63

8  Conclusions

The  population  living  on  the  German  North  Sea  coast  is  familiar  with  the  changing  tides  and
the  occurrence  of  storm  surges  with  high  water  levels  (4-5  m  above  NN,  German  ordnance
datum).  Observing  the  weather,  they  know  when  to  expect  a  storm  surge.  External  surges
generated  outside  the  North  Sea  are  either  insignificant  or,  in  the  majority  of  cases,  they  are
part  of  a  storm  surge  event.  Storm  surges  can  be  predicted  with  high  accuracy.  The
phenomenon  of  waves  "coming  out  of  the  blue",  like  the  tsunami  of  December  2004,  has
raised  concerns  among  the  German  population.  The  BSH,  as  the  authority  issuing  the  official
German  water  level  predictions,  felt  compelled  to  investigate  the  occurrence  of  historical
tsunami  in  the  North  Atlantic  Ocean,  and  to  study  the  special  characteristics  of  tsunami
propagating  in  wide,  shallow  shelf  seas  like  the  North  Sea.
Different  events  might  conceivably  cause  a  tsunami  in  the  North  Sea.  The  risk  of  a  volcanic
eruption  and  its  consequences  is  low  or  non-existent  in  the  North  Sea.  Meteoritic  impacts  in
the  North  Sea  are  relatively  unlikely.  Tsunami  research  has  seriously  considered  the
consequences  of  minor  meteoritic  impacts  in  the  North-East  Atlantic.  Near-earth  asteroids
are  subject  to  continuous  monitoring,  however,  and  are  planned  to  be  either  deflected  or
destroyed  if  they  pose  a  real  threat.
There  exists  historical  and  geological  evidence  for  the  impacts  of  an  earthquake  off  Lisbon
and  of  a  slope  failure  at  Storegga  off  the  Norwegian  coast.  Tsunami  whose  primary  cause  is
an  earthquake  pose  only  a  low  threat  to  the  German  North  Sea  coast.  A  more  realistic
scenario  is  an  earthquake  of  low  magnitude  triggering  a  major  slope  failure  in  the  North-East
Atlantic.  It  is  not  possible  to  estimate  the  probability  of  occurrence  of  slope  failures
attributable  either  to  this  or  to  another  cause.  The  most  recent  slope  failure  that  had  an
impact  on  the  North  Sea  was  an  event  off  Newfoundland  in  1929.  However,  no  evidence  for
a  tsunami  has  been  found  in  historical  records  of  water  levels  at  Cuxhaven.
The  BSH  is  operating  a  model  system  consisting  of  non-linear,  hydrostatic  models.  It
comprises  two  three-dimensional  baroclinic  models  and  one  two-dimensional  barotropic
model  for  the  North  and  Baltic  Seas,  as  well  as  a  two-dimensional  barotropic  model  for  the
North-East  Atlantic  Ocean.  The  question  as  to  whether  this  modelling  system,  which  is  used
operationally  for  storm  surge  predictions  at  the  BSH,  is  suitable  for  simulating  waves
comparable  to  a  tsunami  has  been  discussed.  Tsunami  have  frequencies  and  wave  lengths
ranging  between  those  of  sea/swell  and  tides/storm  surges.  There  is  no  model  that  is  equally
suited  to  all  uses,  but  each  model  is  based  on  reasonable  assumptions  for  the  particular
process  to  be  simulated.  Wave  models  and  tide/storm  surge  models  differ  fundamentally  in
this  respect.  Neither  of  these  models  is  suitable  for  tsunami  simulation  without  prior
modification.  Nevertheless,  only  minor  modifications  were  made  to  existing  models  at  the
BSH  and  other  institutions  before  running  simulations  of  the  propagation  of  tsunami-type
signals  in  the  North-East  Atlantic  and  North  Sea.  Two-dimensional  barotropic  models  were
used  almost  exclusively.  The  North  Sea  model  (North  Sea  2  km)  used  for  the  simulations  in
section  7  is  a  barotropic  non-linear  hydrostatic  model  and  is  thus  suitable  for  computing  the
propagation,  modification,  and  attenuation  of  tsunami  with  periods  of  1800  s  or  more  on  the
shelf.  The  North-East  Atlantic  model  of  the  BSH  was  found  to  have  limited  suitability  for
simulating  the  propagation  of  medium-length  waves  in  the  deep  ocean  and  their  modification
on  the  continental  shelf.  The  BSH  models  in  their  current  configuration  are  not  suitable  to
simulate  near-shore  processes.
According  to  the  model  simulations,  the  prescribed  signal  travels  about  seven  hours  from  the
northern  boundary  of  the  North  Sea  (e.g.  Shetland  Islands)  to  Esbjerg,  and  about  nine  hours
to  Cuxhaven.  A  secondary  signal  that  forms  in  the  North  Sea,  mainly  due  to  diffraction  and
reflection  on  coastal  features,  is  higher  in  Cuxhaven  than  the  primary  signal  and  arrives  a
few  hours  later.
The  standard  signal  used  in  the  simulations  (3  positive  single  waves,  T  1,800  s,  H5m,  from
the  north)  leads  to  a  water  level  of  0.5  m  in  Cuxhaven  for  the  primary  signal,  and  1  m  for  the