L12 Design Document for Land Model Benchmarking
The Design Document page provides a description of the algorithms, implementation and planned testing including unit, verification, validation and performance testing. Please read Step 1.3 Performance Expectations that explains feature documentation requirements from the performance group point of view.
Design Document
In the table below 4.Equ means Equations and Algorithms, 5.Ver means Verification, 6.Perf - Performance, 7. Val - Validation, - competed, - in progress, - not done
Title: Create a Land Model Benchmarking Capability for ACME
Requirements and Design
ACME Land Group
Date:
Summary
A land model benchmarking package will be implemented, leveraging the ongoing ILAMB development in the RGCM-funded BGC Feedbacks SFA, for routine and systematic assessment of land model fidelity through comparison with observational data. Initially, a recent version of the ILAMB prototype will be installed and tested using ACME Land Model (ALM) output. New model–data comparisons will be added by the team of ACME benchmark developers for evaluating the impact of new feature implementation on the overall scientific performance of ALM in coupled and uncoupled configurations. The Workflow Team will be encouraged to adopt this evolving benchmarking package for operational use for evaluating relevant contemporary land and Earth system model simulations. A next-generation version of ILAMB, currently under development and offering parallel processing capabilities, will be used after v1 model activities are complete.
Requirements
Requirement: Install and test the ILAMB prototype code and processed observational data on rhea.ccs.ornl.gov
Date last modified:
Contributors: Forrest M. Hoffman (Unlicensed)
A recent version of the ILAMB prototype code (nominally version 1.2.1) and the associated processed observational data will be installed on rhea.ccs.ornl.gov. This package will be tested using sample output from CLM4.0CN, CLM4.5BGC, and a corresponding offline ALM simulation.
Algorithmic Formulations
Design solution: Scoring metrics will be developed as needed for model–data comparison
Date last modified:
Contributors: Forrest M. Hoffman (Unlicensed)
All new observational data will be accompanied by metrics for comparison with model output. These metrics will be clearly documented in the evolving Description of Scoring Metrics and Data Processing.
Design and Implementation
Implementation: Install and test the ILAMB prototype code and processed observational data on rhea.ccs.ornl.gov
Date last modified:
Contributors: Forrest M. Hoffman (Unlicensed)
ILAMB code will be copied to the target system along with the observational data package and example model output. The ILAMB code will be tested and ported as necessary for correct operation on the target system.
Planned Verification and Unit Testing
Verification and Unit Testing: Install and test the ILAMB prototype code and processed observational data on rhea.ccs.ornl.gov
Date last modified:
Contributors: Forrest M. Hoffman (Unlicensed)
Resulting diagnostics and scores will be compared with the same model and code set at http://redwood.ess.uci.edu/mingquan/www/ILAMB/AllDiagnostics/.
Planned Validation Testing
Validation Testing: Install and test the ILAMB prototype code and processed observational data on rhea.ccs.ornl.gov
Date last modified:
Contributors: Forrest M. Hoffman (Unlicensed)
Model output from ALM will be run through the system and compared with that from CLM4.0CN and CLM4.5BGC. If all of the diagnostics and scores are reasonable for all three model outputs, the package will be considered to have passed validation.
Planned Performance Testing
Performance Testing: Install and test the ILAMB prototype code and processed observational data on rhea.ccs.ornl.gov
Date last modified:
Contributors: Forrest M. Hoffman (Unlicensed)
No performance testing is planned for the ILAMB prototype. Considerable performance testing is being performed for the next-generation package.