Based on immunocomputing (IC), this paper proposes a new way for geoinformation fusion, spatio-temporal modeling, and forecast. The approach includes mathematically, rigorous mapping of high-dimensional spatio-temporal data into a scalar index, discrete tree transform (DTT) of the index values into states of cellular automata (CA), and identification of CA by IC. Numerical examples use official data of International Association for the Development of Freediving (AIDA), World Health Organization (WHO), as well as time series of Solar Influences Data Analysis Center (SIDC) and National Aeronautics and Space Administration (NASA). Anomaly index is also proposed using special the case of DTT. Recent results suggest that the IC approach outperforms (by training time and accuracy) state-of-the-art approaches of computational intelligence.