Subseasonal Forecast Skill of Snow Water Equivalent and Its Link with Temperature in Selected SubX Models

Accurate and skillful subseasonal forecasts have tremendous potential for sectors that are sensitive to hazardous weather and climate events. Analysis of prediction skill for snow water equivalent (SWE) and near-surface air temperature (T2m) is carried out for three (GEPS, GEFS, and FIM) global models from the subseasonal experiment (SubX) project for the 2000-14 period. The prediction skill of SWE is higher than the skill of T2m at week-3 and week-4 lead times in all models. The GEPS forecast tends to yield higher (lower) prediction skill of SWE (T2m) compared to the other two systems in terms of correlation skill score. The snow-temperature relationship in reanalysis is characterized by a strong negative correlation over most of the midlatitude regions and a weak positive correlation over high-latitude Arctic regions. All forecast systems reproduced well these observed features; however, the snow-temperature relationship is slightly weaker in the GEPS model. Despite the apparent lack of skill in temperature forecasts at week 4, all three models are able to predict the sign of temperature anomalies associated with extreme SWE conditions albeit with reduced intensity. The strength of the predicted temperature anomaly associated with extreme snow conditions is slightly weaker in the GEPS forecast compared to reanalysis and the other two models, despite having better skill in predicting SWE. These apparent disparities suggest that weak snow-temperature coupling strength in the model is one of the contributing factors for the lower temperature skill.