This collaborative effort aims at conducting comprehensive diagnostics on precipitation variability, elucidating differences between observed and modeled precipitation data. Dynamic mechanisms and uncertainties associated with these differences will be explored and uncertainties will be quantified for both reconstructed and modeled data. The deliverables include (i) error estimation for global precipitation fields; (ii) precipitation dynamics, data analyses and corresponding uncertainty estimates; (ii) benchmark datasets of the reconstructed precipitation data on daily, monthly, seasonal, and annual scales over the globe and specific regions; (iii) feedback and guidance on reduction of systematic biases in model precipitation; and (iv) professional training of postdoctoral researchers, graduate students, undergraduate students, and high-school students with special attention to minority and underrepresented groups. This research effort has significant broader impacts: in addition to the training and outreach to underrepresented minority students, this activity has the potential of enabling stakeholders interested in precipitation optimize decisions and mitigate impacts due to droughts and floods.
The outcome of our research project includes (A) significant contributions to climate research through the creation of new datasets, new research tools (e.g., mathematical methods and computer software), and new scientific results on global precipitation variations and their mechanisms; (B) training 13 graduate students and one postdoctoral researcher, among whom six are from underrepresented groups (females and Hispanic); and (C) educational partnership development with Ocean Air Elementary School and the Carmel Mountain Preserve. Specifically, our key science outcomes are (i) 16 journal papers; (ii) a suite of reconstructed global precipitation datasets, including (a) global annual precipitation data, (b) global monthly precipitation data, and (c) daily Tibetan Plateau Snow Cover (TPSC 1.0) data Version 1.0; and (iii) a user friendly software engine called the Spectral Optimal Gridding for Precipitation Version 1.0 (SOGP 1.0) capable of producing tailored reconstruction datasets. Our research outcome makes contributions in the following five aspects: (1) A suite of climate data reconstruction software, datasets and videos; (2) Uncertainty analysis for various kinds of precipitation and temperature data; (3) Diagnostic and evaluation analysis on general climate model data and reconstructed data; (4) A suite of mathematical methods for climate data analysis, such as regression reconstruction and error analysis, optimal blending for non-Bayesian data, and time-frequency analysis for non-stationary nonlinear precipitation data; and (5) Educational and public outreach through school projects, public lectures, TV and media interviews, and product release news. Our work has improved the reconstruction of global precipitation since 1900 with better accuracy. Our reconstructed global and regional precipitation products and their uncertainties provide critical information for constructing constraints for future more accurate climate models. Our advanced mathematical methods have enabled a comprehensive error estimation of the uncertainties the historical climate data. Our Spectral Optimal Gridding for Precipitation Version 1.0 (SOGP 1.0) software engine, which is technologically advanced and user-friendly and does the mathematical and statistical heavy lifting for researchers, makes it possible for a much larger community of scientists and other interested individuals to easily reconstruct and access historical precipitation data. For example, a reinsurance company may use SOGP 1.0 to assess agricultural risks based on customarily SOGP-generated historical precipitation information and to offer a highly competitive underwriting policy to its clients.
