Science and Engineering at The University of Edinburgh |
||||
|
||||
![[University Logo]](/icons/unilogo_25476c.gif "Link to The University of Edinburgh Home Page"){kind=icon}
Model Validation
Validation Data
In a recent paper by Thomas et al (2004), a study of the pCO2 in the North Sea was conducted. The North Sea can be divided into 2 regions - a shallow southern region which does not stratify and a deeper northern region which is stratified in summer. In the northern region measurements of dissolved inorganic carbon (DIC) were taken through the water column for each season giving a seasonal mixing pattern. This is ideal for validating the mixing of the bio-chemistry in the interface modelling tool.Forcing Data
To run the model, meteorological data were needed along with initial conditions for each of the state variables. ECMWF Operational data were used for wind speeds (U10 and V10), relative humidity, air temperature and air pressure at six hourly intervals. Monthly atmospheric pCO2 data were taken from Keeling's data series and scaled for this region using Takahashi et al's (2002) climatology. No data were available for cloud cover so this was set to be 50% at all times. Initial conditions for sea surface temperature were taken from ECMWF but there were no data for initialising the biological or chemical variables so these were set to the standard HadOCC initial conditions (i.e. Nutrient=10, Phytoplankton=0.01, Zooplankton=0.01, Detritus=0.0 (all in units of nitrogen (umol N/m3)), Alkilinity=2396.0 umol/l, DIC=2058.0 umol/l).Results
The simulation was run for 2 years (to allow the mixing to fully develop) with the second year coinciding with the North Sea experiment. The figure shows the observed and modelled results. Because the bio-chemical initial conditions were not correct we could not hope to reproduce the exact values measured by Thomas et al - the aim was to see if the model captured the seasonal mixing pattern. The figure shows that although the remineralisation rate in this region appears to be underestimated by the model, the decrease in DIC in the surface layers due to photosynthesis in summer (red line), and the total mixing through the water column in winter (dark blue) is clearly captured, indicating that the mixing of the bio-chemical variables in the model is functioning correctly.Figure showing validation.
Back to Modelling Tool info.