Outputs of AMIP experiments using ECHAM-6.3.05p2 with the concurrent radiation scheme

To improve the scalability and performance of the atmospheric model ECHAM6, we extend the concurrency within the model by running the radiation component in parallel with other atmospheric processes (adiabatic and diabetic processes). We performed experiments to evaluate the computational performance and stability of the model with both concurrent and classical radiation schemes. The experiments are based on ECHAM-6.3.05p2 and follows the  Atmospheric Model Intercomparison Project (AMIP) protocol.

There are two types of data sets in this repository: performance_measurements.tar and AMIP_Data_For_Heidari_et_al.tar

(A) performance_measurements.tar

The data set "performance_measurements.tar.gz" provides the performance measurement results collected from multiple AMIP experiments at coarse resolution and include:

1. The total runtime of the model with the classical radiation scheme.

2. The total runtime of the model with the concurrent radiation scheme

3. The runtime of the classical radiation scheme

4. The runtime of the concurrent radiation scheme

5. The cost of data exchange between the radiation component and the main model when the model uses the concurrent radiation scheme

6. The idle time of the MPI processes when the model uses the concurrent radiation scheme

(B)AMIP_Data_For_Heidari_et_al.tar

This data set provides the output from the AMIP runs at CR and LR configuration using the concurrent and classical radiation scheme. Using reanalysis and observational reference data, we compare and assess the simulated mean climate state and climate variability by the concurrent radiation scheme and its counterpart in the original model. We find that the concurrent radiation scheme retains the major features of the simulated  mean climate and internal variability as in the original model, despite some minor differences.