to larger scales poses several questions: What is the spatial variability of the process measured at the small scale (Wood, 1995)? Given some degree of spatial variability, has the small-scale process been sampled sufficiently to provide stable statistics for integration up to large scale analyses? Do processes retain their relevance across spatial scales (Baldocchi et al., 1991)? What other variables affect the process under investigation and what is their spatial variability (Jarvis and McNaughton, 1986)? What is the spatial correlation between the relevant variables? What is the correct algebraic expression to relate measurements made at one scale to variables computed at another (Avissar, 1995; Baldocchi et al., 1991)? Because these questions are hard to answer, computations involving variables measured at small scales that are used to make large-scale forecasts should be verified by concurrent measurements at larger spatial scales. The Green Book methods (III.D) use empirically derived functions to describe the seasonal changes in transpiration per leaf area (Tlf) and LAI. The proposed study will determine whether the up-scaling of leaf-scale transpiration and empirical curves accurately describe the observed seasonal change in transpiration, and direct future improvements in these methods.

There are various methods of directly measuring ET using micrometeorological measurements (Brutsaert, 1982; Shuttleworth, 1993). The common theme among all micrometeorological methods is that measurements of meteorological variables are made near the land surface (i.e. at a few meters above the plant canopy) to determine fluxes of energy, momentum, or trace gases. Micrometeorological methods have been used in previous cooperative studies to measure water vapor fluxes from phreatophytic vegetation (Duell, 1990; Gay, 1992; Duell, 1992; Stannard, 1992), but these measurements were not made concurrently with the necessary vegetation transect data to undertake the proposed comparison between micrometeorological and vegetation-based estimates of transpiration. The micrometeorological method proposed for this study is the eddy correlation method. The eddy correlation method has gained predominance among micrometeorological methods recently because of its minimal theoretical assumptions and improved instrumentation (Shuttleworth, 1993).

Eddy correlation measurements are based on the correlation between turbulent motions of the air, and the abundance of constituents being transported by the turbulent motions (e.g., heat or water vapor) (Campbell and Norman, 1998). The primary transport mechanism by which heat and water vapor move out of the canopy and up into the atmosphere is by the turbulent motion of the air near the surface (Garratt, 1992). The average vertical wind speed above a flat land surface is considered to be zero, because the ground surface is neither a source nor sink for air; therefore, for heat to move from the surface up into the atmosphere, the upward motions of the turbulence must be warmer than the downward motions. Similarly, for water vapor to undergo turbulent transport from the land surface up into the atmosphere, upward air motions must be more humid than the downward motions. The correlation between fluctuations in vertical wind speed and humidity is positive during evaporation, and during dew or frost formation, the correlation is negative. The eddy correlation method uses high frequency (~10 Hz) measurements of vertical wind speed, temperature, and humidity to compute the correlation between vertical air motions and the constituent of interest. The flux is then computed directly from this correlation (Shuttleworth, 1993).

The main advantage of the eddy correlation method is that it is the most direct measurement of sensible and latent heat fluxes that is possible with micrometeorological methods. No assumptions are made about the land surface properties such as aerodynamic roughness or zero-plane displacement, and no corrections for atomospheric stability are necessary. This is especially advantageous in sparse heterogeneous vegetation canopies and the widely varying stability conditions that exist in semi-arid environments. Hence, Stannard (1994) used eddy correlation measurements as a standard by which to assess the accuracy of other methods of estimating ET in semiarid rangeland. Montieth and Unsworth (1990) discuss the advantages and disadvantages of eddy correlation, Bowen ratio, and aerodynamic methods of measurement of turbulent fluxes. They state that the main disadvantage of eddy correlation methods has historically been that the instrumentation is relatively expensive and fragile. The reliability of instrumentation suitable for eddy correlation measurements has steadily improved, and off-the-shelf eddy correlation systems have recently become available (Campbell Scientific, 1999). Progress in instrumentation has resulted in the adoption of eddy correlation as the method of choice for assessing other methods turbulent flux measurement, e.g., Chebouni et al. (1999) compared eddy correlation measurements of sensible heat flux with large aperture scintillometer measurements to assess the utility of large aperture scintillometers over variable vegetation cover at semi-arid sites in Arizona and Mexico.

Among other micrometeorological methods of measuring ET, numerous studies of semiarid rangeland ET have relied on the Bowen ratio method to measure ET from phreatophytic shrubs. Because the Bowen ratio and eddy correlation methods rely on different theoretical bases, considerable work has been done comparing them. Locally, such comparisons have been made by Duell (1990) and Wilson et al. (1992) during previous cooperative studies, by Tyler et al. (1997) on Owens Lake playa, and by Cznarecki (1990) on Franklin Lake playa in eastern Inyo County. Reginato and Jackson (1992) compared porometer and normal curve methods (Groeneveld and Warren, 1992), similar to those in the Green Book, to Bowen ratio and eddy correlation measurements; however, the LAI and percent-cover measurements were not made concurrently with the micrometeorological measurements, rendering the evaluation of the Green Book method inconclusive. The comparison between the eddy correlation and Bowen ratio methods was also somewhat tentative, because the measurements spanned only a few days. To estimate groundwater discharge at a playa in eastern Inyo County, Cznarecki (1990) compared eddy correlation to various subsurface methods and concluded that eddy correlation provided the most reliable results. Tyler et al.’s (1997) comparison of Bowen ratio and eddy correlation methods was done on the Owens Lake lakebed for establishing a water budget and estimating salt flux from the lakebed. They found that the eddy correlation method agreed better with lysimeter measurements than did Bowen ratio measurements, and concluded that low evaporation rates from the playa surface resulted in such low humidity gradients that the Bowen ratio method was unreliable. Therefore, they based their water budget on eddy correlation measurements of evaporation.

Nichols (1992) used the Bowen ratio method to compare theoretical and measured resistances to sensible and latent heat transfer in a sparse Sarcobatus vermiculatus community in central Nevada, but he had to discard a significant fraction of his data because the humidity was so low that the chilled-mirror hygrometer he used did not provide accurate humidity measurements. In later studies, Nichols (1993; 1994) developed relationships between depth to groundwater, percent cover, and phreatophytic transpiration, but the measurement of vegetative cover was not quantitative. Sorenson et al. (1991), in a previous Inyo Co./LADWP/USGS cooperative study, attempted to relate water-table drawdown to vegetation condition, but their success was somewhat limited because their pumping scheme induced less drawdown than expected and the duration of their study was not sufficient to deduce long-term effects.

As water vapor moves from leaves to the atmosphere, turbulence mixes it with the air in the overlying atmosphere; therefore, the measurements made at a single station represent ET integrated over a plot of vegetation. The size of the plot depends on atmospheric properties such as stability and wind speed, as well as land surface properties such as surface roughness and canopy geometry. In the context of testing current plant-water requirement calculations, this is desirable because it represents a measurement of plant-water use averaged over a spatial scale similar to the scale of a monitoring site. Gash (1986) and Scheupp et al. (1990) give guidelines for the necessary upwind fetch necessary for representative measurements under various conditions.

Measurement of all the components of land surface energy balance has two advantages for the current study: (1) it allows correction for air density effects, and (2) it provides a check in the internal consistency of the data. The energy balance for the land surface is:

Rn + H + lE + G = 0 (1)

where Rn is the net radiation (W m-2), H is the sensible heat flux (W m-2), l is the latent heat of vaporization of water (2450 J g-1), E is evapotranspiration (g m-2 s-1), and G is the soil heat flux (W m-2). Eddy correlation measurements of trace gases (e.g. water vapor) should be corrected for fluctuations in air density, which requires measurement of H; therefore, fluctuations in air temperature should be made, as well as wind speed and humidity (Webb et al., 1980). By measuring Rn and G concurrently with eddy correlation measurements of H and lE, the closure of the energy balance can be checked using equation 1.

Alternatively, if any three terms of equation 1 are measured, the fourth term can be solved for using equation 1. Carman (1993) used such a strategy to estimate transpiration from Great Basin phreatophyte scrub, but his results were somewhat uncertain, because use of equation 1 to compute an unmeasured flux precludes using equation 1 as an independent check on the mutual consistency of all four energy balance terms.

In summary, the studies cited above have tried to relate phreatophyte transpiration, water table depth, and vegetation conditions; however, success has been limited by lack of rigor in monitoring vegetation cover, ineffective manipulation of the water table, and lack of success with Bowen ratio methods under arid conditions. A further limitation has been the use of equation 1 to estimate fluxes as residuals, rather than by direct measurement. These limitations can be remedied by eddy correlation measurements of H and lE with concurrent measurement of Rn and G. The study proposed here will use vegetation monitoring methods from the Green Book (III.C) and eddy correlation methods to evaluate the Green Book-based plant-water requirement computations.

A secondary goal of this study is to provide guidance for ET parameterization in groundwater flow models.

Methods

Site selection: Eddy correlation ET measurements will be made throughout the growing season (March through September) at three sites. Concurrent point-frame measurements of LAI will be made so that estimates of plant-water requirements can be made.

Site selection criteria are:

The Green Book estimate is for non-water-stressed plants; therefore, the micrometeorological measurements should be conducted at sites where the vegetation is subirrigated throughout the growing season.

The sites should be chosen to minimize bare soil evaporation.

The sites should have sufficient fetch of uniform vegetation and hydrologic conditions surrounding the site. Shuttleworth (1993) provides the expression

for estimating the fraction (F) of the measured evaporation originating within and upwind fetch equal to XF under unstable (daytime) conditions, where z is the instrument height, z0 is the surface roughness, and d is the zero plane displacement. Approximations for z0 0 and d are and , where hc is the canopy height.

  • The sites should be undisturbed by water spreading, irrigation, or other land use practices that affect ET.
  • It is desirable to locate stations in both mixed and monocultural sites. Data from mononcultural sites can be compared to vegetation based estimates without the extra complicating factor of mixed species cover. Data from mixed cover sites will be used to test the Green Book method estimating plant-water requirements in the typically mixed cover of the wellfield management areas.
  • The sites should be representative of the vegetation and conditions within wellfield management areas.
  • The sites should be secure from vandalism.
  • Sites with monitoring wells will be given preference.

The proposed study will continue for four growing seasons, and measurements will be made at different monitoring sites each year. Moving the measurement stations each year will provide data from a variety of vegetation assemblages, allowing evaluation of the accuracy and robustness of the currently used methods to estimate plant-water requirements.

Permanent vegetation monitoring sites (Green Book, Table I.A) will be considered for eddy correlation measurements, because a detailed site history is known (in terms of water table fluctuations, soil water content, and vegetation condition) and they are already equipped with piezometers and neutron probe access tubes. The primary interest of this study is transpiration, as opposed to evapotranspiration; therefore, because micrometeorological measurements will include both transpiration from plants and evaporation from bare soil, sites will be chosen such that the transpirational contribution to ET is predominant. Factors that contribute to this desirable condition are dry soil; a high percentage of plant cover; and a water table that is within the root zone of the plants present at the site, but not so near the surface that evaporation from the soil surface is significant. For scrub sites, approximately 3 m depth to water is desirable, and for meadow sites, approximately 2 m is desirable. ET measurements will span the growing season, soil moisture and depth-to-water measurements will be made biweekly throughout the growing season, and point-frame transects will be made monthly during the growing season.

Because each site has a unique vegetation assemblage, soil profile, hydrology, and microclimate, it is not expected that micrometeorological data from heterogeneous plant canopies will be invertible to obtain Tlf on a species by species basis. However, eddy correlation measurements at monocultural sites with minimal bare soil evaporation would yield approximate Tlf for species present. Subirrigated monocultures of important Owens valley plants will be located and ET measurements will be made at those sites. Numerous sites in the Independence and Symmes-Shepard well fields appear to be monocultures of Atriplex torreyi, the area near monitoring site TA4 offers Chrysothamnus nauseous dominated sites, and there is a meadow located between monitoring sites TS2 and TS3 that is a shrubless mixture of Distichlis spicata, Sporobolus airoides, and Juncus balticus. Initially, sites will be chosen in these three areas, unless other more appropriate sites are identified by LADWP and Inyo County staff. Future sites will be located pending the initial results. An additional site criterion that relates to the proposed soil water balance cooperative study is proximity to a monitoring well suitable for regression modeling of the water table. The best area for meeting this requirement is the Symmes-Shepard wellfield.

Materials:
Great Basin Air Pollution Control District (GBUAPCD) has two eddy correlation systems that are currently not in use. We have negotiated and reached an agreement for its use, wherein we will recalibrate and repair the equipment and return it to GBUAPCD at the end of the study. We propose that a third eddy correlation station be purchased to facilitate rapid resolution of this study, and to ensure continuity of measurements, should GBUAPCD need the equipment on short notice. We also propose that GBUAPCD’s equipment be supplemented with fine-wire thermocouples to provide data for measurement of sensible heat flux. This additional equipment is necessary to determine density corrections to evaporative flux, and energy balance closure calculations.

In order to obtain the necessary data in a timely manner, an additional eddy correlation system is required. Campbell Scientific, Inc. supplies sonic anemometers, krypton hygrometers, fine-wire thermocouples, data loggers, power supplies, enclosures, and instrument towers for conducting the proposed measurements. GBUAPCD’s system was purchased from Campbell Scientific, Inc., and we propose that the new system and additional components be purchased from Campbell as well. This provide for interchangability of parts and uniformity of protocols, which will assure both ease and reliability of measurements. Campbell Scientific has been contacted regarding recalibration and repair of GBUAPCD’s equipment, and estimates of those costs are provided in the next section.

Prior to deployment of the equipment, all three stations will be set up at a site of homogeneous cover and run simultaneously. The data from the three stations will be compared to assure that all three stations generate comparable data, and provide an assessment of inter-station precision.

Point-frame transects, soil water content, and water table monitoring will be conducted, as per the protocols of the Green Book (III.C, F) and current routine monitoring practices. The sites will either be equipped with time-domain reflectometry (TDR) probes or gravimetric methods will be used to measure shallow soil moisture. The determination to use TDR or gravimentric methods will be site dependent, because high soil salinity interferes with TDR measurements. Determination of shallow soil moisture will aid in assessing bare soil evaporation, because when a layer of dry soil forms, evaporation becomes limited by soil moisture transport properties, and the relative contribution of bare soil evaporation to total ET diminishes substantially (Kutilek and Nielsen, 1994).

Time-table: We propose that the study be initiated as soon as possible, so that data for the 2000 growing season can be collected.

Table 1. Time table

Year

2000

2001

2002

2003

2004

Quarter

1

2

3

4

1

2

3

4

1

2

3

4

1

2

3

4

1

2

Site selection
Equipment preparation and set up
Data collection
Data analysis
Annual progress report
Final report

Table 2. Schedule for 2000.

ET tower locations flagged in field

Jan 20

Inyo/LA

Soil described and collected

Feb 9-10

Inyo

Soil water equipment installation/calibration

Feb22-Mar 3

Inyo

Arrival of ET equipment

Late Feb

Inyo

ET equipment testing

ASAP following recpt. of equipment

Inyo

Vegetation transects established

Mar 6-10

Inyo/LA

Soil water monitoring

start Mar 10, biweekly

Inyo

ET tower fencing

prior to Mar 17

LA

ET tower set up

ASAP following fencing

Inyo/LA

Vegetation transects

April 3-6, May 10-12, June 7-9, July 5-6, Aug 1-3, Sept. 5-7

LA

ET tower shutdown

Late Oct

Inyo/LA

Recalibrate sensors

Winter 2000/2001

Inyo

Products

Data:
Over the course of four years, the study will produce twelve concurrent sets of surface energy balance, ET, vegetation transect, soil moisture, and depth-to-water data throughout four full growing seasons.

Analysis:
Each data set will be reduced into micrometeorological and vegetation based estimates of transpiration spanning the growing season. Instrument error will be assessed using factory calibration data and by comparing data collected during a pre-deployment concurrent and collocated test run of all three eddy correlation systems. Computation of the energy balance closure (equation 1) will ensure that the measurements are self-consistent. A further quality control measure will be to compare direct measurements of lE to lE computed as a residual using the energy balance closure. Propagation of instrument error, and the energy balance closure will be used to assess the uncertainty in the results presented in the final report.

For mixed cover sites, total growing-season transpiration for the site will be compared to Green Book-based plant water requirements for the site (Green Book III.D.2.c) and to suggested revisions to the parameters used in the Green Book (Steinwand, 1999) using a paired t-test (Snedecor and Cochran, 1980). This will provide an overall assessment of whether the Green Book method of assessing ET at a given site is accurate. As a further diagnostic analysis, daily ET by each method will be examined by analyzing the difference between eddy correlation measurements, daily values of transpiraiton estimated from transpiration coefficient curves and potential ET (Steinwand, 1999), and transpiration estimates from page 58 of the Green Book. This examination will be done by plotting the eddy correlation values and Green Book values as time series; plotting the eddy correlation values against the Green Book values, and testing the statistical hypothesis that the mean difference between the two methods is zero.

For monocultural sites, estimates of transpiration for the monoculture species will be produced and compared to Green Book-based estimates. This will provide an assessment of the accuracy of an individual transpiration curve (Green Book III.D.1.a). The procedure for comparison will be similar to that given above for mixed cover sites, but because the eddy correlation measurement will reflect transpiration from a single species, we can examine the validity of an individual transpiration curve, which will help determine the source of deviations between eddy correlation measurements and Green Book estimates made using the equation given in Green Book section III.D.1.a.

Reporting:
The final report will consist of a comparison of the analysis described above, and, if possible, recommendations will be made for improvements in the Green Book-based plant water requirement forecasting method. Also included will be documentation of the quality control measures described above. Additionally, if linear combinations of species-wise LAI and Tlf prove to be an accurate portrayal of local transpiration, recommendations for parameterizing and spatially distributing ET for groundwater models will be made. Annual progress reports will be issued.

References
Avissar, R., Scaling of land-atmosphere interactions: an atmospheric modeling perspective, Hydrological Processes, 10th anniversary issue, 3-19, 1995.

Baldocchi, D. D., R. J. Luxmoore, J. L. Hatfield, Discerning the forest from the trees: an essay on scaling canopy stomatal conductance, Agri. and For. Meteorol., 54, 197-226, 1991.

Brutsaert, W., Evaporation into the atmosphere, D. Reidel, 1982.

Campbell, G. S. and J. M. Norman, An Introduction to Environmental Biophysics, second edition, Springer, 1998.

Campbell Scientific, Inc., 1999 U.S. Price List.

Carman, R. L., Measurement of evapotranspiration in phreatophyte areas, Smith Creek Valley and Carson Desert, west-central Nevada, 1983, USGS WRI 89-4118, 1993.

Chebouni, A., Y. H. Kerr, C. Watts, O Hartogensis, D. Goodrich, R. Scott, J. Scheildge, K. Lee, W. J. Shuttleworth, G. Deidieu, and H. A. R. De Bruin, Estimation of area-average sensible heat flux using a large-aperture scintillometer during semi-arid land-surface-atmosphere (SALSA) experiment, Water Res. Resour., 35(8), 2505-2511, 1999.

Czarnecki, J. B., Geohydrology and evapotranspiration at Franklin Lake playa, Inyo County, California, USGS WSP 2377, 1990.

Danskin, W. R., Numerical evaluation of the hydrologic system and selected water-management alternatives in Owens Valley, California, USGS WSP 2370-H, in press.

Duell, L. F. W., Estimates of evapotranspiration in alkaline scrub and meadow communities of Owens Valley, California, using the Bowen-ratio, eddy- correlation, and Penman-combination methods, USGS WSP 2370, 1990.

Duell, L. F. W., Eddy-correlation measurements at site L, Chapter D, USGS WRI 91459, 1992.

Garratt, J. R., The atmospheric boundary layer, Cambridge University Press, 1992.

Gash, J. H. C., A note on estimating the effect of a limited fetch on micrometeorological evaporation measurements, Boundary-layer Meteorol., 35, 409-413, 1986.

Gay, L. W., Bowen ratio measurements at sites C and L, Chapter A, USGS WRI 91-4159, 1992.

Groeneveld, D. P., and D. C. Warren, Total transpiration from land areas estimated from hand-held porometer measurements, Chapter E, USGS WRI 91-4159, 1992.

Jarvis, P. G., and K. G. McNaughton, Stomatal control of transpiration: scaling up from leaf to region, Adv. in Ecological Res., 15, 1-49, 1986.

Kutilek, M, and D. R. Nielsen, Soil Hydrology, Catena Verlag, 1994.

Montieth, J. L. and M. H. Unsworth, Principles of Environmental Physics, second edition, Edward Arnold, 1990.

Nichols, W. D., Energy budgets and resistances to energy transport in sparsely vegetated rangeland, Agri. and For. Meteorl., 60, 221-247, 1992.

Nichols, W. D., Estimating discharge of shallow groundwater by transpiration from Greasewood in the Northern Great Basin, Water Resour. Res., 29(8), 2771-2778, 1993.

Nichols, W. D., Groundwater discharge by phreatophyte shrubs in the Great Basin as related to depth to groundwater, Water Resour. Res., 30(12), 3265-3274, 1994.

Reginato, R. J., and R. D. Jackson, Comparison of evaporation measurements, Chapter H, USGS WRI 91-4159, 1992.

Schuepp, P. H., M. Y. Leclerc, J. I. Macpherson, and R. L. Desjardins, Footprint prediction of scalar fluxes from analytical solutions of the diffusion equation, Boundary-layer Meteorol., 50, 355-373, 1990.

Shuttleworth, W. J., Evaporation, in Handbook of Hydrology, ed. D. R. Maidment, McGraw-Hill, 1993.

Snedecor, G. W., and W. G. Cochran, Statistical Methods, Iowa State University Press, 1980.

Sorenson, S. K., P. D. Dileanis, and F. A. Branson, Soil water and vegetation responses to precipitation and changes in depth to groundwater in Owens Valley, California, USGS WSP 2370-G, 1991.

Stannard, D. I., Bowen ratio measurements at sites C and F, Chapter B, USGS WRI 91-4159, 1992.

Stannard, D. I., Comparison of Penman-Montieth, Shuttleworth-Wallace, and Modified Priestley-Taylor evapotranspiration models for wildland vegetation in semiarid rangeland, Water Resour. Res., 29(5), 1379-1392, 1993.

Steinwand, A., Transpiration coefficients for three Great Basin shrubs, report to the Inyo County/LADWP technical group, 1999.

Tyler, S. W., S. Kranz, M. B. Parlange, J. Albertson, G. G. Katul, G. F. Cochran, B. A. Lyles, and G. Holder, Estimation of groundwater evaporation and salt flux from Owens Lake, California, USA, J. Hydrology, 200, 110-135, 1997.

Webb, E. K., G. I. Pearman, and R. Leuning, Correction of flux measurements for density effects due to heat and water vapor transfer, Quart. J. R. Meteorol. Soc., 106, 85-100, 1980.

Wilson, D. H., R. J. Reginato, and K. J. Hollett, Evapotranspiration measurements of native vegetation, Owens Valley, California, USGS WRI 91-4159, 1992.

Wood, E. F., Scaling behavior of hydrological fluxes and variables: empirical studies using a hydrological model and remote sensing data, Hydrological Processes, 10th anniversary issue, 21-36, 1995.