Sensitivity and Uncertainty of a Long-Term, High-Resolution, Global, Terrestrial Sensible Heat Flux Data Set

Sensible heat flux directly influences local and regional climate and can be estimated using remotely sensed satellite observations. Although significant efforts have been made to estimate sensitivity and uncertainty in energy flux estimates at the local and regional scales using both models and algorithms compatible with remotely sensed satellite data, few studies quantify the sensitivity or uncertainty at the global scale, enabling a global comparison among uncertainty drivers. This study uses the 10 percentile change from the mean value in the empirical cumulative distribution function for the distribution of each input data set to calculate the sensitivity of the unconstrained, terrestrial sensible heat flux to change in the input data sets and uses this sensitivity in a first-order analysis of the uncertainty in the sensible heat flux. The largest sensitivities to the Zilitinkevich empirical constant (C_zil) are in the Amazon, northern Australia, and the plains of North America, while the sensitivity of the sensible heat flux to the temperature gradient is largest in dry regions of shorter vegetation. The C_zil contributes most to the uncertainty of over 50–100 W/m² in the Amazon and Indonesia, while the temperature gradient contributes most to the uncertainty elsewhere, producing an overall global average uncertainty of 24.8 W/m². Future work should reduce the uncertainties in the temperature gradient and the C_zil to reduce the uncertainty in sensible heat flux estimates.