College of Engineering | Apply to CSU | Disclaimer | Equal Opportunity Statement | Privacy | Search CSU

The Spatial and Temporal Properties of Precipitation Uncertainty Structures over Tropical Oceans

You must be on the CSU network—either physically or using VPN—to watch this or any of the videos on this site.

March 13, 2015
Jianbo Liu
Hosted by Chris Kummerow (advisor) Chris O'Dell Steven Reising (Electrical and Computer Engineering)

Abstract

The global distribution of precipitation has been measured from space using a series of passive microwave radiometers for over 40 years. However, our knowledge of precipitation uncertainty is still limited. While previous studies have shown that the uncertainty associated with the surface rain rate tends to vary with geographic location and season, most likely as a consequence of inappropriate and inaccurate microphysical assumptions in the forward model, the internal uncertainty structure remains largely unknown. Hence, a classification scheme is introduced, in which the overall precipitation uncertainty consists of random noise, constant biases, and region-dependent cyclic patterns. It is hypothesized that those cyclic patterns are the result of an imperfect forward model simulation of precipitation variation associated with regional atmospheric cycles. To investigate the hypothesis, differences from ten years of collocated surface rain rate measurements from TRMM Microwave Imager and Precipitation Radar are used as a proxy to characterize the precipitation uncertainty structure. The results show that the recurring uncertainty patterns over tropical ocean basins are clearly impacted by a hierarchy of regionally prominent atmospheric cycles with multiple time scales, from the diurnal cycle to multi-annual oscillation. Spectral analyses of the uncertainty time series have also confirmed the same argument. Moreover, the relative importance of major uncertainty sources varies drastically not only from one basin to another, but also with different choices of sampling resolutions. Following the classification scheme and hypothesis proposed in this study, the magnitudes of un-explained precipitation uncertainty can be reduced by 68% and 63% over the equatorial central Pacific and eastern Atlantic, respectively.