F. Große et al.: Looking beyond stratification 
2515 
www.biogeosciences.net/13/2511/2016/ 
Biogeosciences, 13, 2511-2535, 2016 
is provided by Chen et al. (2013). The HAMSOM simulation 
was carried out with a 10 min time step. 
ECOHAM was run off-line with a time step of 30 min 
using the 24 h averages of the hydrographic and hydrody 
namic fields generated by HAMSOM. In the model setup 
used, short wave radiation is attributed to the first layer (sur 
face) only and the specific effect of light attenuation due to 
SPM and planktonic self-shading on the thermal structure 
is not taken into account. A sensitivity study allowing for 
deeper light penetration and feedback on the thermal struc 
ture confirmed this effect to be only of minor importance (not 
shown). 
For the biogeochemical state variables a climatology of 
depth-dependent monthly averages was prescribed at the 
boundaries and solely for DIC yearly changing data were 
provided (Lorkowski et ah, 2012). To include the effect of 
SPM on the light climate, a daily climatology from Heath 
et al. (2002) was used. Data for atmospheric N deposition 
were compiled using a hybrid approach. This was required 
since the overall simulation period (1977-2012) exceeds the 
period of data available from the EMEP (Cooperative pro 
gram for monitoring and evaluation of the long-range trans 
missions of air pollutants in Europe) model (1995-2012). 
First, the EMEP results for total deposition of oxidised 
(NO*) and reduced nitrogen (NH3) were interpolated to the 
model grid. Second, we calculated the average annual depo 
sition rates for the NO, and NH3 for each grid cell, based 
on the 1995-2012 EMEP data. The resulting spatially re 
solved arrays of average deposition rates were subsequently 
normalised by the spatial average of the entire domain to 
yield the spatially resolved anomaly fields. Finally, gridded 
deposition rates for individual years were obtained using 
(1) the gridded anomaly fields, (2) EMEP’s spatially aver 
aged (over our model domain) deposition rates for year 2005, 
and (3) long-term trends (normalised towards year 2005) for 
the temporal evolution of European emissions of NO, and 
NH3 (Fig. 2 in Schôpp et ah, 2003). The output of the bio 
geochemical simulation was stored as daily values (cumu 
lative fluxes, state variable snapshots) for the entire domain 
and simulation period. 
2.2 Extracting stratification parameters from model 
results 
Stratification constitutes the prerequisite for the potential 
evolution of low O2 conditions in the North Sea. In this study 
(1) its duration and (2) the mixed layer depth (MLD) are used 
to describe stratification. Seasonal stratification in the North 
Sea is mainly T-driven (Burt et ah, 2014), except for the re 
gions of haline stratification along the Norwegian coast. As 
observations do not cover the entire model domain and simu 
lation period we determined the duration of stratification and 
the MLD from the simulation results. For this purpose we 
developed a simple 2-step algorithm based on T. First, the 
stratified period is determined using a temperature difference 
criterion: 
5’sii.h (x,y,f) 
1 AT\°_ H (x, y, t) > 0.05K 
0 otherwise 
(1) 
S'strat is a switch defining if a water column at location (x, y) 
and time t is stratified (/»strat = 1) or not (S^trat = 0) de 
pending on the temperature difference A T between the sur 
face and bottom depth H. The critical temperature differ 
ence A7( M | = 0.05 K was determined by evaluating different 
Aigrit against the temporal evolution of simulated bottom O2 
at different locations within the model domain. In addition, 
periods of stratified conditions are only considered as such, 
if they last for at least 5 days without any interruption. Oth 
erwise bottom waters are considered to be ventilated again. 
In the second step, in the case of stratification the MLD 
of a model water column is determined using the vertical T 
gradient AT/Az: 
MLD(x, y, t) 
D(max(AT/Az)) 
0 
5’slral (x. V. t ) — 1 
otherwise 
(2) 
AT /Az is calculated for each grid cell interface within the 
considered water column. A T represents the T difference be 
tween two vertically adjacent model layers and Az represents 
the distance between the centre points of these two grid cells. 
D is then defined as the depth level of the interface where 
AT/Az has its maximum. As the described stratification and 
MLD criterion differ significantly from common MLD crite 
ria (e.g., Table 1 in Kara et al., 2000), an evaluation is pro 
vided in Appendix A. 
2.3 Validation data 
For the validation of the model results we used observation 
data from different sources. The data sets can be subdivided 
into two types: (1) temporally resolved, localised data and 
(2) spatially resolved “snapshots”. The first type was used 
for the validation of the seasonal evolution of O2, whereas 
the second type was used to validate the general spatial pat 
terns and year-to-year variability of bottom O2 during late 
summer. 
2.3.1 Localised, temporally resolved data - 
Cefas-SmartBuoy and MARNET 
Cefas operates a network of SmartBuoys to provide au 
tonomous in situ measurements of physical, chemical and 
biological parameters (Mills et al., 2005). A SmartBuoy was 
located directly north of the Dogger Bank (“North Dogger”) 
at 55°41 / N, 2° 16.80'E (see Fig. 2, region 2) between 24 
February 2007 to 15 September 2008 in 85 m water depth 
(Greenwood et al., 2010). O2 concentrations were contin 
uously recorded with a frequency of 5 Hz at 31 and 85 m. 
These autonomous O2 measurements were corrected for