The efficient global primitive equation climate model SPEEDO V2.0

• Published in: Geoscientific Model Development
• Main Authors: C. A. Severijns, W. Hazeleger
• Format: Article in Journal/Newspaper
• Language: English
• Published: Copernicus Publications 2010
• Subjects: Geology; QE1-996.5; Sea ice
• Online Access: https://doi.org/10.5194/gmd-3-105-2010; https://doaj.org/article/e8ad2a4f998c40c3a6cbdcc5c15a5c2c

The efficient primitive-equation coupled atmosphere-ocean model SPEEDO V2.0 is presented. The model includes an interactive sea-ice and land component. SPEEDO is a global earth system model of intermediate complexity. It has a horizontal resolution of T30 (triangular truncation at wave number 30) and 8 vertical layers in the atmosphere, and a horizontal resolution of 2 degrees and 20 levels in the ocean. The parameterisations in SPEEDO are developed in such a way that it is a fast model suitable for large ensembles or long runs (of O (10 4 ) years) on a typical current workstation. The model has no flux correction. We compare the mean state and inter-annual variability of the model with observational fields of the atmosphere and ocean. In particular the atmospheric circulation, the mid-latitude patterns of variability and teleconnections from the tropics are well simulated. To show the capabilities of the model, we performed a long control run and an ensemble experiment with enhanced greenhouse gases. The long control run shows that the model is stable. CO 2 doubling and future climate change scenario experiments show a climate sensitivity of 1.84 K W -1 m 2 , which is within the range of state-of-the-art climate models. The spatial response patterns are comparable to state-of-the-art, higher resolution models. However, for very high greenhouse gas concentrations the parameterisations are not valid. We conclude that the model is suitable for past, current and future climate simulations and for exploring wide parameter ranges and mechanisms of variability. However, as with any model, users should be careful when using the model beyond the range of physical realism of the parameterisations and model setup.