ComDA, a Common Software for Nonlinear and Non-Gaussian Land Data Assimilation, Released and Implemented Online Applications for Big Data Platform

[2020-05-21]   Author:LIU Feng   Source :  Research team led by Prof. LI Xin at Institute of Tibetan Plateau Research, Chinese Academy of Sciences

A Common Software for Nonlinear and Non-Gaussian Land Data Assimilation (ComDA) was recently released by the research team led by Prof. LI Xin at National Tibetan Plateau Data Center (TPDC), on the platform of the TPDC and the big data analysis module of the platform "A Big Earth Data Platform for Three Pole". ComDA is a common platform for data assimilation for land surface, hydrological and other dynamic models developed by Prof. LI Xin's research team based on long-term land surface data assimilation research. It integrates land surface and hydrological models such as Lorenz, CoLM, SiB2, VIC, SHAW, radiation transfer models such as AIEM, MEMLS and ProSail, and non-Gaussian nonlinear data assimilation algorithms such as ensemble Kalman filter and particle filter. It can also easily introduce new dynamic models, observation operators and data assimilation algorithms, while supporting high-performance computing technologies. It provides solutions for fast and customized applications of multi-disciplinary data assimilation.

A fast, easy-to-use, and multidisciplinary application-oriented assimilation platform for land data assimilation has not been achieved before. Therefore, the research team led by Prof. LI Xin developed and published ComDA, which is a multisource observation and land surface application–oriented software platform that has the ability to run distributed algorithms and work on multiple operating systems. ComDA provides a standardized integration approach and a modular development approach (as shown in the figure), and its high-performance computing can improve computing efficiency by at least 75%. Its launch effectively solved the problem of localization of data assimilation software.

General framework of ComDA.

The use of ComDA can significantly reduce the development cycle of data assimilation system applications. Currently, ComDA has provided support for the research and development environment for the Chinese land surface data assimilation system. Recently, ComDA has also played an important role in the dynamic simulation, parameter estimation and uncertainty analysis of the COVID-19 infectious diseases (e.g., Using the contact network model and Metropolis-Hastings sampling to reconstruct the COVID-19 spread on the “Diamond Princess”: https://doi.org/10.1016/j.scib.2020.04.043). With the support of the National Tibetan Plateau Data Center, ComDA will provide a complete set of service processes for researchers, including non-linear non-Gaussian data assimilation, land surface process model application, high-dimensional multi-parameter optimization, ensemble prediction and other extensive requirements.

The introductory paper of ComDA was published in Environmental Modeling and Software with first author Dr. LIU Feng (Northwest Institute of Eco-Environment and Resources, Chinese Academy of Sciences) and corresponding authors Prof. LI Xin (Institute of Tibetan Plateau Research, Chinese Academy of Sciences) and Dr. WANG Liangxu (Shanghai Normal University). The online application and source code of the software are released and shared on the TPDC, the "A Big Earth Data Platform for Three Pole" platform and GitHub, respectively.

Full-text available at: https://doi.org/10.1016/j.envsoft.2020.104638

Software online applications (currently provides a collection of Lorenz models with Kalman filtering assimilation examples, more applications will be released later): http://analysis.tpdc.ac.cn/index/ComDA

Source code available at: https://github.com/uniliufeng/ComDA