Kenya Coastal Zone - Scientific Cruise

The Eastern African Coastal and Marine Environment Resource Database
is a comprehensive 1:250,000-scale vector database of the Kenya
Coastal Zone. It consists of geographic, attribute, and textual data
which can be accessed, queried, displayed, and modified. The database
was developed under the Eastern African Action Plan, with the
collaboration of UNEP/GRID-PAC, UNEP/OCA-PAC, and KEMFRI. The primary
sources of data are the Survey of Kenya 1:250,000 series, National
Museums of Kenya, Kenya Ports Authority, Coastal Development
Authority, Kenya Wildlife Service, and Kenya Government Departments:
Fisheries, Agriculture, Meteorology, Mines and Geology.

Data Sources

The primary sources of data are the Survey of Kenya 1:250,000 series,
Landsat Thematic Mapper images, and socio-economic data from various
Government ministries, departments, and institutes.

Naming Conventions

A naming convention was established to allow users to identify the
contents of layers based on their name. Coverage names begin with a
two letter abbreviation representing the country code (for example,
KE:-Kenya) followed by a descriptive term representing the theme. The
country code being two characters in length plus the descriptive term,
which is usually six characters long, conforms to the Direct Operating
System (DOS) naming convention. An example of this would be as
follows:- the elevation coverage would be called KEELEVAT, the KE
representing the abbreviation for Kenya and the ELEVAT representing
elevation. A full list of country codes for the East African region
is as follows:-

List by country name
Country Name Data Code
Comoros CN
Kenya KE
Madagascar MA
Mauritius MP
Mozambique MZ
Reunion RE
Seychelles SE
Somalia SO
Tanzania TZ

Edgematching
Edgematching was done manually working from the top most map sheet
(Kolbio: SA-37-8) to the bottom most (Mombasa: SB-37-3). In this way
any errors would be distributed in a systematic way. The greatest
errors are in the order of 750 meters on the ground which is 3
millimeters on the map. These errors occurred between the following
map sheets:- Voi (SA-37-14) / Kilifi (SA-37-15)and Voi (SA-37-14) /
Lushoto (SB-37-2).

Annotation
All feature names which include points, lines or polygons have an
entry in the attribute table describing the feature. Additional
attributes may also exist for the particular feature, however this
varies from feature to feature.

General production process
Coverages developed for the database were derived from three sources:
1:250,000 scale paper maps prepared by the Survey of Kenya, Landsat
Thematic Mapper Data, and socio-economic data from the Kenya Marine
Fisheries Institute (KMFI), the Kenya Wildlife Service (KWS) and other
Government Ministries and NGO's.

1:250,000 paper maps
Twelve TIC's (control points) were selected from the map sheet, based
on the latitude/longitude grid on the sheet. The reason for this is
that one of the other sheets only has a latitude/longitude grid where
as the others have both a latitude/longitude grid as well as a UTM
grid. In this way, consistency is being maintained between all the
map sheets covering the Kenya coastal zone.

Arc/info conversion MACRO PROGRAMME <DEGUTM.SML> was used to convert a
standard ASCII text file of Latitude/Longitude coordinates into
Universal Transverse Mercator coordinates.

SML TO PROJECT DECIMAL DEGREES FILE INTO UTM ZONE 37
INPUT
PROJECTION GEOGRAPHIC
SET INPUT UNITS AS DECIMAL DEGREES
UNITS DD
SPHEROID CLARKE1880
PARAMETERS
OUTPUT
PROJECTION UTM
UNITS METERS
YSHIFT 10000000
PARAMETERS
39 00 00
00 00 00
END
[ARC] Createcov ***tic
{ARC} Tables
> Select ***tic.tic
> Add
> TICID = ?
> XTIC = ?
> YTIC = ?
This is done for all 12 UTM Tic Ids and coordinates.
>List
This is done to check that the Tic Ids and coordinates are correct.
> Q stop
A tile boundary was then added to the coverage to help in the
digitization of the map sheet. This coverage <***tic> was then used
in the creation of all the other coverage layers that were digitized
from the map sheet.

1:250,000 paper maps were used due to lack of stable mylar or acetate
copies. All features:- points, lines, and polygons were digitized
using PC Arc/Info 3.4D Plus software running on a Gateway 2000 P5-90
PC and a CalComp 9100 digitizing board. Editing was carried out to
eliminate obvious errors, after which the data was plotted at scale
using a Hewlett Packard Design Jet 650 C inkjet plotter and then
compared to the source for positional accuracy, completeness, and
topological correctness. All of the data layers were checked using
this method and all edits were verified.

Landsat Thematic Mapper Data
Two full scenes and one quad were used in the land cover
classification, more details concerning the methodology used in the
classification can be found in Annex 3.

Line and Point attribute codes were assigned interactively at the time
of initial data capture. Polygon attributes were assigned after
topology had been reached. Editing was carried out to eliminate
obvious errors, after which the data was plotted at scale and then
compared to the source correctness. All of the data layers were
checked using this method and all edits were verified.

All the data was finally plotted at scale in a single composite and
attribute code value validity, attribute code consistency, topology
errors, attribute field definition correctness, and internal data
structure correctness were checked. In addition Arc/View 2.1 was used
to carry out a quick visual check on the data.