Sensitivity of the solar resource in solar tower plants to aerosols and water vapor

We make a sensitivity study of the solar resource in a solar tower plant, to aerosols, water vapour and solar position, considering typical atmospheric conditions observed at Ouarzazate, Morocco. Four solar resource parameters were defined: the direct normal irradiance (DNI), incident at the heliosat, the slant path attenuation between the heliostats and the receiver (Asp), the solar irradiance incident at the receiver (SIR), and the solar irradiance lost between the heliostat and the receiver (SIRloss). The aerosol optical thickness (AOT), which varies strongly from winter to summer at Ouarzazate, is the main factor on DNI and Asp. However, despite a factor of 10 in AOT between typical winter and summer conditions, Asp varies by a factor of only 2.7 and SIRloss by a factor of only 1.9, for two reasons: 1) the aerosol layer height (ALH) is correlated to AOT and has an opposite effect on Asp; 2) DNI and Asp are anti-correlated. Consequently, despite 100% standard deviation on AOT observed during 2012, the relative standard deviation of Asp is only 60%, and only 47% for SIRloss. Consequently, to estimate solar resource in a solar tower plant, it is recommended to compute directly SIR and not combining DNI and Asp derived from different sources. Also, it is recommended to make computations by considering observed variability of most atmospheric parameters at high time resolution. Computing the solar resource with the annual average of ALH instead of its seasonally dependent value generates an over estimation of Asp from 5.9% to 6.2%. With the annual average of AOT or the Angstrom exponent instead of the hourly-varying values, SIR is changed by ±3%.