Die  Küste,  81  (2014),  273-290

278

of  the  former  V3-NOKU-setup  calculated  by  the  previous  version  of  the  BSHcmod
model,  which  covers  die  14  years  from  2000  to  2013.  The  resolution  of  that  setup  was
half  of  the  resolution  of  the  currently  used  V4  NOKU-setup.  Data  from  die  V4-NOKUsetup
  calculated  by  die  BSHcmod  V4  is  archived  from  2008  onwards.
Because  bodi  die  NA-  and  the  NOKU-setup  of  die  described  BSH-HBM  model  system ­
  are  still  in  pre-operational  mode,  only  results  from  the  EL-setup  calculated  by  HBM
is  archived  at  die  moment.  From  diis  setup  die  output  is  available  since  April  2013.
3  Validation
A  quantitative  comparison  of  model  results  with  different  kinds  of  observations  —  often
referred  to  as  validation  -  is  an  important  step  in  the  operational  model  development  cycle. ­
  Even  though  the  model  version  presented  here  is  still  at  a  pre-operational  stage  and
some  calibration  steps  need  further  iteration,  die  results  presented  below  already  give  an
estimate  of  die  lower  limit  of  the  quality  of  the  upcoming  operational  model  version.
The  year  2008  was  chosen  as  the  main  validation  period  because  a  comprehensive
observational  data  set  was  already  available  for  diis  year.  A  hindcast  run  initialized  in
November  2007  was  carried  out  as  basis  for  die  validation.  Some  results  of  die  validation
of  water  level,  currents,  water  temperature,  salinity  and  sea  ice  will  be  presented  in  this
section.
3.1  Water  level
To  analyse  die  simulated  water  levels  it  is  sensible  to  split  die  analysis  into  two  parts,  respectively ­
  subsections:  The  first  one  on  die  Nordi  Sea,  where  tides  are  die  main  component ­
  of  the  local  sea  level  elevation,  so  tiiat  die  quality  of  die  model  output  depends
mainly  on  die  quality  of  the  simulated  tides.  Therefore  botii  tides  and  total  water  levels
were  analysed  in  this  subsection.  The  second  subsection  is  focused  on  the  Baltic  Sea,
where  tides  are  virtually  absent  and  only  die  total  water  levels  are  considered.
3.1.1  North  Sea
As  outlined  in  section  2.3  die  tidal  boundary  conditions  in  die  applied  NOKU-setup  are
based  on  die  19  dominating  tidal  constituents.  Moreover  die  validation  period  is  only  one
year,  so  tiiat  a  complete  tidal  analysis  of  die  data  is  not  feasible.  Instead  die  analysis  will
be  restricted  to  die  two  dominating  semi-diurnal  tidal  constituents  M2  and  S2.  A  harmonic ­
  analysis  of  die  modeled  data  yields  the  results  given  in  Tab.  1  when  compared  with
harmonic  constants  analysed  from  observations  by  the  French  Service  Hydrographique  et
Océanographique  de  la  Marine  (SHOM  1982).