NCWMS: A FAST WEB MAP SERVICE FOR MULTIDIMENSIONAL ENVIRONMENTAL DATA
We present an implementation of the OGC Web Map Service (WMS) specification that is designed for serving four-dimensional gridded data efficiently, supporting fast visual exploration of large datasets from the environmental sciences. We focus on the visualization of data held in NetCDF files (hence the name "ncWMS"), although the software can handle other gridded data types (such as HDF) through plug-ins. ncWMS is a Java web application, released as free and open source software under a liberal licence, and has an active user and developer community, comprising members from research institutions, government agencies and private industry.
ncWMS is supplied with a dynamic interactive website based upon OpenLayers, known as "Godiva2". The ncWMS/Godiva2 system has been applied and customized in a number of projects for the visualization of large datasets, particularly the outputs of ocean forecast models and high-level satellite products. Based on the experience of these projects we will discuss the strengths and weaknesses of the WMS specification for supporting scientific visualization, and we will outline the non-standard additional functionality we implemented in ncWMS to fill key gaps in capability. (ncWMS is fully backward-compatible with the WMS 1.3.0 and 1.1.1 specifications.)
Considerable effort has been expended to make the generation of imagery as fast as possible, which is particularly challenging when handling large, high-resolution datasets on complex grids. We will discuss some of the strategies we employ to ensure that exploration of large datasets at multiple resolutions is smooth.
Finally, we will outline our future plans for ncWMS, including extending its capabilities to support the visualization of non-gridded data, such as in-situ observations and low-level satellite products. The new system will provide important new capabilities in the visual intercomparison of different environmental datasets, supporting scientific activities such as model validation and data assimilation.
Jon Blower - University of Reading
Alastair Gemmell - University of Reading
Keith Haines - University of Reading
Adit Santokhee - University of Reading