This works in developing an integrated data management, retrieval and visualization system for earth science datasets with extensibility, scalability, uniformity, transparency and heterogeneity. XML based metadata mechanism is the foundation of data management in our system. Dynamically generated query GUI makes it easy and convenient for scientists to access and retrieval diverse datasets. Scientific visualization toolkits display huge amount of data graphically to help researchers have better understanding of the data and gain valuable insights of the datasets under investigation. This system helps earth scientists use, share and visualize data more efficiently. Without knowing any information of the physical storage location, content, structure and format of each dataset instance, and without programming a single line of codes, scientists can now query heterogeneous data easily, and view and understand the retrieved data in analytical and graphical ways. DATA QUERY AND RETRIEVAL 1. Dynamically Generated Query GUIs To allow users to be able to query data system without background knowledge and training, it provides query GUIs for all datasets in the system. Its approach is to create a data query system that dynamically creates dataset query GUI for diverse datasets based on characteristic of diverse datasets. These characteristics are described in XML metadata. By using stored metadata to create the query interfaces, a standardized yet dynamic system is created that allows querying of assorted datasets. By this way, the system eliminates the need to create custom programs for different datasets. When adding new datasets to the system, query GUI for these datasets will be dynamically generated if the characteristic of these datasets has been specified in metadata. Therefore, the system provides an extensible and scalable query GUI framework. Through dynamically generated query GUI, scientists can specify search conditions, customize the format and the resolution type of the result files as needed. 2. Query Categories To make it convenient, users are able to search for data by the category of physical data types or data sources. Users may also define new categories by adding their definitions into metadata and the system will then dynamically add them into GUIs. 3. Data Search and Retrieval The data retrieval is based on the metadata system, that is, the description of the physical datasets. After user submit query request through GUI, the query can be then devised to search over the metadata catalogue, which is implemented in hierarchy. Once all data files that possibly satisfy the query have been identified by searching the available metadata, the system will retrieve these files holding the actual data and obtain useful result data. If the resolution type is not the same as original request, it will do computation and obtain data in new resolution. Then it organizes the result data file in the format that users specified.