Robert T. McMillen

An eddy correlation technique with extended applicability to non-simple terrain

A system is described which is intended to calculate vertical fluxes of heat, moisture, momentum, and certain atmospheric pollutants at sites that are less than ideal. Fluxes, along with other turbulence statistics, are computed in real-time and printed at the end of each averaging period. The main elements of the program are (1) ‘detrending’ (by use of running mean removal), (2) calculation of the entire stress tensor (which allows a three-dimensional coordinate rotation to be performed on the covariances), (3) software-adjustable timing delays for each instrument channel, and (4) real-time graphic presentation of the raw data as stripchart images. The first two of these program elements tend to relax the normal site and sensor-leveling requirements. Sample results are presented, and the sensitivities of the calculated quantities to coordinate rotation and to mean removal time are examined for both ideal and non-ideal sites.

Solar radiation within an oak—hickory forest: an evaluation of the extinction coefficients for several radiation components during fully-leafed and leafless periods

Global shortwave, photosynthetically active, net and allwave radiation was measured above, and at several levels within an oak—hickory forest with instruments mounted on a moving tram system. Profiles of radiation flux densities were quantified using extinction coefficients based on the Beer—Bouguer law. Data are reported here from periods when the forest was both fully-leafed and leafless. In the fully-leafed forest the solar radiation components are attenuated exponentially in the following manner: PAR > Q∗ = K↓ > Q↓;, where PAR is photosynthetically active radiation, Q∗ is net radiation, K↓ is shortwave radiation and Q↓ is allwave radiation. PAR attenuation is greater than that for the other components because leaves preferentially absorb PAR. This preferential absorption causes the ratio, PAR/K↓, to decrease from 0.49 above the canopy to 0.27 at the forest floor. During the leafless phenoseason, the radiation components are attenuated exponentially as follows: Q∗ > K↓ = PAR > Q↓. K↓ and PAR are attenuated in a similar manner during this phenological phase because no leaves are present to absorb PAR preferentially. The magnitude of the attenuation coefficients for Q∗, K↓ and PAR is much greater during winter leafless period because solar elevation angles are lower and the canopy consists of dark, opaque, woody biomass. Shortwave beam radiation is not attenuated in an exponential manner. Consequently, extinction coefficients for beam radiation, γ(S), were computed separately for the upper canopy and lower canopy. A comparison between measured and modeled γ(S) show periods of reasonable agreement and disagreement. Deviations from theory are attributed to clumping and gaps in the canopy.

Land–Atmosphere Interaction Research, Early Results, and Opportunities in the Walnut River Watershed in Southeast Kansas: CASES and ABLE

Abstract This paper describes the development of the Cooperative Atmosphere Surface Exchange Study (CASES), its synergism with the development of the Atmosphere Boundary Layer Experiments (ABLE) and related efforts, CASES field programs, some early results, and future plans and opportunities. CASES is a grassroots multidisciplinary effort to study the interaction of the lower atmosphere with the land surface, the subsurface, and vegetation over timescales ranging from nearly instantaneous to years. CASES scientists developed a consensus that observations should be taken in a watershed between 50 and 100 km across; practical considerations led to an approach combining long—term data collection with episodic intensive field campaigns addressing specific objectives that should always include improvement of the design of the long—term instrumentation. In 1997, long—term measurements were initiated in the Walnut River Watershed east of Wichita, Kansas. Argonne National Laboratory started setting up the ABLE ar...

A micrometeorological investigation of surface exchange of O3, SO2 and NO2: A case study

Data obtained in an intensive field study of the dry deposition of sulfur dioxide, ozone, and nitrogen dioxide, conducted in 1985 in central Pennsylvania, are used to illustrate the factors that must be considered to assure that high quality results are derived. In particular, the quality of the site must be such that flux measurements made above the surface are representative of surface values. For this purpose, tests involving momentum transfer and the surface energy budget are especially useful. In addition, conditions must not be changing rapidly, and the statistical uncertainty associated with flux measurement must be low. For the set of data presented here, conservative quality-assurance guidelines are used to reject potentially erroneous flux data. For ozone, most of the measured fluxes are of use in deriving surface resistances. For SO2, far fewer data points are available. For NO2, fluxes appear to lack the order of the O3 and SO2 fluxes, and do not enable surface resistances to be computed. The highest-quality SO2 and O3 data yield surface resistances in fair average agreement with model predictions for SO2, but substantially higher than predictions for O3.

Seasonal variations in the radiation regime within an oak-hickory forest

Abstract Seasonal variations in solar elevation, canopy phenology and leaf pigmentation result in a dynamic variation in the radiation regime within a deciduous forest. Measurements of several insolation components taken within an east Tennessee oak-hickory forest during eight combinations of season and phenological conditions of the canopy are presented. Insolation measurements were made with instruments mounted on a moving tram system at seven levels within, and at one level above the canopy. The attenuation of solar radiation is least during the leafless phenoseason since only woody biomass is present to intercept and absorb it. Net radiation is attenuated to a greater extent than shortwave and photosynthetically active radiation, which are attenuated to a similar degree, since the sparse woody biomass does not effectively trap the outgoing reflected and radiated radiation. Solar elevation did not influence the attenuation of insolation within the leafless forest under the conditions studied. The attenuation of insolation within the leafing, spring canopy increases progressively with increasing leaf area. The attenuation among the components, however, varies. Shortwave attenuation progresses at a faster rate with expanding leaf area than photosynthetically active radiation since newly expanded leaves are low in chlorophyll. The penetration of light into the leafing canopy is independent of solar elevations less than 40° and increases linearly with greater solar elevations. The radiation regime within the fully-leafed canopy is rather static. For daily mean values, no effects of leaf age or seasonal variation in solar elevation are evident. Penetration of light into the canopy, however, increases when solar elevation exceeds 65°. The attenuation of insolation components diminishes during the autumnal senescence—abscission phenoseason, in response to changes in leaf pigment and loss of foliage. Attenuation of insolation is greater in the autumn than during spring periods with similar leaf area since the sun is lower in the sky in the autumn. The depletion of photosynthetically active radiation with depth is minimal during the leafless, early leafing and autumnal phenoseasons. Moderate depletion occurs during the late leafing period and maximal depletion occurs during the fully-leafed phenoseason.

A comparison of estimated and measured SO2 deposition velocities

A nested-network program for obtaining data on the dry deposition of SO2 and SO4− has been initiated at a small array of locations (6 in 1985, presently 13) across North America. The procedures involved rely on the availability of models for deriving dry deposition rates from observations of air concentrations and of meteorological and surface properties known to influence the deposition velocity. At a subset of locations (i.e., 3), the results obtained by this indirect method are tested by comparison against more direct methods. One of the first comparison experiments of this series was conducted at Oak Ridge in July 1985 when the fluxes determined by inferential methods were compared to those measured by eddy correlation. The results obtained suggest that initial computer routines, developed to estimate deposition velocity for SOz on a routine basis, overestimate the deposition velocity by about 20% to a mixed-species deciduous forest. The difference is possibly due to the omission of water stress as a contributing factor in the initial computer routines, but might also be associated with chemical processes at the substomatal level.

Seasonal variation in the statistics of photosynthetically active radiation penetration in an oak-hickory forest

Abstract The seasonal variability in the frequency distribution (FD) and higher order moments (e.g., standard deviation, skewness and kurtosis) of photosynthetically active radiation (PAR) penetration in an oak—hickory forest were examined. In a leafless, wintertime forest the FD is unimodal and skewed. In a leafing forest, the FD is highly variable and is strongly skewed. This results from higher solar elevation angles enhancing PAR penetration and new, expanding leaves causing penumbral shade and scattering the incoming radiation. The FD of PAR in the fully-leafed canopy is unimodal in the upper canopy, bimodal below crown closure and unimodal in the subcanopy trunkspace. Low solar elevation angles and leaf senescence cause the FD of PAR in the autumnal canopy to be uniformly distributed, as is indicated by its low kurtosis. FDs of PAR also provide a means of estimating sunlit leaf area. Estimated values of sunlit leaf area below crown closure agree within 15% with theoretical estimates derived with a negative binomial distribution.

Effects of sky brightness distribution upon penetration of diffuse radiation through canopy gaps in a deciduous forest

Abstract The penetration of diffuse sky radiation via canopy opening into a fully-leafed tulip poplar forest was estimated from canopy structure data obtained from canopy photographs and from sky brightness distribution approximations, using techniques originally developed by Anderson (1964). Small differences were found among mean daily penetration fractions predicted assuming an isotropic sky (UOC), a standard overcast sky (SOC), or an approximation of the weighted daily average predicted using a clear sky brightness distribution. For shorter time periods, penetration of diffuse radiation estimated using the UOC and SOC approximations differed substantially from estimates made using the clear-sky brightness distribution. Most diffuse radiation passes through openings that are within 10 degrees of the solar disk. Hence the directional distribution of diffuse radiation in the forest is strongly controlled by solar position on clear to partly cloudy days. Amounts of diffuse radiation observed in the forest under clear skies agree fairly well with those predicted by the model developved by Reifsnyder et al. (1971). In both cases, penetration increases with increasing solar elevation on clear days. We conclude that the SOC assumption produces acceptable results for time periods of a day or longer, but for shorter times with clear skies the use of the clear sky brightness distribution is necessary for most simulation or prediction modelling.

Correcting airborne flux measurements for aircraft speed variation

Airplane aerodynamic characteristics correlate aircraft speed with vertical wind velocity, making the time average inappropriate for estimating the ensemble average in airborne eddy-correlation flux computations. The space average, the proper form, is implemented as a time integral by a transformation of variables, which can be interpreted as a ground-speed correction to the time average. The mathematical forms are presented, and the importance of the speed correction is illustrated with airborne data. The computed correction is found to be highly variable, depending on both the turbulent flow encountered and the aircraft used. In general, the speed connection becomes more important as airplane size is reduced. For a small, single-engine Long-EZ airplane, used as an example, the straight time average erred, half the time, by 12%, 10%, 20%, and 15%, respectively, for computed fluxes of momentum, heat, moisture, and CO2. For a much heavier Twin Otter airplane, also used as an example, the straight time average erred, half the time by only 1%. These errors increased with decreasing altitude for the Long-EZ and with increasing altitude for the Twin Otter.

A national critical loads framework for atmospheric deposition effects assessment: III. Deposition characterization

Methods are discussed for describing patterns of current wet and dry deposition under various scenarios. It is proposed that total deposition data across an area of interest are the most relevant in the context of critical loads of acidic deposition, and that the total (i.e., wet plus dry) deposition will vary greatly with the location, the season, and the characteristics of individual subregions. Wet and dry deposition are proposed to differ in such fundamental ways that they must be considered separately. Both wet and dry deposition rates are controlled by the presence of the chemical species in question in the air (at altitudes of typically several kilometers in the case of wet deposition, and in air near the surface for dry). The great differences in the processes involved lead to the conclusion that it is better to measure wet and dry deposition separately and combine these quantifications to produce “total deposition” estimates than to attempt to derive total deposition directly. A number of options for making estimates of total deposition to be used in critical loads assessment scenarios are discussed for wet deposition (buckets and source receptor models) and for dry deposition (throughfall, micrometeorology, surrogate surfaces and collection vessels, inference from concentrations, dry-wet ratios, and source-receptor models).