AOSC Seminar
February 2, 2017

Environmental data fusion: combining satellite remote sensing and data assimilation techniques for NWP and situational awareness applications


Sid Boukabara
NOAA NESDIS/STAR
Abstract:  

The volume and diversity of environmental data obtained from a variety of Earth-observing systems, has experienced a significant increase in the last couple years with the advent of high spectral, high- spatial and temporal resolutions sensors. At the same time, users-driven requirements, especially for nowcasting and short-term forecasting applications, strongly point to the need for providing this data in a user-friendly, consistent, comprehensive and consolidated fashion, combining space-based, air-based and surface-based sources. This trend is expected to continue further with the emergence of commercial space-based data from multiple industry players. We present in this study the results of a pilot project’s effort to fuse data from many sources including satellites, conventional data and airborne data. The outcome is a 5D-cubeset of parameters to describe the state of the Environment, useful for multiple applications including hydrology, cryosphere, atmosphere, land, oceanography, clouds and hydrometeors. The approach relies on combining the techniques of data assimilation (DA) used in Numerical Weather Prediction (NWP), with physically-based space-based remote sensing (RS) inversion techniques. To be clear, the approach does not extend assimilating more products originating from RS algorithms into the DA system, but rather integrate the remote sensing inversion itself within the DA system. This is done in order to ensure consistency of the data fusion outputs, increase the accuracy of the outputs by benefiting from a wider variety of observations for a single grid output and more importantly, enrich the DA outputs to include parameters currently not analyzed (such as precipitation, surface emissivity, all-weather sounding, trace gases, etc). The present study demonstrates the proof of concept of this data fusion approach and highlights the need (and usefulness) of converging data assimilation and remote sensing techniques. The Data Fusion of Environmental Observations (DFEO) presented in this study generates global fields of geophysical parameters at flexible temporal and spatial resolutions; presented here at sub-hourly update and 13 Kms resolution. One of the characteristics of the data fusion as implemented here, is that the observation (e.g. satellite measurements) is fitted even in the case of an inconsistency between the background and observations (as happens sometimes when the forecasted field is misplaced), therefore enabling background adjustments when appropriate. The outputs of the Data Fusion are assessed by comparing them to accurate geophysical analyses.