Accessing multidimensional data pools

While an approach for time series located at points has been in operation at PIK for years, PIXDAT’s subproject xDat aims to provide the research community with functional access also to time series of spatially distributed data on (regular) grids or polygonal structures.

Facing the heterogeneity of data and users, an iterative process is followed, focusing on the development of

• data models for data warehouse-based management of time series on regular spatial grids in different spatio-temporal resolutions and on polygonal structures;
• graphical user interfaces allowing access to and visualisation of the stored data

Products (I) - The database

Currently, the project's database integrates gridded time series from PIK-CDS Global, PIK-CDS Germany, LPJ, ATEAM, CRU and IPCC; resulting in an overall amount of 6x10⁹ data values. In addition, a pilot application for the management of (polygonal) socio-economic data from the Global Change & Social Systems Dept. has been developed.

• for more information on database structure and current contents see the dynamic, Oracle-APEX-based documentation interface (PIK-internal only)

Products (II) - xDat access clients

To allow remote and widely platform independent access to both the gridded and the polygonal time series data, 'rich' Java clients have been developed. Functionality covers e.g. selection of spatio-temporal subsets, derivation of aggregated data, interactive maps including picking, zooming, recolouring etc. Both clients are available for beta testing and will be released in the PIK intranet in 2007.

• cf. the status report on xDat for Regular Spatial Grids
• see also xDat for station data (operational at PIK since 2001)

Products (III) - (derived) 'lightweight' access interfaces

In addition, 'lightweight' access interfaces with HTML fronends have been developed, re-using central modules of the xDat access clients

• PDDD (Platform for Data Display & Dissemination):
  PDDD is a servlet-based platform designed to facilitate internet-wide access to spatial gridded data from PIK. Functionality covers subset selection, quicklook generation and data export to different formats. PDDD already provides access to time series simulated by the LPJ Dynamic Global Vegetation Model; more datasets will follow.
  • PDDD HOME
  • access LPJ time series via PDDD
• HTML-Interface to the socioeconomic database
  a HTML frontend to the socioeconomic database, combining dynamic HTML generated with Oracle-APEX and a servlet based map renderer.
  • access the interface (PIK-internal only)

Products (IV) - Climate Diagram Generator (CDG)

Java-based, interactive visualization of temperature and precipitation time series using diagrams e.g. according to Walter and Thornthwaite.

• Climate diagrams for observational and scenario data (applet).

Prospects

Among others, it is planned to

• increase the number of PIK datasets made available through PDDD;
• integrate more gridded and socio-economic data into the project's database;
• provide enhanced functionality to derive new date by e.g. temporal and spatial aggregation;
• establish mechanisms to derive polygonal from gridded data and vice versa;
• extend the database to a broad range of vulnerability indices.

Presentations / Reports

• Wrobel, M., Glauer, J., Flechsig, M., Rachimow, C.: xDat for Regular Spatial Grids. A pilot application for web-based access, processing and visualization of database-managed geo-referenced time series. Status Report, PIK Potsdam, December 2005
  download
• xDat for Regular Spatial Grids: talk by Markus Wrobel, ToPIK Days 2006, 09.Feb. (PIK intranet)
  download (PIK-internal only)