Canonical Workflows in Simulation-based Climate Sciences  

作　　者：Ivonne Anders Karsten Peters-von Gehlen Hannes Thiemann 

机构地区：[1]German Climate Computing Center(DKRZ),BundesstraBe 45a,D-20146 Hamburg,Germany

出　　处：《Data Intelligence》2022年第2期212-225,共14页数据智能（英文）

基　　金：funded by the Deutsche Forschungsgemeinschaft(DFG,German Research Foundation)under Germany's Excellence Strategy-EXC 2037 CLICCS-Climate,Climatic Change,and Society-Project No.390683824.

摘　　要：In this paper we present the derivation of Canonical Workflow Modules from current workflows in simulation-based climate science in support of the elaboration of a corresponding framework for simulationbased research.We first identified the different users and user groups in simulation-based climate science based on their reasons for using the resources provided at the German Climate Computing Center(DKRZ).What is special about this is that the DKRZ provides the climate science community with resources like high performance computing(HPC),data storage and specialised services,and hosts the World Data Center for Climate(WDCC).Therefore,users can perform their entire research workflows up to the publication of the data on the same infrastructure.Our analysis shows,that the resources are used by two primary user types:those who require the HPC-system to perform resource intensive simulations to subsequently analyse them and those who reuse,build-on and analyse existing data.We then further subdivided these top-level user categories based on their specific goals and analysed their typical,idealised workflows applied to achieve the respective project goals.We find that due to the subdivision and further granulation of the user groups,the workflows show apparent differences.Nevertheless,similar"Canonical Workflow Modules"can be clearly made out.These modules are"Data and Software(Re)use","Compute","Data and Software Storing","Data and Software Publication","Generating Knowledge"and in their entirety form the basis for a Canonical Workflow Framework for Research(CWFR).It is desirable that parts of the workflows in a CWFR act as FDOs,but we view this aspect critically.Also,we reflect on the question whether the derivation of Canonical Workflow modules from the analysis of current user behaviour still holds for future systems and work processes.

关 键 词：Canonical Workflow Fair Digital Objects HPC infrastructures Data services Climate science 

分 类 号：P4[天文地球—大气科学及气象学]