Subseasonal Prediction With and Without a Well-Represented Stratosphere in CESM1
The stratosphere is believed to provide predictability at subseasonal time scales, however few studies have been done to quantify this contribution. In particular, it is not well understood whether the improved representation of the stratosphere leads to a higher prediction skill of key surface variables. Here we evaluate the subseasonal prediction skill in two versions of the Community Earth System Model, version 1 (CESM1): one with the default (poorly resolved, 30-level) stratosphere and one with a well-resolved (46-level) stratosphere. We find that a better-resolved stratosphere improves stratospheric but not surface prediction skill for weeks 3–4. The stratosphere is believed to provide predictability at subseasonal time scales, however few studies have been done to quantify this contribution. In particular, it is not well understood whether the improved representation of the stratosphere leads to a higher prediction skill of key surface variables. Here we evaluate the subseasonal prediction skill in two versions of the Community Earth System Model, version 1 (CESM1): one with the default (poorly resolved, 30-level) stratosphere and one with a well-resolved (46-level) stratosphere. We find that a better-resolved stratosphere improves stratospheric but not surface prediction skill for weeks 3–4.
There is a growing demand in society for understanding sources of predictability on subseasonal to seasonal time scales. In this work we demonstrate that the CESM1 research Earth system model can be utilized as a subseasonal prediction model and show that its prediction skill is comparable to that of operational models. We also show that the inclusion of a well-resolved stratosphere does not improve the subseasonal (week 3–4 averaged) forecast of temperature and precipitation at the surface. The prediction skill of the Madden-Julian Oscillation (MJO) and the North Atlantic Oscillation (NAO) is similar between the two versions of the model
The role of the stratosphere in subseasonal prediction is not well quantified. We examined here the impact of a better-resolved stratosphere on the subseasonal prediction skill using two versions of CESM1. We demonstrate that the subseasonal skill of CESM1 for surface temperature, precipitation, the MJO and the NAO is comparable to that of operational models. We find that a better-resolved stratosphere improves stratospheric but not surface prediction skill for weeks 3–4.