In this study, monthly S2S reforecast for 20 years obtained from the European Centre for Medium-Range Weather Forecast (ECMWF) were evaluated against the in-situ data. The spatial results show that the two datasets agree with the rainfall pattern throughout the year. The S2S ECMWF realistically simulates the mean annual cycle skillfully by identifying the wet season from November to March (NDJFM), and the driest season as June to September (JJAS), in agreement with the observations. However, the depicted slight wet bias in estimating the observed variation during the wet season. Nevertheless, the ECMWF S2S exhibits a better performance during wet months compared to dry months. This study provides insights into the performance of S2S forecasts and their potential application over Zambia. Future studies need to focus on explaining the observed discrepancies and improvement of S2S forecasts in the region, particularly by the modelling centers.

$\bf{Highlights}$

$\bullet$ Monthly S2S reforecast for 20 years obtained from the European Centre for Medium-Range Weather Forecast (ECMWF) were evaluated against the in-situ rainfall data over Zambia using 1-month lead time forecast.

$\bullet$ The two datasets agree with the rainfall pattern throughout the year thereby, the S2S ECMWF realistically simulates the mean annual cycle skillfully by identifying the wet season from November to March (NDJFM), and the driest season as June to September (JJAS), in agreement with the observations.

$\bullet$ The S2S ECMWF exhibits a better performance during wet months compared to dry months though depicted a slight wet bias in estimating the observed variation during the wet season.

$\bullet$ The lowest spatial average value of RMSE and MAE during the wet season (NDJFM) was observed in March while the highest RMSE and MAE were recorded in December.

$\bullet$ The CRPSS in most of the stations during the wet season (NDJFM) had positive values, while during dry months the results show most of the stations having negative CRPSS values.