Toward a Satellite Algorithm Test Bed

1. Toward a Satellite Algorithm Test Bed Ingrid Guch (NOAA/STAR), Steve Goodman (NOAA/GOES-R Program Office), Andrew S. Jones (CIRA), And James G. Yoe (NOAA/NPOESS Data Exploitation) Requirement:Advance space-based data collection capabilities and associated platforms and systems Science : A Satellite Algorithm Test Bed will provide a centralized framework for STAR and its partners to conduct research and develop algorithms enabled by the growing national and international constellation of earth observing satellites and in-situ data available. Benefit: Improved algorithms and ultimately improved understanding of environmental dynamics and climate NESDIS and its Cooperative Institutes – A Long History of Algorithm Support The SATB concept was initiated by the NESDIS Cooperative Institutes. NESDIS and its Cooperative Institutes (CIs) have long worked collaboratively to transition research algorithms to operational code. There are many examples of algorithms developed at the CIs for both geostationary and polar orbiting satellites. In 1980 NESDIS and the Cooperative Institute for Meteorological Satellite Studies used GOES-4 to generate the first geostationary satellite temperature and moisture profiles. In 1989 NESDIS and the Cooperative Institute for Climate Studies used the High-resolution InfraRed Sounder on operational polar satellites to develop a highly accurate technique to measure Earth’s outgoing longwave radiation. In 1993 NESDIS and the Cooperative Institute for Research in the Atmosphere installed the first GOES-8 digital satellite display system at a NWS Office. In 1994 NESDIS and CIMSS implemented the highly acclaimed GOES-8 wind products operationally and continue to provide updates such as using rapid scan imagery and improving height information. The successes of the NESDIS CI program continue today with the formation of the Cooperative Institute for Oceanographic Satellite Studies in 2003 and strong interactions with the NOAA Cooperative Remote Sensing Science and Technology (CREST) Center, a NESDIS focused Cooperative Science Center formed in 2001 by NOAA/OAR. Transitions to operations continue to improve at NESDIS due to more interactions between CIs and CREST with STAR’s Operational Products Development Branch which works closely with Satellite Operations personnel to tailor research code and ideas to operational standards and use. The primary gap to fill according to SATB stakeholders is related to the early stages of R2O. Once a research concept is proven useful there are several existing opportunities to move it to operations (other NOAA Testbeds or transition programs). GOES-R Prototype The GOES-R Risk Reduction program is supportive of the SATB concept. In FY10 five new projects began that incorporate SATB concepts such as the use of multiple satellite sensors and/or in-situ and model data to conduct research/exploration of relevance to major societal issues. The PIs for the selected projects are being asked for their ideas to improve the existing SATB Development Plan (available at http://www.star.nesdis.noaa.gov/star/documents/corp/SATB/TowardSatelliteAlgorithmTestBed.pdf ) Components of Research to Operations (R2O) and associated phases. The Satellite Algorithm Test Bed encompasses the early stages of R2O and facilitates the orderly progression of science into later stage R2O processes. The Satellite Algorithm Test Bed (SATB) will consist of currently existing satellite research funding opportunities (examples: GOES-R Risk Reduction and GOES Improved Measurements and Product Assurance Program), currently existing infrastructure at STAR and STAR’s Cooperative Institutes plus new research opportunities to allow for multisensor and POES-related applied research, new infrastructure to connect and increase sharing of data, tools, algorithms and techniques, and enhanced coordination to standardize scientific review, selection and oversight of work conducted. Combined Geo/Leo High Latitude Atmospheric Motion Vectors Geostationary satellites provide Atmospheric Motion Vectors (AMV) equatorward of ~60° latitude; polar satellites provide AMVs poleward of ~70° latitude. Existing Research Opportunities While transitions into operations have been improving with time for current GOES and POES algorithms, the research that should be done prior to deciding to transition the algorithms is frequently missing. This is the case for POES algorithms because they have a robust Product Systems Development and Integration (“PSDI”) program to allow algorithms to move into operations but they do not have the Product Assurance Program that is currently in place for GOES to provide support for scientific studies. In addition, algorithm development funds are currently piecemeal from sources that are dominated by acquisition-centric programs such as POES, GOES, NPOESS, GOES-R, and NPP. This “stovepipe” structure does not facilitate consistency of products across platforms or a truly science-centric development of multi-instrument, multi-platform products. In order to obtain the highest quality products in the future the use of blended products (e.g. combinations of polar and geostationary data), merged products (composites reflecting the ‘best’ spatially, spectrally, and temporally integrated products) and ‘same science’ approaches (e.g. use the same physics techniques to process similar observations from geostationary, polar, aircraft, etc.) are necessary. Algorithms should be funded in a manner that provides optimum flexibility in approach, delivers consistent product processing, and improves product accuracy. Developing novel ways to fill this gap is the next step in providing complete wind coverage for NWP applications. Data from a variety of satellites are blended and used for AMV generation. The images are composites of the Geo (GOES, Meteosat-7 and -9, FY-2C, MTSAT-1R, Kalpana-1) and Leo satellites (NOAA-15 through NOAA-19, Metop-A, NASA’s Terra and Aqua). Investigators: Matthew Lazzara – PI (SSEC), Dave Santek (CIMSS), Chris Velden (CIMSS), Jeff Key (STAR), Jaime Daniels (STAR) Sample GOES-R Project funded in FY10 incorporating GEO/LEO satellite data streams for algorithm explorations. Slide courtesy of Matthew Lazzara/SSEC Science Challenges:Increasing Complexity and Number of Satellites Next Steps: Refine SATB development plan based upon GOES-R Prototype projects and other new information. Transition Path: Continue to work with partners to identify funding and/or leverage appropriately related activities to fully implement SATB. An SATB development plan was created with numerous input from stakeholders Existing Infrastructure The infrastructure for generating algorithms from several types of satellite data streams is currently available at the CIs, CREST and NESDIS. However each location collects data from a different set of satellites using different software and techniques. Additional infrastructure is needed to allow access to all the primary earth observing satellite data streams for research and development of algorithms by the CIs, CREST and NOAA. Tools to manipulate the data and techniques to move algorithms from research into operations (and from operations into research) also need to be made accessible and shared. A workshop was held at the National Satellite Operations Facility in Suitland, MD, on February 26 - 27, 2008 to define the coordinated activities for the SATB to be established, demonstrated, and maintained. Participants included those who have experience and insight to share regarding both the technical and programmatic aspects of operating test beds, developing, implementing, and supporting environmental satellite data products, and using or coordinating the use of satellite data. Existing Coordination Coordination for algorithm development typically is conducted at the working levels via Satellite Product Oversight Panels. In the 1990s there was a Satellite Research Council providing high-level coordination. However in the 2000s funding for satellite research became so scarce that the Satellite Research Council was disbanded. This coordination Is now primarily done less formally via the STAR management structure.