F.  Große  et  al.:  Looking  beyond  stratification

2531

www.biogeosciences.net/13/2511/2016/

Biogeosciences,  13,  2511-2535,  2016

Appendix  A:  Evaluation  of  the  stratification  and  MLD
criterion
Figure  A1  shows  the  Hovmoller  diagram  of  simulated  T  at
station  North  Dogger  (see  Fig.  2,  region  2)  for  the  year  2007,
including  the  MLD  (dashed  magenta  line)  derived  from  the
simulated  T  field  according  to  Eqs.  (1)  and  (2).  Regarding
the  onset  of  stratification  in  late  March,  it  is  shown  that  the
near-surface  (0-25  m)  T  starts  to  increase  relatively  to  T  in
the  deeper  layers.  r CI it  =  0.5  K  is  reached  on  26  April,  marking ­
  the  beginning  of  the  stratified  period  according  to  Eq.  (1).
The  maximum  vertical  T  gradient  at  the  onset  occurs  in  25  m
depth.  From  that  moment  stratification  according  to  Eq.  (1)
persists  until  31  October,  which  may  represent  a  slight  overestimation ­
  as  the  maximum  T  gradient  is  found  in  60  m  depth
already,  indicating  deep  mixing.
During  the  first  two  months  of  the  stratified  period
(April/May)  the  MLD  shows  stronger  fluctuations  in  terms
of  its  actual  depth.  This  results  from  the  relatively  weak  nearsurface ­
  stratification,  and  thus,  the  stronger  effect  of  mainly
wind-induced  mixing  reaching  depths  of  up  to  35  m.  These
events  are  indicated  by  the  episodic  increase  and  decrease  of
surface  T.  From  early  June  to  end  of  July,  the  MLD  is  less
variable  in  depth  due  to  the  persistent  surface  heating  and  less
strong  wind  events.  In  late  June  a  short-term  decrease  in  surface ­
  T  indicating  enhanced  mixing  occurs  which  also  results
in  a  deepening  of  the  MLD.  From  August  until  the  end  of  the
stratified  period  the  MLD  shows  a  deepening  trend  which  is
caused  by  the  decreasing  surface  heating  and  increasing  wind
activity.

Figure  Al.  Hovmoller  diagram  of  simulated  T  and  MLD  according ­
  to  Eqs.  (1)  and  (2)  at  Cefas  station  North  Dogger  (see  Fig.  2,
region  2)  in  2007.  Depth  levels  represent  the  centre  depth  of  model
layers.
The  main  assumption  behind  the  rather  small  critical  T
difference  of  0.05  K  is,  that  even  if  the  surface  mixed  layer  is
interrupted  due  to  mixing,  this  does  not  necessarily  result  in
a  complete  overturning  of  the  water  column.  Thus,  even  a  minor ­
  difference  in  T  indicates  a  bottom  layer  unaffected  from
vertical  mixing.  Despite  this  significant  difference  to  common ­
  MLD  criteria  (e.g  Kara  et  al.,  2000,  therein  Table  1),
the  criterion  applied  in  this  study  represents  the  stratification
conditions  quite  well.  However,  it  should  be  noted  that  the
end  of  the  stratified  period  may  be  slightly  overestimated.  In
addition,  in  regions  with  a  less  pronounced  onset  of  stratification, ­
  i.e.,  a  less  distinct  increase  in  surface  T,  the  determined
timing  of  the  onset  may  be  slightly  too  early.  The  use  of  the
maximum  T  gradient  to  determine  the  MLD  under  stratified
conditions  yields  reasonable  results,  and  is  closely  related  to
real  conditions  as  the  thermocline  is  defined  as  the  layer  with
the  maximum  T  gradient.