Aerosols like smoke, oceanic haze, air pollution, or smog are suspensions of fine solid particles or liquid droplets in a gas. They play a key role in building clouds and are supposed to have impacts on rainfall and the climate system. Some scientists like Svensmark see a relationship between solar activity, aerosols, cloud cover, and global warming.

For describing and predictiing the global aerosol index the same data set as for ozone concentration modeling has been used for self-organizing modeling:

1. ‣Global Ozone concentration [DU] (Dobson Units) (x1),
   

2. ‣Global Radiative Cloud Fraction (x2),
   

3. ‣Global Aerosol Index (x3),
   

4. ‣Global CO2 concentration [ppm] (x4),
   

5. ‣Sunspot Numbers (x5).
   

The model shown in the image below was developed from data of the period Nov 1978 to Oct 2008 using a maximum time lag of 36 months. The data till Dec 2010 has been used ex post (out-of-sample) for model evaluation. The model represents a non-linear difference equation of 9 self-selected input variables:

x3(t) = f(x1(t-i), x2(t-j), x5(t-k)),

with i = {18, 28, 35, 36}, j = {6, 16}, and k = {29, 34, 36}. In other words, global aerosol index at a time t is described by ozone concentration, radiative cloud fraction and sun activity at certain previous points in time.

The accuracy of this best model is 81% (R2, coefficient of determination, using leave-one-out cross-validation) at a Descriptive Power of 40% and a high model robustness within the forecast horizon of Jan 2011 to Oct 2017.

The data are available on request.