2         Typology Approch and Products

2.1       Typology Database

2.1.1        Database Description

The LOICZ typology database is developed and maintained with the joint support of the partner project, Biogeography of the Hexacorallia.  It is based at the Kansas Geological Survey of the University of Kansas, operates in Oracle® database software, and is served to users on the World-Wide-Web through an interface designed in Cold Fusion® software.  The features of the database, described below, include the geographic grid and cell structure of the basic design, the present contents of the database in terms of environmental variables (listed in Appendix IV), and the user interface for selecting, modifying, acquiring and analyzing the data.

2.1.1.1            Geographic grid and typology cell structure

The objectives of the LOICZ cell structure are to provide a geographically structured basis for database design that will:

• classify coastal environments in a conceptually useful fashion;
• permit integration and consistent analysis of available global data sets dealing with land, sea, air, and human dimension variables;
• operate at a scale of resolution useful for the data and applications envisioned; and
• provide a manageable number of data points for analysis and global upscaling or extrapolation.  

Half-degree resolution has been selected as the most realistically useful compromise between desired resolution and the available data and methods.  Of the 259,200 half-degree cells on the earth's surface, a total of 47,057 have been identified as primary typology cells (coastal, terrestrial and ocean-I).  The remaining oceanic cells (145,989) are included to provide complete coverage for both the larger LOICZ budget sites and for the NOPP/OBIS partner program, Biogeography of the Hexacorallia.  This cell structure also serves the needs of other partner projects. 

Land or Terrestrial cells (T) are typically defined as the cells containing only land (or fresh water).  Coverage typically extends two cells (one degree) inland from the most landward coastal cell, although they have been extended farther inland in a few areas (e.g., estuaries not well represented by the shoreline data set).  These do not have ocean variables. 

Coastal cells (C) are defined as those containing a significant length of the World Vector Shoreline.  These cover significant areas of both land and (marine or estuarine) water, and are populated with all classes of variables. 

Oceanic I cells (O-I) are those extending seaward from the coastal cells the greater of (a) one degree, or (b) the 50 or 100 m isobath in areas of a broad shelf, or (c) to include all of a biogeochemically budgeted area.  These will have only oceanic and atmospheric variables. 

Oceanic II cells (O-II) are those additional cells needed to complete the in-filling of relatively enclosed water bodies or coastal seas that might be the target of future up-scaled biogeochemical budgeting exercises.

Oceanic III cells (O-III) cover all remaining oceanic areas not included in the other classes.

Inland Cells (I) cover all remaining land areas not included in other classes.

Geographic conventions: NORTH latitude and EAST longitude are positive numbers; SOUTH latitude and WEST longitude are negative numbers.

Cell ID numbers are based on a sequential global grid of half-degree cells (Table 2.1).  Numbering (from 1 to 259,200) begins with the cell centered at 89.5 degrees N latitude and 179.5 degrees W latitude (89.5, -179.5) and proceeds from west to east.  When a full circle (or row, in a planar projection) of 720 cells (360 degrees/0.5 degrees/cell) is completed, the numbering steps one cell south along the -180 meridian and continues sequentially west to east.

Table 2.1  Statistics on cell types and numbers

2.1.1.2  Data contents

Appendix IV provides a listing of the environmental variables, the type of data, cell assignments, and a summary identification of the source.  Biogeochemical budget variables contained in the database are discussed in Section 2.2 and Appendix V of this report.  For a complete review and overview of current database contents and form, go to www.kgs.ukans.edu/Hexcoral/Envirodata/envirodata.html.  Additional examples and illustrations may be found in the CD-ROM accompanying this report.

The environmental data contained in the database are derived primarily from public-domain sources, and metadata listings are provided so that users may access and refer to the original data source.  In order to meet the needs of global synthesis, only datasets that are global or near-global in extent are included in the database at present.  The data are divided into categories for ease of searching and reference: these are (in addition to database structure and location references) Atmospheric, Geomorphic, Human Dimension, Oceanic, Terrestrial, Basin and Budget.

The Basin variables are supplied by Charles Vörösmarty, University of New Hampshire, under the aegis of the IGBP-BAHC project.  These consist of modeled annual and monthly outflow (runoff) for all world river basins above a size threshold of a few tens of thousands of km², plus supporting data in the form of basin population, size, and land cover.  The Budget variables, recently included in the database, are the site and biogeochemical budget characteristics assembled by the LOICZ Biogeochemical Budgets activity (see Section 2.2, Appendix V, and http://data.ecology.su.se/mnode/).

2.1.1.3  User interface features

The WWW-based data access site is publicly accessible for review and occasional use by a guest log-in option.  More extensive users are issued a password on request; this provision is designed to regulate traffic and avoid electronic conflicts, not to restrict access. 

The sequence of normal use proceeds in two phases, first:

·         Selection of database version – selected (short form) or full environmental data, with or without budget data.  This option was instituted to provide archival access and variety in the full database option, while keeping the list of variables that have to be reviewed by the typical user to a workable number.

·         Selection of geographic region – by pre-determined zone or by user input of latitude and longitude.

·         Selection of variables.

·         Assembly (automatic) of the user dataset.

Second, after the data set is assembled in the Oracle system, the interface offers the user a variety of options for review and modification.  These include:

·         Reviewing the selection criteria.

·         Viewing the statistics (tabular) and distribution (histogram) of any variable; the histogram display is user-controlled for range and number of classes, and offers some data transformation options (e.g., log, square root).

·         Excluding cells with null values (no data).

·         Modifying the dataset by excluding or resetting numerical values for any variable above, below, or within a user-specified range.

·         Modifying the dataset by transforming any variable (e.g., log, ln, absolute value, square root).

·         Creating a correlation matrix for the selected variables (with modifications). 

·         Generating a report specifying the data selected and all operations performed.

When the final dataset is assembled, the user may view or download it as a comma-separated text file, and/or send it directly to the Web-LOICZView clustering tool via an internet link.

2.1.2        Database Access

2.1.2.1  Internet-based

The use of the internet to maintain and serve the data is considered the only cost-effective and reliable way to make up-to-date data readily available to a widespread user community.  It is recognized, however, that connection, hardware, and software limitations may make access problematic from some parts of the world.  To reduce these problems, all of the interface tools have been designed to run “server-side” to minimize the amount of material that must be actually exchanged between the site and the user.  However, since even these measures will not address all problems, alternative approaches to data dissemination are also used and under development.

2.1.2.2    Portable

Distribution of CD-ROM versions of websites, datasets, and products has been used by LOICZ as part of its workshop process, and is available on request to people who have problems with internet access.  A particular issue is the provision of remote database equivalents, including the selection functionality.  A product under development by Dr L.T. David and existing in prototype form is a database front-end designed to operate in Microsoft Access®, a widely available microcomputer database system that is compatible with the structure of the central Oracle system.

This product would permit distribution of subsets of the database (on CD-ROM), or provision of the entire system on DVD disks, in a form that would support the same basic functionalities and output products as the WWW-based system.