Analyzing Austral Swells in La R´eunionIsland Using Microseismic Noise and Ocean Bottom Data in Fringing Reef Environment E. Cordier (1), E.J. Rindraharisaona (2), G. Barruol (3), F.R. Fontaine (2), A. Gonzalez (2), M. Singh (4) (1) OSU-R`eunion,(2)Laboratoire G`eoSciencesR`eunion(3) Institut de Physique du Globe Paris, (4) Mauritius Oceanography Institute

ABSTRACT

The R´eunionIsland in the South-West Indian Ocean is often affected by austral or cyclonic swells that may have strong impacts on coastal infrastructures. The ReNovRisk project that started in 2017 has been developed by M´et´eo-France and the University of La R´eunionto improve cyclone forecasts in the South-West Indian Ocean, by developing and improving coupled ocean-wave-atmosphere numerical modeling and by increasing direct observations that may help to calibrate and to validate the swell activity. In the present study, we first analyze the microseismic noise recorded by permanent (OVPF (PF) and GEOSCOPE (G) networks) and temporary (Rivi`erede Pluies (ZF) network) seismic stations operating in La R´eunion.Together with microseismic noise, we also analyze data issued from instruments deployed on the ocean bottom in coastal areas: OSSI pressures gauges, sensors and ADCP sensors from which wave height and water currents are derived. These sensors were operated during the period March-May 2017 and will be installed twice per year, with an operational time of 3 months. The sensors were and will be installed in and outside of the fringing reef, at a depths ranging from 1 m at the reef up to 80 m outside. Our results for the austral swells occurring in March-May 2017 show that the swells measured on both sides of the reef have the same patterns, except that the amplitudes outside the lagoon are ten times higher than those inside, clearly showing the strong attenuation of the swells by the coral reef. Comparing the amplitude of the primary microseism (0.05 Hz-0.1 Hz) to swell’s amplitude, we observe a good correlation between the two datasets, confirming it is generated by swell-coast interaction. The Secondary Microseisms (SM) that develop within the storm at several thousands of km in the Austral Ocean clearly arrive 2-4 days before the swell hit the coast of La R´eunionIsland, suggesting that SM may represent good precursors of a swell event.

SEISMIC AND OCEANOGRAPHY STATIONS DISTRIBUTIONS WAVES POWER SPECTRAL DENSITY MIDDLE REEF (LEFT) and OFF REEF (RIGHT) SM (3-10S) vs WAVE HEIGHT MID/OFF REEF

−20˚48'

PRO RIV1 3.0 5 CAM RIV6 RIV1 RMA1 SALI MAID PRA offreef wave height MRF RIV6 RMA2 CAM MVL PRO midreef wave height −21˚00' RMA1 MVL RMA2 2.5 4 0.4 OFF MAID SALI )

m 2.0 −21˚12' (

e

Reunion d 3 0.3 u t i PRA l p 1.5 m a

)

−21˚24' s 0 2 0.2 1 - 3

( 1.0 M Wave height offreef (m) Wave height midreef (m) 55˚00' 55˚12' 55˚24' 55˚36' 55˚48' 56˚00' S

PF (Permanent station, OVPF) ZF (Temporary Network) 1 0.1 Mid_Reef pressure station Off_Reef pressure station 0.5

0.0 0 Mar 19 2017 Apr 02 2017 Apr 16 2017 Apr 30 2017 • The PF is a permanent network oper- ated by the Observatoire volcanologique du Piton de la Fournaise (OVPF), Insti- • Dark gray shows the arrival time of SM, light gray shows arrival time swell tut de Physique du Globe Paris (IPGP). and red background indicates when a SM was recorded but no swell have • Power spectral density of the Gravity (top) and Infragravity waves (bottom) of each plot. arrived at the coast. • The ZF network is a temporary network • The spectrum outside the reef are dominated by the gravity waves, while they are dominated by infragravity waves at the middle of the • SM always recorded before any swell swell hit the coast. belongs to the Rivi`eredes Pluies Project. reef. • Delays between the PM/swell and the SM depends on the source and the • The reef acted as low pass filter for the waves, thus the gravity waves are filtered when crossed the reef. • The OSSI Pressure stations were operated group velocity of the swell. between Mars−May 2017

ATTENUATION OF THE WAVE SPECTRAL ANALYSIS METHODS 0.5 5 PM (10-20s)/SM (3-10) vs WAVE HEIGHT MID-REEF infragravitary offreef sea swell 3.0 • PSD analysis seismic data: (1) Selected 1hr data, with overlapping 50 %, 0.4 very low frequency 4 4*PM (10-20s) (2) Each 1-hr time series record is divided into 13 segments, overlapping by total SM (3-10s) • Figure shows the attenuation by the reef. Straight line: wave height outside the 75%, (3) Transformed into the time-frequency domain using the S transform midreef 2.5 (McNamara & Raymond, (2004)) 0.3 3 reef & continue line, wave height inside the reef. Different colors indicate the 0.4 type of wave.

• PSD analysis waves data: (1) Gravity wave 0.04−0.45 Hz, (2) Infragravitaire ) wave 0.003 − 0.04 Hz and Very Low Frequency: < 0.003 Hz. • The two waves have comparable pattern but it attenuated by ∼10 times when m 2.0

0.2 2 (

hsout ee f e 0.3 crossed the reef ie hs = r . d u

midree f 10 t i l Wave height offreef (m) p 1.5 Wave height midreef (m) 0.1 1 • The infragravity and very low frequency swells are almost zero outside the reef. m a

M 0.2 A /

MICROSEISM POWER SPECTRAL DENSITY M 1.0

0.0 0 P wave height midreef (m)

0.1 2017-03-19 2017-03-26 2017-04-02 2017-04-09 2017-04-16 2017-04-23 2017-04-30 0.5

wave height 0.0 PRIMARY MICROSEISM (10-20s) vs WAVE HEIGHT MID-REEF PRIMARY MICROSEISM (10-20s) vs PRESSURE

0.8 0.8 Apr 20 2017 Apr 23 2017 Apr 26 2017 Apr 29 2017 May 02 2017 Infragravity PM (10-20s) PM (10-20s) Pressure )

m 0.7 Very low frequency 0.7 ( Swell 0.4 16000 e • Blue Arrows indicate the arrival time of the SM and Red arrows show arrival d 0.6

u Total 0.6 t )

i time of the PM/swell. l m p 0.5 0.3 m ( 0.5 14000

a

• SM produced at the noise source and propagate at velocity of ∼3 km/s, thus e d m 0.4 s

u recorded at the seismic stations shortly after they occurred. i t i e 0.4 l

s 0.2 p o 12000 r 0.3 m c • The swell propagate at velocity between 15 and 35 km/h, thus arrive after i Wave height (m) a 0.3

Pressure (Pa) M

the SM. M y 0.2 r

0.1 P a 0.2 10000 m

i • Delays between the SM and the PM/swell depends on the distance between

r 0.1 P source-seismic station and the group velocity of the swell. In this example 0.0 0.1 0.0 8000 the swells hit the cost between ∼2.5 and ∼3 days after the SM recorded at 0.0 the seismic station. 2017-04-13 2017-04-16 2017-04-19 2017-04-22 2017-04-25 2017-04-28 2017-05-01 2017-05-04 • Primary microseism (PM): 0.05 to 0.1 Hz & Generated in shallow water • The PM produced and recorded at the seismic stations almost the same time through pressure variation as the swell hits the coast and can be used as a proxy of the swell amplitude. • The different waves have the same pattern but with different amplitude. 2017-04-13 2017-04-16 2017-04-19 2017-04-22 2017-04-25 2017-04-28 2017-05-01 2017-05-04 • Secondary microseism (SM): 0.1 to 0.33Hz & Occurs in deep and shallow • The SM can be used to predict the swell. waters & Generated by more distance sources • Correlation between the PM (10-20s) amplitudes and wave heights. • Correlation between the PM and sea floor pressurization. • Intensification of microseisms during the occurrence of the swell. • Pearson correlation coefficient: 0.95. • Interactions between the atmosphere and the sea floor verify by the presence of the microseisms.