Towards the implementation of a fully operational HF coastal radar network operated by Puertos del Estado

Transcription

1 Towards the implementation of a fully operational HF coastal radar network operated by Puertos del Estado Pablo Lorente, Javier Soto-Navarro, Maria Isabel Ruiz, Enrique Alvarez- Fanjul Coastal Ocean Department Puertos del Estado Madrid, SPAIN plorente@puertos.es Silvia Piedracoba Applied Physics Department University of Vigo Pontevedra, SPAIN spiedra@uvigo.es Pedro Montero Ocean Modeling Department Intecmar Pontevedra, SPAIN pmontero@intecmar.org Abstract Puertos del Estado has made noticeable efforts to articulate a HF coastal ocean radar network with the aim of strengthening its observational infrastructure. In recent years, four HF radar Codar SeaSonde systems have been deployed along the Spanish coastline. Active and planned development efforts to implement a fully operational radar network include: homogenization of maintenance practices, quality control checks, validation works with independent in-situ sensors, standardization of data access and development of customized visualization tools. A dedicated online website has been developed to operationally monitor radar system health in real time. This automated quality control application analyze a number of diagnose parameters to obtain estimates of their standard ranges and evaluate radar site performance according to them. Abrupt or gradual degradation and failure problems can be easily detected, triggering alerts for troubleshooting. Keywords HF radar; operational network; quality control; validation. I. INTRODUCTION HF radar networks have grown worldwide, rapidly becoming an essential component of coastal ocean observation and prediction systems. This novel shore-based remote sensing technology produces near real-time 2-D surface current maps, which have a broad range of practical applications for Search and Rescue (SAR) operations, oil spill modelling or the calibration and validation of operational ocean forecasting systems. In recent years, Puertos del Estado (PdE hereinafter), the Spanish public institution devoted to operational oceanography and in charge of coordination of the state-owned port system, has made noticeable efforts to implement a HF coastal ocean radar network with the aim of strengthening its observational infrastructure and developing tailored products to be used not only in port operations but also in emergency situations. In this context, PdE has deployed four HF radar Codar SeaSonde systems along the Spanish coastline in the frame of different EU-projects or as a result of cooperative programs with other institutions. These four radar systems are shown in Fig. 1 and briefly described in Table 1. Fig. 1. HF coastal radar network managed by Puertos del Estado, composed by four different radar systems. TABLE 1. DESCRIPTION OF THE HF RADAR COASTAL NETWORK OPERATED BY PUERTOS DEL ESTADO AND OTHER PARTNERS. System Location Frequency Project Operated by (sites) Galicia PdE and NW Spain 5 MHz (4) Intecmar Gibraltar S Spain 26.5 MHz TRADE 1 PdE (3) Ebro E Spain 13.5 MHz RIADE 2 PdE, Magrama Delta (3) and Acuamed Huelva SW Spain 13.5 MHz TRADE 1 PdE and HIP 3 (3) 1 TRADE: Trans-regional RADars for Environmental applications 2 RIADE: Red de Indicadores Ambientales en el Delta del Ebro 3 HIP: Hydrographic Institute of Portugal

2 With the main purpose of articulating a fully operational HF radar network, a variety of tasks is mandatory to be carried out. Active and planned development efforts include the implementation of: A defined set of operational best practices for radar systems operations, deployment and maintenance. A common protocol for data processing and the standardization of a single HF radar interoperable data format, including a unified list of metadata descriptors. A scalable data management, a long-term storage in relational databases and a distribution following internationally accepted standards. Unified procedures for quality assurance - quality control (QA-QC) in real-time. Offline validation methodologies with independent in-situ instrumentations. Web services to access and deliver georeferenced data in real time following agreed dissemination policies. User-friendly visualization interfaces to foster the integration and usability of oceanographic knowledge. The primary goal of the present work is to provide an overall overview of PdE main achievements and activities, carried out in the designed roadmap for implementing a fully operational HF radar network. II. MAINTENANCE AND OPERATIONS A set of best practices for radar operations and maintenance must be defined to articulate a totally integrated network. Recommendations encompass, among others: 1) a proper selection of placement, free of radio interferences and obstructions; 2) protection against lightning strikes and external attacks of animals or human vandalism; 3) provision of onsite electricity and Internet connectivity; 4) accurate antenna beam response patterns (APMs) for each station in order to provide precise radial vectors subsequently used to generate total current vectors; and 5) routine maintenance tasks to ensure consistent and sustained operation. In this context, selected locations for PdE radar systems deployment are lighthouses (8 stations), wide beaches (3) and marinas (2). Routine maintenance, performed by the contractor (Qualitas Remos, partner of CODAR Ocean Sensors COS hereinafter-), includes automated remote checks of site status and regular on-site inspections on a quarterly basis together with annual APMs. III. QUALITY CONTROL OF HF RADAR DATA The development of quality assurance quality control metrics (QA-QC), implemented at various stages of processing, constitutes an ongoing work line. A number of parameters has been examined for their applicability as a diagnostic tool for measuring HF radar performance, according to previous works ([1], [2]). Such parameters include hardware diagnostics (e.g. temperature and power) of receiver and transmitter systems, antenna (e.g. amplitude corrections for loops 1 and 2 to the monopole, AMP1 and AMP2, respectively), radial (e.g. range, bearing and number of radials provided) and total (e.g. temporal and spatial coverage) parameters. A. Radial quality control Automated techniques for data quality check must be firstly applied at the radial level. A quality control online website has been developed to operationally monitor the aforementioned parameters (Fig. 2). The real-time monitoring (with hourly updates) allows operators to inspect data, obtain estimates of standard ranges of parameter values and evaluate radar site performance according to them. Abrupt or gradual degradation and failure problems can be easily detected, triggering alerts for troubleshooting. Since the signal-to-noise ratio at the monopole (SNR3) has been analyzed as valid proxy for radar data quality ([3]), its monthly evolution is routinely monitored. As an example, the two northernmost stations of Galicia HF radar system present different performances (Fig. 3, a): PRIO exhibits a sharp decrease in SNR3 by the 27 th of February 2015 whereas VILA station shows a regular performance within the tolerance levels defined by the mean and standard deviations values. Accordingly, the anomalies detected in the monthly evolution of AMP1 and the number of radials (measured) provided by PRIO (Fig. 3, b-c) shows a non-normal status and the potential existence of a malfunction in PRIO station since the indicated date. Complementarily, hourly maps of spatial coverage and temporal availability are plotted for each radar station to inspect the existence of limited radial coverage, flipped radial fields and outlier vectors ([4]) since they are considered as indicators of possible malfunctions or abnormal status. In addition, comparison between ideal (not calibrated) and measured (generated with the APM) radial vectors are performed in order to detect any distortion of the receive antenna pattern due to the geometry of the surrounding environment or the presence of metal items ([1]). Fig. 2. Online website developed to operationally monitor HF radar site status in real time. The performance of the four radar systems operated by PdE is routinely evaluated through the analysis, on different frequencies, of a variety of diagnose parameters.

3 temporal data availability in the northernmost area of the radar domain (Fig. 4, a) impact on the evolution of hourly averaged areal coverage, with the mean value barely reaching the 45% (Fig. 4, b). Finally, a low temporal versus spatial availability (35% of spatial coverage for the 80% of the time for the analyzed period) is observed as a result of PRIO malfunction since the 27 th of February 2015 (Fig. 4, c). Fig. 3. (a) Monthly evolution of signal-to-noise ratio for the monopole (SNR3) at the two northernmost stations of Galicia HF radar system, PRIO and VILA (see Fig.4, a). (b-c) Monthly evolution of AMP1 and the number of radials (measured) provided by PRIO, respectively. B. Total quality control Automated quality control checks have been implemented to ensure that HF radar systems are accurately mapping the 2D surface current field. To this purpose, the temporal and spatial coverage of the radar systems are analyzed on different frequencies (from hourly to monthly). The temporal versus spatial coverage are also depicted to check if HF radar systems operate within tolerance ranges, fulfilling the recommended level of data provision: 80% of the spatial region over the 80% of the time ([1]). To illustrate this methodology, an analysis of Galicia radar status is presented in Fig. 4, where weekly results confirm a quality degradation in PRIO station performance: the poor Fig. 4. Statistical analysis of total current measurements provided by Galicia HF radar system for the period: / 03-20, (a) Temporal coverage. (b) Weekly evolution of spatial coverage (referred to its maximum value). (c) Temporal versus spatial coverage.

4 C. Alert system The implemented alert algorithm is based on a ternary flag system. When established thresholds, defined in terms of average and standard deviation, are abruptly and continuously exceeded for more than 24 hours, data are categorized as suspect and corresponding radial files are flagged as 1 (yellow color). If such anomaly persists for a longer period of time, site status evolves to alert (flagged as 2, with red color) since data are either not released or provided but with a highly compromised quality. Values within the tolerance level are classified as a regular performance (flagged as 0, with green color). At this preliminary development stage, the described methodology is not able to eliminate errors (remove spikes or spurious values) in the data stream in real time as [3] but only to detect anomalies and categorize them in order to create a historic radial database similar to [1] for a later offline reprocessing. Future efforts should be devoted to improve radial data quality in real-time prior to the vector combination process and also to assign quality descriptor flags for each grid point data in total current fields. Information System ( to foster the management and usability of oceanographic knowledge. PORTUS is fully operational and accessible 365/7/24 from any web-enabled system. Animations of hourly total vector maps are visualized in an interactive approach through an intuitive georeferenced web interface, thanks to a calendarbased access (Fig. 5, a). In addition, a variety of interactive features are enabled for each grid point within the radar domain, as the display of data in graphical and/or numerical format by means of tables, diagrams and charts (Fig. 5, b). VI. APLICATIONS The success of a HF radar network relies on the usability of tailored products, delivered in accordance with the user community demands. The positive contribution of HF radar to enhance marine safety and oil pollution preparedness has been proved in previous works. [8] showed in the Galicia HF Radar Experience that simulated trajectories integrating HF radar currents are more accurate than those obtained considering only wind numerical data, with the mean SAR area reduced by approximately a 62% in comparison with the search area calculated without these data. IV. VALIDATION OF HF RADAR MEASUREMENTS Direct comparisons against other in-situ sensors permit an independent assessment of HF radar performance together with a quantitative estimation of uncertainties in radar observations. Extensive validation studies with pointwise current meters have been conducted to routinely evaluate the accuracy of total vectors derived from Gibraltar ([5]) and Galicia ([6]) radar systems, using zonal (U) and meridional (V) components. Special attention has been also focused to the evaluation of direction-finding capabilities and the determination of bearing errors in radial current velocities. Complementarily, selfconsistency checks at the midpoint of overwater baselines have been performed (when possible) since radar-to-radar comparisons present several benefits like the nonexistence of horizontal scale or depth mismatches. Comparison analyses have reported satisfactory correlation coefficients, RMSE and bearing offset values in the ranges , cm.s-1 and [-15, +15 ], respectively, which are consistent with previous studies ([7]). V. DISSEMINATION AND VISUALIZATION A near real time dissemination service has been implemented in the PdE s central repository machine. Hourly total vectors are freely available via both graphical display tools and middleware services like THematic Real-time Environmental Distributed Data Services (THREDDS), providing access to georeferenced HF radar datasets through a XML-based web catalog, which is routinely updated to aggregate the last HF radar files generated. Complementarily, radar data are inserted into a relational database for long-term storage and later ingested by PORTUS Fig.5. (a) Snapshot of PORTUS: hourly map of total current vectors provided by Gibraltar HF radar system. (b) 1-week time series of hourly current direction and speed, respectively, estimated at a selected radar grid point (orange filled dot).

5 With the goal of optimizing SAR activities, HF radar estimations are readily ingested by the Environmental Data Server (EDS) managed by SASEMAR (the Spanish Maritime Safety Agency) to enhance the emergency planning process for a prompt response. Complementarily, radar data provide assistance to maritime navigation by identifying large eddy features or extremely intense current pulses. In this context, the HF radar deployed in the Strait of Gibraltar is used as ancillary tool in the measurement and forecast local services from the SAMPA (Sistema Autonomo de Medicion, Prediccion y Alerta) Project of Port Authority of Algeciras Bay and PdE ( Other practical application currently underway is the routine validation of the MyOcean IBI operational circulation system [9, in press], obtaining a set of objective metrics that characterize IBI strengths and weaknesses and providing useful information employed to transition the forecasting system to upgraded versions. VII. CONCLUSIONS In this paper, a number of active and planned efforts to transform individual radar systems into a fully operational HF radar network operated by PdE have been described, including: homogenization of maintenance practices, validation works with independent in-situ sensors, standardization of data access and development of customized visualization tools. A dedicated online website has been developed to operationally monitor radar system health in real time. This automated quality control application analyze a number of diagnose parameters, considered as indicators of possible malfunctions or abnormal status, to evaluate site performance. Abrupt or gradual degradation and failure problems can be easily detected, triggering alerts for troubleshooting. Short-term future plans should address the improvement of QA-QC protocols to deliver new practical quality metrics for both radial and total vectors. Additional development efforts should focus on the validation and scientific exploitation of HF radar-derived wave data, which have not been as extensively explored. Finally, the assimilation of HF radarderived currents into numerical circulation models still remains as a priority for the HF radar research community. ACKNOWLEDGMENT The authors gratefully acknowledge Xunta de Galicia and Intecmar (Technological Institute for the Control of the Marine Environment in Galicia) for providing HF radar measurements from VILA and PRIO sites. We would also like to thank Qualitas Remos Company (partner of CODAR) for their support and useful suggestions. REFERENCES [1] H. Roarty, M. Smith, J. Kerfoot, J. Kohut, S. Glenn, Automated Quality Control of High Frequency Radar Data, OCEANS 2012, 2012, pp 1-7. [2] B.M. Emery, L. Washburn, Evaluation of Seasonde Hardware diagnostic parameters as performance metrics, UCSB, technical report. [3] S. Cosoli, G. Bolzon, A. Mazzoldi, A Real-Time and Offline Quality Control Methodology for SeaSonde High-Frequency Radar Currents, Journal of Atmospheric and Oceanic Technology, 29, , [4] C. Evans, H. Roarty, M. Smith, J. Kerfoot, S. Glenn, P. Hardik, C. Whelan, Improvement of surface current measurements with spectra reprocessing for 13 MHz SeaSonde systems, OCEANS 2013, 2013, San Diego, pp 1-5. [5] P. Lorente, S. Piedracoba, J. Soto-Navarro, E. Alvarez-Fanjul, Accuracy assessment of high frequency radar current measurements in the Strait of Gibraltar, Journal of Operational Oceanography, 7 (2), pp 59 73, [6] P. Lorente, S. Piedracoba, E. Alvarez-Fanjul, Validation of highfrequency radar ocean surface current observations in the NW of the Iberian Peninsula, Continental Shelf Research 92, pp 1-15, [7] J.D. Paduan, K.C. Kim, M.S. Cook, F.P. Chávez, Calibration and validation of direction-finding high-frequency radar ocean surface current observations, IEEE Journal of Oceanic Engineering, 31 (4), pp , [8] A.J. Abascal, S. Castanedo, R. Medina, I.J. Losada, E. Álvarez-Fanjul, Application of HF radar currents to oil spill modelling, Marine Pollution Bulletin, 58, pp , 2009 [9] M.G. Sotillo, S. Cailleau, P. Lorente, B. Levier, R. Aznar, G. Reffray, A. Amo-Baladrón, E. Alvarez-Fanjul, The MyOcean IBI Ocean Forecast and Reanalysis Systems: Operational products and roadmap to the future Copernicus Service, Journal of Operational Oceanography, 2015 (accepted).