The Need for Outside Source Hydrographic Data

Quality Assurance Methods for Utilizing Outside Source Hydrographic Data to Update NOAA s Nautical Charts Lieutenant Edward J. Van Den Ameele, NOAA NOAA Pacific Hydrographic Branch, Seattle, Washington, USA Abstract Multibeam echosounders, side-scan sonar, LIDAR, and GPS were not long ago tools belonging primarily to the hydrographer conducting surveys to update nautical charts. Currently, however, organizations such as fishery management agencies and university research programs have significantly expanded their use of these technologies in support of their missions. Although the surveys conducted are usually not intended for navigational purposes, the resultant data potentially meet International Hydrographic Organization (IHO) S-44 standards for hydrographic surveys and may be useful for application to nautical charts. As survey resources continue to be limited and budgets tight for many hydrographic offices, it only makes sense to actively solicit and evaluate these sources of data to incorporate into bathymetric databases and update charts, as they represent a cost-effective method for obtaining new hydrographic data. Ideally, a thorough assessment of data quality can be conducted so that the areas can be considered to have modern survey coverage and obviate the need to send a hydrographic resource there - thus permitting more efficient use of an HOs limited survey assets. However, there are many challenges with Outside Source Data (OSD). Because the data are often collected for purposes other than charting, data acquisition and processing procedures vary widely and the accuracy of the data may be difficult to assess. Additionally, HOs can quickly get buried under the volume of the available data, and processing and evaluating OSD requires new tools, procedures, and personnel resources. While the utilization of data from outside sources on nautical charts is not a new concept, evaluating the data against hydrographic standards and fully superseding existing chart data with data from outside sources is something which HOs have historically avoided. However, it is now arguable that the high quality sources of data can no longer be ignored in the mission for the most accurate, up-to-date nautical charts. NOAA has initiated a program to evaluate and incorporate OSD into its nautical charts and has begun to confront some of these challenges. This paper explores NOAA s initiative, procedures, and challenges in utilizing hydrographic data from non-traditional sources to update nautical charts, as well as on overview of how many of the latest technological developments such as Total Propagated Error modeling and the Navigation Surface have aided in the assessment of data quality. The Need for Outside Source Hydrographic Data The statutory mandate of the National Oceanic and Atmospheric Administration (NOAA) authorizes NOAA to provide nautical charts and related hydrographic information for the safe navigation of maritime commerce as well as provide basic data for engineering, scientific and other commercial and industrial activities. This mandate covers all U.S. territorial waters and the U.S. Exclusive Economic Zone (EEZ), a combined area of 3.4 million square nautical miles (snm) which extends 200 nautical miles offshore from the nation s coastline. The production of high-quality navigation charts to support the safety of marine transportation depends on the availability of up-to-date, reliable hydrographic survey data. 1 Van Den Ameele 1 U.S. Hydro 2005

To accomplish this hydrographic surveying mission NOAA utilizes a mix of government-owned and contracted survey platforms. With the reactivation of the NOAA Ship FAIRWEATHER in 2004, NOAA s hydrographic fleet increased to four ships, while the number of Navigation Response Teams (mobile hydrographic units which conduct rapid-response surveys) increased to six in the same time period. For fiscal year 2005, NOAA was appropriated $27 million for contract hydrographic surveys. In order to effectively allocate these hydrographic resources, NOAA s Office of Coast Survey developed NOAA s Hydrographic Survey Priorities to identify the most significant areas in need of new hydrographic surveys. Out of the 3.4 million square nautical miles for which NOAA is responsible for charting, 42,000 snm were originally identified as critical areas and approximately 500,000 were identified as navigationally significant. NOAA based these categorizations on factors such as vintage of survey data, volume and type of marine traffic, underkeel clearance and bottom characterization, and input from the maritime community. Since classification of the critical and navigationally significant areas in 1994, NOAA s hydrographic survey platforms and survey contractors have surveyed 14,800 square nautical miles, an average of approximately 1200 snm per year. 2 With additional hydrographic survey resources and contracts recently added to NOAA s capability, annual survey production is expected to increase, with a goal of reducing the remaining critical area survey backlog by 50% in the next five years. Nonetheless, with new emerging critical areas and areas continually in need of resurveying (e.g. due to frequently shifting shoals, or increase in volume or size of traffic), NOAA cannot meet its hydrographic survey requirement for all navigationally significant areas with its current resources. The goal of utilizing Outside Source Data (OSD) is to augment traditional hydrographic data acquisition with data which may have been acquired for other applications but may still be usable to update nautical charts. OSD, also known commonly as third-party data, is defined as hydrographic data which was acquired outside of the administrative control of a Hydrographic Office (HO) and was not acquired specifically for the purposes of revising nautical charts. 3 Traditional hydrographic surveys, whether conducted using government or contract resources, must adhere to a strict set of specifications and requirements such as those outlined in International Hydrographic Organisation (IHO) Special Publication 44 (S-44). Documents such as contract statements of work, specification manuals, and cruise instructions prescribe these requirements for hydrographic survey units. By contrast, OSD is often acquired for a purpose other than supporting navigation, and therefore may be acquired under a different set of requirements (or sometimes none at all). Evaluating OSD for application to nautical charts can be especially challenging when the data were acquired to adhere to a different set of specifications. Nonetheless, it may be incumbent upon NOAA and other HO s to actively solicit and evaluate OSD to augment their traditional hydrographic survey resources and provide the most accurate, up-to-date nautical charts. As multibeam, side-scan sonar, LIDAR, and GPS systems become more affordable, their use will become more widespread for uses such as defining essential fish habitat (EFH), supporting harbor construction and improvement projects, supporting improved coastal zone management (CZM), or mapping critical coral reef resources. These uses may also produce data sets potentially suitable for charting. However, the amount of data being acquired promises to increase as the use of these technologies becomes more widespread, and HO s can quickly find themselves buried under the massive volumes of data available for potentially updating nautical charts. Thus, it is essential that HO s develop efficient and practical means of evaluating these data sources utilizing the personnel and technical resources available. There are also legal implications, as hydrographic offices are liable for the data presented on nautical charts. Therefore, it is essential that rigorous methods of assessing the quality of these data are developed to support the decision to chart any OSD. This paper describes some of the methods NOAA has employed to date for evaluating Outside Source Data and discusses possible future directions for better utilization of OSD on nautical charts. Van Den Ameele 2 U.S. Hydro 2005

Shallow-water multibeam data acquired under contract for a state Department of Fish and Game for essential fish habitat assessment in an area of sparse and outdated hydrographic data. These sources of data are often of very high quality and should be actively solicited and evaluated to create the most accurate, up-to-date nautical charts The Outside Source Data Challenge Utilizing third-party sources of data is not new for NOAA or other Hydrographic Offices. NOAA s Marine Chart Division annually receives approximately 2,400 hydrographic source documents from thirdparty suppliers which it evaluates for application to NOAA charts. Out of these, approximately 2200 are channel surveys from the U.S. Army Corps of Engineers. From the additional 200 hydrographic surveys received, NOAA usually limits application to evaluating the data for new shoals, shoreline construction, or dangers to navigation. NOAA is not alone in this approach; the UK Hydrographic Office in 2002 received approximately 2000 sources of outside source hydrographic data but limited the application of these data to selecting new shoals. The UKHO did not use these data to disprove charted information. 4 This limited approach is taken for the safety of navigation as well as for reasons of legal liability; without an assessment of the accuracy and quality of these data, HO s must take a conservative approach of charting new shoals which may be unverified. HO s cannot defensibly remove shoals, wrecks, obstruction, etc. when the quality of the data is unknown. This approach, however, ultimately degrades the quality of the nautical chart, as unverified and potentially non-existent shoals are added to the chart and must be verified or disproved by a future hydrographic survey. In addition, charted shoals and features which may legitimately no longer exist cannot be removed. The result at best is a cluttered chart which may support safe navigation but does not necessarily provide the mariner with the most accurate and best available information; at worst, critical navigation channels may effectively be closed if unverified shoals or obstructions are applied to the chart. While this conservative approach may have been valid in the past, the quality of much of the OSD available today often allows for the full application of data and the disproval of charted information. Charting agencies should fully assess all sources of data for maximum possible chart application. NOAA aims to utilize OSD, when deemed to be of sufficient Van Den Ameele 3 U.S. Hydro 2005

quality and accuracy, to address its critical area survey backlog, thus allowing more efficient utilization of limited hydrographic resources, and allowing for the maximum possible application of these data to the agency s nautical charts. The ultimate goal is to provide the best possible information to the mariner. The challenge to utilizing OSD is that bathymetric data acquired for differing purposes often meets different requirements. Hydrographic surveys undertaken to update nautical charts must meet very stringent requirements, with special attention paid to vertical control and water level determination, horizontal control and positioning, sensor offsets and alignments, sound speed profiling, bottom coverage and object detection, and precise attitude measurements. Hydrographic surveys are also measured relative to a charting datum, usually a tidal datum in the vertical such as mean lower-low water, and a horizontal datum in the horizontal such as WGS84. By contrast, hydrographic surveys supporting constructions project may be relative to a vertical datum such as NGVD88, while pipeline, cable route, or inspection surveys conducted using Autonomous Underwater Vehicles (AUVs) or Remotely Operated Vehicles (ROVs) strive for high relative accuracy but are often not controlled to a water level datum. Transforming these data sources to support nautical charting applications can be tedious and timeconsuming. Outside Source Data Quality Assurance Process NOAA s Outside Source Data process is diagrammed below. There are three categories of OSD. The first category is data which NOAA receives which does not have supporting documentation, raw data, or metadata. These OSD are simply routed straight to NOAA s Marine Chart Division (MCD) for chart applications since they cannot receive a full hydrographic evaluation. Some of these data are from trusted suppliers and are expected to be of suitable quality, such as channel surveys from the US Army Corps of Engineers, and are applied directly to NOAA charts without further evaluation. Other data lacking documentation may be of unknown quality, such as an ASCII XYZ file of LIDAR soundings received from a state coastal zone management agency. These data may be evaluated for new shoals and navigational dangers but do not supersede and charted information. The second category is OSD for which NOAA has made a prior commitment to utilize, through joint projects or partnerships. This is data for which NOAA recognized the potential value prior to data acquisition and made a commitment to chart provided accuracy requirements were met. The third category is unsolicited OSD, which NOAA receives after it is acquired, and is of varying quality. For this third category of OSD, the first step is to make a cursory determination as to whether or not the data are in a navigationally significant or critical area, and whether the data appear to be of reasonable quality for charting. If not, the data are still evaluated for dangers to navigation. Any data what appear to be of sufficient quality and in a navigationally significant area are checked into NOAA s database for tracking. Van Den Ameele 4 U.S. Hydro 2005

OSD with no supporting information or metadata Prior Commitment OSD Unsolicited OSD Valuable for charting? Register in database and prioritize Hydrographic Evaluation Comparison against specifications and requirements Data Processing Cartographic Application Limited application Database population ENC application Updated Nautical Chart Data do not meet spec Deliverables and recommendations Are there critical corrections Notice to Mariners Letter Valuable to other NOAA? Value added processing NGDC Bathy database NOAA Outside Source Data Process Diagram The next phase of the process is for the data to undergo a hydrographic evaluation to assess the degree of compliance with NOAA s Hydrographic Surveys Specifications and Deliverables Manual (HSSDM) and IHO S-44 (The HSSDM is NOAA implementation of S-44 requirements). During the hydrographic evaluation, the following factors are examined: Sonar (or LIDAR) systems utilized: What are the resolution, frequency, range scales available, and were they appropriate for the depth of water? Is the sonar capable of meeting IHO and NOAA specifications? Positioning: Was DGPS or RTK utilized? What is the horizontal datum? What quality control measures were utilized to ensure adequate positioning? Water level measurements: What type of water level measurement equipment was used? How were the final water level correctors determined and applied? Corrections to echosoundings: How, and how frequently, were sound speed profiles measured and applied to the sonar data? Was an offset and alignment survey conducted? Was a patch test conducted to determine sensor biases? Were draft measurements recorded and applied to the data? Are systematic errors (such as refraction errors) noticeable in the data, and if so, what is their magnitude? Data acquisition procedures: Was full-bottom coverage obtained? Were vessel speed, line spacing, range scales, and ping rates sufficient to meet object detection requirements? Data processing and quality assurance procedures: Were the data bin-averaged or were shoal depths preserved? Which processing software and processed were used to inspect and remove fliers? Internal data consistency: Where overlapping data exist from redundant coverage or check-lines, do the data agree within allowable error bounds? (Although this is not a check on absolute accuracy, internal data consistency can add confidence that the data were properly controlled and corrected, particularly when redundant data exist from different systems or platforms.) Van Den Ameele 5 U.S. Hydro 2005

Comparison with external datasets: Contemporary or prior data sets are compared with the OSD under evaluation within any areas of overlap. Any full assessment of OSD requires thorough documentation and metadata describing systems, field procedures, and processing and quality assurance methods. Raw or full density sensor data are often desirable, rather than an end product such as a selected sounding set or gridded surface, particularly when information regarding acquisition and processing procedures is not available. However, with OSD, since the data are acquired for purposes other than hydrography, complete information or data are often not available. When available, error estimates from the various sensors used are compiled into an estimated error budget which NOAA personnel can use to quantitatively assess the accuracy of the data. When values are not available for components of an error model, the maximum possible error is used for that component of the error budget, leading to a conservative value for that portion of the error estimate. Once the hydrographic evaluation is completed, the data are qualified for application to the nautical chart. This requires a comparison with the currently charted data for the survey area. Ideally, a quantitative assessment of accuracy and uncertainty can be given to charted legacy data so that an analysis of uncertainty can lead to charting selections which are based upon the least uncertain data for any given area receiving preference. It is anticipated that in the future, uncertainty information will accompany hydrographic data in a chart production database so that selection of charted information is matter of selecting the data with the least uncertainty. For now, the data are qualified for application under the following categories, in decreasing order of data quality: The data fully meet IHO S-44 and HSSDM accuracy, seafloor coverage, and object detection requirements and can supersede charted information in common areas. All charted shoals, wrecks, rocks, obstructions, depths, depth areas, and contours can be superseded by the new data. Data meet IHO S-44 and HSSDM requirements for accuracy but not for object detection, or full seafloor coverage was not achieved. Data are sufficient for updating depths, contours, and depth areas showing general bathymetry trending, but not for disproving isolated wrecks, rocks, obstructions, or shoals, because object detection requirements were not met or cannot be determined given the information available. Data meet neither IHO nor NOAA HSSDM accuracy or object detection requirements yet may still represent better information than that charted. Charted data may be from very old or undetermined sources, or the new survey data may be of greater bottom coverage that prior survey sources. The new data may be also be slightly outside of allowable error bounds (e.g. IHO Order 2). New shoals and features are charted and data supplement charted information where sparse. Data do not meet IHO or NOAA requirements and do not represent better information than that charted. Dangers to navigation may still be selected if shoals not represented on the chart are found in the dataset. It is anticipated that NOAA will align these categories of application with IHO S-57 zones of confidence (ZOC) and will populate M_QUAL meta objects in NOAA ENCs as well as source diagrams on paper and raster nautical charts. Conveying to the mariner the quality of information on nautical charts through the use of CATZOC is extremely important when data from outside sources are used to supplement the high quality of data the mariner typically expects to appear on charts from traditional hydrographic surveys. Van Den Ameele 6 U.S. Hydro 2005

Outside Source Data Examples The Outside Source Data which NOAA s Office of Coast Survey has evaluated to date has varied extensively in quality, survey purpose, acquisition and processing methods, specifications met, and accuracy; and therefore the methods used to evaluate these data and resulting chart application have varied widely as well. The following examples describe some of the types and quality of data received as well as the methods used so far to evaluate OSD for nautical chart application in NOAA. Example 1: Disenchantment Bay The Hubbard Glacier in Disenchantment Bay, Alaska, is a popular destination for cruise ships. In May 2002, the glacier and its moraine dammed the entrance to Russel Fiord at the head of Disenchantment Bay. In August 2002 the dam parted and the resulting current pushed a large volume of sediment into Disenchantment Bay. The resulting changes in bathymetry posed a potential hazard to cruise ship navigation. However, NOAA had surveyed the area with shallow-water multibeam in 1999 and had no immediate plans to revisit in the near future for a new survey. The entrance to Russel Fiord before (June 23, 2002) and after (August 10, 2002) the dam parted (Courtesy: USGS) In the weeks following the parting of the dam, glacial investigators from the University of Alaska at Fairbanks (UAF) conducted a single-beam sonar survey of the area. Because changes were noted, UAF offered the data to NOAA to evaluate for charting. NOAA s Pacific Hydrographic Branch in Seattle, Washington, evaluated the data for data quality and potential chart application. A one-page document which briefly described UAF s survey methodology indicated that no corrections for sound velocity were made, vessel draft was not measured, tide application used an estimated zoning, and transducer to GPS offsets were neither measured nor applied. While the data were deemed to not meet NOAA specifications, the significant changes to the seafloor and the hazards they imposed mandated that NOAA report shoal depths. Soundings from the dataset representing dangers to navigation were reported through the US Coast Guard Local Notice to Mariners and applied to NOAA charts. These represented a small number of sounding in comparison with the full dataset which was not utilized. Nonetheless, the hazard presented dictated the evaluation and incorporation of these data to the maximum extent possible until a higher accuracy survey could be conducted. Van Den Ameele 7 U.S. Hydro 2005

Single-beam trackline data (left) at Russel Fiord, and the resulting shoals charted (right). Example 2: Glacier Bay The federal contracting process allows any federal agency the opportunity to attach additional surveying task orders to the hydrographic survey contracts NOAA awards each year provided the agencies have the funding and the contractor has the time and capacity to conduct the survey. Since the cost of mobilization and de-mobilization is paid for with NOAA funding, it provides a cost-effective option for other US agencies to acquire bathymetric data in support of their programs since they need not cover these costs. Because the contract vessel is already outfitted and staffed to accomplish a NOAA hydrographic survey, it also means that they are capable of acquiring data to NOAA hydrographic specifications. Frequently these piggy-back surveys are conducted while in transit at the start or the end of the surveys conducted for NOAA. As an example, the US Geologic Survey (USGS) in 2001 had requirements for a multibeam survey of Glacier Bay, Alaska, which was also identified as a critical survey area in NOAA s Hydrographic Survey Priorities. Full coverage multibeam data were acquired to support both uses. USGS provided these data to Coast Survey to evaluate and update nautical charts as appropriate. Sonar systems, positioning, acquisition procedures, and tidal zoning were consistent with those used for NOAA hydrographic contract surveys. Survey deliverables and data products matched NOAA requirements outlined in the HSSDM. The data were evaluated and deemed to meet NOAA and IHO specifications. Because the new dataset largely confirmed charted information, only a few new shoals were selected and charted. A full recompilation of the chart for this area was not required. It should be noted, however, that in this instance, there were nuances between data acquisition and processing methods required for geologic research and hydrography. For example, the USGS required a bin-averaged dataset as the final product. While this would suffice in providing charted depths representing general bathymetry, charting requirements dictate that least depth on shoals and other significant features be preserved. This required NOAA hydrographers, in limited instances, to compare gridded depths with full resolution multibeam data, and supplant the gridded depth with the measured least depth. Often, Outside Source Data requires additional manipulation and processing since the data may have been acquired and processed for a different purpose and under a separate set of requirements. This can add time to the evaluation of OSD. However, it is important to keep this in perspective, as the time and cost associated with and reprocessing is less than that required to completely survey an area. Van Den Ameele 8 U.S. Hydro 2005

Multibeam data from Glacier Bay, Alaska, acquired under contract for USGS and NOAA Example 3: Stellwagen Bank Hydrographic agencies can also use Outside Source Data to more efficiently task their survey resources. While OSD may not always completely alleviate the requirement to survey a priority area, it can be used to limit the amount of work required to that necessary to validate or augment the OSD. In 2003 the NOAA Ship THOMAS JEFFERSON was tasked hydrographic surveys in the approaches to Boston Harbor. The US Geologic Survey had also funded multibeam surveys of much of the same area between 1994 and 1998. These surveys were conducted through a combination of the Canadian Hydrographic Service and the University of New Brunswick (Canada) aboard the S/V Fredrick Creed, with NOAA hydrographers participating in various phases. Additional data were acquired in the area in the winter of 2002-2003 through a survey contracted by the University of New Hampshire. None of these data were specifically acquired to current hydrographic standards. THOMAS JEFFERSON survey personnel obtained the data from the USGS, and rather than completely resurvey the entire area, limited their operations to that necessary to validate and augment the USGS data. First, the USGS data were converted into CARIS HIPS, the processing software used aboard JEFFERSON. Next, tide correctors consistent with the zoning scheme used for the JEFFERSON survey area were reapplied to the data. The vessel then ran cross lines sufficient to verify the USGS data and obtain error estimates. Item investigations on a handful of noted were conducted to obtain accurate least depths for charting. Additional multibeam coverage was obtained in a few areas needed by NOAA by not surveyed by USGS. The data were combined in CARIS, and processed, evaluated, and submitted as a single dataset. A 0.5 meter vertical offset between the USGS data and NOAA data could not be resolved. This offset was within IHO Order Van Den Ameele 9 U.S. Hydro 2005

1 tolerances. The combined JEFFERSON data and USGS data will be used to update the chart with modern multibeam coverage. What would have required several months of NOAA ship time (the Creed required 86 sea days to survey the area) for a complete survey was reduced to a few weeks of crosschecks and validation. Example 4: Northwest Hawaiian Islands Collaboration When possible, it is desirable to develop partnerships to define survey requirements prior to conducting survey operations and maximize the use of the data. This can alleviate much of the a posteriori quality assurance effort, and limit the verification to ensuring that the specifications, requirements, and procedures agreed to beforehand have been adhered to. In 2002, NOAA Coast survey hydrographers collaborated with scientists from the University of Hawaii as well as other NOAA programs such as the Office of Ocean Exploration and the National Marine Sanctuary Program to acquire multibeam data in the Northwest Hawaiian Islands (NWHI) using the University of Hawaii vessel Kilo Moana. The Kilo Moana is a Small Waterplane Area Twin Hull (SWATH) vessel outfitted with Simrad EM1002 and EM120 multibeam sonars and an Applanix POS/MV positioning and orientation system. The data acquired were to meet IHO S-44 requirements to update NOAA charts where there was scant data, support boundary definition of Reserve Protected Areas within the NWHI Coral Reef Ecosystem Reserve based on the 100-fathom contour, and acquire backscatter and bathymetry data to support ecosystem management activities of the Reserve. While NOAA survey ships in the past have conducted scientific mapping in areas which were also of navigational significance, this was the first time that a NOAA hydrographer had deployed aboard an academic ship with the intent of acquiring data to update nautical charts. 5 At the end of the cruise, the NOAA hydrographer returned with a data set, which is currently undergoing evaluation for application to the chart. The quality assurance task is proving to be much more straightforward than, say, having merely received the data after the cruise, since Coast Survey personnel were involved with the stakeholders in defining the requirements prior to the cruise, and by participating in the cruise to ensure that data acquisition procedures and parameters were consistent with those required for accurate hydrographic data for nautical charting. Example 5: NOAA/NAVOCEANO Partnership In 2001 the US Naval Oceanographic Office (NAVOCEANO) approached NOAA with new requirements for updated nautical charts requiring new hydrographic surveys in the Main Hawaiian Islands, Northern Marianas Islands, and Guam. These areas were not at the time listed as critical in NOAA s National Survey Plan, yet the area is in the US EEZ and included in NOAA s area of responsibility for charting. An agreement was reached between NAVOECEANO and NOAA s Coast Survey that NOAA would update its charts with hydrographic surveys conducted by NAVOCEANO provided that IHO S-44 Order 1 specifications were achieved. NAVOCEANO conducted these surveys in 2001 using a combination of SHOALS LIDAR, single-beam and side-scan sonar using NAVOCEANO Fleet Survey Team resources aboard vessels of opportunity, and shallow-water multibeam aboard US Navy survey vessels. Integrating these three sources of data presented a challenge for NAVOCEANO since the data sets were processed independently of one another using varying sensors and positioning systems, and were processed via differing processing software and procedures. While these differences led to offsets between data sets which needed to be resolved, it also provided an excellent cross check between platforms and revealed systematic errors which might not have been otherwise detected using a single platform 6. The difference between the data sets also presented a challenge for NOAA reviewers, as the data from each platform needed to be independently assessed for accuracy, seafloor coverage, object detection, and suitability for updating NOAA s charts. In the end, the extent to which the charted information was superseded depended largely upon whether a survey area had been covered with LIDAR, multibeam, single-beam, side-scan sonar, or some combination of these systems. For this reason, NOAA reviewers also needed to segregate the fused dataset according to system type so that a NOAA cartographer could make the proper charting application. Van Den Ameele 10 U.S. Hydro 2005

Shortly after the NAVOCEANO surveys of Saipan, a new NOAA survey vessel for fisheries habitat mapping, the R/V AHI (Acoustic Habitat Investigator), was conducting habitat mapping in Saipan when they were approached by the Saipan harbormaster. The harbormaster was concerned about shoaling in the dredged channel leading into Saipan Harbor and asked the crew of AHI to conduct a multibeam survey of the channel. The scientists aboard the AHI agreed to a reconnaissance survey of the channel and approaches to the harbor. The AHI is a 25-foot survey launch equipped with a Reson SeaBat 8101 multibeam sonar, POS/MV positioning and orientation system, and SeaBird SBE-19 CTD for velocity profiling. Since the AHI was still a new vessel, much of the required calibrations for a survey vessel such as a patch test had not yet been performed. The crew of the AHI had a limited window in which to conduct this survey and thus did so prior to performing these calibrations. Additionally, differential GPS was not available in Saipan so C/A mode GPS was used, and a tide gauge was not available in Saipan Harbor. Yet the AHI found several coral heads and other shoals in the Saipan Harbor channel which were shoaler than the controlling depth of the channel and presented hazards to navigation. The data were provided to NOAA s Office of Coast Survey for evaluation for charting. Discrepancies between LIDAR soundings (blue) and multibeam least depths on significant features (red) for Saipan Harbor overlaid on the nautical chart. The underlying sun-shaded DTM is from the multibeam data. Reconciling these two data sets presented a challenge for NOAA evaluators. In the area of overlap, the NAVOCEANO data set consisted of LIDAR, while the AHI data set was full bottom coverage shallowwater multibeam. Based on NOAA s evaluation of the LIDAR, it was determined that given the coralline nature of the harbor and its environs, it could not be determined with absolute certainty that the least depth on any particular feature could be positively resolved with LIDAR alone. This is largely due to the spot spacing used during LIDAR acquisition in comparison with nature of the seabed. Yet NOAA evaluators did conclude that LIDAR data did meet IHO Order 1 accuracy requirements for depth and positioning where soundings were recorded. In other words, depth and position accuracy requirements were met while object detection requirements had not been me with LIDAR. By contrast, the AHI multibeam survey met object detection requirements with extensive multibeam coverage, yet did not meet depth and position accuracy requirements due to the factors mentioned above. The AHI survey detected Van Den Ameele 11 U.S. Hydro 2005

numerous shoal features which had not been adequately resolved with LIDAR, and there was a general vertical offset between the two surveys of approximately 0.5 meters. To address the discrepancies and make sound charting decisions, NOAA evaluators first constructed an error budget for the AHI data (error analysis for the NAVOCEANO LIDAR data had already been accomplished) taking into account the crude alignment calibration, lack of differential positioning, and water level correctors zoned from the station at Apra Harbor, Guam. For flat or evenly sloped, featureless regions, the LIDAR data was selected for charting since it met IHO Order 1 requirements. Over shoals such as coral heads, if the discrepancy between the LIDAR least depth and multibeam least depth was within the computed error, the LIDAR sounding was selected for compilation. When the AHI multibeam data detected a sounding shoaler than the LIDAR and the discrepancy exceeded the error of the multibeam, the multibeam sounding was selected as the most conservative and safe approach. Future Directions for Outside Source Data Error Modeling and Uncertainty Analysis The procedures NOAA has used to evaluate Outside Source Data have been largely qualitative in nature. Although error models have been developed and analyzed for some OSD surveys, they often represent rough estimates of error components and more attention is devoted toward a review of systems and procedures. Quantitative analyses have only involved comparisons of internal data consistency, and external data sources such as prior surveys or overlapping contemporary surveys. However, new developments in hydrography may allow for a more quantitative assessment of data accuracy in the future. The Navigation Surface concept provides the hydrographer and cartographer a more rigorous methodology for making decisions about preceding surveys by modeling, tracking, and reporting uncertainty as well as depth at each model node. 7 The horizontal and vertical uncertainty for each sounding is calculated by modeling all of the contributing error sources. Each resulting node in the Navigation Surface includes values for depth, total propagated error, and standard deviation, among others. The Navigation Surface concept would fit Outside Source Data well, since it was developed as a way of supporting multiple uses of a single dataset 8. Provided enough information is available about the Outside Source Data, the total propagated error could be computed and used in the creation of an attributed Navigation Surface which would result in a better quantitative assessment of the accuracy of OSD data sets. At the time of writing, NOAA hydrographers are working with non-traditional providers of data, such as the NOAA Ship HI IALAKAI, to develop vessel error models which can then be used to quantitatively assess the accuracy of the acquired data. The HI IALAKAI was recently commissioned by NOAA as a multipurpose mapping vessel to support NOAA mapping missions in the Hawaiian Islands. The HI IALAKAI is outfitted with Simrad EM3002D and EM300 multibeam sonar systems and a POS/MV positioning and orientation system, and will support coral reef mapping, ocean exploration, and deep water mapping missions. Integrated Ocean Mapping NOAA and other federal agencies are starting to address the disparities in requirements and mapping missions through the development of Integrated Ocean Mapping (IOM). The goal of Integrated Ocean Mapping is to ensure that any ocean mapping data serves the maximum possible users and applications by defining requirements and specifications prior to data acquisition, federally coordinating all ocean mapping activities, and developing products and datasets which are widely available and serve the maximum possible applications. The need for IOM was addressed by the recent U.S. Commission on Ocean Policy, which noted At least eleven federal agencies, almost all coastal states, and many local Van Den Ameele 12 U.S. Hydro 2005

agencies, academic institutions, and private companies are involved in mapping, charting, and assessing living and nonliving resources in U.S. waters. The report continues: Maps of coastal land areas, and charts of nearshore and offshore areas, are essential for safe navigation and for defining boundaries, mitigating hazards, tracking environmental changes, and monitoring uses. Because many organizations have mapping and charting responsibilities, there are significant overlaps. This situation results in multiple entities within government, industry, and academia undertaking the expensive and time-consuming task of repeating surveys of the same area for different purposes. Furthermore, differences in scale, resolution, projection, and reference frames inhibit the integration of onshore and offshore data. 9 NOAA has begun to advance the Integrated Ocean Mapping concept through the installation of multibeam sonars and other hydrographic equipment on agency vessels which will not primarily support hydrographic surveying for nautical charting. The NOAA Ship HI IALAKAI will support coral reef and deep water mapping missions in the NWHI. The areas in the NWHI of concern to coral reef and fisheries biologists also have sparse hydrographic data. This is but one example of ways in which NOAA is working to integrate requirements between all programs with an interest in ocean mapping activities to ensure that any mapping data acquired ultimately meet all requirements. Conclusions With limited resources and expanding requirements, hydrographic offices such as NOAA s Coast Survey can no longer afford to merely accept data acquired through traditional hydrographic means such as inhouse hydrographic fleets and turn-key hydrographic survey contracts. The expanded use and relative affordability of multibeam, side-scan sonar, GPS, and LIDAR systems and services has increased their application well beyond hydrography for nautical charting. These additional data sources often represent high-quality data sets which can be used to update nautical charts with data in areas which might not receive contemporary hydrographic surveys. The traditional practice of simply selecting shoals and dangers from third-party data sources is no longer a viable solution when full coverage, high quality data may be available and can lead to a better charting product for the mariner. However, a rigorous quality assurance process is required in order to effectively utilize these data sources and legally defend any data selected for charting. Because Outside Source Data can arrive in a variety of formats, with varying accuracies, and varying levels of supporting data and information, the usable must be sorted from the unusable. For these reasons, evaluating OSD for nautical charting applications will always be more time consuming than verification of traditional hydrographic survey data acquired under the administrative control of a hydrographic agency. However, the time and cost required to utilize OSD must be compared with the time and cost which would have been required to conduct a complete survey utilizing an agency s limited hydrographic assets. In most cases, utilizing the OSD can prove to be a cost-effective solution, and lead to better utilization of limited hydrographic survey resources. The key to an effective OSD program is to establish partnerships with other organizations which have requirements for ocean mapping data in similar regions. Separate hydrographic surveys and scientific surveys should be replaced with multi-purpose datasets where possible. Innovations such as the Navigation Surface and Total Propagated Error Modeling will better facilitate the creation of multipurpose data sets in the future. Coordinating all mapping resources on a national level by implementing recommendations outlined in the U.S. Commission on Ocean Policy Report and through the development of Integrated Ocean Mapping will only further this goal. Van Den Ameele 13 U.S. Hydro 2005

1 National Oceanic and Atmospheric Administration, Office of Coast Survey. NOAA s Hydrographic Survey Priorities. October, 2004, available online at http://nauticalcharts.noaa.gov/staff/nhsp.html 2 Square nautical mile figures are through the end of 2003. 3 For the purposes of this paper, nautical chart refers to paper charts, raster nautical charts, and electronic navigational charts (ENCs). 4 Howlett, Chris, UK Hydrographic Office. New Bathymetric Data Assessment at the UK Hydrographic Office. Hydro International, May 2003. 5 Evans, B.K., et al. Collaborative Nautical Charting and Scientific Seabed Mapping Missions: A Case Study in the Northwestern Hawaiian Islands. IEEE Oceans 2003 Conference Proceedings, Vol. 2. 6 Van Norden, Maxim, et al. Lessons Learned in Multi-Platform Hydrographic Surveys. IEEE Oceans 2002 Conference Proceedings. 7 Smith, Shepard. The Navigation Surface: A Multipurpose Bathymetric Database. University of New Hampshire Masters Thesis, May 2003. 8 Smith, 2003. 9 An Ocean Blueprint for the 21st Century: Final Report of the U.S. Commission on Ocean Policy Pre-Publication Copy. Chapter 25, pp. 337-338. Washington, D.C., 2004 Van Den Ameele 14 U.S. Hydro 2005