U.S. Government Open Internet Access to Sub-meter Satellite Data Christopher S.R. Neigha*, Jeffery G. Maseka, and Jaime E. Nickesonb,c aHydrospheric & Biospheric Sciences Laboratory, Code 618, NASA Goddard Space Flight Center, Greenbelt, Maryland 20771, United States bTerrestrial Information Systems Laboratory, Code 619, NASA Goddard Space Flight Center, Greenbelt, Maryland 20771, United States cSigma Space Corporation, 4600 Forbes Blvd, Lanham MD, 20706, United States 1 The National Geospatial-Intelligence Agency (NGA) has contracted United States 2 commercial remote sensing companies GeoEye and Digital Globe to provide very high 3 resolution commercial quality satellite imagery to federal/state government agencies and those 4 projects/people who support government interests. Under NextView contract terms, those 5 engaged in official government programs/projects can gain online access to NGA’s vast global 6 archive. Additionally, data from vendor’s archives of IKONOS-2 (IK-2), OrbView-3 (OB-3), 7 GeoEye-1 (GE-1), QuickBird-1 (QB-1), WorldView-1 (WV-1), and WorldView-2 (WV-2), 8 sensors can also be requested under these agreements. We report here the current extent of this 9 archive, how to gain access, and the applications of these data by Earth science investigators to 10 improve discoverability and community use of these data. 11 Satellite commercial quality imagery (CQI) at very high resolution (< 1 m) (here after 12 referred to as CQI) over the past decade has become an important data source to U.S. federal, 13 state, and local governments for many different purposes. Near global wall-to-wall sub-meter 14 coverage is now available when combining all the archives of U.S. CQI sensors. A coordinated 15 effort was needed to reduce and/or remove image acquisition costs from duplication of requests 16 made by multiple government agencies. The National Geospatial Intelligence Agency (NGA) 17 has been appointed to acquire and archive data from vendors to eliminate duplication costs 1 18 between government organizations. NGA has developed a system to request, archive, and 19 distribute CQI data to all federal agencies. NGA assists all federal branches, departments, 20 agencies and offices to acquire and use CQI at no cost to the supported organization and has 21 developed a series of contracts with GeoEye and DigitalGlobe. OB-3 is no longer operational, 22 but it collected data from 2003 – 2007 (> 180,000 images) and is currently available for free 23 through the United States Geological Survey (USGS) EarthExplorer (earthexplorer.usgs.gov) 24 and will not be discussed in detail. The first contract between NGA and commercial vendors 25 was ClearView which began in 2003, followed by NextView from 2007 - 2010, and currently 26 EnhancedView from 2012 - 2018. These contracts have provided the ability for U.S. 27 government to investigate changes in the Earth’s surface at sub-meter resolution through a 28 negotiated bulk purchase of data. 29 The rapid growth of free global CQI data has been slow to disseminate to NASA Earth 30 Science community and programs such as the Land-Cover Land-Use Change (LCLUC) program, 31 which sees potential benefit from unprecedented access. This article evolved from a workshop 32 held on February 23rd, 2012 between representatives from NGA, NASA, and NASA LCLUC 33 Scientists discussion on how to extend this resource to a broader license approved community. 34 Many investigators are unaware of NGA’s archive availability or find it difficult to access CQI 35 data from NGA. Results of studies, both quality and breadth, could be improved with CQI data 36 by combining them with other moderate to coarse resolution passive optical Earth observation 37 remote sensing satellites, or with RADAR or LiDAR instruments to better understand Earth 38 system dynamics at the scale of human activities. We provide the evolution of this effort, a 39 guide for qualified user access, and describe current to potential use of these data in earth 40 science. 2 41 Who Can Access Data? 42 The current NextView license agreement states that CQI data can be used by all 43 branches, departments and offices of the U.S. Government. With appropriate approval and 44 acknowledgement from NGA, data can also be shared with non-governmental organizations 45 (NGOs), state and local governments, intergovernmental agencies, as well as universities and 46 foreign governments if the use is in support of U.S. government interests when approved by an 47 official legal representative at NGA. Users of CQI must store it offline to ensure it is not openly 48 shared or distributed. All use of CQI must have the appropriate licensing acknowledgements 49 displayed, for example “2012 GeoEYE NextView”. 50 NGA’s online interface to access CQI data is called the web based access and retrieval 51 portal (WARP). WARP provides internet access to NGA’s Unclassified St. Louis Information 52 Library (USTIL). This library is a subset of the vendor archive from prior government agency 53 requests. To gain WARP access users must have a .gov email address and have public key 54 infrastructure (PKI) that allows secure communication on an insecure public network. Users can 55 register for an account via the WARP website (https://warp.nga.mil), and users must have an ftp 56 server for data from WARP to be pushed to from NGA. 57 Data in WARP are provided from vendors in National Imagery Transit Format (NITF), a 58 standard Department of Defense (DoD) format. NITF data are stored in compressed format and 59 metadata of sensor/solar/target/geometry information is imbedded within layers of the file. Most 60 image processing software packages can read NITF at no additional cost to the user, although 61 freely available open source tools from GDAL (Geospatial Data Abstraction Library) can be 62 used to convert NITF to more commonly used geospatial tagged image format (Geotiff). 3 63 Additional imagery collected by the vendors archive not available in WARP, can be 64 requested through an online USGS interface Commercial Remote Sensing Space Policy 65 (CRSSP) Imagery Derived Requirements (CIDR) tool. CIDR registration and requests can be 66 submitted via the CIDR website (https://cidr.cr.usgs.gov/). A form must be provided that is 67 subject for approval, including project title description and justification. The USGS currently 68 acts as a conduit for civilian agency data requests through NGA’s EnhancedView contract. 69 More information about CIDR can be found on the website. 70 These data support many different U.S. agencies, although immediate access to WARP is 71 limited to those with .gov email addresses. Other users include federally funded scientists from 72 Universities, NGO’s, state and local governments who do not have a federal email account and 73 have not been granted special access. NASA is currently exploring options to support NASA’s 74 Land-Cover Land-Use Change, Biodiversity, and Cryosphere science communities providing 75 access the commercial archive data for its investigators. Data for NASA program scientists are 76 coordinated to ease access to WARP/CIDR and are placed on a secure NASA server for 77 download. 78 NGA WARP Data Volume 79 Density of coverage varies by region, with multi-date time series coverage typically 80 limited to urban areas, or areas of long-term interest from customers of CQI data. All sensors 81 have the ability to point off vertical at targets of interest, and this has negative implications to 82 systematic wall-to-wall acquisition. Capacity is increasing rapidly, although annual time-series 83 acquisitions are currently rare in the archive outside of urban locations. Long-term hotspots of 84 environmental change such as tropical deforestation are not well represented in the archive. This 85 is due to both limited cloud free observations, and lack of a supporting acquisition strategy. 4 86 Greater than one half of the WARP archive is post 2007 WV-1 panchromatic imagery 87 due to WV-1’s on board storage and downlink capacity that supersedes any other U.S. 88 commercial sensor. WARP was developed primarily for Department of Defense (DoD) users 89 who do not require scientific quality multi-spectral surface reflectance data. Archived imagery is 90 primarily raw at-sensor radiance and not spatially corrected for terrain artifacts (orthorectified). 91 This reduces viability for ecosystem studies, although new methods and algorithms continually 92 evolve to enable these data to be pan-sharpened or fused with other multispectral sensors 93 [Ehlers, 2008]. Image processing software also can readily read NITF CQI data and require only 94 a DEM with imbedded rational polynomial coefficients (RPCs) to rapidly orthorectify raw non- 95 terrain corrected data. As of mid – 2012 we estimate that > 4 petabytes (4 million gigabytes) of 96 global data currently exist in WARP with much more data available through USGS CIDR 97 requests. 98 How to Query Vendor Archives 99 The complete data collection archives of DigitalGlobe (digitalglobe.com) and GeoEye 100 (geooeye.com) can be searched online with their respective user search and discovery tools. 101 Metadata are available including cloud coverage estimates, corner coordinates, and reduced 102 resolution quick-look images. Cross-referencing archives is difficult as data file naming 103 conventions are not consistent between vendor archives and WARP. If insufficient data are 104 found within WARP, users are encouraged to search vendor archives. If data are found in the 105 vendor archives that are not in WARP, a USGS CIDR request should be submitted. Using CIDR 106 directly would waste limited available resources and is counterproductive to NGA’s CQI 107 distribution goals. 5 108 GeoEye has many search options available through GeoFUSE tools 109 (http://geofuse.geoeye.com). Users can access online maps, use advanced options, such as 110 searching with areas of interest (AOIs) using ESRI Shapefiles or a Google Earth keyhole markup 111 language (KML). Geoeye’s online resource center also provides up to date compressed Esri 112 Shapefiles of IK-2 and GE-1 acquisition coverage freely available for download 113 (http://geofuse.geoeye.com/resources/Default.aspx) that includes metadata information for 114 archive searches. DigitalGlobe data can be searched using a web interface called image finder 115 (http://browse.digitalglobe.com/imagefinder/main.jsp?), where users can search with a map 116 display for their area of interest or upload an ESRI Shapefile. Note that imagery is dynamic and 117 is in constant state of update. 118 Earth Science Applications of Sub-Meter Satellite Data 119 Examples of CQI data use are abundant in the earth science community. This resource 120 provides many opportunities to understand sub-pixel phenomena that occur in other freely 121 available moderate to coarse resolution satellite data used in earth science remote sensing 122 applications. The primary use of data have been for validation of Landsat and Moderate 123 Resolution Imaging Spectroradiometer (MODIS) land products for sub-pixel analysis, although 124 the capabilities of CQI data have been used in other unique and novel ways due to the benefits of 125 sub-meter resolution. We provide examples here of how these data have been recently used. 126 Many different forest applications of CQI include species identification [Han et al., 127 2012]; crown delineation [Palace et al., 2008]; plot-level tree height coupled with lidar data [St- 128 Onge et al., 2008]; canopy surface model generation [Baltsavias et al., 2008]; forest health 129 monitoring [Wulder et al., 2012]; monitoring protected areas [Soares et al., 2011]; and 130 disturbance assessment from insects [Wulder et al., 2009] and storms [Romer et al., 2012]. 6 131 CQI data were used for coastal zone for surveys of mangroves [Satyanarayana et al., 132 2011]; benthic community mapping [Roelfsema and Phinn, 2010]; bathymetric mapping 133 [McCarthy et al., 2011]; and wetland pattern analysis [Peregon et al., 2009] integrated with 134 measurements of CH4 exchange [Flessa et al., 2008]. 135 CQI data have been used for Cryosphere studies mapping changes in glacier extend in the 136 high Alps [Paul et al., 2011]; permafrost extent in the Mackenzie River Delta [Nguyen et al., 137 2009]; monitoring rates of Arctic coastal erosion from melting ground ice [Lantuit and Pollard, 138 2008]; distribution of vertical meltwater conduits (moulin) in West Greenland [Phillips et al., 139 2011]; and monitoring Weddle seals abundance and population trends in remote Erebus Bay, 140 Antarctica [LaRue et al., 2011]. 141 Data have also been used for human-environment monitoring with urban land-cover 142 delineation/characterization [Huang and Zhang, 2012]; urban disaster assessment in Haiti 143 [Kazama and Guo, 2010]; cropland type mapping [Upadhyay et al., 2012]; infectious disease 144 monitoring by larval habitat mapping for malaria transmission [Krefis et al., 2011]; archeology 145 mapping of Neolithic settlements [Alexakis et al., 2009]; and humanitarian aid decision support 146 mapping of internally displaced persons (IDPs) camps in Southern Darfur [Jenerowicz et al., 147 2011] and Sri Lanka [Kemper et al., 2011]. 148 Current WARP Development and Future Opportunities 149 Recent applications of CQI data have been highlighted, and additional unforeseen 150 applications could be revealed in the future as data are used by more of the scientific community. 151 As the data archive grows, multi-temporal high-resolution analysis becomes a possibility. 152 Improvements to the WARP interface are ongoing and speed of access to query and retrieve 153 more data volume will evolve. Graphical user interfaces (GUI’s) are currently under 7 154 development using an interface similar to Google Earth. The release date of these interfaces to 155 users outside of NGA is still to be determined. Note the current WARP system has limitations; 156 large areas (> 500 x 500 km) can be difficult and time consuming to search and discover data. 157 This is due to limited download rates from WARP servers (~1 image per hour from 9 AM - 5 158 PM), and maximum results returned for each search < 250. 159 Commercial remote sensing industry growth has been rapid from the onset of NGA’s 160 contracts. New instruments will be launched in 2013 that have greater resolution and image 161 acquisition capacity, ushering in the next era of CQI. These data provided by NGA at no charge 162 via license agreement with U.S. commercial vendors is a vast resource available to qualified 163 image analysts and Earth scientists that have yet to reveal their full benefit to the research 164 community. 8 U.S. commercial sub-meter image archives from GeoEye and DigitalGlobe displayed as color coded cloud cover percentage by individual image bounds by sensor. Overlapping bounds show earliest image acquisition from the archive and data is primarily post 2007. 9 165 References: 166 Alexakis, D., A. Sarris, T. Astaras, and K. Albanakis (2009), Detection of Neolithic Settlements 167 in Thessaly (Greece) Through Multispectral and Hyperspectral Satellite Imagery, Sensors-Basel, 168 9(2), 1167-1187. 169 Baltsavias, E., A. Gruen, H. Eisenbeiss, L. Zhang, and L. T. Waser (2008), High-quality image 170 matching and automated generation of 3D tree models, Int J Remote Sens, 29(5), 1243-1259. 171 Ehlers, M. (2008), Multi-image Fusion in Remote Sensing: Spatial Enhancement vs. Spectral 172 Characteristics Preservation, Advances in Visual Computing, Pt Ii, Proceedings, 5359, 75-84. 173 Flessa, H., A. Rodionov, G. Guggenberger, H. Fuchs, P. Magdon, O. Shibistova, G. 174 Zrazhevskaya, N. Mikheyeva, O. A. Kasansky, and C. Blodau (2008), Landscape controls of 175 CH4 fluxes in a catchment of the forest tundra ecotone in northern Siberia, Global Change 176 Biology, 14(9), 2040-2056. 177 Han, N., K. Wang, L. Yu, and X. Y. Zhang (2012), Integration of texture and landscape features 178 into object-based classification for delineating Torreya using IKONOS imagery, Int J Remote 179 Sens, 33(7), 2003-2033. 180 Huang, X., and L. P. Zhang (2012), Morphological Building/Shadow Index for Building 181 Extraction From High-Resolution Imagery Over Urban Areas, Ieee J-Stars, 5(1), 161-172. 182 Jenerowicz, M., T. Kemper, and P. Soille (2011), An automated procedure for detection of IDP's 183 dwellings using VHR satellite imagery, Image and Signal Processing for Remote Sensing Xvii, 184 8180. 185 Kazama, Y., and T. Guo (2010), House Damage Assessment Based on Supervised Learning 186 Method: Case Study on Haiti, Image and Signal Processing for Remote Sensing Xvi, 7830. 187 Kemper, T., M. Jenerowicz, L. Gueguen, D. Poli, and P. Soille (2011), Monitoring changes in 188 the Menik Farm IDP camps in Sri Lanka using multi-temporal very high-resolution satellite data, 189 Int J Digit Earth, 4, 91-106. 190 Krefis, A. C., N. G. Schwarz, B. Nkrumah, S. Acquah, W. Loag, J. Oldeland, N. Sarpong, Y. 191 Adu-Sarkodie, U. Ranft, and J. May (2011), Spatial Analysis of Land Cover Determinants of 192 Malaria Incidence in the Ashanti Region, Ghana, Plos One, 6(3). 193 Lantuit, H., and W. H. Pollard (2008), Fifty years of coastal erosion and retrogressive thaw 194 slump activity on Herschel Island, southern Beaufort Sea, Yukon Territory, Canada, 195 Geomorphology, 95(1-2), 84-102. 196 LaRue, M. A., J. J. Rotella, R. A. Garrott, D. B. Siniff, D. G. Ainley, G. E. Stauffer, C. C. Porter, 197 and P. J. Morin (2011), Satellite imagery can be used to detect variation in abundance of 198 Weddell seals (Leptonychotes weddellii) in Erebus Bay, Antarctica, Polar Biol, 34(11), 1727- 199 1737. 200 McCarthy, B. L., R. C. Olsen, and A. M. Kim (2011), Creation of bathymetric maps using 201 satellite imagery, Ocean Sensing and Monitoring Iii, 8030. 10