This paper presents the Simple Ocean Data Assimilation (SODA) reanalysis of ocean climate variability. SODA 1.4.2, a reanalysis spanning 44 years from 1958 to 2001, improves upon previous versions by incorporating data assimilation to correct model forecasts with observations. The reanalysis uses a model with 0.25° × 0.4° horizontal resolution and 40 vertical levels, continuously updated with observations every 10 days. The observation set includes historical hydrographic profiles, ship measurements, moored observations, and satellite SST data. A parallel run, SODA 1.4.0, uses the same surface boundary conditions without data assimilation. SODA 1.4.2 shows improved estimates of eddy kinetic energy and sea level variability, more closely matching independent observations. The analysis shows that at near-annual frequencies, the model forecast has a strong influence, while at decadal frequencies, observations dominate. As a result, interannual variability in SODA 1.4.2 closely resembles that in SODA 1.4.0, but decadal anomalies of 0–700-m heat content in SODA 1.4.2 are more accurate. The SODA system uses a sequential approach with a numerical model providing a first guess of the ocean state, which is corrected using linear Kalman equations. The system assimilates data every 10 days, with corrections applied incrementally. The reanalysis includes temperature, salinity, and velocity data, remapped to a uniform global grid. The system compares two experiments: SODA1.4.0, which uses ERA-40 winds without observations, and SODA1.4.2, which uses ERA-40 winds with all temperature and salinity observations. The reanalysis includes a large dataset of temperature and salinity profiles, with quality control applied to ensure accuracy. The reanalysis shows that the model's forecast error is weakly biased, with significant variability in the thermocline depths. The SODA1.4.2 reanalysis shows improved accuracy in mean transports, such as the Gulf Stream and Antarctic Circumpolar Current, compared to observations. The reanalysis also shows improved accuracy in variability, including sea level, vertically averaged temperature, and near-surface currents. The SODA1.4.2 reanalysis shows higher eddy kinetic energy in the tropical Indian Ocean and more accurate decadal trends in heat storage. The reanalysis also shows that the decadal change in heat storage in the upper ocean is significant, with a linear trend of 0.0061°C yr⁻¹ in the northern band, equivalent to a heat gain of 1.3 × 10²¹ J yr⁻¹. The reanalysis also shows that the southern band hasThis paper presents the Simple Ocean Data Assimilation (SODA) reanalysis of ocean climate variability. SODA 1.4.2, a reanalysis spanning 44 years from 1958 to 2001, improves upon previous versions by incorporating data assimilation to correct model forecasts with observations. The reanalysis uses a model with 0.25° × 0.4° horizontal resolution and 40 vertical levels, continuously updated with observations every 10 days. The observation set includes historical hydrographic profiles, ship measurements, moored observations, and satellite SST data. A parallel run, SODA 1.4.0, uses the same surface boundary conditions without data assimilation. SODA 1.4.2 shows improved estimates of eddy kinetic energy and sea level variability, more closely matching independent observations. The analysis shows that at near-annual frequencies, the model forecast has a strong influence, while at decadal frequencies, observations dominate. As a result, interannual variability in SODA 1.4.2 closely resembles that in SODA 1.4.0, but decadal anomalies of 0–700-m heat content in SODA 1.4.2 are more accurate. The SODA system uses a sequential approach with a numerical model providing a first guess of the ocean state, which is corrected using linear Kalman equations. The system assimilates data every 10 days, with corrections applied incrementally. The reanalysis includes temperature, salinity, and velocity data, remapped to a uniform global grid. The system compares two experiments: SODA1.4.0, which uses ERA-40 winds without observations, and SODA1.4.2, which uses ERA-40 winds with all temperature and salinity observations. The reanalysis includes a large dataset of temperature and salinity profiles, with quality control applied to ensure accuracy. The reanalysis shows that the model's forecast error is weakly biased, with significant variability in the thermocline depths. The SODA1.4.2 reanalysis shows improved accuracy in mean transports, such as the Gulf Stream and Antarctic Circumpolar Current, compared to observations. The reanalysis also shows improved accuracy in variability, including sea level, vertically averaged temperature, and near-surface currents. The SODA1.4.2 reanalysis shows higher eddy kinetic energy in the tropical Indian Ocean and more accurate decadal trends in heat storage. The reanalysis also shows that the decadal change in heat storage in the upper ocean is significant, with a linear trend of 0.0061°C yr⁻¹ in the northern band, equivalent to a heat gain of 1.3 × 10²¹ J yr⁻¹. The reanalysis also shows that the southern band has