Assessment of 2019 Sub-seasonal to Intra-annual Arctic Sea Ice Forecasts

The rate of climate change in the Arctic has been much faster than the global average over the past decades. The ongoing increase of surface temperatures and the decline of sea ice cover are considered the direct manifestation of the Arctic warming. This "new normal" regime has already affected, and it is expected to further impact, not only on the Arctic environment, but also local to regional economic development, resilience and natural resource management. Hence, it is becoming increasingly vital to know in advance near-future sea ice and climate states to assist surging human operations, in parallel to the declining sea ice cover. Despite such demands, the advancement of reliable sea ice forecasts up to months ahead remains challenging. This presentation summarizes our results on the 2019 sub-seasonal to intra-annual (up to 6 months) forecasts of Arctic sea ice conditions, using the Regional Arctic System Model (RASM) for dynamical downscaling of global forecasts. RASM is a fully coupled regional climate system model, consisting of the atmosphere, ocean, sea ice, land hydrology, and streamflow routing components, coupled through the Community Earth System Model flux coupler. The ocean and sea ice configurations include the horizontal resolution of 1/12 degree with 45-vertical levels and 5 sea ice thickness categories, respectively. The atmosphere and land hydrology components are configured on a 50-km grid with 40-vertical levels and 3 soil layers, respectively. For dynamic downscaling, we have used the Climate Forecast System (CFS) Reanalysis (CFSR) and version 2 (CFSv2) output for 1979-present. The respective initial conditions for ensemble forecasts are derived from the RASM 40+ year hindcast forced with CFSR/CFSv2, to assure physical and internal consistency across all the RASM components. RASM intra-annual ensemble forecasts have been initialized on the 1^st day of each month in 2019. Their boundary conditions have been derived from the CFSv2 9-month forecasts, 24-hour apart within the month preceding the RASM ensemble forecast start date. The skills of the RASM ensemble forecast and hindcast as well as the CFSv2 global reanalysis are assessed against available satellite observations. In particular, the effects of lead time and initial conditions on the skill of Arctic sea ice predictions are investigated.