International Institute for Environment and Development

Report Part Title: METHODS OF LAND EVALUATION AND LAND USE PLANNING
Report Title: SURVEYS, PLANS AND PEOPLE
Report Subtitle: A Review of Land Resource Information and its Use in Developing
Countries
Report Author(s): Barry Dalal-Clayton and David Dent
International Institute for Environment and Development (1993)

Stable URL: http://www.jstor.com/stable/resrep18107.9

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

International Institute for Environment and Development is collaborating with JSTOR to

digitize, preserve and extend access to this content.

This content downloaded from

103.143.6.18 on Tue, 12 Dec 2023 03:14:32 +00:00
All use subject to https://about.jstor.org/terms
 CHAPTER FIVE

METHODS OF LAND EVALUATION AND LAND USE PLANNING

The demands for natural resources information have widened - from emphasis on specialist data; to land
evaluation, predicting the potential of land for one or more uses; to land use planning involving
consideration of land use problems and opportunities, generation of a range of land use options, and
making choices between these options (Dent 1991, FAO 1993 in press). Early impetus for this direction
came from soil conservation issues and from land settlement and irrigation schemes. Most
interpretations of natural resources data have been physical ones - either predicted physical response to
various land uses or in terms of constraints on such uses. Where financial appraisal has been demanded,
this has been tacked on without much change in procedure, either from natural resources specialists or
economists.

The step from land evaluation to land use planning is a big one and goes beyond the physical and
financial assessment of natural resources. Where conflicts of interest arise, conventional methods have
been found wanting, and possible future directions are largely untried. In this chapter, the logic and
assumptions of conventional methods of land evaluation and land use planning are reviewed according
to their principles - from trial-and-error, to the transfer of experience by analogy, to simulation models
(McKenzie 1991, Dreissen and Konijn 1991). Chapter Six deals with what now seems to be the
stumbling block to more effective use of natural resources information and rational land use planning -
the need for participation in the process by all the stakeholders.

5.1 Trial and Error

This is the oldest way of adapting land use to changing circumstances; the only way in the absence of a
scientifically-based program of resource survey and evaluation; and still the usual way. However, the
economic, social and environmental costs can be very high and it is difficult to develop a rational
strategy for new areas, new problems and new opportunities.

Usually, the experience gained is not recorded and most is lost when managers change.

5.2 Transfer of Experience by Analogy

The results of a land use trial, be it a farmer's experience, a field experiment or a catchment study, are
strictly applicable only to that site and, may be, only to that time. The assumption that underlies soil
survey, land systems survey, and most methods of land evaluation is that this experience can be
transferred to other sites by analogy - by assuming that all occurrences of a particular class of land (the
land analogue) will respond in the same way to the same treatments.

Transfer by analogy needs a classification of land that permits the prediction of performance or
behaviour over a wide range of land uses. The analogues have to be identified in the field even though
the attributes of the land that determine performance may not be known or not visible in the field.

Analogues may be defined by classifications of soil, e.g. Soil Series, or land, e.g. Land Systems. For
example, if trials are carried out on Mondha Series, a Typic Chromustert (NBSS & LUP 1987), we
assume that these results apply to all soils of Mondha Series and, to some extent, all Usterts. The method

This content downloaded from

103.143.6.18 on Tue, 12 Dec 2023 03:14:32 +00:00
All use subject to https://about.jstor.org/terms
relies on detailed characterisation of the trial sites and tightly-defined analogues that are rigorously
correlated in the field, across the country and within the higher levels of the taxonomy. This is a
substantial and demanding task that has been achieved by only a few survey organisations.

It is possible to improve a system of classification in the light of experience. The present amendments to
Soil Taxonomy (Soil Survey Staff 1990) are the fifth published amendments since the definitive version
of 1975. However, the combinations and values of land charactersitics that determine the performance
of one crop (or system of management) in one environment are not the same for other crops and/or other
environments (Gibbons 1961), so it becomes difficult to define a manageable number of land analogues
(about 10 000 Soil Series are established in the USA). For precision in the transfer of experience, we are
forced to adopt narrowly-defined mapping units (e.g. phases of Soil Series) that can only be mapped at
large scales. National soil survey coverage at this level has proved not to be a practicable proposition for
poor countries, and for most rich ones too.

5.3 Empirical Land Evaluation Using Expert Knowledge

Expert knowledge of land response to management can be used to interpret natural resources data to
provide information that is, at once, more accessible and more focussed on land management than the
basic data. Four approaches have been widely adopted: Land Capability Classification, FAO Framework
for Land Evaluation, USBR (United States Bureau of Reclamation) land classification system for
irrigation schemes, and Parametric Indices. More recently, decision trees have been advocated as a more
transparent way of using expert knowledge so that it can be built upon by the manager or decision-
maker in the field.

5.3.1 Land Capability Classification

The best known and most widely used method is Land Capability Classification, originally developed
by the United States Soil Conservation Service in the 1930s to interpret conventional 1:20 000 soil maps
for farm planning (Hockensmith and Steele 1949). The definitive account of the system is given by
Klingebiel and Montgomery (1961). It has been adopted and, sometimes, modified by survey
organisations in many developing countries.

Land is classified according to the degree of its limitations for sustained use and the soil conservation
measures necessary to maintain it in productive use (Table 7). The limitations to use that are considered
are those not feasible for the farmer to correct (e.g. climate, slope, soil depth, liability to flooding) and
these are recognized at subclass level. A third level of classification, within the subclass, is the capability
unit which groups soils that require similar management and are suitable for similar crops. Land
allocated to a particular class has capability for the defined land use (e.g. Class I: arable, no restrictions)
and all uses allowed for lower classes (so Class I is also suitable for grazing, forestry and wildlife). This
does not necessarily indicate which use is most productive or profitable.

The classification assumes the technical means, capital resources and skills of the better farmers under
local conditions and a favourable ratio of outputs to inputs. Access to markets, shape and size of land
parcels and land ownership are not considered.

This content downloaded from

103.143.6.18 on Tue, 12 Dec 2023 03:14:32 +00:00
All use subject to https://about.jstor.org/terms
Table 7: Land Capability Classes of the USDA Soil Conservation Service

CLASS DESCRIPTION

I Soils with few limitations

II Soils with limitations that reduce the choice of crops or require simple soil conservation practices

Soils with severe limitations that reduce the choice of crops and/or require special conservation
III practices

Soils with very severe limitations that restrict the choice of crops and/or require very careful
IV management

Soils with little or no erosion hazard but with other limitations that limit their use largely to pasture,
V range, woodland or wildlife

Soils with very severe limitations that restrict their use to pasture, range, woodland or wildlife
VI
Soils with very severe limitations that restrict their use to range, woodland or wildlife

VII Soils and landforms with limitations that preclude commercial crops and restrict their use to
recreation, wildlife and water supply
VIII

There is an in-built assumption that the most desirable land use is arable cropping requiring no special
conservation practices. This determines the choice of limiting factors and the values of limiting land
characteristics assigned as class boundaries:

Class

I Arable (all crops, no conservation practices) Most desirable

¦
II,III,IV Arable (increasingly costly conservation practices ¦
and or restricted choice of crops) ¦
¦
VI Improved pastures ¦
¦
VII Grazing of natural range, or forestry ¦
Ú
VIII Recreation, wildlife, water catchment Least desirable

Class V is an oddball for limitations other than erosion; in practice it is used for wetland.

This content downloaded from

103.143.6.18 on Tue, 12 Dec 2023 03:14:32 +00:00
All use subject to https://about.jstor.org/terms
Expert judgement and field testing are needed to specify the criteria for classes and subclasses, and the
strength and wide application of the system lies in this flexibility - criteria appropriate to local
conditions should be specified. Table 8 shows the conversion table drawn up for Malawi by Shaxson et
al. (1977), who also prescribe the soil conservation practices to be followed in each subclass.
Astonishingly, the original USDA criteria have sometimes been adopted without change in quite
different technological and physical environments.

Land capability classification is relatively easy to do and easy to use. Shaxson (1981) shows how class
and subclass can be mapped quickly and directly, without bothering with time-consuming soil survey.
Users are much more comfortable with land capability maps than soil maps or other complex land
resource surveys (e.g. Woode 1981). After all, class I is obviously the best, class III is less good and
class VI is rubbish, isn't it ? Well, no, but the assumptions and shortcomings of the system are not
immediately obvious.

The most obvious shortcoming is that land cannot be graded from best to worst irrespective of the kind
of management. Some kinds of use have special requirements and tolerances that others do not have, for
example:

- Rice requires a soil with a high available water content and tolerates prolonged flooding;
other cereals will not tolerate waterlogging during their period of active growth;

- Tea, sugar cane and oil palm need efficient transport to processing plants and so have a
minimum area requirement; grain grown for subsistence does not;

- For mechanised operations, stones and rock outcrops are limiting, but with oxen or hand
hoeing you can work round them.

The arable bias of Land Capability Classification and the very generalised nature of the information
does not help choice between alternative uses, except to eliminate the grossly unsuitable. Land use,
productivity and prof itability are often poorly correlated with land capability class (Smit et al. 1984,
Burnham et al. 1987). No one-shot land evaluation can provide the information needed to choose
between two or several land use options and, thus, match land use closely with land suitability. These
objections were addressed in the FAO Framework for Land Evaluation (1976).

5.3.2 FAO Framework for Land Evaluation

The first principle of the Framework is that evaluation is for a specified land use type, relevant to local
conditions in terms of the physical environment and social acceptability. The first step is, thus, to
identify and define promising land use types and establish their land requirements (and also
requirements for labour, capital and infrastructure - so this becomes a substantial, inter-disciplinary
task).

The structure of the Framework is outlined in Box 16. The various land suitability classes are arrived at
by matching the requirements of the land use type with the qualities of the land. The Framework
separates land characteristics - attributes of land that can be measured, like slope angle; and land
qualities - complex clusters of land characteristics that act in concert to determine the performance of a
land use type. Sufficiency of water is a land quality. It is a dynamic attribute of land resulting from the
interaction of several demand-side and supply-side characteristics (Figure 12) so it cannot be measured
directly, though it can be modelled.

This content downloaded from

103.143.6.18 on Tue, 12 Dec 2023 03:14:32 +00:00
All use subject to https://about.jstor.org/terms
Table 8: The Malawi System of Land Classification (Source: Shaxson et al. 1977)

The Special Arable class is restricted to areas with more than 1150 mm mean annual rainfall

This content downloaded from

103.143.6.18 on Tue, 12 Dec 2023 03:14:32 +00:00
All use subject to https://about.jstor.org/terms
 Box 16: Structure of the FAO Framework for Land Evaluation

The target set by the FAO Framework is a four category evaluation:

♦ Land suitability orders . The first categorization is into SUITABLE/NOT SUITABLE for a specified
land use type - dubbed land suitability orders.

Suitable means that sustained use of the kind under consideration will yield benefits which
justify the inputs without unacceptable risk of damage to land resources.

Not suitable means that the kind of land use is technically impracticable, or would cause
unacceptable degradation of land resources, or that the value of expected benefits does not
justify the exp ected costs of needed inputs.

♦ Land suitability classes . These reflect degrees of suitability. Experience of testing land suitability
evaluations against crop performance is very limited. Such experience as there is does not support a
detailed sub-divis ion. Two classes may be enough; FAO recommends not more than three, as follow:

Class S1 land having no significant limitations to sustained use, or with only minor limitations that will
not significantly reduce productivity or benefits and will not raise inputs above an
acceptable level;

Class S2 land having limitations that, in aggregate, will reduce productivity or benefits and will
increase required inputs so that the advantage to be gained from the land use -
though still attractive - will be less than that expected on Class S1 land;

Class S3 land having limitations that, in aggregate, are so severe that expenditure on the land use will
be only marginally justified.

Within the order Not Suitable, there are two classes:

Class N1 Currently not suitable. This land could be used for the purpose under consideration but the
social or economic cost is, at present, unjustified;

Class N2 Permanently not suitable. Land having limitations that appear so severe that sustained use is
not possible.

♦ Land suitability subclasses . These reflect the kinds of limitations, e.g. water deficiency, erosion hazard
- S2w, S3e. There are no subclasses within S1.

♦ Land suitability units . These are subdivisions of a subclass which differ in their response to
management. Units are significant at the farm level. They are distinguished by arabic numbers, e.g.
S2e - 1, S2e - 2.

This content downloaded from

103.143.6.18 on Tue, 12 Dec 2023 03:14:32 +00:00
All use subject to https://about.jstor.org/terms
Figure 12: Schematic Model for the Land Quality Sufficiency of Water

The clarity of thinking behind the FAO Framework has had a profound influence on land evaluation,
paving the way for quantitative modelling. But early applications were qualitative and used the same
kind of matching technique between land and land use requirements as in Land Capability. Typically,
one or more diagnostic land characteristics have been used as surrogates for land qualities (e.g. Table 9).

A refinement is to begin with a cropping calendar for the land use type and, from this, establish critical
periods of the year for different qualities, e.g. traffic ability at sowing and harvest, a dry spell for ripening
and so forth (see Hackett 1988 who also outlines the expert procedure for judging the severity of
individual limitations on crop performance).

In the absence of short cuts that can be taken only by an experienced 'old hand', the procedures
recommended by the series of FAO Guidelines for land evaluation (for rainfed agriculture, FAO 1984a;
forestry 1984b; irrigated agriculture 1985; and extensive grazing 1991) are time-consuming and opaque.
The end product of a qualitative evaluation - that a parcel of land is, say, suitability class S2s for
sorghum but class S3s for cotton - is not a lot more interesting to a farmer or a policy maker than news
that it is in land capability class IIIs.

Another principle of the FAO Framework is the comparison between land use and land in terms of
benefits yielded with inputs needed. If this comparison is to be made in financial or economic terms,
then suitability classes have to be calibrated in terms of yield or other outputs and the inputs needed.
This has not been done often. The magnitude of the task is illustrated by the recent agroecological land
resource assessment for agricultural development planning for Kenya (Kassam et al. 1991, Box 17).
Even though the scale of mapping was only 1:1 million, the study involved a GIS with some 91 000
unique records (in terms of the land and climatic characteristics considered), five principal computer
programs, and a nine volume final report weighing 3.3.kg !

This content downloaded from

103.143.6.18 on Tue, 12 Dec 2023 03:14:32 +00:00
All use subject to https://about.jstor.org/terms
Table 9: Qualitative Application of the FAO Framework for Land Evaluation: Example of Land Requirements
for Bunded Rice, Sri Lanka

(Source: Dent & Ridgway 1986)

LAND QUALITIES DIAGNOSTIC LAND LIMITING VALUES FOR LAND CHARACTERISTICS

CHARACTERISTICS
S1 S2 S3 N

Sufficiency of energy Mean annual temp, o C >24 21 - 24 18 -21 < 18

Elevation, m* 0 - 600 600 - 1200 1200 - 1800 > 1800

Sufficiency of water 75% probability rainfall, mm > 1300 900 -1300 500 - 900 < 500

Soil drainage class Poorly drained Imperfectly Well drained, Excessively

drained Moderately well drained
drained

Soil texture C,SC,SCL,L SC,SCL,SL SL S,LS

Soil depth, cm > 80 60 -80 40 - 60 < 40

Sufficiency of nutrients pH of flooded soils 6- 7 5- 6 4.5 - 5 < 4.5

7- 8 8 - 8.5 > 8.5

Salinity hazard ECe, mS cm-1 <3 3- 5 5- 7 >7

Ease of water control Slope angle, degrees <1 1- 2 2- 6 >6

Ease of cultivation Stones and rock outcrops, % nil 1- 5 5 - 10 > 10

*
Elevation is used to assess sufficiency of energy where temperature data are not available.

This content downloaded from

103.143.6.18 on Tue, 12 Dec 2023 03:14:32 +00:00
All use subject to https://about.jstor.org/terms
 Box 17: Semi-Quantitative Land Evaluation for Kenya

Computer databases for agroclimate and soils were established. Then for 64 crop types, pastures and two types
of fuelwood at high, middle and low levels of inputs, yields were estimated using the FAO agroecological zones
method (Kassam 1977, FAO 1978) as follows:

(i) Maximum attainable yield was modelled using data for photosynthetically active radiation, crop life
cycle, and the crop's photosynthetic efficiency less its respiration requirements. Total biomass
production was partitioned between harvested and non-harvested parts. References to the derivation
and testing of the models are not presented, so values for maximum attainable yield have to be taken
on trust.

(ii) The maximum attainable yields were then de-rated according to the extent to which the thermal and
growing period requirements of the crop are met. Length of growing season was calculated as the
period during which rainfall is ≥ half evapotranspiration calculated by the Penman method.

(iii) The climatically-attainable yields were further de-rated according to soil constraints that affect
sufficiency of aeration, water and nutrients; salinity/sodicity; and slope.

(iv) Hazard of soil erosion was modelled for each crop management/soil combination and further de-rating
of yields applied according to the forecast effect on nutrient and water supply.

For each de-rating step, constraints were allocated by expert judgement into five classes:

S1 yield reduction < 20%

S2 20-40%
S3 40-60%
S4 60-80%
N yield reduction of > 86%

For livestock production, further assumptions were made about conversion of pasture production to livestock
products.

The interpretations of the soil data are problematic in two ways. First, the soil mapping units of the Exploratory
Soil Map of Kenya, scale 1:1 million (Sombroek et al. 1982) are associations of several contrasting kinds of
soil. For each of the 392 mapping units, the percentage composition of each component was estimated (van der
Pouw 1983) but this proportion has to be applied equally to every part of that unit. Secondly, the soils were
defined and mapped according to their morphology, not according to water-supplying and nutrient-supplying
capacity. The key land qualities were estimated from derived information on soil, texture, stoniness, depth,
salinity/sodicity and slope.

No evidence of local calibration of crop, fuelwood and livestock production is presented. The study appears to
be a desk exercise. However, the production estimates for each land/crop/management combination have been
made on an consistent basis and do permit a crude economic evaluation of land use options over the whole
country. Storage of the data in a GIS facilitates recombination of different elements of the information and any
part of the computer programs and corresponding data sets can, in principle, be modified in the light of new
knowledge and/or new objectives.

Source: Kassam et al. 1991

This content downloaded from

103.143.6.18 on Tue, 12 Dec 2023 03:14:32 +00:00
All use subject to https://about.jstor.org/terms
Unavoidable problems of economic evaluations are that costs and prices are themselves ephemeral, and
that performance depends on management as well as land qualities. The effectiveness of management is
difficult to forecast and its ability to cope with problems - physical, social and economic - is not assessed
in land evaluation.

These problems can be overcome to an extent if the data about land qualities is stored in a GIS and
computer software is written to recalculate physical and economic suitability from updated information.
A simple and readily-available example is the ALES computer program (Rossiter and van Wambeke
1993) that can perform the computation of physical and financial suitability from decision rules entered
by the investigator. By this means, any new or revised parameter can be taken account of and the re-
evaluation undertaken quickly. Similarly, evaluations can be performed for a range of possible future
scenarios.

5.3.3 The USBR System

The Land Classification System of the Bureau of Reclamation of the US Department of the Interior
(USBR 1953) was developed for planning irrigation projects. It classifies land in terms of its payment
capacity - the money remaining for the farmer after all costs except water charges are met and after
making an allowance for family living costs. This was an early attempt to integrate physical and
financial criteria of land suitability and was the standard method of evaluation for irrigation projects for
more than 30 years.

Once again, expert judgement is brought to bear to identify specific limits of land, soil and water
characteristics that determine the payment capacity. This has to be done separately for each individual
irrigation scheme, according to the technical and economic setting (Table 10).

Classes 1 to 3 have progressively lower positive payment capacities; class 4 designates restricted land
use or special engineering needs; class 5 is a holding class pending further investigation; and class 6 is
not suitable for irrigation, as it doesn't pay. Classification proceeds directly by survey of the relevant
land characteristics and USBR prescribe the scale, accuracy and survey intensity for different purposes.
Mapping units are identified by a compound symbol (Figure 13). The USBR system works only for a
single use within a specific scheme. The parameters change from scheme to scheme and, also, as costs
and farm gate prices change with time.

Figure 13: USBR Compound Mapping Symbol

This content downloaded from

103.143.6.18 on Tue, 12 Dec 2023 03:14:32 +00:00
All use subject to https://about.jstor.org/terms
Table 10: USBR Land Class General Specifications

(Source: Landon 1984)

This content downloaded from

103.143.6.18 on Tue, 12 Dec 2023 03:14:32 +00:00
All use subject to https://about.jstor.org/terms
5.3.4 Parametric Indices

Parametric methods of land suitability assessment consider just a few key properties of the land and
assign to each property a numerical value. The valuations of individual factors are then combined in a
mathematical equation that produces a single numerical expression of performance or a relative index.
The criteria for selection of key properties, their valuation and combination are, again, defined by
experts.

The underlying assumption is that land suitability or performance is determined by only a few
significant factors. The effect of each individual significant factor is expressed as a response function.
Driessen and Konijn (1992) give the example of the single land characteristic depth of soil which is
positively correlated with production, strongly so when the soil is shallow and tending to an asymptote
when the depth approaches the unrestricted rooting depth of the crop. An index which expresses the
sufficiency of depth of soil on a scale of 0 to 1 could be:

SDI = (1 - exp(-t. SD)

where SDI is soil depth index,

t is a crop-specific coefficient (cm-1 ),
and SD is depth of soil (cm).

All relations and the values of all coefficients have to be established by experiment or field calibration.
Once each significant factor has been evaluated, all the single factor indices are combined, either by
multiplication or some more complex mathematical function to produce a single expression of system
performance.

The best known parametric system is the Storie Index (Storie 1933, 1978) (Box 18), but there is a host
of systems and many have provided locally useful results. The Bonitierungskala (Fackler 1924) was
established to rate land for taxation purposes, originally in Bavaria, later throughout Germany. The
Storie Index classifying land for irrigated citrus in California has been revised locally many times and
there are versions for, e.g. Columbian and Spanish orchards.

Subjective, even arbitrary, decisions are taken by the expert at several stages: the selection of properties
to be used, the valuations of each factor, the formulation of the equation and, not least, the translation of
the final numerical value into planning or operational terms. The Storie Index illustrates common
shortcomings of these methods: it uses compound factors, such as character of soil, which include
factors that are used again and are not independent variables. Above all, the functions are developed and
tested for one application, in one area and at one time. They do not travel well. Tests must be done anew
before each application but, often, they are not.

From the point of view of the land use planner, the system is easy to apply by a non-specialist. But the
final numbers appear as if by magic; the assumptions are hidden; the logic is difficult to retrace. If
performance really is dependent on just a few characteristics, surely it is better to say so explicitly ?

Probably, parametric indices represent an evolutionary dead end in land evaluation, although they
represent the beginnings of calculation. Future developments lie in, on the one hand, more transparent
expert systems and, on the other hand, quantitative modelling of physiological processes.

This content downloaded from

103.143.6.18 on Tue, 12 Dec 2023 03:14:32 +00:00
All use subject to https://about.jstor.org/terms
 Box 18: The Storie Index

The Storie Index Rating is given by:

SIR = A x B x C x X

Character of Texture of Slope Miscellaneous

soil profile surface soil factors

Each factor is scored as a percentage but multiplied as a decimal. The final index is expressed as a percentage.
Where more than one property is considered, as in factor X, each is also scored as a percentage, then all are
multiplied together as decimals and expressed as the combined percentage for that factor.

Examples of ratings from the most recent revision (Storie 1978) follow:

FACTOR A - Rating on character or physical profile

Percent

I Soils on recent alluvial fans, flood plains, or other secondary deposits having
undeveloped profiles 100
x shallow phases (on consolidated materials) 60 cm deep 50 - 60
x shallow phases (on consolidated materials) 90 cm deep 80
g extremely gravelly subsoils 80 - 90
s stratified clay subsoils 80 - 95

II Soils on young alluvial fans, floodplains, or other secondary deposits having slightly
developed profiles 95 - 100
x shallow phases (on consolidated material) 60 cm deep 50 - 60
x shallow phases (on consolidated material) 90 cm deep 70
g extremely gravelly subsoils 80 - 95
s stratified clay subsoils 80 - 95

III Soils on older alluvial fans, alluvial plains or terraces having moderately developed
profiles (moderately dense subsoils) 80 - 95
x shallow phases (on consolidated material) 60 cm deep 40 - 65
x shallow phases (on consolidated material) 90 cm deep 60 - 75
g extremely gravelly subsoils 60 - 90

IV Soils on older plains or terraces having strongly developed profiles (dense clay
subsoils) 40 - 80

V Soils on older plains or terraces having hardpan subsoil layers

at less than 30 cm 5 - 20
at 30 to 60 cm 20 - 30
at 60 to 90 cm 30 - 40
at 90 - 120 cm 40 - 50
at 120 - to 180 cm 50 - 80

VI Soils on older terraces and upland areas having dense clay subsoils resting on 40 - 80
moderately consolidated or consolidated material

This content downloaded from

103.143.6.18 on Tue, 12 Dec 2023 03:14:32 +00:00
All use subject to https://about.jstor.org/terms
Box 18: Continued.

VII/ Soils on upland areas underlain by hard igneous bedrock/consolidated sedimentary

VIII rocks
at less than 30 cm 10 - 30
at 30 to 60 cm 30 - 50
at 60 to 90 cm 50 - 70
at 90 to 120 cm 70 - 80
at 120 to 180 cm 80 - 100
at more than 180 cm 100
IX
Soils on upland areas underlain by softly consolidated material
at less than 30 cm 20 - 40
at 30 to 60 cm 40 - 60
at 60 to 90 cm 60 - 80
at 90 to 120 cm 80 - 90
at 120 to 180 cm 90 - 100
at more than 180 cm 100

FACTOR X: EXAMPLES:

Alkali: according to degree 5 - 100

Nutrient (fertility level): according to degree 60 - 100

Acidity: according to degree 80 - 95

Erosion: according to degree 30 - 100

An example of the calculation of the SIR for the Altamount soil map unit in California is:

FACTOR A Altamount series brown upland soil, shale parent material,

bedrock at 90 cm. Profile group VIII. 70 %
FACTOR B Clay loam texture. 85%
FACTOR C Rolling topography. 90%
FACTOR X Moderate sheet erosion with shallow gullies. 70%

Index rating = 0.70 x 0.85 x 0.90 x 0.70 = 0.37, reported as 37%

A soil map can be annotated with SIRs, if necessary weighted according to the areal proporation of different
soils within compound map units. SIR ratings can also be converted to ranked categories.

Source: McRae and Burnham 1981

This content downloaded from

103.143.6.18 on Tue, 12 Dec 2023 03:14:32 +00:00
All use subject to https://about.jstor.org/terms
5.3.5 Knowledge-based Systems

Dent (1993) and Bouma et al. (in press) have argued that more transparent knowledge-based or expert
systems should be built to answer specific land use questions. Each step in the decision-making process
is flagged, the rules for making the decision are laid down explicitly, and the evidence on which the
current decision is based is also presented. Then, by using the system, it is expected that the user will
become more expert and can change the rules and bring in additional land evidence to reach better
decisions than the expert who originally devised the system.

While computing power greatly enhances the scope of expert systems, it is not a prerequisite. Figure 14
illustrates a manual system to provide a practical guide to afforestation in North Yemen. PLANTGRO
(Hackett 1991) is an example of a computer-based system.

5.4 Process Models

The Penman-Monteith formula is an expression of a physically-based process model that relates

evaporation from short grass to solar radiation, wind and relative humidity. Many other physical models
find application in the various specialist disciplines of natural resources science. More recently, process
models have been developed to predict crop production, risk, hazards of use or inputs needed for a land
use type by combining a large number of individual sub-models like the Penman-Monteith model.
Models with a sound physical basis have a wider potentia l application than analogue models or
empirical (expert) knowledge. In particular, they are quantitative and dynamic. They can, in principle,
provide probabilistic predictions of, say, crop yield or profit under a range of management practices
(Jones 1990). For example, Huygen et al. (1990) have used the WOFOST (Centre for World Food
Studies) simulation model (Box 19) to assess the riskiness of improved technology maize production in
Zambia.

At present, simulation models are complex; their data requirements exorbitant. Their development and
maintenance consume scores of man years of research time in well-found institutions. A great deal of
work is needed to develop and validate the models, and their thirst for quantitative data on specific land
qualities cannot yet be matched by data available from natural resources surveys or knowledge of
farming systems. However, these are not good reasons to dismiss their relevance to decision-support
systems in developing countries with limited data bases and cash-starve d scientific institutions. If the
data are not available for simulation models, they are not available either for empirical models, hybrid
methods like the FAO Agro-ecological zones approach (FAO 1978), or expert judgement.

There is a need for innovative research into simpler yet realistic models. There is also need for surveys
to provide quantitative data on the spatial distribution of relevant individual soil characteristics (like
texture and soil depth) rather than taxonomic units; topographic characteristics (like slope angle,
roughness and length of slope) rather than land systems or landform units; specific agrometeorological
and hydrological characteristics, and so on.

The present state of the art in simulation modelling is comparable to that in geographic information
systems. They are active and exciting research fields needing scarce and costly staff and equipment.
Linking them will greatly increase their applications in land use planning - data can be re-analysed and
presented anew almost instantly when conditions change, or to meet specific client needs, and to
incorporate estimates of risk into assessments (e.g. for cropping in areas with marked climatic
variability). Oh, brave new world!

This content downloaded from

103.143.6.18 on Tue, 12 Dec 2023 03:14:32 +00:00
All use subject to https://about.jstor.org/terms
Figure 14: Key to Forest Planting Sites in the Central Highlands, Montane Plains of North Yemen

(Source: Dent & Murtland 1990)

See explanatory notes

This content downloaded from

103.143.6.18 on Tue, 12 Dec 2023 03:14:32 +00:00
All use subject to https://about.jstor.org/terms
Explanatory notes to Figure 14

a mean annual rainfall <400 mm

fs slight frost
fms mod/severe frost
c continuous calcic horizon
vs steep (slope > 11°)

1 S1 Most suitable. Frost free, rainfall over 400 mm

2 S1a Suitable with limitations due to rainfall (less than 400 mm). Frost free
3 S2 fs Suitable with limitations due to slight frost. Rainfall over 400 mm
4 S2 a fs Suitable with limitations due to slight frost and rainfall (less than 400 mm)
5 S3 fms Limited suitability due to frost (moderate to severe). Rainfall over 400 mm
6 S3 a fms Limited suitability due to frost (moderate to severe) and rainfall (less than 400 mm)
7 N1 c Currently not suitable due to calcic horizon. Rainfall over 400 mm, frost free
8 N1 a c Currently not suitable due to calcic horizon and rainfall (less than 400 mm); frost
free
9 N2 c fms Not suitable due to calcic horizon and frost (moderate to severe). Rainfall over 400
mm
10 N2 c fms Not suitable due to calcic horizon, frost (moderate to severe) and rainfall (less than
400 mm)
11 N2 vs Not suitable due to steep slope and lack of soil. Rainfall over 400mm
12 N2 a vs Not suitable. Rainfall less than 400 mm, steep slope and lack of soil

Data from several years of species trials were correlated with site characteristics. Key site characteristics such as
frost incidence and severity and sufficiency of water were established; then locally-observable surrogates were
identified - qat cultivation for frost hazard, indications of former terracing and landform for sufficiency of water.

Sites can be keyed out by local technical staff. Suitable species and their planting and management
recommendations are provided for each of the site classes.

This content downloaded from

103.143.6.18 on Tue, 12 Dec 2023 03:14:32 +00:00
All use subject to https://about.jstor.org/terms
 Box 19: The WOFOST Crop Simulation Model

WOFOST simulates the production of annual crops by modelling their growth from emergence to maturity.
Three situations are defined:

(1) potential production under optimum water and nutrient supply under a given temperature
regime;

(2) water-limited production under that temperature regime and optimum nutrient supply;

(3) nutrient-limited production or, alternatively, fertilizer requirement for optimum production.

For potential production, the net daily increase in dry matter as a result of assimilation and respiration is
calculated according to agroclimatic conditions, green leaf area, biomass and stage of phenological
development.

Water-limited production is calculated through models of water sufficiency, according to the crop's tolerance of
drought.

For both of these situations, calculations are performed on a daily basis throughout the growing season.
Allowance for nutrient limitation is empirical and made for the growing season as a whole.

The data requirements, in summary, are:

Crop species, climatic and soil type (the package comes with data from representative weather stations,
soils and crops);

Site information, e.g. initial water status, depth to groundwater, nature of soil surface and soil nutrient
status;

Agroclimatic data converted to daily values of minimum and maximum temperature, potential
evaporation, rainfall;

Soil physical data including soil water retention and unsaturated hydraulic conductivity;

Nutrient data, for example uptake rates of N, P and K from unfertilized soils;

Crop data, especially properties determining assimilation and respiration rates and rate of phenological
development.

Further assumptions have to be made to translate the modelled yield into a realistic estimate of production, for
example incidence of disease, losses to pests, water losses or gains by runoff and runon.

Sources: Rappoldt et al. (1986), van Diepen et al. (1989)

This content downloaded from

103.143.6.18 on Tue, 12 Dec 2023 03:14:32 +00:00
All use subject to https://about.jstor.org/terms
5.5 From Land Evaluation to Land Use Planning

Land users have always made land use plans, whether formally or informally. There is also a long
history of sectoral planning. All this has been ad hoc, for individual farms, plantations, forests or parks,
or narrow sectoral purposes - water resources development, roads, etc. While no clearly-defined
planning methodology is discernable, conventional approaches have developed within individual
sectors. For instance, meeting objectives for both wildlife conservation and development has been
greatly assisted through designating particular management categories. Most countries follow the
categories defined by IUCN (1982):

I Scientific reserve/strict nature reserve

II National park
III Natural monument/natural landmark
IV Managed nature reserve/wildlife sanctuary
V Protected landscape or seascape
VI Resource reserve
VII Natural biotic area/anthropological reserve
VIII Multiple use management area/managed resource area
IX Biosphere reserve
X Natural world heritage site

It is usual to prepare manage ment plans for protected areas designated within these categories. These
tend to adopt a standard approach providing baseline information (e.g. on landscapes, fauna, flora) and
including prescriptions for use and development covering, for example, zoning, visitor management,
resource management (e.g. fire control, game cropping, hunting quotas), research, infrastructure,
education, etc. The master plans for national parks and wildlife management in Malawi are an excellent
example (Clarke 1983).

In the forestry sector, many countries have designated areas in various categories (e.g. protected forest
areas, forest reserves, plantations) and have prepared management plans for such areas. Recently, legal
categories have been defined for forests and forest lands by functions and by conditions (ITTO 1993):

1 Protection forests
1.1 Protection forests on fragile land
1.2 Forests set aside for plant and animal species and ecosystem preservation
1.3 Totally protected areas
2 Production forests
3 Conversion forests

Frequently, protected forests are also designated as protected areas under the IUCN protected area
categories. There are no universally agreed criteria for preparing management plans for these and other
forest categories. Usually, plans describe the forest types and percentage cover, the standing stock,
management objectives, logging quotas, controls and plans for regeneration/replanting.

Such planning is often effective within its own narrow terms of reference, so long as the planning
agency is also the executive agency and has some freedom of action; and so long as there are few
conflicts of interest. Land use planning has been most effective and, sometimes, successful in settlement
of empty land and in new plantation and irrigation developments. In these cases, administrative or
engineering concerns have been paramount. Limited natural resources information has been used to
guide the physical layout of farms, roads and the water distribution system and in the choice of crops. A
good example is the work by Hunting Technical Services (1980/81) for the Mahaweli irrigation
development in Sri Lanka (Dent and Goonewardene 1993). Social aspects have received generally less
attention, in some cases leading to conflict, e.g. the case of Barabaig pastoralists affected by the
Tanzania -Canada wheat production scheme in Hanang District, Tanzania (Box 20).

This content downloaded from

103.143.6.18 on Tue, 12 Dec 2023 03:14:32 +00:00
All use subject to https://about.jstor.org/terms
Management of natural resources is much more difficult when there are many independent decision-
makers and management units, several sectoral and administrative authorities, and conflicts of interest
between the various stakeholders. This situation is typical of local-level land use plans undertaken by
district or provincial teams within the Department of Agriculture or an Integrated Rural Development
Plan.

The techniques have been adapted from farm planning and small-scale engineering development for
settler farms, often based on a soil or land capability survey. Several national or departmental manuals
have been drawn up and these show strong family relationships. For example, the manuals of the Land
Use Services Division (1970) - later the Land Use Branch (Zambia Dept.Agric. 1977) in Zambia, and by
Shaxon et.al. (1977) in Malawi, evolved out of earlier manuals prepared by the Department of
Conservation and Extension (CONEX) of the Federation of Rhodesia and Nyasaland. In Zambia, for
example, land use plans follow a conventional format covering description of area, history, physical
conditions, resources, population, communications, present land use, assessment of agricultural
potential, proposals for land use, plot demarcation, bush clearing, extension program, roads, water
supplies, soil conservation, staff and housing, economic appriasal, etc. Good examples are the North
Nyamphande Settlement Scheme (Wilson and Bourne 1971) and the Msandile Catchment Plan (Wilson
and Priestley, 1974).

The difficulties of implementing land use plans in settled areas that have inherited many and complex
problems of land use have already been mentioned in Chapter Three. These difficulties stem from
conflicts of interest within local communities, between government and local people and, not least, the
failure of professional planners and administrators to comprehend and respect these different goals.
These various problems are exemplified by the case of the Barabaig in Tanzania (Box 20). There have
also been difficulties in getting several agencies to work together and, often, a lack of technical solutions
to land use problems that are practicable, profitable and easily-incorporated into existing farming
systems.

This content downloaded from

103.143.6.18 on Tue, 12 Dec 2023 03:14:32 +00:00
All use subject to https://about.jstor.org/terms
 Box 20: The Case of the Barabaig

In 1970, in response to an expected increase in demand for wheat in Tanzania, and with financial and technical
support from Canada, the government - through the National Agricultural and Food Corporation - initiated a
wheat production project which covers 100 000 acres of the volcanic Basotu plains, 12% of Hanang district (see
Figure 15). The scheme is highly mechanised and based on the mono-cropping of hybrid wheat varieties along
the lines of prairie wheat farming in Canada.

The areas of land appropriated by the scheme are also used for dry season communal grazing by the Barabaig, a
tribe of semi-nomadic pastoralists who number more than 30 000 in Hanang district. Each household manages
its herd to maximise production of milk, meat and occasionally blood. But they do not exist on a purely pastoral
diet and maize is obtained through exchange or sale of livestock, and from shifting cultivation by households
with the help of communal labour provided by relatives and neighbours.

The Basotu plains are characterised by drought and lack of permanent water supplies. Therefore, the use of
these plains by the Barabaig involves trade-offs between the productivity and stability of grassland production
in different areas and by constraints of water availability and the incidence of tsetse flies. The sustainability of
the pastoral system is critically dependent on a flexible response to changing patterns of resource availability;
opportunistic natural resource use that allows for the exploitation of key resource patches (e.g. wet depressions,
river and lake margins) at particular periods. The Barabaig have developed their own natural resource
management strategy which includes: traditional seasonal grazing rotation; grazing management; tsetse control
measures (e.g. burning bushland); controlling resource access (common property management); and customary
regulations to control degradation (e.g. banning settlements in certain areas).

However, there are limits to the ability of traditional Barabaig practices to cope with changing circumstances.
The removal of access to muhajega grazing resources (depressions on the plains which provide important dry
season fodder) through the appropriation of land for the wheat project has increased pressure on other resources,
resulting in land degradation. Despite the fact that some Barabaig were resident in the area appropriated, and
that the muhajega were a vital forage resource, this land was described as "idle" during the project assessment
(Young 1983).

According to the Barabaig, cattle populations in areas adjacent to the wheat farms have declined by about 30%
over the past seven years, whilst productivity of the remaining herds has declined due to the loss of land.
Traditional burial sites have been ploughed up, causing considerable cultural discontent.

The loss of muhajega has forced the Barabaig to adopt a new grazing pattern and to rely more heavily on the
remaining forage areas, particularly in times when they would otherwise be rested from grazing. As a result,
they are more intensively used during the critical regeneration period. There is increasing pressure to utilise
more intensively those areas which have low resource potential, e.g. the rift escarpment which has shallow soils
on steep slopes, and the tsetse infested bushland. This will inevitably result in further soil erosion and reduced
production.

The mechanised wheat farming, with no provision made for soil conservation measures, has also caused
considerable soil erosion and siltation of water courses.

The wheat project assessment was inadequate on two important counts. Firstly, it overestimated the potential
economic returns of the scheme. Secondly, the assessment ignored the opportunity costs of reductions in land to
the pastoral system.

This content downloaded from

103.143.6.18 on Tue, 12 Dec 2023 03:14:32 +00:00
All use subject to https://about.jstor.org/terms
Figure 15: Map of Hanang District, Tanzania, Showing Major Land Use Features

Box 20: Continued.

The changes in pastoral land use induced by the establishment of the wheat scheme have resulted in a range of
environmental impacts:

RESOURCE TYPE IMPACT

Upland Range Resources

Plains Decreased perennials

Hills/mountains Soil loss
Bushland Bush clearance

Bottomland Key Resources

Muhajega (depressions or plains) Gully erosion in cultivated land

River and lake margins Heavy grazing and soil erosion, puddling

Source: Lane & Scoones (1991)

This content downloaded from

103.143.6.18 on Tue, 12 Dec 2023 03:14:32 +00:00
All use subject to https://about.jstor.org/terms
5.5.1 FAO Guidelines for Land Use Planning

Guidelines for Land Use Planning (FAO 1993 in press) lays out a planning procedure that can be
applied at field or global level, and drawing on the often painful experience of formal land use planning
over the last 25 years (Dent 1988, 1991a & b). It lays out how to define goals, identify opportunities and
constraints; integrate diverse strands of physical and socio-economic information to devise a range of
land use options and to choose between these. Box 21 outlines the steps involved. The diagram implies
steady progress from one step to another but, in practice we often have to retrace over stepsto take
account of new information, changing conditions or new goals. Figure 16 may be nearer to the truth.

The crucial question remains, who shall be responsible for each step ? As Brinkman (1993) has pointed
out, there has to be a platform for negotiation between the different interest groups where goals can be
agreed, problems prioritized and responsibilities for action allocated by consensus. And this requires
democracy of information, including information about natural resources (Dent et al. 1993).

5.5.2 Integrated Land Use Models

The technical difficulties of handling and combining large amounts of diverse data have severely limited
the use made of natural resources information. If decisions are to be taken on rational grounds, decision-
makers must weigh the natural resources information along with economic and social imperatives.
Usually, the balance is struck intuitively according to the information available to, and understood by,
the decision-maker at the time.

The power of economics in decision-making lies in its reduction of many variables to a single measure -
money - and the general acceptance of a limited range of measures of project worth (net present value,
benefit-cost ratio, internal rate of return). Natural resources information has not been fully used by
decision-makers in government and international aid institutions because it is not easily condensed into
single financial terms. Air, water and soil, for example, are often treated as free resources in economic
planning. This problem has quite recently been addressed by the growing discipline of environmental
economics which seeks to develop economic measures of environmental resources and ways of
incorporating these measures into decision-making. For accessible introductions, see Pearce et al.
(1990), Pearce (1991), Turner (1985) and Turner et al. (1993).

The problem of dealing with a welter of detailed land resource information has been addressed through
mathematical programming techniques that can be carried out by computers and can be linked to
geographic information systems. The LUPLAN package, developed in Australia, has three components
(Ive and Cocks 1983, 1988; Ive et al. 1985; Kessel 1990). First, the area is divided into mapping units or
planning areas and information is assembled on these. Secondly, mapping units are rated in terms of
their attractiveness for a range of land uses on the basis of different policy guidelines. These ratings
express the relative contribution that each mapping unit can make to the achievement of each specified
policy. Third, the relative importance of different policy guidelines is assessed. This means that the more
important (preferred) policies are given greater weight.

LUPLAN can hold individual characteristics of each mapping unit, provide attractiveness ratings for
each potential land use, determine the most attractive land use for each mapping unit, and allocate land
use on this basis. Output is in the form of percentages which express the extent to which individual land
use plans achieve a given policy. Results can be reviewed and votes assigned to

This content downloaded from

103.143.6.18 on Tue, 12 Dec 2023 03:14:32 +00:00
All use subject to https://about.jstor.org/terms
 Box 21: Steps in Land Use Planning

(Source: Dent 1988)

DEFINE PROBLEMS

Step 1: Decide what you want to achieve. Establish the present land use situation; find out the needs of the people; agree
and specify the goals.

Step 2: Plan to plan. Organise the work needed.

Step 3: Structure the problems and opportunities. Socio-economic and bio-physical considerations should be given equal
weight.

MODEL SOLUTIONS

Step 4: Devise alternative solutions. Identify or design alternative land use types that might achieve the goals.

Step 5: Evaluate land suitability. For each promising land use type, establish its land requirements and match these with
the qualities of the land.

Step 6: Appraise alternatives. For each well matched combination of land use and land, assess its
environmental, economic and social impact.

DECISION

Step 7: Choose the best achievable land use.

Step 8: Draw up a land use plan, allocating land to land use and making provision for appropriate management.

TEST SOLUTIONS

Step 9: Implement the plan. Action by decision-maker, implementing agencies and land users.

Step 10: Learn from the plan. Monitor progress. Revise the plan in the light of experience and to accommodate new goals.

This content downloaded from

103.143.6.18 on Tue, 12 Dec 2023 03:14:32 +00:00
All use subject to https://about.jstor.org/terms
Figure 16: An Experienced Practitioner's View of the Steps in Land Use Planning

Source: Dent (1988)

different policies may be changed. Further runs of LUPLAN can be made until broadly acceptable
results are achieved.

Another approach developed by the Canadian group of University of Guelph and Agriculture Canada
(e.g. Smit et al. 1984, Brklacich et al. 1989, Smit et al. 1991) also starts with land mapping units and
sets of policy objectives or land use scenarios. Objectives are quantitatively specified as production
targets. Scenarios specify supply-side conditions such as land availability, quality and productivity. The
relationships between production potential and needs can be expressed in terms of resource use
feasibility, flexibility and sensitivity. Where an area's potential for agricultural

This content downloaded from

103.143.6.18 on Tue, 12 Dec 2023 03:14:32 +00:00
All use subject to https://about.jstor.org/terms
 Box 22: Canadian Integral Land Evaluation Procedure: Example of the Use of
Quadratic Programming to Appraise Flexibility and Criticality of Land Use

A set of simultaneous functions represent:

(1) Constraints Limitations on resource availability and use. For example availability of land, fuel,
fertilizer; goals or targets of production. These are set out first.

(2) An objective function The example chosen simultaneously minimises the proportional allocation of
each land use type to each land unit and also the differences between these allocations. If it were found
that a large proportion of a land unit must be used in some particular way, this would indicate a very
narrow range of feasible alternatives, given the constraints and production targets. Conversely, if only
a small proportion of the land unit need be allocated, this would indicate a high degree of flexibility of
use.

A general model can be specified in mathematical notation for a planning unit comprising m land units in which
each land unit may be devoted to one or more of n land use types.

For xij = amount of land unit j allocated to use i;

Pij = productivity of land unit j for use i;
Xj = supply (or availability) of land unit
Pi = goal (or target or requirement) for activity i.

The variables xij are subject to certain constraints:

a) xij cannot be less than zero: xij ≥0(1≥i-n; 1≥j≥m)

b) Availability constraints, since the amount of land unit j utilized cannot exceed the supply:

n
Σ xij ≤Xj (1≤j≤m)
i=1

c) Goal constraints, which mean that targets must be met for the total output of each land use type:

n
Σ Pij xij ≥Pi (1≤i≤n)
j=1

d) The objective function, which selects one solution from the feasible set and measures
overall flexibility:

n m
Min Z = Σ Σ xij 2/Xj
i=1 j=1

This content downloaded from

103.143.6.18 on Tue, 12 Dec 2023 03:14:32 +00:00
All use subject to https://about.jstor.org/terms
 Box 22: Continued

A measure of the critical importance of a land unit for a land use type is derived from the solution variables:

aij = xij /Xj

Any aij approaching zero means that the production targets could be attained without a large proportion of the
land unit being needed for a particular land use. Any aij approaching 1.0 means that it would be very important
to have most of that land unit devoted to a particular use to meet the production targets. In the simplified
example below, land unit 1 is already critical for agricultural production; land unit 2 becomes critical for
agriculture under a substantial growth scenario, and unit 3 under modest growth and substantial-growth
scenarios.

LAND UNIT LAND USE INDEX OF CRITICAL IMPORTANCE

Scenario 1 Scenario 2 Scenario 3

(base scenario) (modest growth in (substantial
demand) growth in
demand)

1 grains 0.55 0.48 0.18

forage 0.33 0.34 0.76
fruit, vegetable 0.12 0.18 0.06
total agriculture 1.00 1.00 1.00

2 grains 0.00 0.00 0.00

forage 0.06 0.42 1.00
fruit, vegetable 0.00 0.00 0.00
total agriculture 0.06 0.42 1.00

3 grains 0.16 0.41 0.46

forage 0.17 0.58 0.54
fruit, vegetable 0.00 0.01 0.00
total agriculture 0.33 1.00 1.00

(Source: Smit et al. (1986)

production surpasses projected needs (feasibility), the magnitude of the excess provides a measure of the
options for production (flexibility). Should the productive capacity be only a little greater than the need,
then there would be little flexibility but to use all resources efficiently. However, if there is considerable
excess capacity in the production system, there would be many options (much flexibility). If the capacity
for production is less than the need, it is important to estimate the magnitude of the shortfall, so that
policies can be developed to make it good.

Neither production potential nor needs are static. The degree of flexibility of a system depends on the
conditions and targets specified. Sensitivity refers here to the degree to which flexibility is changed
given a change in some variable that influences production potential, needs, or both.

The major characteristics of a production system are specified in the form of constraints and an objective
function (Box 22). A general model can be specified of a production system comprising resource or land
units of different type or quality, where each land unit can be used for one or more activities or land uses

This content downloaded from

103.143.6.18 on Tue, 12 Dec 2023 03:14:32 +00:00
All use subject to https://about.jstor.org/terms
to attain specified levels of production. The solutions or allocation variables are subject to constraints
on resource availability and product demands and the objective function selects from a feasible set a
solution that maximises or minimises some function of land allocation. To predict or prescribe the
allocation of land use to land, the objective function is set to simulate the dominant allocation process,
e.g. maximising profit or minimising outlay.

The procedure has also been used to assess the possible impact of soil erosion (Smit et al. 1988), acid
rain (Ludlow and Smit 1987), land drainage (Brklacich et al. 1987), and climatic change (Brklacich et
al. 1989).

5.5.3 Local Evolution of Practicable Procedures in Developing Countries

Land use planning procedures range from the sophisticated, as described in the preceding section, to the
summary, even arbitrary. Examples of both have been tried in Tanzania, where land use problems arose
both from pressure on the land and resettlement of people in new villages located by administrative fiat.
The problems were addressed first by an externally-funded resource inventory of Tabora Region.
Subsequently, a rigorous village land use planning procedure was developed based on algorithms of
carrying capacity, economic viability, livestock carrying capacity and fuelwood availability (Corker
1982). The procedure proved too ambitious in terms of the time and expertise demanded (at least 45
days by an interdisciplinary team per village plan) and also in terms of the resources available to
implement desired developments. With the benefit of this experience, a simplified procedure has been
evolved (Wheeler et al. 1989) involving a much reduced professional input to produce a framework plan
that can be fleshed out and implemented by the local Village Council (Box 23).

A problem of this approach is that the villagers themselves are not involved from the outset - the
initiative is external - nor in the active gathering and appraisal of data. Furthermore, Village Councils in
Tanzania are essentially organs of government and Village Chairmen are political appointees. As a
consequence, Council decisions do not necessarily reflect the needs, wishes or aspirations of the village
community as a whole.

In Tabora, the planning teams bring in considerable expertise in land resources and planning built up by
the externally-funded land use project. Extension of the simplified procedure to
other areas without this benefit is unproven. In the absence of external funding, the cost of SPOT
imagery has proved prohibitive.

Here we see two continuing problems of local level land use planning:

1) an acute shortage of both professional expertise and of resources for both planning and
implementation;

2) a growing awareness of the need for people's participation in planning without grasping
the far-reaching implications of this for political development and professional
procedures.

This content downloaded from

103.143.6.18 on Tue, 12 Dec 2023 03:14:32 +00:00
All use subject to https://about.jstor.org/terms
 Box 23: Procedure for Framework Village Plans in Tanzania

I Quick appraisal

i Using air photos or 1:50 000 enlargements of SPOT imagery, delineate village boundaries and
measure areas suitable for cultivation and areas actually cultivated.

ii Collect basic data on population and farming systems.

iii Discuss local land use problems with village leaders to arrive at a crude appraisal of the match between
village land resources and village needs.

This takes a three-man team about four days working in the village. On the basis of such appraisals, priority
villages for the next stage of planning can be identified at district level.

II Framework plan

i Sketch landforms, land use, soils, eroded areas, water sources and tracks on 1:50 000 imagery.

ii Field check, especially of soils and water sources.

iii Survey village and sample households to determine population distribution and growth, land holdings,
livestock ownership, levels of production and other economic activities.

iv Assess land suitability and draw up an indicative land use plan. Discuss its implications with the
people concerned.

The whole procedure, including the production of a framework plan, involves a three-man planning team over
about two weeks, living in the village.

Implementation relies on devolution of authority to the Village Council which resolves conflicts and determines
the priorities for development. The resulting framework plan (Figure 17) is its responsibility. It can allocate land
according to customary law, lay out individual farm plots, and manage communal land uses such as woodlots
and grazing reserves. Locally-developed plans can be implemented because they do not rely on major external
inputs.

Source: Wheeler et al. (1989)

This content downloaded from

103.143.6.18 on Tue, 12 Dec 2023 03:14:32 +00:00
All use subject to https://about.jstor.org/terms
Figure 17: Land Suitability, Present Land Use and Framework Plan for a Village in
Tanzania

(After Wheeler et al. (1989)

This content downloaded from

103.143.6.18 on Tue, 12 Dec 2023 03:14:32 +00:00
All use subject to https://about.jstor.org/terms