spheric variables was described. It could compute the covariance in a frequency band governed at
high frequencies by the sensor response times and at low frequencies by capacitive filters. This
instrument, using vertical wind velocity as one input, has been successfully used for measurements
of Reynolds stress and sensible and latent heat fluxes.
Though the necessity of measuring the vertical flux of heat and vapor, brought about by eddy
measurement in the lower atmosphere, has long been recognized, Swinbank (1951) was the first to
design an apparatus providing a continuous record (over a 5-minute interval) of the detailed
structure of temperature, vapor pressure, and wind speed as well as its vertical component. He
described the derivation of the vertical fluxes and presented an analysis of error estimation.
Dyer (1961) reported measurements of evaporation and heat transfer in the lower atmosphere by
an automatic eddy-correlation technique that he proved to be practical and accurate for this type of
measurement over a reasonably homogeneous surface. Under conditions of near-neutrality, the
accuracy of the technique is not limited by response time over the range of wind speed normally
encountered. With increasing atmospheric instability, the wind speed range can be further expand-
ed and, unlike in the aerodynamic method, optimum performance is obtained at maximum instabil-
ity. For measurements made at 4 m from the surface, Dyer indicated that minor surface irregulari-
ties of up to several tens of centimeters are clearly of no consequence, and if sufficient upwind fetch
is available, even greater irregularities can be alleviated by increasing the height and period of
observation.
Dyer (1967), in a study on the turbulent transport of heat and water vapor in an unstable
atmosphere, indicated that the transfer mechanism for heat and water vapor are identical for a
freely evaporating surface (i.e., φh = φw where φh and φw are MoninObukhov universal functions
on the values of the stability parameter z/L (where z is the vertical distance from the surface at
which the measurement is made, and L is the Obukhov length). He concluded that this finding is
consistent with that from shape function analysis of the same data by Swinbank and Dyer (1967).
Hicks and Martin (1972) reported experimental results on atmospheric turbulent fluxes over
snow. Their experiments were conducted under light wind along with a highly stable atmosphere
near the snow surface. Fluxatrons developed by Hicks (1970) were used to determine these eddy
fluxes. Hicks and Martin indicated that the use of eddy correlation technique to measure the fluxes
of momentum, water vapor, and sensible heat over a snow or even an ice surface has been proved to
be practical and to have a degree of accuracy satisfactory for many studies. They showed that snow
surfaces are generally extremely smooth (not different from aerodynamic smoothness) with corre-
sponding low friction velocities, so large deviations from atmospheric neutrality may be rather
common. They claimed that direct measurement should be preferable to alternative techniques,
based on assuming either the surface roughness or the similarity of transport mechanism. Low bulk
transfer coefficients, based on data collected over Lake Mendota, Wisconsin, were expected after
considering the influence of atmospheric stability on eddy diffusivity, and they reported fluxes of 9
W/m2 for sensible heat and 22 W/m2 for latent heat.
Swinbank (1968) derived a number of relationships based on dimensional analysis that connect
the vertical fluxes of heat and horizontal momentum in the constant flux layer with other relevant
variables. Experimental data taken under unstable conditions are used to test the correctness of the
developed prediction relationship and its functional form. Swinbank found many dimensionless
groups fit the data extremely well, with correlation coefficients in the range of 0.980.99. He
reported that θ / u is height-dependent, indicating dissimilarity in respective transfer mecha-
nisms, and concluded that any H, u*, θ / z , and u / z at one level are sufficient to prescribe the
heat transfer, the shearing stress, and the temperature and wind gradients throughout this constant
flux layer over any uniform surface. Based on the predictive relationships and experimental data of
half-hourly mean values, Swinbank concluded that up to a height of 16 m, the constant flux layer is
2
Previous Page
