Abstract
A new algorithm called Lagrangian Simulation (LSIM) has been developed that enables the interpolation uncertainty present in Lagrangian satellite rainfall algorithms such as the Climate Prediction Center (CPC) morphing technique (CMORPH) to be characterized using an ensemble product. The new algorithm generates ensemble sequences of rainfall fields conditioned on multiplatform multisensor microwave satellite data, demonstrating a conditional simulation approach that overcomes the problem of discontinuous uncertainty fields inherent in this type of product. Each ensemble member is consistent with the information present in the satellite data, while variation between members is indicative of uncertainty in the rainfall retrievals. LSIM is based on the combination of a Markov weather generator, conditioned on both previous and subsequent microwave measurements, and a global optimization procedure that uses simulated annealing to constrain the generated rainfall fields to display appropriate spatial structures. The new algorithm has been validated over a region of the continental United States and has been shown to provide reliable estimates of both point uncertainty distributions and wider spatiotemporal structures.
