Based on the delineation of watersheds at different pixel levels (cf. appendix II), ALCOM developed the database that accompanies the watershed polygon file. The different steps in the development of this database include: ordering, naming, downstream and upstream sequence generation, assignment of colours for mapping and calculation of surface areas.
A database with four fields was created based on manual ordering. Most names were based on a map for Central and Southern Africa from J. Bartholomew & Son LTD at a 1:5000000 scale from 1988. Names for the South Africa watersheds were based on the digital river map from the IWQS (Roodeplaat, ZA). Some names were taken from national maps, preferably the most recent hydrological maps when available. In certain cases the Operational Navigation Charts from the US Defense Mapping Agency provided supplementary information on river names.
Database fields from the first step of ordering and naming:
| 1. | WS ID | Integer Identifier for the watershed, as in WS_SADC.BNA. |
| 2. | NAME_1 | Name of the main river in the watershed (coastal drainage if no unique outflowing river) |
| 3. | NAME 2 | Name(s) of secondary rivers in the watershed |
| 4. | D1 | Identifier of the watershed just below this watershed |
The identifiers of all downstream basins (D2 to D39) were generated automatically by repetitive linking of field 4 with field 1. This resulted in the database file WS_D.DBF. A maximum sequence of 39 watersheds was found in the generated database.
Based on the created downstream basins, the watershed was built up again from the lowest basin in each mega-watershed. In most cases this was the ocean (with identifier 0) but it could also be an internal basin (with identifier 9999) or an edge basin (identifier 8888). This was done by aligning all cells to the right with a macro in a spreadsheet program. The result of this process was stored in:
fields UP37 to UPO with:
| UP0 | ID of the drainage of the lowest basin (e.g. 0 for the ocean, 9999 for internal drainage, 8888 for edge basin) |
| UP1 | ID of the lowest watershed in the megabasin (after which the whole basin is named) |
| UP39 | ID of the highest basin in the watershed |
The UP(i) fields were stored in WS_UP.DBF
Based on the former classification the following fields were generated:
| LEVEL | level of the basin: no of watersheds above lowest watershed in the megabasin |
| MEGABASIN | name of the megabasin: typically the name of the lowest watershed |
These fields were stored in the core database, WS_SADC.DBF together with the fields that were created in the first step (of. 2.1).
Colour codes were assigned to megawatersheds in order to make sure that two adjacent megabasins would have different colours. A small colour code database determines what colours were used for mapping. These data are stored in WS_CLR.DBF and CLR_CODE.DBF.
All the downstream ID's (D1 to D39) were combined in one text string, delimited by commas in order to allow combined searching for these watersheds in a database. This was done by concatenating D1 to D39. The result was stored in WS_DOWN.DBF. The exact purpose of this string is explained in part 4.
Surface areas (in km2) were calculated in Idrisi 2.0 for Windows and linked to the main database file. This field is called AREA. It allows to calculate total surface areas for megawatersheds or upstream watersheds.
