1: \begin{abstract}

2: We consider Aerosol Index (AI) time-series extracted from TOMS

3: archive for an area covering Italy  $(7-18^o E ; 36-47^o N)$. The

4: missing of convergence in estimating the embedding dimension of

5: the system and the inability of the Independent Component Analysis

6: (ICA) in separating the fluctuations from deterministic component

7: of the signals are evidences of an intrinsic link between the

8: periodic behavior of AI and its fluctuations.  We prove that these

9: time series are well described by a stochastic dynamical model.

10: Moreover, the principal peak in the power spectrum of these

11: signals can be explained whereby a stochastic resonance, linking

12: variable external factors, such as Sun-Earth radiation budget and

13: local insolation, and fluctuations on smaller spatial and temporal

14: scale due to internal weather and antrophic components.

15: \end{abstract}

16: