The Challenge of Arctic Sea Ice Thickness Prediction by ECMWF on Subseasonal Time Scales. Issue 8 (18th April 2022)

• Record Type: Journal Article
• Authors: Xiu, Yongwu; Luo, Hao; Yang, Qinghua; Tietsche, Steffen; Day, Jonathan; Chen, Dake
• Is Part Of: Geophysical research letters. Volume 49:Issue 8(2022)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2022-04-18
• Subjects: Geophysics -- Periodicals; Planets -- Periodicals; Lunar geology -- Periodicals; 550
• DOI: 10.1029/2021GL097476
• Languages: English
• ISSNs: 0094-8276

Abstract

Abstract: A recent study has provided the first thorough assessment of subseasonal Arctic sea ice edge prediction in operational forecast systems. However, the corresponding assessment of the Arctic sea ice thickness (SIT) is still lacking. Here, the Arctic SIT reforecasts from the European Centre for Medium‐Range Weather Forecasts (ECMWF) on subseasonal time scales are evaluated for the first time using a well‐developed SIT reanalysis data set. The results show that ECMWF forecasts of Arctic SIT are more skillful than persistence forecast (PFs) during the transition seasons for lead times longer than 30 days, indicating the advantages of dynamical prediction. However, ECMWF suffers from large initial SIT errors and has lower skill than a PF, especially from March to June. Thus, subseasonal Arctic SIT predictions still face many challenges, and improving the mean SIT in the model and assimilating satellite‐estimated SIT in the ocean analysis would significantly improve SIT forecasts. Plain Language Summary: The demand for sea ice forecasts on subseasonal time scales is continuously increasing. Recent studies evaluate the sea ice edge prediction on subseasonal time scales in the main operational forecasting systems, which shows that the European Centre for Medium‐Range Weather Forecasts (ECMWF) performs best. Despite sea ice thickness (SIT) being a very important parameter in the Arctic climate system, little is known about how well it is predicted by operational forecasts on subseasonal time scales. Here, we evaluate the only SIT reforecasts from ECMWF based on a well‐developed SIT reanalysis. The results show that ECMWF makes skillful SIT forecasts for July/August and November/December with lead times longer than 30 days. However, during other times, ECMWF suffers from the large initial SIT error and thus exhibits poor predictive skill. The large initial SIT error is presented by an overestimation in the Beaufort Sea and an underestimation in the Central Arctic. Our results imply that further progress on subseasonal time scales prediction can be made by improving the forecast model and the SIT initial conditions. Key Points: Arctic sea ice thickness (SIT) predictions by European Centre for Medium‐Range Weather Forecasts (ECMWF) are evaluated on subseasonal time scales with a well‐developed SIT reanalysis ECMWF makes more skillful Arctic SIT predictions during the transition seasons for lead‐times longer than 30 days ECMWF's SIT predictions suffer from a large initial error, and efforts in SIT in both the coupled model and initial conditions are required