166

Geo-Mar  Lett  (2017)  37:163-170

*£)  Springer

algorithm  for  natural  oil  dispersion  in  PADM  was  modified  assuming ­
  that  at  each  time  step  100%  of  surfaced  oil  becomes
dispersed,  regardless  of  its  viscosity.
For  each  element  of  the  ensemble  of  simulations  (both  with
and  without  the  assumed  use  of  dispersants),  the  percentage  of
the  released  pollutant  that  crossed  sensitive  tidal  flats  (blue  areas
in  Fig.  2)  was  evaluated  at  hourly  time  steps  for  the  7  day  drift
period.  Additionally,  times  until  the  first  10%  of  oil  polluting
tidal  flats  arrived  were  monitored.  Given  the  large  ensemble  of
pair-wise  simulations  with  and  without  dispersant  application,
only  those  hypothetical  releases  were  counted  as  being  successfully ­
  combated  using  chemical  dispersants  that  fulfil  two
conditions;
1.  Mean  significant  wave  heights  prevailing  at  the  respective
release  position  were  within  the  range  of  0.5-3.0  m  during
the  first  6  h  after  the  assumed  accident.
2.  The  application  of  dispersants  decreased  the  amount  of  oil
entering  sensitive  areas  by  at  least  95%.
The  first  condition  refers  to  Allen  (1988,  cited  in  Fingas
2004),  who  specified  wave  heights  between  about  0.5  and  3  m
as  a  prerequisite  for  successful  dispersant  application.  Although

chemical  dispersion  needs  a  certain  level  of  wave  energy  for
mixing,  the  application  of  dispersants  is  unnecessary  and  ineffective ­
  if  the  wave  energy  is  high  enough  for  sufficient  natural
dispersion.  The  ocean  wave  information  was  taken  from  hindcast
simulations  (Groll  and  Weisse  2016)  for  the  North  Sea  within  the
framework  of  coastDat  (Weisse  et  al.  2015).  The  underlying
model  was  version  4.5.4  of  the  third-generation  spectral  wave
model  WAM  (WAMDI-Group  1988).  The  model  calculated  directional ­
  wave  spectra  on  a  0.050x0.075°  latitude-longitude  grid
(approx.  3x3  NM)  with  35  frequencies  and  24  directions.  Wave
parameters,  such  as  the  significant  wave  height,  were  stored  on
an  hourly  basis.  At  the  lateral  boundaries,  wave  spectra  from  a
North  Atlantic  simulation  were  used  to  take  into  account  wave
fields  generated  outside  the  North  Sea.  Wind  forcing  for  the  wave
data  originated  from  a  regional  atmosphere  hindcast  simulation
using  the  NCEP/NCAR  reanalysis  (Kistler  et  al.  2001)  with  a
spectral  nudging  scheme  (Geyer  2014).
The  second  of  the  above  conditions  seems  to  be  very  restrictive, ­
  but  it  must  be  seen  in  the  context  of  the  assumption  of  a
100%  effective  dispersant,  ff  efficacies  depending  on  oil  type,
weathering  processes  and  environmental  conditions  were  taken
into  account,  then  the  threshold  of  pollution  reduction  would
probably  have  to  be  adjusted  differently.

Fig.  2  Probabilities  that
application  of  a  100%  effective
chemical  dispersant  would
prevent  at  least  95%  of  the
released  oil  from  entering
sensitive  areas  (blue),  considering
a  maximum  drift  time  of  7  days.
Lines  are  the  10  m  (grey)  and  20
m  (blue)  depth  contours

7°0'0"E

7°30'0"E

8°0'0"E

8°30'0"E

9°0'0"E

Probability
successful
dispersant
application
65%
55%
45%
35%
25%
15%
5%