Abstract.A 20-year retrospective reanalysis of the ocean state in the Baltic Sea is constructed by assimilating available historical temperature and salinity profiles into an operational numerical model with three-dimensional variational (3DVAR) method. To determine the accuracy of the reanalysis, the authors present a series of comparisons to independent observations on a monthly mean basis.In the reanalysis, temperature (T) and salinity (S) fit better with independent measurements than the free run at different depths. Overall, the mean biases of temperature and salinity for the 20 year period are reduced by 0.32 • C and 0.34 psu, respectively. Similarly, the mean root mean square error (RMSE) is decreased by 0.35 • C for temperature and 0.3 psu for salinity compared to the free run. The modeled sea surface temperature, which is mainly controlled by the weather forcing, shows the least improvements due to sparse in situ observations. Deep layers, on the other hand, witness significant and stable model error improvements. In particular, the salinity related to saline water intrusions into the Baltic Proper is largely improved in the reanalysis. The major inflow events such as in 1993 and 2003 are captured more accurately as the model salinity in the bottom layer is increased by 2-3 psu. Compared to independent sea level at 14 tide gauge stations, the correlation between model and observation is increased by 2 %-5 %, while the RMSE is generally reduced by 10 cm. It is found that the reduction of RMSE comes mainly from the reduction of mean bias. In addition, the changes in density induced by the assimilation of T/S contribute little to the barotropic transport in the shallow Danish Transition zone.The mixed layer depth exhibits strong seasonal variations in the Baltic Sea. The basin-averaged value is about 10 m in summer and 30 m in winter. By comparison, the assimilation induces a change of 20 m to the mixed layer depth in deep waters and wintertime, whereas small changes of about 2 m occur in summer and shallow waters. It is related to the strong heating in summer and the dominant role of the surface forcing in shallow water, which largely offset the effect of the assimilation.

Abstract. The spatial averaged correlations are presented in 1.5 • × 1.5 • bins for the North and Baltic Sea region. The averaged correlations are computed based on the proxy ocean data generated by the operational forecast model of Danish Meteorology Institute (DMI). It is shown that the spatial distribution of the averaged correlations could reflect the overall influence of the local atmospheric forcing, complex topography, coastlines, boundary and bottom effect, etc. Comparisons with the satellite SST data demonstrate that the proxy ocean data reproduce realistic results at the surface. Based on the spatial bin-averaged correlations, a general correlation model is assumed to approximate the spatial and temporal correlation structure. Parameters of the correlation model are obtained on the standard Levitus levels. It is found that the correlation model is not the typical Guaussian-type function. For instance, the exponents of the correlation model vary in the longitudinal direction from 0.75 at the surface to 1.33 at the depth of 250 m for temperature. For salinity, the temporal correlation can be approximated with an exponential function.Two complementary quality-indicators, effective coverage rate and "explained" variance, are defined based on the correlation models obtained above. The two indicators are able to identify the "influence area" of the information content in a given observation network and the relative importance of observations at different locations. By these indicators, the 3-D temperature and salinity observational networks are assessed in the Baltic Sea and North Sea for the period 2004-2006. It is found that the surface level is more effectively covered than the deep waters with existing networks. In addition, the Belt Sea and the Baltic Proper also show good coverage for both temperature and salinity. However, more observations are required in the Norwegian Trench and Kattegat. In the

Abstract. This paper describes the implementation and evaluation of a pre-operational three dimensional variational (3DVAR) data assimilation system for the North/Baltic Sea. Univariate analysis for both temperature and salinity is applied in a 3DVAR scheme in which the horizontal component of the background error covariance is modeled by an isotropic recursive filter (IRF) and the vertical component is represented by dominant Empirical Orthogonal Functions (EOFs). Observations of temperature and salinity (T/S) profiles in the North/Baltic Sea are assimilated in the year of 2005. Effect of the 3DVAR scheme is assessed by a comparison between data assimilation run and control run. The statistical analysis indicates that the model simulation is significantly improved with the 3DVAR scheme. On average, the root mean square errors (RMSE) of temperature and salinity are reduced by 0.2 • C and 0.25 psu in the North/Baltic Sea. In addition, the bias of temperature and salinity is also decreased by 0.1 • C and 0.2 psu, respectively. Starting from an analyzed initial state, one month simulation without assimilation is carried out with the aim of examining the persistence of the initial impact. It is shown that the assimilated initial state can impact the model simulation for nearly two weeks. The influence on salinity is more pronounced than temperature.

A 20-year retrospective reanalysis of the ocean state in the Baltic Sea is constructed using three dimensional variational (3DVAR) data assimilation combining an operational numerical model with available historical temperature (<i>T</i>) and salinity (<i>S</i>) profiles. To determine the accuracy of the reanalysis, the authors present a series of comparisons with independent observations on a monthly mean basis. The performance of the assimilation in deep/shallow waters is investigated. <br><br> With assimilation, temperature and salinity in the reanalysis fit better than the free run with independent measurements at different depths. Overall, the mean biases of temperature and salinity are reduced by 0.32 °C and 0.34 psu, respectively. Similarly, the mean root mean square error (RMSE) of the reanalysis is decreased by 0.35 °C and 0.3 psu compared to the free run. In space, the model error is inhomogeneous and strongly steered by the model error dynamics. Seasonally varying error of the modeled sea surface temperature is mainly controlled by the weather forcing, and shows the least improvements due to sparse observations. Deep layers, on the other hand, witness significant and stable model error improvements. In particular, the salinity related to saline water intrusions into the Baltic Proper is largely improved in the reanalysis. The major inflow events such as in 1993 and 2003 are captured more accurately in the reanalysis as the model salinity in the bottom layer is increased by 2–3 psu. Sea level is also improved due to an improved density field. The correlation between model and observation is increased by 2 %–5 %, and the RMSE is generally reduced by 10 cm in the reanalysis compared to the free run. The reduction of RMSE is mainly due to the reduction of mean bias. Assimilation of <i>T/S</i> contributes little to the barotropic transport in the shallow Danish Transition zone. <br><br> The mixed layer depth exhibits strong seasonal variations in the Baltic Sea. The basin-averaged value is about 10 m in summer and 30 m in winter. In addition, assimilation of <i>T/S</i> profiles results in changes of about 20 m for the mixed layer depth in the Baltic Proper region in winter. Comparisons of mixed layer depth show that the assimilation induces more changes in deep water of winter time whereas the mixed layer depth is changed only about 2 m in summer time and shallow waters. One reason could be that the effect of the assimilation is counterbalanced by the effect of heating in summer and the dominant role of the surface forcing in shallow water. The significant impact in deep waters suggests that the <i>T/S</i> assimilation mainly adjusts the baroclinic transport by redistributing the density field

When altimetric data is assimilated, 3DVAR and Ensemble Optimal Interpolation (EnOI) have different ways of projecting the surface information downward. In 3DVAR, it is achieved by minimizing a cost function relating the temperature, salinity, and sea level. In EnOI, however, the surface information is propagated to other variables via a stationary ensemble. In this study, the differences between the two methods were compared and their impacts on the simulated variability were evaluated in a tropical Pacific model.Sea level anomalies (SLA) from the TOPEX/Poseidon were assimilated using both methods on data from 1997 to 2001 in a coarse resolution model. Results show that the standard deviation of sea level was improved by both methods, but the EnOI was more effective in the central/eastern Pacific. Meanwhile, the SLA evolution was better reproduced with EnOI than with 3DVAR. Correlations of temperature with the reanalysis data increased with EnOI by 0.1-0.2 above 200 m. In the eastern Pacific below 200 m, the correlations also increased by 0.2. However, the correlations decreased with 3DVAR in many areas. Correlations with the independent TAO profiles were also compared at two locations. While the correlations increased by up to 0.2 at some depths with EnOI, 3DVAR generally reduced the correlations by 0.1-0.3. Though both methods were able to reduce the model-data difference in climatological sense, 3DVAR appears to have degraded the simulated variability, especially during El Niño-Southern Oscillation events. For salinity, similar results were found from the correlations. This tendency should be considered in future SLA assimilations, though the comparisons may vary among different model implementations.