CORRECTING ERRORS ASSOCIATED WITH SOIL HEAT-FLUX MEASUREMENTS AND ESTIMATING SOIL THERMAL-PROPERTIES FROM SOIL-TEMPERATURE AND HEAT-FLUX PLATE DATA

This paper deals with some of the problems associated with measuring the soil heat flux at a semi-arid and sparsely vegetated site on the Colorado short-grass steppe. A linear model of soil temperature and heat flux is used to show that errors in the measurements of soil heat flux by the combination method can occur whenever the mean time rate of temperature change of a soil slab is estimated from a set of discrete temperature measurements. A synthesis of model and data is used to demonstrate how these errors can be avoided. The same technique is extended to estimate the soil diurnal damping depth, thermal conductivity, and volumetric specific heat capacity from the soil temperature and heat flux measurements. The present techniques close the surface energy balance to within +/-150 W m-2 (at worst) on a half-hourly basis. To keep remaining errors in the energy balance in perspective, long-duration comparisons are made between two types of net radiometers and between estimates of the evapotranspiration as determined by eddy correlation and a weighing lysimeter.