Traditionally, the measurement of surface radiation has been made to quantify the downward flux incident on a horizontal surface. This is used for studies involving the radiative transfer of flux through the atmosphere and at the earth's surface.
The downward flux is the total radiation integrated over the celestial hemisphere weighted with the cosine of the zenith angle.
Most radiation monitoring instruments, including the Brewer, are designed to replicate the cosine response to incident radiation. The Teflon diffuser of the Brewer approximates the cosine response; however, at large zenith angles, the response falls below the cosine function (Fioletov et al., 2002). Measurements of global irradiance are therefore in error by values that range up to -10% depending on wavelength, SZA (solar zenith angles), and clarity of the sky. The cosine response is instrument dependent (Bais et al., 2005) and, with proper measurement, correction for the departure from the cosine response is applied to the data with the aid of a radiative transfer model or another parameterization technique (Bais et al., 1998; Fioletov et al., 2002). The uncertainty of the corrected values is about 2%. Also, work has been carried out to improve the cosine response of the sensor (Grobner, 2003), and these new diffusers are available.
It has been demonstrated that there is a sensitivity dependence on SZA for measurements of direct radiation due to polarization effects between components of Brewer spectrophotometers (Cede et al., 2004; 2006a). Unpolarized radiation from direct sun becomes increasingly polarized as the incident angle through the quartz window increases from normal transmission (SZA = 35°), and the direction of enhanced polarization is along the axis parallel to the length of the window. This polarization interacts with the grating, which is more sensitive to radiation polarized on the axis perpendicular to the grating grooves. At SZA = 0°, the polarization axis of the grating is aligned parallel to the polarization axis transmitted through the window. With increasing SZA, the angle between the two axes of polarization increases. Thus, for window transmission normalized to SZA = 35°, the apparent transmission is enhanced for SZA < 35° and reduced for SZA > 35°.
Corrections can be applied to direct radiation data provided the degree of polarization of the grating sensitivity is known. It should be noted that the degree of polarization is not necessarily the same for all instruments. Calculations by Cede et al. (2004; 2006a) assume 100% polarization (extreme case) for the grating response. In general, the degree of polarization for the grating is less than 100%, and for single spectrometers (Mark II and Mark IV), the polarization is significantly less than 100% since there is only one grating.
Corrections can also be achieved by the use of hardware modifications. Use of a cylindrical quartz window with the axis centered on the axis of rotation for the zenith prism would ensure that direct radiation always passes through the window at right angles. Also, a depolarizer inserted in the fore-optics removes the polarization effect (Cede et al., 2006a). Use of a Teflon diffuser instead of the ground quartz diffuser for measurements of direct radiation also avoids this effect since the
Teflon has been found to be an effective depolarizer, whereas the ground quartz removes very little polarization.
The internal polarization effects do not impact the operational measurements of ozone and other atmospheric gases that use DOAS techniques. These techniques measure the relative absorption at different wavelengths and the wavelength dependence of the polarization effect was found to be negligible (Cede et al., 2006a). However, corrections for the polarization effects should be applied for absolute measurements of the extraterrestrial spectra, aerosol optical depth, aerosol single scattering albedo, and estimates of actinic flux (Webb et al., 2002). Measurements of global UV irradiance are not affected since the Teflon diffuser on top of the instrument has been found to be an effective depolarizer.
Was this article helpful?