Estimation of evaporation using a dual-beam surface layer scintillometer and component energy balance measurements

摘要

A dual-beam surface layer scintillometer (SLS), for the estimation of sensible heat flux density H for a path length of 101m, was used in a mixed grassland community in the eastern seaboard of South Africa for 30 months. Measurements also included Bowen ratio (BR) and eddy covariance (EC) estimates of H. Acceptable SLS data between 0600h and 1800h, judged by the percent of error-free 1kHz data exceeding 25% and an inner scale of turbulence exceeding 2mm, showed little seasonal variation and was consistently high--between 86.7% and 94.8%. An analysis of the various Monin-Obukhov similarity theory (MOST) empirical dimensionless stability functions used for estimating H from the SLS measurements showed percent differences in H that varied from -30% to 28% for neutral to unstable conditions, respectively and for stable continuous conditions the differences in H were within 60Wmpo with much larger differences for stable sporadic conditions. The good agreement in measurements of H over an extended period for the SLS, BR and EC methods demonstrates the applicability and robustness of the SLS method and the associated MOST empirical functions used for estimating H for a range of canopy heights, stability conditions and diurnal and seasonal weather conditions. Furthermore, there was no evidence for an underestimation in EC sensible heat compared to SLS and BR measurements, which implies that any lack of energy balance closure points to possible latent energy EC underestimation or due to energy fluxes not included in the shortened energy balance if the net irradiance and soil heat flux components are correct. A sensitivity analysis was used to determine the relative importance of the SLS data inputs of air temperature, atmospheric pressure, beam path length and beam height on H estimates. Worst-case errors in air temperature, atmospheric pressure, beam path length and beam height resulted in errors in H within 1.0%, 1.3%, 3.0% and 4.0%, respectively. Overall, the worst-case total percent error in SLS-estimated H is within 5.3% and the typical percent error is within 3.9%. Accounting for the error in net irradiance and soil heat flux measurements, the seasonal variation in the error in daily evaporation estimated as a residual of the energy balance is generally less than 0.2mm (0.49MJmpo) in winter when the daily evaporation was about 1mm (2.45MJmpo) and typically less than 0.4mm (0.98MJmpo) when the evaporation exceeded 4mm (9.8MJmpo). Soil heat flux density measurements can contribute significantly to the overall error.