OPERATIONAL USE OF SAR-DERIVED OCEAN PRODUCTS ARE WE THERE YET

OPERATIONAL USE OF SAR-DERIVED OCEAN PRODUCTS:

ARE WE THERE YET?

Pablo Clemente-Colón1, William G. Pichel, Douglas Lamb2, Michael Van Woert1,2, Frank. M. Monaldo3, Donald R. Thompson3, and Christopher C. Wackerman4

1NOAA/NESDIS/ORA

5200 Auth Road, Camp Springs, MD 20746, USA

2National Ice Center

4251 Suitland Road, Suitland, MD 20395, USA

3JHU/APL

Laurel, MD, USA

4General Dynamics Advance Information Systems

Ann Arbor, MI, USA

ABSTRACT

Synthetic Aperture Radar (SAR) data from the Canadian RADARSAT-1 satellite is used operationally by the U.S. National Ice Center (NIC) in the production of sea ice analysis charts. The quasi-operational production of other SAR products has also been demonstrated by the National Oceanic and Atmospheric Administration (NOAA.) National Environmental Satellite, Data, and Information Service, (NESDIS). Since 1998, NESDIS has supported the development of SAR oceanographic and hydrological applications. In particular, the development of automated algorithms to determine ocean surface wind speed and direction, the location of the sea ice edge, vessel locations, and the presence of slicks has been has been a major part of the NESDIS Alaska Synthetic Aperture Radar (SAR) Demonstration Project (AKDEMO). As an example, an algorithm developed for NOAA by the Johns Hopkins Applied Physics Laboratory (JHU/APL) uses model wind directions and a modified CMOD-4 scatterometer wind algorithm to calculate high-resolution (<1 km) wind-speed estimates from RADARSAT-1 data. Similar algorithm and application development has also occurred in Canada and in a number of European countries. Still, the routine use of automated SAR-derived environmental products is very limited and far from established at national operational agencies, both in the U.S. and abroad. It appears that the main constraint to the integration of automated SAR-derived into truly operational activities is the routine availability of the SAR data themselves. Efforts being lead by the Nansen Environmental and Remote Sensing Center (NERSC) and NOAA partners are underway to develop an international consensus on the best tools for operational use of present and future SAR data streams. The approach includes the selection and or merging of the most suitable algorithms for efficient operational processing leading to robust products that can be incorporated into national environmental monitoring programs. SAR data streams under consideration include ENVISAT ASAR as well as upcoming ALOS PALSAR and RADARSAT-2 data.

1. OPERATIONAL USE OF SAR AT THE NATIONAL ICE CENTER

Synthetic Aperture Radar (SAR) satellite data are used by the U.S. National Ice Center (NIC), a joint agency formed by the U.S. NAVY, the National Oceanic and Atmospheric Administration (NOAA), and the U.S. Coast Guard (USCG), to classify ice types and monitor ice conditions mainly in support of maritime traffic in the Great Lakes, the Arctic and the Bering Sea. The use of SAR by NIC is the only true operational application of these data in the U.S. Most of the spaceborne SAR data used in the U.S. are acquired from the Canadian RADARSAT-1 satellite at the University of Alaska Fairbanks Alaska SAR Facility (ASF). RADARSAT-1 C-band (5.3 GHz, 5.6 cm) SAR data is made available to NOAA and NIC as part of a U.S. allocation agreement between the Canadian Space Agency (CSA), NOAA, and the National Aeronautics and Space Administration (NASA). Although RADARSAT-1 has passed its design life, there are no present commitments for RADARSAT data continuity and no U.S. spaceborne SAR missions planned in the near future. An agreement between the National Space Development Agency of Japan (NASDA), NOAA, and ASF on the establishment of an Advanced Land Observing Satellite (ALOS) North America Node at ASF, may provide some of the required coverage with the Phased Array type L-band SAR (PALSAR). A list of past, present, and future satellite SAR missions is shown in Table1.

The station mask at ASF allows for near real-time access to SAR data over the Beaufort Sea, Chukchi Sea, most of the Bering Sea, and the Northern Gulf of Alaska. Additional Arctic SAR data is received from acquisition stations at Gatineau, Canada, Troms?, Norway, and West Freugh, Scotland. These stations provide North America and near complete Arctic coverage using mostly ScanSAR Wide mode data (450-500 km swath and 100-200 m resolution). RADARSAT-1 can in fact image every point on the earth’s surface north of 65ON at least once per day. Near real-time RADARSAT data acquired at the Gatineau station are also available for the Great Lakes region during the winter season as part of an arrangement between the Canadian Ice Service (CIS) and NIC. Other SAR acquisitions of the U.S. northeast region within the Gatineau station mask are done on a non real-time basis with delays in delivery to users of the order of months unless purchased at commercial rates. Collection of data over regions outside station masks may also be requested using the onboard recorder.

SAR imagery is essential to the operational ice analysis provided by NIC, which customers demand detail and reliable information for critical operations. Given its all-day and all-weather capabilities, satellite SAR imagery is the data source of choice for NIC analysts. In fact, SAR is the primary data source used when available. For example, SAR data is used in about 23% of NIC operational Arctic ice products analyzed. In addition to sea ice mapping and characterization, the data is use to monitor ice shelves breaking and iceberg shedding in both the Arctic and Antarctic regions. Most of the operational NIC products depend heavily on an analyst interpretation rather that on the use of automated algorithms. SAR data interpretation is done in synergy with additional passive and active microwave, IR, and visible satellite observations of lower spatial resolution but that can provide better temporal coverage. In fact, the high-resolution SAR data is used in the validation and tuning of automated products from scatterometer and passive microwave sensors.

2. PRE-OPERATIONAL SAR DEMONSTRATION AT NOAA/NESDIS

NESDIS is the line office of the U.S. Department of Commerce NOAA responsible for managing the U.S. civil operational remote-sensing satellite systems, as well as global databases for meteorology, oceanography, solid-earth geophysics, and solar-terrestrial sciences. In addition to the operation of geostationary and polar satellites, NESDIS acquires, archives, and distributes data and geophysical products from other non-NOAA and non-U.S. satellite systems. This includes RADARSAT-1 and to a lesser extent, European Space Agency' (ESA) ERS-2 SAR data. The focal point for SAR oceanic and atmospheric research and research support is the Satellite Oceanography Division (SOD) under the NESDIS Office of Research and Applications (ORA).

A pre-operational demonstration of SAR cryospheric and oceanic high-resolution products was developed for the Alaska region to test the use of SAR by other operational agencies, alone or in synergy with other remote-sensing and in situ environmental data, on a near-real-time basis (Pichel and Clemente-Colón, 2000). The project, also known as the Alaska SAR Demonstration (AKDEMO), uses RADARSAT-1 ScanSAR Wide

B and Standard Mode (100 km swath and 25 m resolution) imagery of coastal Alaska, the Bering Sea, Chukchi Sea, Beaufort Sea, and Northern Gulf of Alaska to produce and provide a number of derived products to a select group of operational users in the region. The ultimate objective of this effort is to develop algorithms and data analysis techniques that will allow all NOAA CoastWatch Regional Sites to make operational use of future SAR data as part of their normal research, advisory, warning, educational, and regulatory activities.

Research and development under the demonstration support identified NOAA environmental parameter requirements. Table 2 lists Environmental Data Records (EDRs) that can be addressed by the use of high-resolution SAR observations. In particular, the AKDEMO system has three main products: (1) the SAR imagery itself, (2) ocean vessel positions from SAR, and (3) ocean surface winds from SAR. Near real time availability of SAR imagery can provide users with a unique indication of the location of the ice edge, which is critical for some Alaskan fisheries. Automatic detection of SAR hard targets can be useful for fisheries management (ship detection) or iceberg monitoring. Automated algorithms for ice edge detection, based on a supervised classification, and for hard target detection have been implemented. SAR imagery is also been used to routinely monitor river ice breakup, jamming, and the development of associated flooding conditions.The most mature SAR application is the routine production of SAR-derived wind maps (Thompson and Beal,2000). A modified scatterometer geophysical model function that takes into account the horizontal polarization of RADARSAT-1 SAR is used in conjunction with prescribed wind directions to derive SAR ocean surface wind speeds. The large amount of SAR data available under the U.S. allocation agreement has allowed for the accuracy of these winds to be well validated against NOAA buoy winds and NASA’s QuikSCAT scatterometer wind vectors collected within a few-hour time window. Statistics against both data types indicate a robust retrieval of SAR ocean surface winds with rms errors of around 1.3 m/s (Monaldo et al., 2003). Since SAR wind speed inversion needs a-priori knowledge of wind direction, wind direction information from forecast models or from near-coincident ancillary observations such as those from scatterometers or buoys is required. An additional algorithm to automatically obtain wind directions from linear features associated wind wind-aligned phenomena such as wind rows or organized large eddies observed on a SAR scene is also available. Development of a strategy to combine into the wind speed algorithm the best wind direction information obtainable for a particular SAR scene, i.e., from atmospheric models, scatterometers observations, or features in the SAR image itself, is underway. In addition,investigation on the polarization effects on the SAR wind algorithms is been conducted as ENVISAT Advanced SAR (ASAR) polarimetric data are made available. Besides wind speed, the ability of SAR to capture mesoscale and small-scale atmospheric circulation patterns associated with boundary layer dynamics is being used as an important new tool by meteorological scientists and weather forecasters.Satellite System

Country Operational Dates (month and year) SAR Band(s) Seasat SAR

USA 6/78 - 10/78 L SIR-A (Shuttle)

USA 11/81 - 11/81 L SIR-B (Shuttle)

USA 10/84 - 10/84 L KOSMOS 1870

Russia 7/87 - 7/89 S ALMAZ-1

Russia 3/91 - 9/92 S ERS-1

Europe 7/91 - 3/00 C JERS-1

Japan 2/92 - 11/98 L SIR-C/XSAR (Shuttle)

USA/Germany 4/94- 4/94 and 10/94-10/94 L,C,X ERS-2

Europe 4/95 - present C RADARSAT-1 Canada 11/95 - present C SRTM (Shuttle) USA 2/00 - 2/00 C,X ENVISAT ASAR Europe 3/02 - present C ALOS PALSAR Japan 2004 L RADARSAT-2 Canada 2005 C TerraSar-X1 Germany 2005 X TerraSar -L1 (?) Germany 2005 L COSMO/SkyMed Italy 2005 X SAOCOM Argentina 2006 L RADARSAT-3 (?) Canada 2007 C NASA InSAR (?)

USA 2009 L NPOESS Ocean

Observer

Operational

SAR (?) USA 2012 L,C or L,X

Table 1. Past, present, and future spaceborne SAR missions available for environmental observation (presently available systems are highlighted).

Table 2. Identified SAR-Related Environmental Data Requirements (EDRs) in the Integrated Operational Requirements Document (IORD), the Ocean Observer User Requirements Document (OOURD), and the NESDIS/ORA Satellite Oceanography Division (SOD).

Routine NOAA Products NESDIS Research Products

SAR patterns can provide important information on physical processes typically not captured by models, in-situ observations, or other spaceborne sensors. The unique ability of SAR to measure the high-resolution spatial structure of coastal wind fields and to capture associated atmospheric boundary layer processes is demonstrated by the images in Figure 1. These SAR wind maps show conditions around the Alaska Peninsula during northerly (1a) and southerly (1b) wind events and reveal the result of the flow interaction with the topography and the development of instabilities in the form of atmospheric internal gravity waves on the leeward side over the Bristol Bay and over the Northern Gulf of Alaska, respectively. The wind speed scale on these products goes from 0 to 25 m/s. The vectors indicate the model direction used in the retrieval and are coded to the model wind speed using the same scale. The high spatial resolution of the SAR data allows for the imaging of wind variability conditions right up to the coast, including embayment areas and relatively small lakes. SAR winds are indeed beginning to be considered as an exceptional tuning and validation tool for high-resolution (2 km) atmospheric models that include realistic topography. The potential use of SAR winds for assimilation into high-resolution coastal ocean models is also under study. Coastal SAR winds can in fact be of tremendous use in environmentally sensitive areas where buoy or other conventional observations are difficult or impossible.

Studies of mesoscale circulation systems such as polar low storms and hurricanes are also been conducted using SAR imagery. In particular, SAR offers an effective way of centering the location of these systems at the ocean surface. SAR imaging of storm systems may also include patterns of convective cell activity, precipitation, cloud ice, and even storm-induced ocean swell. Other ocean features readily imaged by SAR and under active research include ocean fronts and mesoscale circulation, river plume outflow and coastal interaction, oceanic internal gravity waves, upwelling, biological activity, and near shore bathymetric features. SAR data are also being used routinely to detect and monitor accidental oil spills, oil seepage from natural sources or submerged structures, and illegal ocean dumping activities. An example of SAR detection of some of these features is shown in Figure 2.

3. VISION FOR AN OPERATIONAL SYSTEM DEVELOPMENT AND

IMPLEMENTATION

Within NESDIS/ORA, a Sea Surface Roughness (SSR) Science Team has been established. Its primary goal is the development of an integrated end-to-end SAR ocean products system for operational generation of the sea-surface roughness products specified in the National Polar-orbiting Operational Environmental Satellite System (NPOESS) Integrated Operational Requirements Document (IORD) II and the Ocean Observer User Requirements Document (OO URD). SSR Team mission-related activities include 1) requirements development, 2) applications research, 3) establishment of collaboration with academic, Government, and commercial partners, 4) product development and validation, 5) user base development -including training, 6) applications demonstration, and 7) operational code development, documentation and implementation. Except for an operational implementation of SAR ocean products outside of operational sea ice charting at NIC, all of these areas are active and well underway. Present SSR salient activities include continuation of the AKDEMO, expansion into the Gulf of Mexico (GOMEX) using historical SAR data, participation in the NASA GhostNet project to detect ocean fronts, and the publication of comprehensive SAR interpretation manuals. The OO URD was part of an Ocean Observer Study conducted to define and develop cost estimates for U.S. operational L & C-Band Interferometric SAR as well as a Wide Swath & Delayed Doppler Altimeter to meet U.S. ocean data requirements.

In order to assess the operational readiness of state of the art algorithms for winds, waves, current features, and sea ice, a Workshop on SAR Coastal and Marine Applications was conducted in collaboration with the Nansen Environmental and Remote Sensing Center (NERSC) and the Norwegian Space Agency (NSA) on September 8-12, 2003 in Longyearbyen, Svalbard, Norway. International SAR investigators, space agency representatives, and developers met to discuss the status of SAR ocean algorithms and the perspectives for future SAR data availability (Svalbard, 2003). Our proposed strategy for an Operational SAR Products System Development is based on the selection of the best/mature algorithms and approaches identified at Svalbard. Additional collaboration with Norwegians, Canadians, and ASF is envisioned to refine automated versions of these algorithms. NESDIS will continue wind product development to merge SAR winds with scatterometer and passive microwave polarimeter winds for validation and improvement of direction information. Extension of existing algorithms to ENVISAT ASAR & ALOS PALSAR is also planned. Funding has already been identified to develop a prototype of a portable automated operational SAR wind system for deployment at the University of Miami Rosenstiel School of Marine and Atmospheric Science (RSMAS) Center for Southeastern Tropical Advanced Remote Sensing (CSTARS) satellite acquisition station.

Figure 1. Examples of AKDEMO pre-operational RADARSAT-1 SAR-derived wind maps of the Alaska

Peninsula region showing conditions on (1a) 25 January, 2002 04:28 UTC and on (1b) 15 November,

2000 16:50 UTC.

associated slicks

slick pattern

Figure 2. Example of multiple ocean features detected over the Delaware Bay mouth region by RADARSAT-1Standard Mode SAR on 13August,199811:15UTC.?

?Svalbard

?Gatineau

Fairbanks RADARSAT-1/2

ENVISAT

?Miami

ALOS Figure 3. Concept for initial International SAR Ocean Products System acquisition and

processing stations and satellite missions.

Based on the CSTARS wind system and Alaska SAR Demonstration experience, we propose to work closely with several international partners in the development and implementation of a fully automated operational SAR ocean products system to be deployed at selected acquisition stations that provide routine access to SAR. The ultimate goal is the efficient, automated, and routine real time generation of ocean products from each acquired pass that can be made readily available to users without any user need to have access to the original SAR datasets.

4. CONCLUSIONS AND FUTURE PLANS

The NIC and NOAA/NESDIS conduct and support SAR operational, pre-operational, and research activities. SAR data are operationally used to interactively produce sea ice analysis charts. Our research and pre-operational experience indicate that additional operational oceanographic and meteorological use of SAR data can routinely provide critical information not readily available through any other means. Spaceborne SAR can provide high-resolution cryospheric and ocean surface observations suitable for mesoscale as well as high-resolution coastal applications. In fact, significant operational implementation and use of SAR products would be commonplace if increased access and availability to the data were achieved. So, are we there yet? Well, we are almost there if we could only secure long-term operational access to SAR data. Data from the upcoming new generation of SAR satellite systems should provide some continuity and widen the present SAR data streams. New sensors will provide multi-band and multi-polarization data, which are expected to allow for better discrimination of SAR-detected features. Multi-satellite SAR data availability will also provide for much shorter revisit periods over areas of interest.

The RADARSAT-1 satellite is presently running well passed its design lifetime. We are looking beyond RADARSAT-1 for securing continuity of SAR data. The ENVISAT satellite was launched by ESA last year. We have ESA-approved data-only proposals in response to an ENVISAT Announcement of Opportunity program that provides limited access to ASAR data but to date there is no agreement in place between NOAA and ESA for the significantly greater data access that we desire. With the launch of ALOS in mid-2004 and the establishment of the North America Data Node at ASF, a realistic opportunity for operational implementation will open. Still, additional U.S. Government negotiated agreements with the foreign space agencies (ESA, CSA-RSI, CONAE-ASI) for routine access to SAR data should continue to be encouraged. At the same time, we will continue our efforts to gain routine access to at least SAR data products through the implementation of local real time ocean products processing systems at acquisitions stations. Other potential avenues for data access include straight data buys (particularly if pricing levels fall to or below the $100 USD per scene mark) and participation in cross-agency data requirements initiatives and data exchange agreements. Additional data exchange agreements as part of international sea ice centers initiatives such as the North America Ice Service (NAIS) or the International Ice Charting Working Group (IICWG) should also be pursued.

Further down the road, we look forward to the success of present efforts being undertaken by NASA to place an initiative in the federal budget process for a SAR mission to be flown by 2009. We also welcome indications from the Integrated Program Office (IPO) that it may be interested in adding operational SAR capabilities to the NPOESS series in a fourth satellite by 2012. Even if no environmental U.S. SAR satellite is launched, the ample bandwidth, acquisition timeliness, and redundancy that will be afforded by the NPOESS ground segment may provide a very attractive incentive for foreign and/or non-government satellite operators to reach satisfactory SAR data access agreements with NOAA and IPO in the future…keep tune!

5. BIBLIOGRAPHIC REFERENCES

Monaldo, F.M., D.R. Thompson, W.G. Pichel, and P. Clemente-Colón, (2003) A systematic comparison of QuikSCAT and SAR measured wind speeds, in print in IEEE Trans. Geosci. and Remote Sensing, TGRS-00347-2002.R1.

Pichel, W.G. and P. Clemente-Colón, (2000) NOAA CoastWatch SAR applications and demonstration, The Johns Hopkins Univ. Tech. Dig., vol. 21, no. 1, pp 49–57.

Thompson, D.R. and R.C. Beal, (2000) Mapping high-resolution wind fields using synthetic aperture radar, The Johns Hopkins Univ. Tech. Dig., vol. 21, no. 1, pp 58–67.

Svalbard, (2003) SAR Workshop on Coastal and Marine Applications, September 8-12, Svalbard, Norway https://www.wendangku.net/doc/7815298083.html,/meetings/svalbard_2003/