A geostatistical framework for quantifying the temporal evolution and predictability of rainfall fields

We propose a Geostatistical framework for studying the spatial and temporal structure of rainfall fields in a Lagrangian reference system. Using space-time variogram maps at different time lags, we derive important information about the dynamics and structure of the fields like the apparent motion, anisotropy and temporal innovation (including growth and decay). Special emphasis is put on quantifying the temporal innovation of organized mesoscale convective systems and understanding how fast these systems evolve with respect to their size, intensity and spatial organization. We illustrate our method using high resolution radar data for the United States and discuss potential applications to related problems such as the space-time interpolation of rainfall fields and short-term quantitative precipitation forecasting based on radar observations.