International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B7. Istanbul 2004
Land use is dynamic and operations and events take place that
the surveyor cannot personally witness. The land user has that
capability and is (assumed) able to recall important land use
system aspects for at least the past growing season. His/her
knowledge regarding those aspects can only be sampled
through interviews. For land use surveys, interviews at plot
level are thus essential. To complement and verify interview
information, observations by the surveyor must be recorded on
a field form. Interviews can be based on questionnaires or on
checklists; each has its respective strengths and weaknesses.
Use of checklists is recommended for scientific research to
prevent that relevant site-specific operation data or occurrences
are overlooked.
Table 3. Overview of ‘Information Data Sets’ with Specific
Relevance for Land Use Surveys
RS-Image characteristics
e 1D-features (tone, color), as related to: >
e crop calendars, cropping patterns and other land use
operations
e infrastructure
e 2D-features, such as:
e field sizes, shapes and patterns
e internal patterns (textures, grids, mottles)
e line features
e 3D-features (on APs): vertical structure, no. of layers
Observation / measurement data:
e plot size, coordinates, slope, position, etc.
e crops (residues) and infrastructure present in / around the
plot
e land cover data (crop condition, growing stage, weed
incidence, biomass, height, etc.)
e ground cover status (bare soil, mulch, crop residues, etc.)
e specific observations (soil characteristics, tillage condi-
tion, erosion status, hydrological aspects, pests / diseases
incidence, evidence of grazing, etc.)
Interview data:
e holding/holder information (profile)
e site aspects (tenancy arrangement, cadastral no., distance
to holding)
e land use system (plot) aspects for the period considered:
e a-priori land use class
e crops grown / services provided (% of area, numbers,
etc.)
e land use purposes
e operation aspects (the crop calendar and cropping pat-
tern):
e operation name; species involved; % of plot in-
volved; period / periodicity / duration and task
times; main power source
e labor and material inputs and implements used
e products / benefits obtained
e observations by land user and indigenous knowledge:
e soil related (workability, infiltration rate, fertility
- status, etc.)
e weather related (hail storm, dry period, etc.)
e crop related (pests, diseases, lodging, wilting, etc.)
5. LAND USE CLASSIFICATION CONCEPTS
Classifiers
There is enormous variation in land use worldwide. To map
land use, compile land use statistics, and carry out land use
planning, common characteristics in the wide variety of land
uses must be identified. Common land use characteristics can
be identified in two ways:
e By generalizing the descriptions of actual land use systems
to a description that conforms to, e.g. land use names / de-
scriptions in map legends; these descriptions hold only for
specific areas and periods of time.
e By classification of land use descriptions resulting in de-
scriptions that are not limited to a certain area or time
frame.
Land use classification was defined as: "The process of defin-
ing land use classes on the basis of selected diagnostic criteria",
and a land use class as: "A generalized land use description,
defined by diagnostic criteria that pertain to land use purpose(s)
and operation sequence followed; it has no location or time
indications." Land use classes are exclusively based on attrib-
utes of land use in the context of a LUS.
Classification (of land use) must be based on unambiguous
diagnostic criteria that are known as “classifiers”. Often classi-
fiers are not properly documented in land use (classification)
reports; only names of classes are given.
A land use class is a taxon that is solely based on information
on land use purpose and operation sequence. In combination
with attributes of land, the land use class becomes extended to a
LUS-class. Using LUS-classes does not allow assessing the
suitability of a certain land unit for a certain land use or for
monitoring land use changes. In spite of this, land characteris-
tics are sometimes considered as classifiers, resulting in land
use system classes such as "un-used bare soil" or "protected
tropical forest".
Three types of classifiers can be applied to define land use
classes (Figure 6 and the Appendix):
e Land use purpose classifiers
specify aimed at [Species/Service - Product/Benefit] com-
binations in general terms. At least one combination must
be specified for each land use class. No new products or
benefits can be added to define sub-classes, but existing
definitions can be sharpened or split into several new defi-
nitions.
e Land use operation sequence classifiers
specify (one or more) aspects of operations in general
terms. For sub-classes new classifiers can be added; higher
level classifiers remain valid for all sub-classes, or can be
further narrowed down.
e Land use context classifiers
specify (one or more) circumstantial aspects of the land
use in general terms that are not a part of the land use pur-
pose or operation sequence. Context classifiers are better
not used but are mentioned to link up with existing prac-
tices. They can include: Land aspects, ¢.g. infrastructure,
tenancy arrangements, etc.; Holding (context) aspects, e.g.
origins of inputs/implements, destinations of outputs
(market orientation), capital intensity, holder attitude,
goals of holder, credit availability, pricing policies, etc.
The parametric method of defining land use classes employs a
combination of classifiers to define a land use class. Table 4
presents an example of a land use class defined in terms of
independent classifiers. The Appendix contains lists of opera-
tion sequence and context classifiers that can be used to define
1152
Va tc . U
