Mid-summer melt pond fraction–September Arctic sea ice extent relationship in the EC-Earth3

What is it about?

Arctic sea ice is decreasing in volume and extent at a rapid rate, the causes of which are not fully understood. A reasonably reliable prediction of Arctic sea ice is, therefore, much needed. In a recent numerical modelling study, the spring melt pond fraction has been suggested to be a key predictor of subsequent September Arctic sea ice minimum extent anomalies. However, another study based on satellite data did not provide evidence for such a long-range relationship. Here, we explore the existence of this springtime association in the state-of-the-art global climate model EC-Earth3, which includes an explicit treatment of melt ponds on sea ice, for present-day climate conditions (constant forcing of the fix-year 2000 CE). We do not find a statistically significant inverse relationship between September sea ice extent and spring melt pond fraction on the seasonal scale established from a stable control simulation spanning two centuries without transient forcing. Our results support the inferences based on satellite data that the mid-summer (June-July) melt pond fraction highly correlates with the September ice extent. However, the melt pond fraction in springtime (May) is poorly correlated with September sea ice extent anomalies, suggesting limited predictability stemming from this variable at a 5-month lead time.

Featured Image

Photo by Willian Justen de Vasconcellos on Unsplash

Photo by Willian Justen de Vasconcellos on Unsplash

Why is it important?

This paper explored the relationship of September sea ice extent in the Arctic with the springtime melt pond area fraction, under the constant greenhouse gas forcing for the fix-year 2000 CE. Our new perspective as compared to Schröder et al. (2014) is the inclusion of an explicit melt pond scheme with additional ice category dependence in a global climate model (instead of a forced stand-alone ice model, which does not permit ocean-sea ice-atmosphere interactions), and the application of fixed forcing representing the present-day climate (2000 CE) to ensure a causal relationship between the September minimum sea ice extent and melt pond fraction. We find a statistically significant inverse relationship between September sea ice extent and integrated melt pond fraction on the seasonal scale. Our results support the satellite-based inferences made by Liu et al. (2015) and Feng et al. (2022) that the mid-summer (June-July) pond fraction highly correlates with the September sea ice extent suggesting limited predictability stemming from this variable at a 5-month lead time. i.e., springtime. We have only contested the results of Schröder et al. (2014) modelling study. Liu et al. (2015) and Feng et al. (2022) are observation-based studies that disagree with Schröder et al. (2014) modelling results and are in full agreement with our modelling results validated using observations.

Perspectives