Data series corresponds to sea‐level rise recorded by the tide gauge from 9.7.2009 to 31.12.2012 with a sampling frequency of 5 minutes. Fig. 4: Observed sea level time series in Pasajes station: (1) Whole time period covered. (2) One month of records. (3) Zoom over a 8‐days time windows. DATA Fig. 5: Amplitude spectrum of the time series over the tidal bands of 2 cycles per day (left) and 1 cycle per day (right). DETERMINATION OF THE TIDAL HARMONIC CONSTANTS AND TIDE PREDICTION FOR PASAJES HARBOUR (SPAIN) M. Benavent (1) , A. Martin (1) , M. J. Sevilla (1) , E. J. Vélez (2) , J. Zurutuza (3) (1) Fac. CC. Matemáticas, Universidad Complutense de Madrid, Spain (2) Departamento de Geodesia de la Sociedad de Ciencias Aranzadi, Spain (3) Instituto de Geociencias (CSIC‐UCM), Madrid, spain On February 2007, a tide gauge station (TG) was installed in the Pasajes harbour (Bay of Biscay, northern Spain). In addition, a permanent GNSS station is connected to the TG benchmarks allowing us to study local vertical crustal motions and sea level changes. Ancillary data such as atmospheric pressure, air temperature and humidity are also measured in the station in order to correct meteorological phenomena affecting the sea level observations and to support the GNSS analysis. The aim of this work is to study three years (2010‐2012) of sea level records in the Pasajes station. We have performed the harmonic analysis of the time series by means of a set of procedures and programs. The results obtained, the main tidal constituents, are analyzed through the study of the residuals (observed minus predicted tide). As a consequence of this study we provide more accurate harmonic constants for the main tidal waves in Pasajes harbour and tide prediction for upcoming months. LOCATION OF THE TIDE GAUGE STATION Fig. 1: Location maps of the tide gauge in Pasajes Harbour (latitude 43° 19ȇ 18ȇȇ.4N, longitude 1° 55ȇ 53ȇȇ. 5W). Images obtained with GEODAS software and Google Earth. The station is placed in Pasajes Harbour, inside the dependencies of AZTI‐ Tecnalia Technological Research Centre, ensuring good conditions of installation and maintenance. A few meters of the tide gauge station is located a GPS station belonging to the Active Network of the Provincial Council of Gipuzkoa. RESULTS Tidal Harmonic Constants in Pasajes Station INSTRUMENTATION AND INSTALLATION Fig. 2: Instrumentation installed on the station: (1) Depth sensor Digiquartz, Paroscientific Inc. model 8DP015‐GV‐2. (2) Intelligent Electronics, Paroscientific Inc. model 735. (3) Datalogger CR800 (Campbell Scientific). Fig. 3: (4) Station location photograph. (5) External box with datalogger and power supply unit. (6) Detail of the installation of the tide gauge. The installation, carried out on February 2007, is characterized by a double tube system, the external one anchored to the harbour wall and the sensor fixed within the inner tube at a depth of 7 m from the level of the dock. (1) (2) (3) (5) (4) (6) References • Foreman, M.G.G., (1977, revised 2004): ȈManual for tidal heights analysis and predictionȈ. Pacific Marine Science Report, 77‐10. • Foreman, M.G.G., J.Y. Cherniawsky and V.A. Ballantyne, 2009. Versatile Harmonic Tidal Analysis: Improvements and Applications. J. Atmos. Oceanic Technol., 26, 806–817. doi: 10.1175/2008JTECHO615.1. • Sevilla, M.J., Zurutuza, J. and A. Martin, 2011. Three Years of Tide Gauge Measurements in the Pasajes Harbour. FIG Working Week 2011, Marrakech, Morocco, 18‐22 May 2011. • Venedikov, A., Arnoso, J., Vieira, R., 2003. VAV: A program for tidal data processing. Comput. Geosci. 29, 487–502. • Venedikov, A., Arnoso, J., Vieira, R., 2005. New version of program VAV for tidal data processing. Comput. Geosci. 31, 667–669. Acknowledgements This work has been supported by the research project ‘Redes Geodésicas y GPS’ developed by GEOLAN DONOSTI S.L company and Complutense University of Madrid for the Provincial Council of Gipuzkoa. Fig. 6: Tidal residuals obtained in Pasajes station from the tidal harmonic analysis using VAV (blue curve) and VERSATILE_TIDANA (red curve), for the observing period (left) and a period of one month (right). HARMONIC ANALYSIS OF THE TIME SERIES WITH VERSATILE_TIDANA Improvements of VERSATILE harmonic TIDal ANAlysis (Foreman et al., 2009) software to traditional methods are: Analysis of randomly sampled and multiyear data. Direct incorporation of nodal corrections and inference in the least‐squares fit. Resolution of the overdetermined matrix by means the Singular Value Decomposition (SVD) techniques. Calculation of covariance matrices and correlation coefficients, useful for choosing the best set of constituents to be included in the the analysis. Time series analysis methodology with VERSATILE_TIDANA for Pasajes tide gauge We have carried out exhaustive studies to determine the best analysis options. These include: Constituents selection: 38 diurnal and semidiurnal tidal waves and 26 shallow‐water constituents were finally included. Comparison between VERSATILE_TIDANA and traditional harmonic analysis (Foreman, 1977) using the SLPR2 software (University of Hawaii Sea Level Center). Results obtained in the final analysis are showed in Table 1. HARMONIC ANALYSIS OF THE TIME SERIES WITH VAV Harmonic analysis of tide gauge measurements with VAV software (Venedikov et al., 2005, 2003) consists in the transformation of the data from the time domain into a time/frequency domain through a filtration process. The least square algorithm is applied taking into account the colored character of the noise. This allows getting frequency dependent estimates of the precision. Other features of VAV are: An appropriate model of the tidal signal, which allows all theoretical tides to be taken into account. Ability to process data with arbitrary time step and gaps, without any interpolation. Automatic search of anomalies and flexible model of the drift. Time series analysis methodology with VAV for Pasajes tide gauge Constituents selection: A variant of grouping that includes 30 diurnal and semidiurnal tidal waves was selected. Partition of the data in 24‐hours time windows and approximation of the drift by polynomials of power 1 in each separate segment. Results obtained in the final analysis are showed in Table 1. Table 1: Amplitudes (in cm) and phases (in degree, with respect to Greenwich) of the main tidal waves obtained from the harmonic analysis using VERSATILE_TIDANA and VAV. The average of the absolute values of the differences between the predicted tides, obtained from the tidal harmonic analysis with VERSATILE_TIDANA, respect to the de observed tides is 9.7 cm (standard deviation 8.2 cm). With VAV the average of the differences is reduced to 4.1 cm (standard deviation 3.6 cm). Therefore, we select the tidal constants in Pasajes harbour obtained with VAV program. Main differences in amplitudes and phases of the tidal waves in the diurnal and semidiurnal bands obtained using VAV and VERSATILE_TIDANAare: For amplitudes: 0.03 cm (0.45%) for O1; 0.18 cm (2.87%) for K1 , 0.42 cm (1.45%) for N2, 0.146 cm (0.11%) for M2 and 0.11 cm (0.23%) for S2. For phases: 0.48° for O1; 0.34° for K1 , 2.97° for N2, 0.07° for M2 and 0.00° for S2. For the main constituents, differences obtained between VERSATILE_TIDANA and VAV do not exceed significantly the estimated errors in their determination. Comparison of predicted and observed tides (1) (2) (3)