Examining Relative Representativeness through Cross-Estimation of Atoll and TAO/TRITON Monthly Rainfall Data

Gridded historical monthly rainfall estimates, derived from atoll-based observations for locations across the Pacific, are available through the PACRAIN website (pacrain.ou.edu.) To improve spatial and temporal coverage and performance, rainfall data from the TAO/TRITON buoy network will eventually be incorporated into the gridded product. However, as buoy-based rainfall data are collected using different instrumentation than that employed at atoll stations, buoy and atoll data sets are not necessarily interchangeably representative of rainfall at their respective locations. Furthermore, while rainfall patterns at buoy locations are as meteorologically unperturbed as may be for situ collection, literature merely supports the idea that cloud prevalence over atolls is perturbed little, relative to that over surrounding waters. To determine the relative representativeness of atoll and buoy rainfall observations with respect to open ocean conditions, and to determine the similarity of error properties of estimates derived from each class, rainfall observations are interpolated to buoy locations using rainfall observations collected solely at buoys, and then solely at atolls. The procedure is then repeated for atoll locations. Interpolated estimates and the predictive abilities of their respective theoretical errors are compared.