An Error Analysis of the Thornthwaite-Holzman Equations for Estimating Sensible and Latent Heat Fluxes over Crop and Forest Canopies

View More View Less
  • 1 Department of Meteorology, University of Maryland, College Park 20742
© Get Permissions Rent on DeepDyve
Restricted access

Abstract

Computations of sensible and latent heat fluxes over forest and crop canopies using the Thornthwaite-Holzman equations require an a priori knowledge of roughness and data displacement heights. If the values of these parameters are not representative of the wind structure above the canopy, then errors will be induced in the heat flux calculations.

A detailed error transfer analysis is derived and tested against tower data taken in a tropical dry ever-green forest in Thailand. It is shown that significant errors may occur in heat flux computations if mean values of roughness parameter and datum displacement height are used during periods of significant variation of those parameters. Further, the sensitivity of a heat flux calculation to the discrepancy in the profile parameters used may be either suppressed or amplified.

A simple model is derived to estimate this error transfer from estimates of the mean values and variances of the roughness parameter and data displacement height.

Abstract

Computations of sensible and latent heat fluxes over forest and crop canopies using the Thornthwaite-Holzman equations require an a priori knowledge of roughness and data displacement heights. If the values of these parameters are not representative of the wind structure above the canopy, then errors will be induced in the heat flux calculations.

A detailed error transfer analysis is derived and tested against tower data taken in a tropical dry ever-green forest in Thailand. It is shown that significant errors may occur in heat flux computations if mean values of roughness parameter and datum displacement height are used during periods of significant variation of those parameters. Further, the sensitivity of a heat flux calculation to the discrepancy in the profile parameters used may be either suppressed or amplified.

A simple model is derived to estimate this error transfer from estimates of the mean values and variances of the roughness parameter and data displacement height.

Save