Web-Based near Real-Time Surface Energy Balance for Short Grass Using Surface Renewal, Temperature Variance and Dissipation Theory.

Increasingly, near real-time evaporation and surface energy fluxes are required, for example, for comparison with satellite flux estimates or for use when remotely-sensed estimates are compromised by the occurrence of cloud, particularly for summer-rainfall areas. Net irradiance and soil heat flux components of the shortened surface energy balance for short grass were measured with the sensible heat flux (H) determined in near real-time using surface renewal (SR) and temperature variance (TV) methods and using a surface renewal-dissipation theory (SRDT) method. Measurements of air temperature from an unshielded fine-wire thermocouple placed at a height of 0.46 m above the soil surface were obtained at 10 Hz. From these measurements, the following air temperature statistics were determined every 30 min: mean, variance, skewness, and air temperature structure functions of order 2, 3 and 5 for lag times of 0.4 and 0.8 s. For cloudless conditions, the 0.4-s lag corresponded to the maximum of the negative of the structure function of order 3 divided by the time lag and was therefore used for the calculation of H using the SR and SRDT methods. The SR method requires calculation of the air temperature ramp amplitude and the quiescent and ramping periods for a 30-min period. For the air temperature ramp amplitude, the roots of a cubic polynomial were obtained in real-time using a datalogger program employing an iterative procedure for which the ramp amplitude was determined to within 0.005 ^oC from which H was determined using a SR weighting factor of 1. For the TV method, the direction of H was determined from the sign of the third-order air temperature structure function and the magnitude of H determined from the mean, variance and skewness of air temperature with adjustments for skewness applied for positive skewness and unstable events. The SRDT method, which uses the square of the SR ramp amplitude, the ramp period and the variance of air temperature, tended to underestimate H compared to SR and TV methods (for the SRDT vs SR method: slope = 0.771, coefficient of determination (R²) = 0.990, root mean square error (RMSE) = 3.1 W m^-2) using data from 1 June to 28 July 2012. With adjustment for skewness, the TV method showed good agreement with the SR method (slope = 1.035, R² = 0.905, RMSE = 13.2 W m^-2). A shortened surface energy balance was used to determine the latent energy from measured net irradiance from two- and four-component net radiometers, the measured soil heat flux and H, the latter using SR, TV and SRDT methods. The fluxes, updated half-hourly, are displayed on a web-based system: (http://agromet.ukzn.ac.za:5355/?command=RTMC&screen=Energy%20balance). Current daily, weekly and monthly totals are reported as well as a temporal plot of the energy balance components for a five-day period. Half-hourly energy balance components are tabulated.