Experimental design for flux measurements: matching scales of observations and fluxes

Abstract

Listen

Translate article

Practical and conceptual questions of micrometeorological flux measurements over inhomogeneous areas are discussed, especially pertaining to the selection and layout of sites and instrumentation in areas that do not comply with generally accepted standards of fetch and homogeneity. Well-known models for footprint or source area analysis are used to evaluate quantitative criteria for the point-to-area representativeness of each component involved in flux measurements. Based on the acceptance or rejection of these criteria, practical recommendations are formulated about the choice of measurement technique and the height range of instrument deployment for a given situation in the field. The scope and scale of the practical recommendations for experimental design that are presented concentrate on two principal measurement scenarios: (1) the measurement of a vertical flux, such as water vapour transport, from a specific crop of limited size; (2) the estimate of spatially representative regional fluxes over agricultural, savannah or urban areas. For both scenarios the optimal selection and layout of measurement sites are examined with respect to the choice of measurement technique: by eddy correlation, the profile technique, or via the energy balance approach. It is argued that it is essentially unimportant what method is used, as long as the overriding principle is observed that the scale of areal representativeness of each component of the measurements be matched to the relevant scale of the flux. To achieve this, it is convenient to define a criterion for the point-to-area representativeness of the measurements involved in the determination of the flux. Three different (but related) versions of representativeness criteria are presented in this paper, which differ by their rigour and the amount of effort required for their application. The most pragmatic version, the source area case examination method, is described in some detail, and its use is demonstrated by examples that show that the suitability of observations for flux evaluations varies greatly between the different measurement techniques in practice today.