Seamless Synthetic Aperture Radar Archive for Interferometry Analysis

The NASA Advancing Collaborative Connections for Earth System Science (ACCESS) seamless synthetic aperture radar (SAR) archive (SSARA) project is a collaboration between UNAVCO, the Alaska Satellite Facility (ASF), the Jet Propulsion Laboratory (JPL), and OpenTopography at the San Diego Supercomputer Center (SDSC) to design and implement a seamless distributed access system for SAR data and derived interferometric SAR (InSAR) data products. A unified application programming interface (API) has been created to search the SAR archives at ASF and UNAVCO, 30 and 90-m SRTM DEM data available through OpenTopography, and tropospheric data from the NASA OSCAR project at JPL. The federated query service provides users a single access point to search for SAR granules, InSAR pairs, and corresponding DEM and tropospheric data products from the four archives, as well as the ability to search and download pre-processed InSAR products from ASF and UNAVCO. Figure 1. Diagram of SSARA workflow. Grey highlighted boxes show component development as part of the project, green outlines represent tasks already implemented and red outlines represent future implementation. * Corresponding author. This is useful to know for communication with the appropriate person in cases with more than one author. The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XL-1, 2014 ISPRS Technical Commission I Symposium, 17 – 20 November 2014, Denver, Colorado, USA This contribution has been peer-reviewed. doi:10.5194/isprsarchives-XL-1-65-2014 65


Objectives
Develop and implement a federated metadata query and data and data product download capability from distributed airborne (NASA UAVSAR) and spaceborne SAR archives at ASF and UNAVCO/WInSAR by enhancing existing web browser user interfaces and data access web service Applications Programming Interfaces (APIs).Define and make available new QC parameters and products that will enhance the usability of data and data products from these existing NASA-funded collections.A related goal is to develop standard formats for image products such as interferograms, coherence images, tropospheric effects, and terrain corrections.Implement a web services enabled InSAR terrain correction service using NASA SRTM data at SDSC.Enhance ASF SAR interferogram processing service to access distributed NSAR data collections, utilize NSAR terrain correction service, and generate enhanced QC products.Establish processed data products archive (InSAR interferograms and related enhanced QC products) at ASF and UNAVCO with standard formats and metadata.

Significance
InSAR-enabled Science.The InSAR technique provides an excellent remote sensing means of observing motions and deformation over broad areas.It is capable of detecting mm-level changes of the Earth's land and ice surfaces with dekameter-level spatial resolution at monthly or greater intervals.InSAR has proven to be a powerful tool to characterize large-scale deformation associated with active faults.It can resolve small-scale deformation features such as shallow creep, postseismic and interseismic deformation.It is an ideal method for measuring land subsidence and improving digital terrain models.InSAR complements ground-based measurements such as GPS that have higher temporal sampling (seconds to days) and absolute reference frame accuracy, but are considerably more sparsely distributed.When possible, InSAR and GPS are used in tandem such as with the NASA JPL QuakeSim and ARIA (Advanced Rapid Imaging and Analysis) modeling and analysis systems.
While scientific research is not directly proposed here, the technologies and infrastructure developed in this effort will support and benefit the diverse scientific goals of the InSAR communities that utilize UNAVCO and ASF InSAR archives.The science themes being investigated span solid Earth, atmospheric and cryospheric research and include for example: • The earthquake cycle throughout the world including the San Andreas Fault and Basin and Range • Volcanic activity including South America, Hawaii, Aleutians and Cascade/Yellowstone • Groundwater and coastal hazards throughout North America • Mountain building in South America and the Himalayas • Rifting in Iceland and East Africa • Motions and deformation of the cryosphere related to global change • Ionospheric and tropospheric effects • Anthropogenic changes from oil and water extraction • InSAR noise source mitigation and time series analysis.
UNAVCO manages InSAR support (data tasking, ordering, distribution and access) for the WInSAR consortium.The WInSAR consortium currently has 72 US universities and research institutions as full members and 24 non-US (adjunct) institutional members; the consortium goals include acquisition and distribution of SAR data to scientists doing various aspects of solid Earth, hydrological, cryospheric and human impacts research.Through the WInSAR Consortium UNAVCO has assembled references to published InSAR science through the past several years.The rate of InSAR based science being produced is dramatic, with at least 40 publications produced in one year, indicative of a highly productive research community utilizing the data in the UNAVCO, ASF, and other SAR Archives.http://www.unavco.org/pubs_reports/reports/annual/winsar/2010-2011-AnnualReport_WInSAR.pdfNASA Role in the Availability of SAR Data.Access and availability to SAR data is the greatest challenge for U.S. investigators using or developing techniques for scientific and hazards applications of InSAR.Until the NASA DESDynI mission is operational at the earliest in 2017, NASA will not have its own spaceborne SAR mission.All current and past InSAR satellites are operated by foreign space agencies.Unlike most data collected by NASA, which are freely available to anyone with an internet connection, all scientifically-relevant international satellite SAR data have or had significant costs and restrictions for acquisition, distributions, and use.On the horizon, the volume of openly available data will expand rapidly when the proposed NASA DESDynI mission is realized.Recognizing this gap in data availability, NASA has long worked to establish relationships and partnerships with other space agencies to facilitate data access at reduced cost and has funded ASF and the UNAVCO/WInSAR consortium (in partnership with NSF and USGS) to provide infrastructural resources for acquisition, archiving, and distribution of SAR data from foreign space agencies and the ongoing NASA airborne UAVSAR mission.NASA, ASF, UNAVCO and the WInSAR community, in partnership with the other agencies, have to date been successful in adapting to the constantly changing landscape of data availability and complex access rules and the ASF and UNAVCO/WInSAR archives contain a considerable volume of SAR data.
The largest proportion of data acquired by UNAVCO and ASF is from the European Space Agency's (ESA's) European Remote Sensing Satellite (ERS)-1 and ERS-2 and Envisat platforms.ERS-1 is retired and ERS-2 will be shut off in July, 2011.Envisat will be operated in its present 30-day orbit for the next two and a half years.ESA will launch the Sentinel-1a satellite in 2013 and data from this mission is expected to be more openly available.The NSAR team will collaborate closely with ESA to ensure that access Sentinel data are integrated into the federated data repository.Data collections also include the Japan Aerospace Exploration Agency's Japanese Earth Resources Satellite (JERS-1) and Advanced Land Observing Satellite (ALOS-1).Both are now non-operational and ALOS-2 is planned for 2013.NASA shared a joint agreement with Canadian Space Agency for RADARSAT-1 satellite data.There was no such agreement for RADARSAT-2 and higher commercial-only cost precludes wide academic use.The Radarsat Constellation Mission (RCM) is planned for 2014.WInSAR is planning to archive the German TerraSAR-X data from a few areas, and there is hope to get COSMO-SkyMed data from the Italian Space Agency.NASA's Airborne Synthetic Aperture (AIRSAR) and Uninhabited Aerial Vehicle Synthetic Aperture Radar (UAVSAR) are available through ASF.The NSAR project will work to streamline knowledge of and access to these data and help the user negotiate through complex data access and use requirements.
This proposal will enhance the existing NASA-funded SAR archives and services and provide new capability that will facilitate the use of new data to come from DESDynI.With DESDynI and other new spaceborne SAR missions comes the promise of not only large data volumes, but also expanded spatial coverage, more frequent (possibly weekly) scans, and even greater utility of InSAR for hazards research and mitigation.With this wealth of data comes the challenge of providing easy access, efficient transfer, and efficient use of SAR data and integration of results with other observations.The web services-enabled access and product generation tools proposed here form the foundation for enhancing these capabilities and streamlining the work flow.
Relevance to the NASA Earth Science Missions.This project is directly applicable to the Deformation, Ecosystem Structure, and Dynamics of Ice (DESDynl) mission recommended for "Tier 1" priority by the NRC Earth Science Decadal survey (pages 96-99).The proposed seamless SAR archive and enhanced InSAR capabilities will be able to incorporate future DESDynI SAR metadata thereby making data available to a large group of productive users by lowering search barriers and ensuring broader access.The NSAR project will address the following specific strategic goals or research objectives from the ROSES ACCESS Solicitation and the NASA Decadal Survey.This project addresses Improving Earth science user access to web services and service registries and Improving knowledge of NASA's Earth science data quality and production legacy.In particular, it will increase access to existing SAR data resulting in more effective data flow to the scientists who will more readily be able to find and retrieve data that are available for their use, and it will provide InSAR products to scientists who may not wish to process data themselves.
The NRC Decadal Survey sets out several Strategic Goals and subgoals.This proposal addresses strategic subgoal 3A of Strategic Goal 3: Develop a balanced overall program of science, exploration, and aeronautics consistent with the redirection of the human spaceflight program to focus on exploration.Subgoal 3A: Study Earth from space to advance scientific understanding and meet societal needs.SAR data from satellite and aircraft platforms provides information that is unique in the solid Earth sciences in several ways described elsewhere in this proposal.This project provides knowledge about available SAR data for Earth science research.The successful accomplishment of the goals of the proposed project will provide a foundation for mechanisms of discovery and access to DESDynI SAR data.These data address subgoal 3A, in particular providing data that is of extremely high value for Earth system research with respect to natural disasters and climate change.The partner organizations are well positioned to archive and distribute data from the DESDynI mission.The proposed work will enhance the usability of all SAR data from all missions including DESDynI.This project enables new information products to be delivered based on newly accessible data.
Relevance to Applications and Operations Community.The goal of the DESDynI mission is to serve both the research community, which conducts systematic science investigations that are usually not timecritical, and the operational applications community with clear requirements for low-latency data that places additional demands and requirements upon the data distribution system.These needs were discussed at the October, 2008 DESDynI Applications Workshop in Sacramento, CA.Primary areas of interest examined included Geohazard Assessment and Response including the risk from earthquakes, volcanoes, and landslides; Floods, Oceans, and Coastal Applications; Subsurface Reservoirs; and Forest and Ecosystems Management.Hazards applications such as these require timely access to data and products that can quickly be analyzed by decision makers to assess the likelihood of hazards and to assess damage and additional risk in the event of an emergency.The services we will leverage and enhance will not only facilitate these operational applications and scientific research using SAR data from existing and planned NASA air and space missions, but will also provide integrated access to other existing SAR data sets by enhancing advanced and established information technologies.
Broader Impacts for GEO and CEOS.The Group on Earth Observation (GEO) and Committee on Earth Observation Satellites (CEOS) have been working together on an initiative called the Geohazard Supersites.This project is bringing together SAR data, other satellite imagery, and in situ measurements for a set of Supersites where there is a great risk of geologic hazards or where major events have already struck such as the 2010 Haiti earthquake and 2011 Tohoku-oki, Japan earthquake.NASA is contributing their UAVSAR data for the Supersites in the Los Angeles region and in Hawaii, and ESA, JAXA and DLR are all contributing SAR data from their satellites for Supersites.NASA is supporting work that UNAVCO is performing on behalf of the GEO Geohazard Supersites to build a prototype interface to Virtual Archives of the raw data plus extremely popular web pages where many researchers share their preliminary results (http://supersites.earthobservations.org).

Relevance to ACCESS Program Goals. This proposal is relevant to the three principal ACCESS program goals:
Improvements in users' ability to efficiently discover, find, access, and readily utilize useful science content from NASA's increasingly large volumes of multimission, multi-instrument Earth science data.An outcome of the project will be simpler discovery and access to CEOS spaceborne and NASA UAVSAR airborne SAR metadata and data from web browser interfaces and scriptable web service enabled APIs.Users will also be able to access from distributed archives and easily produce useful science content from InSAR processing services and InSAR archives from the NASA-funded ASF and UNAVCO/WInSAR archives.Existing web services will be aligned and federated query and data downloading capability will be added to implement these new capabilities.A new web service for InSAR terrain corrections will be constructed to mosaic and subset NASA SRTM topographic data and integrate this into processing workflows.The tools and technologies for this proposal are TRL 7 or higher, as NSAR is reusing and leveraging software and services.NSAR will integrate these components for a unique purpose useful to the InSAR community.A critical aspect to the effective query screening and use of data and products is knowledge of data quality and noise sources.NSAR will review existing archive data quality parameters for SAR and resulting processed InSAR data and enhanced them providing greater data quality visibility and better utilization of NASA mission data and NASA-acquired CEOS SAR data.Enhanced quality factors will be created and distributed through the NSAR project-developed services.Additionally, tropospheric corrections from other services will be incorporated into the workflow to facilitate efficient evaluation and retrieval of possible data error sources from legacy and new process data and products.
Tools that improve and expand the accessibility and usability of NASA's Earth science observational data for the modeling and model analysis communities.NSAR will develop and implement standard product formats in coordinated with providers of tropospheric corrections (NASA OSCAR project using MODIS data and the Scripps/JPL NASA AIST project using ground-based GPS data, Moore et al., 2010) and users of new InSAR products from this proposed project (NASA JPL QuakeSim and ARIA project).QuakeSim will be able to use NSAR archived InSAR products and GPS time series and velocities from the current UNAVCO ACCESS GSAC project for advanced crustal deformation modeling and analysis for seismic hazards and geophysical research communities.Propagation of quality control information with InSAR products will enable QuakeSim and other advanced modeling and analysis services to improve their ability to estimate errors and evaluate noise contributions.

Other tools and technologies that enhance the accessibility and usability of Earth science data and extend the reach of NASA's Earth Science Division IT investments to new users and communities.
A key suggested action from the 2010 EarthScope Science plan (Williams, et al., 2010) was "Provide InSAR deformation products (interferograms and tools to facilitate time series analysis) for improved spatial resolution of the strain field.This appeal came from earth scientists desiring to integrate multiple data sources (geodetic, seismic and geologic) and who are not experts in InSAR processing.NSAR developments are specifically designed to enhance accessibility to a broader community of users by providing web accessible processing tools to generate interferograms and easily assess quality and possible error sources.

Background Relevant to the Project
Data and Information Technology Infrastructure.Currently UNAVCO and ASF have extensive collections of data (Table 1) paid for by US sponsors (NASA, NSF, and USGS).The intent of the sponsors is to make these data widely accessible for science.

Overview
The NSAR project approach will be to employ standards-compliant web services to provide uniform access to existing data archives at ASF, UNAVCO/WInSAR, and the GEO Supersites.The system architecture will be web services-based and will incorporate middleware to allow federation of queries/responses to/from UNAVCO and ASF by leveraging a modification of the Generic Service Layer (GSL), software previously developed as part of the ACCESS GSAC project.The GSL will also access interferogram product archives that will be established at both archives, the ASF archive for products generated with the ASF InSAR processing service, and UNAVCO/WInSAR archives for contributed PI solutions.As part of this effort, a standard product format will be created that will allow for consistent archiving and post-processing of higher-level InSAR products and related QC/QA products that will also be specified and implemented into the ASF InSAR service and the new JPL/Stanford ISCE next generation InSAR processing application.The new product format, tentatively using HDF5, will also be adapted to NSAR InSAR terrain correction service and existing NASA JPL OSCAR tropospheric correction service.The overall NSAR workflow, the GSL, the NSAR terrain correction service and other developments/enhancements for this project are detailed below.

NSAR Workflow
The proposed NSAR workflow is shown in Figure 2 where existing services are indicated in black and services that will be created, modified or enhanced are drawn in red.The SAR user, the large block on the right, will have several ways to access data and products from distributed archives at ASF and UNAVCO and also to learn of data that can be ordered from the agencies that might not already be in either archive (e.g.ESA).The heart of NSAR will be the federated SAR/InSAR repository middleware that will implement the Generic Service Layer.For data and product access, the user will be able to use the ASF or UNAVCO/WInSAR browser tools or alternatively their own or packaged NSAR query scripts using the NSAR API to make a federated query of the available archived SAR and processed interferograms from both archives.The user will then select data or products and download what they are authorized to access.
The NSAR API can be used as a data acquisition step prior to processing by the JPL-developed application ROI_pac (Repeat Orbit Interferometry Package) or ISCE and can be built into the processing workflow when the user is processing with downloadable software rather than the ASF InSAR processing service which will by necessity be streamlined for production processing options.Alternatively a user might go to the QuakeSim system to model a deformation event such as an earthquake using InSAR products archived in the NSAR and/or GPS data from the NASA ROSES ACCESS Geodesy Seamless Archive Centers.The QuakeSim "QuakeTables" currently holds only a limited set of interferograms that were generated by Paul Lundgren (JPL).NSAR will facilitate researcher and QuakeSim access to other interferograms archived at UNAVCO or ASF or even generate them using the ASF processing system.The NSAR will also provide new enhanced QC/QA products allowing for improved visibility into the quality of the original SAR data, the resulting interferograms and assessment and optional correction of the effects of terrain and troposphere.By adopting the GSL, the NSAR project will leverage this common infrastructure and also serve as a driver to evolve the GSL middleware to encompass new data and metadata models and new domains of use.Furthermore, with the support that the GSL provides for open standards such as OpenSearch and ATOM data casting, the SAR/InSAR repository will be able to integrate with the larger scale data discovery and access efforts that the ESDSWG and ESIP communities are working towards.
Additional details of the GSL are shown in Figure 3.The Federated GSL Repository can interact with multiple external repositories.On initialization it reads a repository descriptor from each of the external repositories.The descriptor contains a specification of what types of search metadata (query capabilities) a repository supports.The Federated Repository aggregates the search metadata to present a unified search interface to the end user.
On receipt of a search request, the Federated Repository, through the GSL, determines which external repositories need to be queried and forwards the queries to this set of external repositories in parallel.The user interaction with the Federated Repository is in terms of a set of standard vocabularies (e.g., file type).However, the base GSL layer has support for managing these standard vocabularies and provides mappings from the standard vocabularies to the internal values a particular repository may use.The search results are returned to the Federated Repository which then aggregates them and presents the results to the end user in the desired format.
For the NSAR project the GSL will be modified to access and handle SAR and InSAR data, products and metadata.GSL software will be installed over existing archive APIs to provide web service access.
Additionally, unlike space geodetic data that are openly and anonymously available, SAR data are typically only available to authenticated users due to non-NASA space agency restrictions.For this reason an NSAR user authentication capability will have to be added to the Federated Repository GSL and ASF and UNAVCO/WInSAR GSL services.
Figure 3. Details of the Federated Repository and Generic Service Layer (GSL) transactions.The incoming request might come from an NSAR-capable browser, API or other GSL web service.The result of the request could be a list of data (and associated metadata) for data and products at each archive repository, or it could be download information for browser or automated file access.

InSAR Product Archive and Processing Service
One of the primary products that can be produced with SAR data is an interferogram, a difference of a pair of SAR scenes from repeated satellite or aircraft radar scans that are typically separated from days to years apart.The resulting phase differences between the two SAR images is called Interferometric SAR, or InSAR, and provides an estimate of surface differences at the subcentimeter level.InSAR from a single pair of SAR images also contains direct topographic effects that must be removed.To provide this correction a terrain correction service will be established at SDSC's OpenTopography and the technology for this will be described later.
Currently the bulk of InSAR processing is done by individual PIs, most using ROI_pac, a NASA-funded software package developed by JPL/Caltech with contributions from the user community and distributed at no charge for non-commercial use.To help broaden the InSAR user community, UNAVCO supports an annual InSAR processing short course that is taught by ROI_pac PI Paul Rosen, Eric Fielding (also a CO-I on this NSAR proposal), and a third WInSAR community member.After several days of training users can generate InSAR images but proficiency can take years to develop.Among the obstacles that users must overcome are the variety of tools for accessing data and complex user authorization rules, software maintenance, the need to access other data (such as digital elevation models needed to perform terrain corrections), and the need to assess effects of varying troposphere conditions.Reducing these obstacles is the primary goal of the NSAR proposal.Products and processes generated through NSAR will be disseminated in part through the InSAR processing course and in the future with ISCE.
As part of NSAR, InSAR product archives will be established from the web processing service at ASF, and from PI contributed products at UNAVCO.These archives will be connected to the GSL for access and retrieval.The archives will follow the prototype developed by the NASA AIST QuakeSim project and associated QuakeTables database (http://quaketables.quakesim.org/insar.jsp).Key products are the phase unwrapped, geo-coded interferogram, metadata file, image rendering file, and kml.NSAR will evaluate the metadata and possible enhancements for QC/QA such as coherence files that will be implemented in the distributed archives.NSAR will work with QuakeSim PI Andrea Donnellan to ensure that NSAR product web services can easily be integrated into the QuakeSim workflow.UAVSAR InSAR products are generated by JPL and the QuakeSim UAVSAR InSAR database will be migrated to ASF as part of ongoing QuakeSim efforts.
The ASF URSA interface currently allows users to identify stacks of SAR granules suitable for interferometric processing, assess the perpendicular and temporal baseline distribution of a stack by interacting with the online baseline plot tool (Figure 4), and select pairs to order a number of beta interferometric products.In principle, data from all satellites in the ASF archive are suitable for InSAR, as long as the image pair adheres to some very basic rules.The data needs to be acquired by the same sensor, in the same beam mode, and with the same look direction.Users who order interferograms from ASF receive a beta product that has been processed using ROI_pac.ASF currently offers four InSAR product packages.Each package includes an interferogram, coherence image, and amplitude image geocoded to UTM projection and stored in GeoTIFF format (Figure 5).Optional packages include the complex master and slave data in ROI_pac format and/or the complex interferograms in ROI_pac format.Through its regular ASF DAAC contract funding, and complementary to this proposal, ASF plans to enhance the InSAR products and services by the first part of 2012.An InSAR browse product will be created for closest temporal neighbor or same-season pairs for each of the datasets in the ASF DAAC archive.The resultant phase and coherence products will be made available via a search and download mechanism as a quality measure for data selection.The search and order interface will be enhanced to exploit these new quality measures, along with the proposed tropospheric product.The goal of the interface changes is to assist in finding useful interferometry products more quickly and without the need for on-demand computationally expensive processing and sorting at the time of order.Modifications to the current interferometric processing chain are also planned, with the goal of creating a custom processing interface and algorithms that result in a phase-unwrapped terrain-corrected, geocoded GeoTIFF product with topography removed.This product would be especially useful for new users unfamiliar with interferometry, but familiar with GIS and other remote sensing datasets.

Enhanced Quality Control for SAR and InSAR
Part of this project will be to review and enhance QC/QA information for SAR data and resulting InSAR products.Two basic types of quality control can be defined: (1) QC for a single SAR acquisition, which can include basic checks on whether the data can be processed to a SAR image, the orbital baseline of that scene relative to other scenes, and information about the atmospheric conditions at the time of the acquisition that can affect InSAR measurements made with that SAR scene; and (2) QC for interferogram products, which is made from a pair of SAR scenes and can include InSAR coherence estimates and other quality factors.Interferogram products will inherit the quality control factors estimated for the individual scenes that make up the InSAR pair since those factors will also affect the quality of the interferograms.
The ASF processing system already provides calculations of the spatial distance between the orbit of a given SAR scene and scenes of possible InSAR pairs, the orbital baseline (Figure 4).For InSAR analysis, the perpendicular component of the orbital baseline (perpendicular to the radar line of sight), determines the sensitivity of an InSAR pair to errors in the topographic phase corrections and strongly affects the overall coherence of the interferogram.The NSAR system will provide orbital baseline calculations for SAR data in the WInSAR archive in addition to the ASF baseline metadata.
One of the largest sources of errors in InSAR measurements are variations in radar wave propagation through the atmosphere.For satellite InSAR, the radar signal passes through a large part of the ionosphere and through the entire troposphere.For airborne InSAR, only the troposphere is included in the radar path.Increases in the total electron content (TEC) of the ionosphere cause delayed propagation of radar waves and rotated polarization of the radar (Faraday rotation).These effects are both strongly dependent on the radar wavelength ( ), proportional to 2 , so the ionospheric effects are much stronger for the 24 cm L-band data (ALOS PALSAR and JERS-1) than for the shorter wavelengths, including 6 cm C-band (ERS, Envisat, Radarsat) and 3 cm X-band (TerraSAR-X and COSMO-SkyMed).The ASF already calculates an estimate of the average Faraday rotation for each ALOS PALSAR scene from a global model of the TEC.The global TEC maps do not provide information on the small-scale variations of TEC smaller than a typical SAR scene that could cause variable phase delay and Faraday rotation within the scene, but the small-scale variations are typically correlated with the average level.
In the troposphere, the propagation of radio waves is affected by the total mass of atmosphere along the path (often called the "dry" delay) and the integrated amount of water vapor along the path (usually called the "wet" delay).The total mass of the atmosphere is easily calculated from the surface pressure.The atmospheric water vapor content can be measured by satellite instruments, such as the MODIS instruments on Terra and Aqua (Li et al., 2005) or estimated from weather models.GPS ground station data processing for precise locations includes estimation of the tropospheric path delay due to both the dry and wet components, which is exactly the same delay that affects InSAR measurements (Li et al., 2006).The Scripps/JPL GPS AIST project is routinely producing these tropospheric delay products for the continuous GPS stations that they process.
We will build upon the web services provided by the JPL OSCAR (Online Services for Correcting Atmosphere in Radar) project (on which Fielding is Co-I) to provide the OSCAR estimates of radar propagation delay for each scene, both as a quick-look product that can be viewed to see the level and pattern of water vapor and as data files that can be used to apply the correction in InSAR processing.The OSCAR server is providing a set of RESTful web services that will be easily integrated into the NSAR system.We will calculate simple statistics from the tropospheric water vapor maps (such as the mean and root mean square variation of values within the scene) and provide this as a quality control factor for individual dates.
Interferogram products fundamentally measure the phase difference between two SAR scenes.The accuracy of ground deformation measurements from the InSAR phase depends on several factors, including: (1) the coherence of the radar return, (2) propagation of the radar through the atmosphere described above, (3) errors in the topographic correction, (4) orbit knowledge errors, and (5) phase unwrapping errors.We propose to provide quality estimates based on factors 1-3.Coherence is estimated in standard InSAR processing from the spatial correlation of the InSAR phase over small window (Fielding et al., 2005).When the correlation is very low, then the InSAR phase measurements are essentially random and contain no information, because the ground conditions have changed so completely that the radar return is incoherent.We will provide quality estimates for each interferogram from the coherence, including the percentage of the scene that is above a coherence threshold and the average coherence for the scene, in addition to providing the coherence image (e.g., Figure 5b) with the interferogram.

Figure 6abc.
The ALOS-PALSAR image in Figure 5 from the ASF processing service was not corrected for topographic effects.To facilitate terrain corrections, the proposed NSAR project will provide InSAR-ready topographic data through OpenTopography.Shown in 6a above is the terrain correction (EGM96 removed) via GMTSAR from NASA SRTM data.The terrain corrected differential interferogram unwrapped phase in 6b from the same ALOS-PALSAR pair was processed using ROI_pac.Red star shows epicenter of April 2010 Mw 7.2 earthquake.The apparent range change variation is 30 cm. 6c shows the zenith path delay difference from OSCAR MODIS zenith path delay maps.The path delay difference map shows no large gradient due to the troposphere in this case.The ASF URSA catalog reports very high values for the Faraday rotation in the ionosphere for these two scenes, which would be consistent with large ionospheric delays.The NSAR project will standardize product and corrections/QC formats and facilitate this type of product quality evaluation and access to products critical to the interpretation of interferograms for earth surface motions and deformation.
The atmospheric effects on the quality of an interferogram (e.g. Figure 6b) will be estimated by taking the difference between the atmospheric delay maps (Figure 6c) that come from the OSCAR server for the two scenes that make up the interferogram.This zenith path delay difference map is useful both for assessing the quality of the interferogram before correction and for correcting the atmospheric effect.The path delay difference is always smaller than the path delays of the two dates, but the values can be positive or negative depending on which scene has the greater delay.The standard deviation of the path delay difference map is a simple estimate of the variability of the path delay affecting the interferogram.

InSAR Terrain Correction Service
As noted in Figure 6, in order to construct an interferogram from a single pair of SAR images, it is necessary to first remove the direct topographic effects contained in each scene.This is derived from the perpendicular component of the baseline and the DEM source.To perform this terrain correction, access to appropriate global digital elevation data are required.Sandwell et al. (2011) provide digital elevation model (DEM) data through a simple web-interface (http://topex.ucsd.edu/gmtsar/demgen/)associated with the open source GMTSAR InSAR data processing toolkit.The GMTSAR DEM website provides access to NASA SRTM1 (inside USA) and SRTM3 (global) DEMs for areas of up to two by two degrees.The EGM96 gravity model is removed from the topography data to return ellipsoidal elevations as required for SAR processing.For the NSAR project, the SDSC OpenTopography team will leverage the existing GMTSAR backend for mosaicing and subsetting of DEM data to develop a high-performance Web service interface that will allow rapid access to DEMs for terrain correction for larger areas Working with the Sandwell group (also at UCSD), the SDSC team will migrate the back-end GMTSAR data mosaicing and subsetting tools to a high-performance server currently used for OpenTopography's lidar data processing system, This migration will provide improved performance and reliability for data access through the NSAR project.To enable access to this DEM processing system from the NSAR generic service layer, an Opal Toolkit-based (http://www.nbcr.net/software/opal/)Web service wrapper will be added to the GMTSAR backend.Opal is designed for deployment of scientific data analysis codes as web services and provides a simple SOAP API through which clients can interact with the service.Initial work on the terrain correction service will emphasize NASA SRTM1 and SRTM3 data.
The SDSC team will also explore incorporation of the ASTER GDEM and the USGS National Elevation Dataset (NED) products into the terrain correction service.

Management Plan
The development of the proposed NASA Seamless Synthetic Aperture Radar (SAR) Archive for Interferometry Analysis (NSAR) will take place over two years and involve investigators and collaborators from UNAVCO, ASF, JPL and SDSC.The tasks will include information technology development as well as science product evaluation.The participating institutions have strong ties to the InSAR community and collectively provide data and archiving services, software and scientific contributions to this community.The project team members have experience with distributed project development and currently or have in the past worked on similar IT projects such as GEON, NASA ROSES ACCESS (NASA LiDAR ACCESS SYSTEM project), NASA DAAC Technology Infusion, and EarthScope and WInSAR data acquisition and services.Annual face-to-face project team meetings and monthly teleconferences are planned and budgeted.A project wiki will be established to facilitate collaboration among the project partners.Standard project task management and cost/schedule controls and regular reporting will be implemented to ensure project time lines and goals are met.
Dr. Charles Meertens, UNAVCO, will perform the overall project management task.Dr. Jeff McWhirter, UNAVCO Senior Software Developer, will be project technical lead.Both are currently in the same roles in the current UNAVCO ROSES ACCESS LiDAR project (Meertens et al., 2010).Dr. McWhirter will also represent the NSAR project in the ESDSWG, a role he currently fills in the UNAVCO ROSES GSAC project.Co-PI Dr. Jeremy Nicoll (ASF) will direct the activities of the ASF system engineering effort and participate in the web services and QC/QA definition team.Co-PI Dr. Eric Fielding will serve as the primary science team partner and user interface evaluator.Fielding will also work closely with the NSAR team to help define the InSAR data product formats and assess data quality.Co-PI Dr. Chaitan Baru (SDSC) will direct the activities of the SDSC software engineer implementing the InSAR terrain correction service.

Special Matters
Discovery of Web Services.Cyberinfrastructure resources produced by this ACCESS project will be advertised through established outreach and information exchange avenues already utilized by the project partners.ASF, through its website, newsletter, advisory committee and user community and UNAVCO through its website, and the WInSAR consortium and UNAVCO list serves, provide established conduits to InSAR communities.UNAVCO and ASF also host short courses in InSAR processing and data access that help outreach to new users.UNAVCO and ASF participate in the Federation of Earth Science Information Partners (ESIP).NSAR services will be advertised on their ESIP Atom "Service Casting" feeds, will be advertised on the ESIP Available Services wiki, and will be registered on the NASA ECHO data registry, web services registry and Global Change Master Directory.Metadata about the data holdings at ASF and UNAVCO will be provided to ECHO through the same mechanism currently used by ASF for its standard holdings.We intend to follow the development and evolution of the EOSDIS Science Data Process Segment (SDPS) and, if appropriate, explore making NSAR services available as Tool Adapters in the SDPS workflow.Connections to the Open GIS Consortium's Catalog Services for Web will be made via OpenTopography.
Operations concept for continuance of the tools and services developed for the ACCESS Program.
The ASF and UNAVCO SAR archives have established longer-term funding through NASA and the NSF that will ensure sustainability of the proposed NASA ACCESS NSAR effort.UNAVCO and ASF have a history of developing and supporting data access tools and will incorporate these new tools into their software life cycle maintenance strategies.Management of SAR data is a key part of the UNAVCO Strategic plan and is expected to be fully integrated into UNAVCO core services in the near future.The issue of sustainability of the system is addressed by ensuring that the foundation of software and services will reside with these two archives responsible for the long term curation and distribution of these data.Also, all software and services will be documented and provided openly to encourage standardization and sustainability by use with a broader community.Future missions like DESDynI can be expected to follow suit and in the long term significant barriers to the use of data from non-standard formats will be removed and advantages of improved QC metrics will be realized without additional cost.

Figure 1 .
Figure 1.NASA UAVSAR interferogram showing deformation from the 2010 Mw 7.2 earthquake in Baja California, Mexico.Fringe contours of deformation are 11.9 cm each showing up to 80 cm of motion.Enhanced access to UAVSAR interferograms is one objective of our project.

Figure 2 .
Figure 2. The proposed NSAR workflow.Federated Repository Middleware As part of the ACCESS GSAC project a generic data repository middleware infrastructure (shown as the large red box in the center Figure 2) was developed called the Generic Service Layer or GSL.The GSL is intended to be domain agnostic and provides a number of basic repository services including: Metadata query capabilities; Extensible metadata model; Standard search forms; Output encoding (HTML, CSV, XML, RSS, ATOM), Standardized vocabularies; Repository aggregation; OpenSearch; and ATOM Data and Service casting.

Figure 4 .
Figure 4.An example of the URSA baseline plot.Users are provided with the capability to observe the perpendicular and temporal baseline distribution of an InSAR stack based on the selection of a desired master, filter granules, and select granules to order.

Figure 5abcd .
Figure 5abcd.ALOS-PALSAR amplitude image, coherence image, interferogram, and interferogram overlaid on the amplitude of the master image (left to right).Master image: acquired 2006-11-05.Slave image acquired 2008-11-10 over Baja, California, Mexico.The copyright for the scenes used to create this image (and those in Figure 6) is held by the Japan Aerospace Exploration Agency/Ministry of Economy, Trade and Industry.

Table 1 . UNAVCO and ASF Data Collections.
The WInSAR archive was originally constructed to handle a limited ERS-1 and ERS-2 satellite data collection with relatively small data volumes.Current total holdings include approximately 12,600 SAR frames, and the collection has expanded to include Envisat.UNAVCO also operates the EarthScope SAR Archive, which holds approximately 106,500 frames of ERS-1, ERS-2, Envisat, and Radarsat-1 data.UNAVCO and WInSAR worked with ESA to promote and develop the GEO Supersites data sharing concept and to prototype infrastructure for SAR data.UNAVCO orders the data from ESA and provides access through the Supersites website (http://supersites.earthobservations.org).Data are distributed through Level3 cloud computing infrastructure.Detected products from the ASF Datapool.The recently created Datapool contains 500 Terabytes of online readily-downloadable data.Those datasets useful for interferometry can be searched publicly, including browse products for selected datasets, and kml files of locations for all datasets, but require user authentication to access the data themselves.This authorization is through on-line initiated proposal submission services, coordinated by ASF, with access granted through NASA.The ASF archive is accessed through both a web GUI and scriptable mechanisms that directly query an Oracle database.Geographic and direct data searches of the ASF archive are available to all potential users through the ASF on-line web interface URSA, for approved users through scripted index queries and data pulls, and through a multi-format accessible RESTful API.The ASF API provides direct user access to the ASF Datapool through a simple design and implementation, allowing for fast user authentication, database queries through URSA and command line interfaces, multi-format data returns and downloads via http access.The primary metadata of the ASF Datapool, accessed through minimal spatial and textual parameters, identifies quality data returns of the API in URSA, csv, kml, or Metalink (xml) format; all formats have direct download access to the ASF Datapool for approved authenticated users, with bulk download capabilities through URSA and Metalink formats.Recent architecture improvements have allowed scoping for web services that are under development with anticipated implementation in late 2011.
The WInSAR, EarthScope, and GEO Supersites archive holdings can be searched through a php web GUI frontend with a mysql backend; spatial selections can be made through GoogleMaps and additional metadata criteria can be applied.Users can request Google Earth-ready kml files to be generated in response to searches.The architecture currently does not utilize web services.To download data WInSAR data users must be affiliated with a consortium member institution and registered with UNAVCO, and must sign data use agreements.Users of EarthScope data must become registered by submitting a "Minicat" proposal and must sign data use agreements.Authentication is required to gain access to data.The WInSAR system also offers a mechanism for Consortium members to request ordering of particular archived scenes for ERS and Envisat data that are not yet in WInSAR holdings.During 2010, the European Space Agency revised their policy for data use for scientific purposes, eventually waiving data acquisition (tasking) fees and costs associated with ordering archived data.Early in 2011 the WInSAR Executive Committee voted to redirect unused data purchase budget towards infrastructure investments, to include storage and processing hardware upgrades, development of RESTful web services, an API, database and web upgrades, and expansion of capacity to handle additional SAR platform data types.The sponsors approved this plan for utilization of the budget and these infrastructure upgrades are currently underway.Alaska Satellite Facility Archives.The Alaska Satellite Facility (ASF) of the Geophysical Institute (GI) at the University of Alaska Fairbanks, downlinks, archives, and distributes satellite-based data to scientific users around the world (Table1).ASF's NASA-sponsored Distributed Active Archive Center (DAAC) specializes in SAR data collection, processing, archiving, and distribution.These data are processed on-demand or can be downloaded as a standardized Level 0, Level 1 Complex, or Level 1