An Enterprise Environmental Data Fusion and Assimilation System for Nowcasting and NWP Applications

The key requirement for operational nowcasting of extreme weather is timely and reliable environmental intelligence gathered from a variety of sources, including various remotely sensed observations from space, surface-based radar, in-situ observations, and model analyses and forecasts, to name a few. The volume of data available along with the disparities across data sources presents many challenges to the forecaster. For instance, traditional global data assimilation (DA) analyses are only available, at best, every 6 hours, exclude many satellite observations (e.g. radiances impacted by precipitation), and are highly weighted toward the forecast background. Although regional DA analyses provide better temporal and spatial resolution compared to the global systems, often they exclude even more satellite observations. Also, although remote sensing products provide observations directly from the satellite to the user, the volume of data from a large number of sources and algorithms make it difficult to parse for information relevant to specific weather threats and understand error characteristics with the products. Current data fusion and blending approaches have mitigated these issues to some degree, but at the expense of losing physical consistency across parameters and decreasing the fit to the original observations. To address these challenges, a new Enterprise Environmental Data Fusion (EDF) system has been developed at NOAA/NESDIS/STAR with the objective of providing half-hourly, global, high-resolution analyses which unifies together remote sensing and in-situ observations with model fields. The EDF concept involves a preprocessing framework developed to integrate remote sensing algorithms to precondition a background forecast, executing an enhanced global DA system, and a post-processing framework developed to integrate remote sensing algorithms which derive products from the DA analysis. The result is a comprehensive, 4D cube EDF analysis where all environmental parameters are physically consistent with both the input observations as well as model dynamic and thermodynamic fields. In this study, we will present the methodology of the preprocessing using the Multi-Instrument Inversion and Data Assimilation Preprocessing System (MIIDAPS) algorithm which preconditions the background forecast by performing a displacement correction of temperature, water vapor, and cloud fields, and also provides quality control (QC) and other variable constraints (e.g. surface emissivity) to the DA system; as well as show the impacts on the DA analysis from the preprocessing. Additionally, the benefits for operational nowcasting will be illustrated. This includes demonstrating the “one-stop shop” for environmental data with consistent and well-defined QC and error characteristics; and the consistency between EDF analysis parameters themselves, the model fields, and the input satellite observations which span both the LEO and GEO constellations.