Satellite-based forecasts for seeing and photometric quality at the European Extremely Large Telescope site

In this article, we describe a new algorithm for short-time satellite-based forecasts for seeing and photometric quality at the European Extremely Large Telescope (E-ELT) site (Armazones) and we analyse the correlation between the Paranal and Armazones sites. The algorithm uses data from the polar satellite Aqua's Moderate Resolution Imaging Spectroradiometer (MODIS) and the Geostationary Operational Environmental Satellite (GOES 13). We have analysed 13 years (2003-2015) of cloud coverage data from MODIS in order to obtain the cyclical perturbations through Fourier analysis. Then we have developed the forecast model using GOES 13 d data (2015). Monthly calibration atmospheric-layer temperature thresholds have been achieved through the daily temperature range detected by the satellite. The algorithm works through conditional probability. This allowed us to extrapolate the main frequency of the cloud-coverage perturbations, achieving three results: there are two major seasonal meteorological frequencies at Armazones and a short one of 14 days. This result allows us to improve the rate of the prediction algorithm by introducing a new threshold function. The correlation of 98 per cent found between the pixel above Paranal and the pixel above Armazones allows us to use the Paranal ground data to validate the prediction model. We analysed the 2015 data at Armazones and reached a correlation of 97 per cent for the short-time photometry and seeing quality forecast.