Microwave tomographic inversion technique based on stochastic approach for rainfall fields monitoring

The microwave tomographic inversion technique (MTIT) proposed in 1991 for reconstruction of rainfall fields at ground through microwave attenuation measurements is reconsidered. A new algorithm for data inversion is presented [referred to as stochastic reconstruction technique (SRT)] that generally performs better than the one originally adopted [referred to as arithmetic reconstruction technique (ART)]. Improvement is achieved in spatial definition and general reliability of rainfall field reconstruction. The new model adopted to represent the reconstructed rainfall fields leads to a completely different strategy for the inversion problem, and this strategy is based on a global optimization stochastic technique (GOST). Results obtained through the SRT-MTIT are presented in the paper and compared to those obtained by employing the ART-MTIT. It also is shown that, based on the SRT-MTIT approach, fast and reliable time tracking of rainfall events is made possible by exploiting previous reconstructions and by the improved long-term physical consistency of the model adopted for rainfall field decomposition