Home

About

CODATA Membership

Resources

Task and Working Groups

Archives

Abstract:

A Case Study of the Utilization of Archived Data:
Methodology of Data Integrating and Mining for 1:1 M Land Type Mapping of China

LIU Chuang, SHEN Yuancun
Global Change Information and Research Center
Institute of Geography and Natural Resources
Chinese Academy of Sciences
Beijing 100101 China
lchuang@igsnrr.ac.cn , shenyc@igsnrr.ac.cn

It is the most important progress for a scientist to use the archived data and add its value by data mining in the information society. The data utilization is also the point of data archiving. The methodology for data integrating and mining will play a critical role in most cases of the data utilization. The paper presents an example how a new methodology of data integrating and mining together with the global and local archived data could support the 1:1M land type mapping of China.

The term of “land type” here is defied as an identifiable group or class in the solid ground surface layer and inland water of the earth, which has common characteristics and distinguished geobio-process functioned by interactive actions of climate, landform, vegetation, soil, water, as well as human activities. The land is considered as a spatial concept, it could reach to lithic contact in bottom and uppermost canopy of the ground in top.

The land type study has been one of the most of traditional and comprehensive studies in the long history of China. The earliest historical records can be cased back to the Tribute of Yu (the 5 th century B.C.) and the Record of the Rite of Zhou. The most resent important outcomes of the land type studies are the classification system for the 1:1M land type mapping of China in 1987, and 8 sample sheets among the 64 sheets in 1:1M land type map of China in 1990. Limited by data, information and technology, the 1:1M land type map covering the broad area of China would not be able to be completed in the 20 century, although a very solid foundation had made by hundreds of scientists in China during the 15 years from 1970’ to 1990’.

With the data support from WGISS/CEOS and related data holders, including MODIS archived in NASA/CAS, TM/ETM in USGS/CAS, DEM in NGCC/CAS, climate data in CMA/CAS, as well as the ground truth survey data archived in CAS, a Hierarchical Land Type Reference Model (HLTR Model) for 1:1M Land Type Mapping of China has been developed. The HLTR Model is designed in five levels: the first level is designed for regionalization (or zoning) and initial version creation of land type classification system for 1:1M land type mapping of China. Three natural realms and 7 natural divisions had been identified. A hierarchical land type classification system in three levels in the initial version also recognized based on that created in 1990. The second level is designed for the data (elements) selection and the weight analysis, with which, the role of each elements in land type future and functions will be analyzed. The third level is the data integration level, all data selected will be compatible in code system, time period, spatial resolution, projection, parameters, units as well as the data formats. The fourth level will focus on the data mining for each of land type identification. A series of analysis models and reference parameters for each of land types have been developed, with the support of the computational methodology, a series of analysis software have been and will continue to be developed. The fifth level is for the land type validation and the final version of land type classification system for 1:1M land type mapping of China will be completed.

The HLTR Model provides a new methodology of data integrating and mining for 1:1M land type mapping of China. By which, together with the time series of data from dynamic earth observations, the innovative knowledge for understanding the processes of materials and energies transformation in each land types will be created.

Working to improve the quality, reliability, management and accessibility of Data for Science and Technology

| home | about | codata membership | resources | task and working groups |
| archives | newsletter | contact | members area |

| | RSS Feed subscription instructions |