208

P.  Poli  et  al.:  SVP-BRST:  genesis,  design,  and  initial  results

Ocean  Sci.,  15,199-214,  2019

www.ocean-sci.net/15/199/2019/

Table  4.  Similar  to  Table  1  but  for  two  prototype  SVP-BRST  buoys  (each  buoy  is  fitted  with  a  HRSST  and  static  pressure  probe).

WMO  identifier

Deployment  basin

HRSST  sensor  model  and  S/N

Start  date

End  date

6102622

Mediterranean  Sea

Digital  MoSens  4658

26/04/2018

12/06/2018

6102623

Mediterranean  Sea

Digital  MoSens  4656

26/04/2018

-

Figure  6.  Sketch  of  the  SVP-BRST  (for  the  drogue,  only  the  tether
attachment  is  shown  here),  with  the  HRSST  sensor  unplugged
shown  in  zoom  panel  (b).  Note  each  SST  sensor  is  protected  from
solar  radiation  by  a  cap.

for  investigation.  Furthermore,  the  distribution  of  SST  observed ­
  within  the  5  min  is  transmitted  at  coarse  resolution
(10th  percentile,  30th  percentile,  50th  percentile  or  median,
70th  percentile,  and  90th  percentile).  This  non-parametric  information ­
  makes  no  assumption  about  the  shape  of  the  SST
distribution:  it  can  be  used  to  drive  an  ensemble  of  applications, ­
  rather  than  using  solely  the  mean  SST,  and  to  assess,
for  example,  whether  the  SST  distribution  is  symmetric.
Second,  the  HRSST  sensor  is  removable  from  the  buoy
with  simple  tools  (see  Fig.  6),  and  includes  a  co-located  pressure ­
  sensor  that  allows  reporting  static  pressure  with  an  accuracy ­
  of  5  cm  in  calm  waters.  Even  if  the  instrument  is  affected ­
  by  accelerations  in  wavy  conditions,  and  the  depth  is
only  valid  in  rather  calm  conditions  (when  the  sensor  depth  is
already  known  by  design),  information  can  be  derived  about
the  hydrostatic  water  pressure  variability  (within  5  min).
Third,  all  SST  sensors  are  insulated  to  shield  them  from
unwanted  effects  caused  by  the  non-water  surrounding  environment. ­
  This  aims  to  avoid,  for  the  SST  sensors,  exchanges
by  conduction  with  the  buoy  hull,  exchanges  by  radiation
with  the  Sun  and  the  atmosphere,  and  radio  interference  from
the  buoy  electronic  board  and  antenna.  This  is  done  in  practice ­
  by  using,  respectively,  insulating  material  between  the
sensor  and  the  buoy,  a  small  cap  to  shield  the  SST  sensor

from  radiation,  and  a  metal  plate  underneath  the  buoy  electronic ­
  board  and  antenna.
Fourth,  the  HRSST  sensor  is  defined  with  a  calibrating
housing  and  protocol.  Calibration  coefficients  are  determined
for  each  HRSST  sensor  individually  so  that  their  expanded
calibration  uncertainty  can  be  assessed.  These  uncertainties
are  calculated  according  to  the  Guide  For  Uncertainty  of
Measurement  (BIPM,  2008).  They  are  found  to  be  smaller
than  0.01  K  for  each  buoy.  Response  time  and  systematic  errors ­
  related  to  the  integration  in  the  buoy  have  been  assessed
on  two  prototypes.  The  details  of  these  laboratory  measurements ­
  will  be  the  subject  of  another  paper.
4  Results
Initial  testing  was  conducted  in  the  Brest  area  (see  the  Supplement). ­
  The  results  presented  hereafter  are  based  on  data
collected  by  the  two  prototypes  in  the  Mediterranean  Sea  between ­
  27  April  and  11  June.  The  data  are  available  in  open
access  (see  the  section  on  data  availability).
4.1  Deployment
Two  SVP-BRST  prototypes  were  deployed,  as  shown  in  Table ­
  4.  At  the  time  of  writing,  the  second  prototype  is  still
operating.  Before  deployment  for  release,  the  buoys  were
deployed  briefly  on  23  April  for  comparison  in  the  seawater ­
  with  an  SBE-35  thermometer.  The  SST  differences  were
then  found  to  be  —0.006  K  for  one  buoy  and  —0.001  K  for
the  other  buoy,  thereby  meeting  the  0.01  K  claimed  uncertainty. ­
  In  comparison,  the  SST  difference  between  the  regular ­
  (or  analogue)  SST  sensor  with  the  SBE-35  was  found  to
be  —  0.05  K  (for  both  buoys).
4.2  Analysis  of  the  data  collected  at  sea
Once  deployed  on  26  April  2018,  the  buoys  have  followed
the  tracks  shown  in  Fig.  7.  The  separation  distance  between
the  two  buoys,  initially  under  1km,  remained  under  10  km
until  23  May.  After  that,  the  two  buoys  quickly  diverged  until
the  first  one  ran  ashore.
The  buoy  reports  data  to  shore  using  Iridium  according  to
a  binary  data  format  number  no.  091  documented  by  Blouch
et  al.  (2018).  Besides  the  usual  parameters  reported  by  SVP-B
  buoys  (position,  time,  strain  gauge,  air  pressure,  analogue
SST,  and  other  technical  parameters  such  as  battery  voltage
and  GNSS  TTFF),  one  notes  the  following  key  additions:  the