Comparisons of several technologies at the Spanish test site

1
Comparisons of several technologies at the
Spanish test site
Belén Martín Begoña Pérez
2
Introduction

• Original aim upgrading of REDMAR network
• Test site established by PE within ESEAS-RI
  project
• Equipments acquired by PE and borrowed from
  several companies and institutions

3
Test site Port of Vilagarcía

• Mean tidal range (max. 4.2 m)
• Relatively sheltered location yet submitted to
  varying atmospheric conditions
• 24 h surveillance by harbour personnel

4
Port of Vilagarcía
5
Tide Gauges Types

• Acoustic AQUATRAK / SRD
• Radar MIROS / RADAC / SEBA / VEGA (Geonica)
• Pressure AANDERAA (discarded) / SEABIRD
• Float THALIMEDES (discarded)
• Total 7 instruments

6
Radar
7
First results
8
Criteria

• SRD gauge is the reference
• Sampling strategy as similar as possible to SRD
• Focus on the time series provided by the tide
  gauges QUALITY of data

9
Preparation of data
1 - Change of units
2 - Sign inverted (when the measurement indicates
distance to water)
3 - Substraction of mean value of the raw data
10
Sources of error

• Data from all the sensors are recorded by
  independent data loggers or computers, that have
  their own clock. There can arise time differences
  that must be compensated for. In certain
  occasions data have been shifted.
• Impossible to average over the same period
  exactly due to differences in systems of
  measurement.
• AQUATRAK - Effect of temperature in the tube not
  considered
• SEABIRD - The tidal gauge only records
  temperature, pressure is obtained from a nearby
  station and salinity is not known.
• RADAC - The sensor was not properly installed and
  that implies a loss of valid data which
  influences final results

11
Variables evaluated

• Correlation
• Scatter diagram
• slope

• Root Mean Square Error (RMS)

12
Periods of comparison
13
MARCH - Correlation

• Coefficient of determination R2

14
MARCH - Linear regression

• Slope

15
MARCH - Linear regression

• Slope

16
MARCH - Deviation

• Root mean square error (RMS, cm)

17
MAY/JUNE - Correlation

• Coefficient of determination

18
MAY/JUNE - Linear regression

• slope

19
MAY/JUNE - Deviation

• Root mean square error (RMS, cm)

20
CONCLUSIONS

• All equipments studied have shown a good
  performance and provided data of enough quality
  during this first stage of the experiment.
  However, longer time series are needed to perform
  a better evaluation.

21
FUTURE WORK

• New equipment (bubbler from POL)
• Harmonic analysis
• Study of trends (mean sea level)
• Cross-Spectral analysis
• Detailed study of maintenance and calibration
  requirements