...
In the table below, 4.Equ means Equations and Algorithms, 5.Ver means Verification, 6.Perf - Performance, 7. Val - Validation, 
- completed, 
- in progress, 
- not done
| HTML Comment |
|---|
|
| Page Properties | ||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ||||||||||||||||||||||
|
| Expand | |||
|---|---|---|---|
| |||
|
...
Requirements
Requirement 1: Fast track implementation and evaluation were done successfully. We found a couple of implementation errors and fixed. Also, we found an issue with new tracers in UNICON and SE dycore. A fix was implemented.
Date last modified:
Contributors: Jin-Ho Yoon (Unlicensed), Sungsu Park (Unlicensed), Hailong Wang, Po-Lun Ma, Mark Taylor, Phil Rasch (pnl.gov)
Requirement 2: Slow track implementation is discussed.
Date last modified:
Contributors: Jin-Ho Yoon (Unlicensed), Sungsu Park (Unlicensed), Hailong Wang, Po-Lun Ma, Shaocheng Xie, Wuyin Lin, Balwinder Singh, Phil Rasch (pnl.gov)
Requirement 3: Slow track implementation.
Date last modified:
Contributors: Jin-Ho Yoon (Unlicensed), Sungsu Park (Unlicensed), Hailong Wang, Po-Lun Ma, Shaocheng Xie, Wuyin Lin, Balwinder Singh, Phil Rasch (pnl.gov)
Algorithmic Formulations
Design solution: UNICON design is well documented in Sungsu Park (Unlicensed)'s papers.
Date last modified:
Contributors: Jin-Ho Yoon (Unlicensed), Sungsu Park (Unlicensed)
Design and Implementation
Implementation: Fast track implementation
Date last modified:
Contributors: Jin-Ho Yoon (Unlicensed)
Planned Verification and Unit Testing
Verify against UNICON's implementation in CESM1 code base and its behavior.
Verify/Test its compatibility with Polar mods
Date last modified:
Contributors: Jin-Ho Yoon (Unlicensed), Hailong Wang, Po-Lun Ma,
Planned Validation Testing
Validation Testing: short-desciption-of-testing-here
Date last modified:
Contributors: Jin-Ho Yoon (Unlicensed), Sungsu Park (Unlicensed), Phil Rasch (pnl.gov), Mark Taylor, Hailong Wang, Po-Lun Ma, Hui Wan, Kai Zhang
The initial implementation has been tested for without considering B4B capability with default ACME, because we are taking a fast track approach for quick evaluation. However, we know how to maintain B4B with default ACME. On the other hand, we rigorously tested for consistencies with NCAR simulation in terms of FV and SE dycores. An earlier issue in SE dycore has been fixed. 10 year fixed SST simulation has been done for UNICON in CESM1 branch with FV dycore (used as a guideline), UNICON in ACME with FV dycore, and UNICON in ACME with SE dycore to ensure all the features are preserved.
Planned Performance Testing
Performance Testing: short-desciption-of-testing-here
Date last modified:
Contributors: Jin-Ho Yoon (Unlicensed), Phil Rasch (pnl.gov), Po-Lun Ma, Hailong Wang, Balwinder Singh
Early results show that UNICON improves simulations of the mean climate, tropical variability, diurnal precipitation, and cloud properties. An early test with UNICON on high vertical resolution shows much less sensitivity, meaning good scale adaptivity and easier tuning. UNICON costs extra 40-50%. We'll work with the performance team to evaluate/improve performance.