The GTOPO30 database (EROS Data Center, 1998) provides digital elevation data in a regular grid of 30 arc-seconds latitude/longitude. A distribution of terrain slope classes for each 5 arc-min grid-cell of the Digital Soil Map of the World (DSMW) has subsequently been derived from the 30 arc-second DEM and the DSMW as follows:
1. Calculation of ‘quadrant’ slope classes for each 30 arc-second grid-cell of GTOPO30 using 7 classes: 0-2%, 2-5%, 5-8%, 8-16%, 16-30%, 30-45%, and > 45%. Since the computation of mathematical slope angles from 30 arc-second grid-cells gives unrealistic results, exaggerating the prevalence of flat terrain, the following rule-based procedures have been applied:
1a Determine for each 3 by 3 sub-grid the difference in altitude of the central grid-cell with regard to all neighboring grid-cells, e.g., Alt(a) – Alt(e), Alt(b) – Alt(e), etc.
1b Select the orthogonal pair defined by the neighboring grid-cell with the largest altitude difference, i.e., a-e-i, c-e-g or b-e-h, d-e-f.
|
a |
b |
c |
|
d |
e |
f |
|
g |
h |
i |
1c For the selected orthogonal pair of neighboring grid-cells and corresponding altitude differences dAltj, j=1,...,4, determine ‘quadrant’ slope classes according to specified ranges of elevation differences. The following thresholds were used for classification of terrain conditions in the Global AEZ study: £ 2m, 6m, 16m, 40m, 125m, 250m, and > 250m
2. Aggregation of the ‘quadrant’ slope classes by 30 arc-second respectively to: 5 arc-min latitude/longitude grid-cells of the DSMW, and to individual soil association map units.
3. Establishing ‘priority’ soil unit/slope relationships by defining upper limits of slope classes for each soil unit. For use in GAEZ, the soil units have been allocated to five groups with regard to maximum slopes of respectively 2% (max. 8%), 8% (max. 16%), 30%, 30-45%, and one unlimited class. For instance, it is known that the major part of the Fluvisols are related to the flat slopes (0-2%). This approach was first elaborated in South East Asia (Bruggeman and Nachtergaele, 1986) and was further refined by Sys (Sys, 1990) for China. The approach was also used for FAO’s study on World Agriculture: Towards 2010 (FAO, 1995d). In GAEZ the five groups of soils for priority slope classes are defined as follows:
CLASS |
FAO ‘74 |
FAO ‘90 |
Class 1 |
< 2% / max. 8%: |
< 2% / max. 8%: |
Class 2 |
< 8% / max. 16%: |
< 8% / max. 16%: |
Class 3 |
< 30%: |
< 30%: |
Class 4 |
< 30-45%: |
<30-45%: |
Class 5 |
all slopes: |
all slopes: |
4. Establishing for each soil association map unit of the DSMW a consistent allocation of slope classes and soil units, accounting for:
· slope distribution calculated in step 2 above by grid-cell and soil association map unit.
· Slope classes as defined for the soil units in the soil association map unit expansion table of the DSMW.
· the ‘priority’ slope relationships (as defined in 3 above).
5. Allocation of soil units within soil association map units to 5 min grid-cells, according to calculated slope distributions, slope ‘priorities’, and soil association map unit composition.