Evolving Capabilities and Expectations for the GPCP Precipitation Products

Abstract:

The Global Precipitation Climatology Project (GPCP) was established in the mid-1990’s to determine whether the then-new passive microwave sensors could reliably depict quasi-global variations in precipitation, in particular those associated with El Niño Southern Oscillation (ENSO) events. This international community-based effort was successful, and its data record now extends past 35 years. In the process, we have found ways to use estimates from sensors flying before the passive microwave era, and estimates for polar regions in which the passive microwave retrievals are not useful, as well as providing pentad (5-day) and daily estimates (the latter for a subset of the 1979-present record). Several versions have been released as lessons learned and new input datasets have been applied to the computation. Among these lessons are the importance of maintaining consistency in input data sources, requiring consistent processing for the entire record of input datasets, and ensuring completeness in data coverage. The resulting data sets were not originally designed to be formal Climate Data Records (CDRs), but they do emphasize homogeneity, high quality, and stability in the input data sets.

The success of the datasets has raised user expectations and encouraged a variety of analyses that were not envisioned in the original design. In particular, the GPCP data contain approximations that make it difficult for the data to adequately represent fine-scale “extremes”. As well, improved input data sets and algorithms have been developed which are not accommodated in the current version. In response, the GPCP is working under NASA MEaSUREs funding to create a next-generation version at finer scale that is better suited for a wider range of analyses. The widening circle of non-expert users has widened the range of formats and services that are needed. These developments increase the utility of such data sets to users, with the unintended effect that the cost of getting a data set into operation is raised. We find that data archive facilities play a key role in providing these user-oriented services. Community-based standards offer the promise of uniform access to Earth science data, but in the short term we are seeing more diversity in data set presentation and more confusion among users, not less.