e. the MEDAR and NODC datasets), and the results

are illustrated in Figure 5. The modelled seasonal and interannual variations in the surface temperatures and salinities realistically follow the observations. However, the observations indicate periods of high surface salinity that are underestimated by the model. Yearly averaged temperatures and salinities for the surface (0–150 m), intermediate (150–600 m) and deep (below 600 m) layers are presented in Figure 6. The modelled surface temperature follows the reanalysed temperature closely with a correlation (R) of 0.98 and a standard error of 0.7 ° C. The mean modelled and reanalysed surface temperatures over the study period were calculated BIBW2992 to be 20.65 ± 3.7 and 20.3 ± 3.7 ° C respectively. Average modelled and reanalysed surface salinities were calculated to be 38.34 ± 0.14 and 38.39 ± 0.14 PSU respectively, with a correlation of 0.6 and a standard error of 0.11 PSU. In the intermediate layer, the yearly simulated temperate and salinity are over- and underestimated by 0.7 ° C and –0.37 PSU respectively, indicating that local processes such as deep-water convection need to be considered. Moreover, the MEDAR data set shows an insignificant trend of intermediate BMS-354825 solubility dmso layer salinity content, while our model results indicate a small negative

trend. This could be explained by the horizontal averaging for the whole EMB, which leads to reduced deep water formation. However, there is only a negligible bias between the simulated and calculated deep layer temperatures/salinities. To investigate the heat balance in some detail, PROBE-EMB modelled

evaporation rates were compared with meteorological modelled evaporation data. This is an important test of the forcing fields and the modelling, as the evaporation rates were calculated independently using both methods. For the meteorological data, we used the NCEP reanalysed data, an independent dataset. Figure 7 depicts the monthly and yearly average values of modelled evaporation rates based on the PROBE-EMB simulations. Figure 8 depicts the scatterplot Temsirolimus in vivo of modelled and NCEP reanalysed evaporation rates for the EMB. Over the study period, modelled evaporation rates ranged from 0.2 to 1.3 mm day− 1, with an average of 3.1 ± 1.5 mm day− 1. The monthly average evaporation rates over the study period ranged from 4.95 ± 1.8 mm day− 1 in August 1985 to 1.31 ± 0.45 mm day− 1 in May 1993, while the yearly average evaporation rates ranged from 3.26 mm day− 1 in 1961 to 2.74 mm day− 1 in 1972. The reanalysed and modelled monthly evaporation rates agreed fairly well, with a correlation of 0.76 and a standard error of 0.5 mm day− 1. The PROBE-EMB model results for surface temperature, salinity and evaporation rates were also calculated as monthly means (Figure 9): the monthly average surface temperature ranged from 15.8 ± 0.32 ° C in March to 25.98 ± 0.