Analysis of daily solar irradiation variability and predictability in space and time important for energy resources planning, development, and management. The natural variability
of solar irradiation is being complicated by atmospheric conditions (in particular cloudiness) and
orography, which introduce additional complexity into the phenomenological records. To address
this question for daily solar irradiation data recorded during the years 2013, 2014 and 201511 stations measuring solar irradiance on La Reunion French tropical Indian Ocean Island, we use set of novel quantitative tools: Kolmogorov complexity (KC) with its derivative associated measures
and Hamming distance (HAM) and their combination to assess complexity and corresponding
predictability. We find that all half-day (from sunrise to sunset) solar irradiation series exhibit
high complexity. However, all of them can be classified into three groups strongly influenced trade winds that circulate in a “flow around” regime: the windward side (trade winds slow down),
the leeward side (diurnal thermally-induced circulations dominate) and the coast parallel to trade
winds (winds are accelerated due to Venturi effect). We introduce Kolmogorov time (KT) that
quantifies the time span beyond which randomness significantly influences predictability.
