New Zealand’s National Vegetation Survey Databank:  improving access and interoperability

WISER S., SPENCER N. & BURROWS L.

Landcare Research, PO Box 69, Lincoln 8152, Christchurch, NEW ZEALANDE
Email: wisers@landcareresearch.co.nz

The National Vegetation Survey databank (NVS) is New Zealand’s primary archive for plot-based vegetation data.  NVS holds data from 52,000 releves and 14,000 permanent plots.  The principal NVS goals are to be a secure repository for such data and to ensure quality data are readily available to end-users in different organisations.  Computerisation of NVS data began in the 1970s.  Until recently, data entry and analyses relied on outdated DOS-based systems, un-normalised file-based data structures led to internal inconsistencies, file formats were too rigid to store the range of data being collected, data could not be readily exported, and it was difficult to link NVS data with other New Zealand databases.  Since 2000 considerable progress has been made on modernising our system and making data more accessible.  A website (http://nvs.landcareresearch.co.nz) provides general background information on NVS, protocols for data deposit and use, a FAQ page, downloadable data sheets, and searchable metadata.  The NVS databank metadata (information about the datasets held) provides the key for determining whether the databank holds specific datasets of interest and whether they are appropriate to meet user needs.  Data storage was redesigned and implemented in MS SQL Server, a flexible client-server relational database management system. New linkages between metadata and plot data were enabled. A MS-Access interface for data entry and viewing was developed as well as an Internet-based text searchable interface for accessing metadata.    A parallel project developed a geo-spatial oriented interface using map-based GIS technology (ESRI ArcIMS), to allow users to view plot locations and request data over the Internet.  About 2/3 of the actual plot data has been migrated from the DOS-based system to the relational system in SQL-Server, which supports data entry and viewing interfaces in MS-Access and linkages to any ODBC/OLE DB compliant software.  Current developments emphasise improved Internet access to information, including a NVS data dictionary and a database of publications based on NVS data, and the development of interfaces to allow Internet-based management and analysis of plot data.  We are now refining a more robust data model that will have flexibility to store a wider range of data.  This is based on the VegBank model for releve data, but expanded to accommodate data on permanently marked individuals and to incorporate an alternative approach to concept taxonomy.  Future plans include developing a system for automated retrieval of datasets requiring permission for use, to redevelop software for summarising data and allowing ready export to other vegetation data storage and analysis software, and to provide electronic access to media describing plot locations (e.g., maps, aerial photos etc.).  To support our aim of working towards harmonized data storage and interoperable data querying and transfer, we are collaborating with organizations holding similar datasets in the US, specifically the VegBank project and the US Forest Service.