Report Number




An upgraded spectral radiation model called SMARTS2 is introduced. The solar shortwave direct beam irradiance is calculated from spectral transmittance functions for the main extinction processes in the cloudless atmosphere: Rayleigh scattering, aerosol extinction, and absorption by ozone, uniformly mixed gases, water vapor, and nitrogen dioxide. Temperature-dependent or pressure-dependent extinction coefficients have been developed for all these absorbing gases, based on recent spectroscopic data obtained either directly from the experimental literature or, in a preprocessed form, from MODTRAN2, a state-of-the-art rigorous code. The NO2 extinction effect, in both the UV and visible, is introduced for the first time in a simple spectral model. Aerosol extinction is evaluated using a two-tier Angström approach. Parameterizations of the wavelength exponents, single-scattering coefficient, and asymmetry factor for different aerosol models (proposed by Shettle and Fenn, Braslau and Dave, and also in the Standard Radiation Atmosphere) are provided as a function of both wavelength and relative humidity. Moreover, aerosol turbidity can now be estimated from airport visibility data using a function based on the Shettle and Fenn aerosol model.

An improved approximation to the extraterrestrial solar spectrum, treated at 1 nm intervals between 280 and 1700 nm, and 5 nm intervals between 1705 and 4000 nm, is based on recent satellite data in the UV and visible. The total irradiance between 280 and 4000 nm (1349.5W/m2) is obtained without scaling and is in good agreement with the currently accepted value of the solar constant (1367 W/m2). Incident diffuse radiation and global radiation for any plane orientation at ground level are also calculated by the model, with provision for both multiple scattering effects and ozone absorption intricacies in the UV. SMARTS2 also has an optional circumsolar correction function and two filter smoothing functions which together allow the simulation of actual spectroradiometers. This facilitates comparison between modelled results and measured data.

Detailed performance assessment of the model is provided. It consists of prediction comparisons between the proposed model and rigorous radiative transfer models for different specific atmospheric conditions in the UV, visible, and near IR. High quality measured datasets in the UV and in the intervals 290-620 nm and 300-1100 nm are also used to validate the model by direct comparison. The resultant overall accuracy appears a significant improvement over existing simplified models, particularly in the UV. Therefore, SMARTS2 can be used in a variety of applications to predict full terrestrial spectra under any cloudless atmospheric condition.

Date Published




Rights Statement

In Copyright