Increases in the demand for agricultural produce and for space to meet non-agricultural needs are provoking rapid changes in the use of land. These changes have stimulated a critical examination of our methods of looking at land. Most useful is a land evaluation that predicts the inputs, outputs, and other favourable as well as adverse effects resulting from specified uses of the land that is being evaluated (Chapter 1).
LAND UTILIZATION TYPES
Thus, relevant uses need to be identified at an early stage (Chapter 2). This has not always been satisfactory. To help in land evaluation, the concept 'land utilization type' (LUT) has been introduced. This is defined as a specific way, actual or alternative, of using the land, described in terms of produce, labour, capital, management, technology and scale of operations. The principal objective of this thesis is to strengthen the philosophical base of land evaluation by explaining the LUT concept. Many similarities exist between this concept in land evaluation and other land-use defining concepts such as production and farming systems. Due to the complicated interactions that occur between their many constituent parts, the analysis of farming systems cannot fully account for the variation in physical land conditions. Land evaluation contributes to solving this problem by making preliminary and partial analyses of the variability of the land and of its influence on the performance of present and alternative land uses. To this end, land use is arbitrarily subdivided into two elements: the land (LU), mostly described by land evaluators in terms of land (mapping) units, and the use (LUT). Thus it should be possible to predict the performance of different LU, LUT combinations, called 'land use systems' (LU S) in this report. Such a 'landuse systems approach' should permit easy extrapolation of the land evaluation results to farming systems research and land use planning.
LAND REQUIREMENTS AND LAND QUALITIES
In agronomy the term 'requirement' is commonly used when referring to the specific land conditions required for the successful growth of a crop or the functioning of an agricultural implement, e.g. the water requirements of wheat or the soil workability requirements of a tractor-driven plough. These land requirements (LR) are the most fundamental aspects of the land utilization types for purposes of land evaluation (Chapter 3). A very critical aspect of land evaluation is the availability of information
about these LRs, especially in developing countries. Most useful are the descriptions of LR expressed in terms of relationships between different levels of specified land conditions and the corresponding levels of output, e.g. a table or graph that relates different levels of soil salinity with yield.
The land requirements of a LUT determine to a great extent which land resources data need to be studied and in how much detail. Early identification of LUTs and their land requirements may considerably reduce the cost and duration of the land resource studies by focussing attention on those land characteristics that may not wholly meet these requirements. In any event, land resource studies result in an enormous amount of data about soil, climate, hydrology etc. But, because of the way data are collected according to the academic discipline of the researcher, important relations and interactions between different land attributes are often overlooked, particularly those between climate and soil. To synthesize the overwhelming volume of data into a more comprehensible form, the construction of simple functional models of the physical environment (LU) is proposed, based on the concept of land qualities. A land quality is a component of the land which acts as a separate factor on the land-use performance. The following broad types of land qualities have been distinguished:
- ecological qualities; e.g. available water, length of growing season;
- management qualities; e.g. the possibility of using specified types of implements or transportation;
- conservation qualities; they represent the land's unique capacities to maintain the status of the land qualities, in particular the productive capacity;
- improvement qualities; land units differ in behaviour when certain physical inputs are applied for their improvement: they have a different 'input application efficiency', e.g. in their response to fertilizers or irrigation water.
There is still much to be achieved in the quantitative measurement of land qualities. They are usually ranked on an ordinal scale: high-medium-low-very low. Statistical methods, such as multiple regression and principal component analysis, are also used as a means of rating land qualities, e.g. soil fertility or soil erosion susceptibility. The success of these statistical methods for describing land qualities seems to be attributable to the small number of factors taken into account. The prospects of using mathematical and analog models for characterizing and simulating dynamic land qualities influenced by the weather, e.g. the soil workability, oxygen contained in the soil, soil nitrogen, are very important. The timing of land-use activities and processes of the LUT - the cropping calender - affects the way in which the time intervals need to be chosen for measuring and simulating dynamic land qualities and component properties.
Land evaluation should be able to predict the impact of land use proposals not only for single land (mapping) units but also for combinations of land units and for the physical environment as a whole. Also, interactions occurring between different land
uses operating on different land units should be foreseen. For this purpose, a distinction is proposed between internal land qualities of individual land units and overall land qualities of major landscape elements, internal land requirements of individual land utilization types and overall land requirements encompassing the sum of individual land requirements made by the different land utilization types that operate simultaneously.
APPROACHES TO LAND EVALUATION, LATIN AMERICA
In Chapter 4 the different approaches to land evaluation are presented. At the highest level, a distinction is made between general purpose and specific purpose land evaluation. General purpose land evaluation represents a standardized approach for all lands to evaluate their capability to support a generally defined land use. The best known example is the USDA Land Capability System. Specific purpose land evaluation represents a pragmatic approach: not only the land but also the use possibilities (LUT) are explicitly studied. The use (LUT) becomes as much a determinant of land suitability as the land itself. Many land suitability classifications for specific crops belong to this category.
To compare the performances of different LU-LUT combinations, not only an analysis of the physical factors is needed (physical land evaluation), but also a socioeconomic analysis. The approach to land evaluation that includes socio-economic analysis has been named 'integral land evaluation'.
Application of the proposed concepts and procedures of land evaluation requires close contact with the farmer: his operations need to be observed, and his achievements, attitudes, and expectations taken into account. During field surveys, one should always be on the look-out for potentially constraining land qualities. Observation of present land use and discussions with farmers will improve the correspondence between the real land conditions and their descriptive models in terms of qualities and properties.
To illustrate the above concepts and procedures in land evaluation, methods from Venezuela, Nicaragua, Mexico, Brazil, and Chile are described. In Latin America land evaluation is relied on as a fundamental source of information for agricultural development. Land evaluation methods that evolved in other countries, especially the USDA Land Capability System, have not been rigidly followed. New systems are being developed to suit local needs. The willingness of national scientists to abandon established methods of land capability classification is encouraging the introduction of new approaches that pay more attention to the biological, technical, and socioeconomic aspects of land use, and in particular to the farmer himself, In such specific.purpose land evaluations, the dynamic aspects of land and land use can no longer be ignored; this is making land evaluation more complex, but not insurmountably so, given today's data- handling techniques. This idea is elaborated in Chapter 5, where the possibilities for using systems analysis are explored.
LAND-USE SYSTEMS ANALYSIS
Land-use systems analysis in land evaluation must be understood as 'simulation', defined by De Wit and Goudriaan (1974) as the building of a dynamic model and the study of its behaviour. The land-use model only includes that part of reality that is needed to answer the questions asked: to predict inputs (I), outputs (Y) and changes in the values of land qualities (LQ), on-site or off-site, that would arise if a particular LU were to be combined with a specific LUT.
To be able to provide this information the relation structure of the land-use system must be known. This consists of three fundamental relations:
Y = F(LQ); LQ = F(I); Y = F(I)
These relations are interrelated; one relation can be derived from the two others through the elimination of one variable, usually LQ A graphical method of co-axial analysis is shown for expressing the I-LQY relations.
A further simplification is the tabular presentation of the relation structure of a landuse system, presenting only a few input-land quality-output combinations. Two multiple- entry tables are proposed: the land quality table (Table 5.4A) expressing the input-land quality relations for land units(LU) with different land improvement qualities, and the output table (Table 5.4B) expressing land quality-output relations for land utilization types (LUT) with different land requirements. Combination of the two tables permits the identification of several alternative input-land quality-output combinations for each LU- LUT combination.
A distinction is made between descriptive and prescriptive land-use systems analysis. During the descriptive analysis, physical inputs for ameliorating constraining land qualities, their management and conservation, are compared with their effects on the land qualities and the outputs: 'descriptive input-output analysis'. This information is needed for the next step, when the suitability of a particular land unit (LU) for combination with a particular land utilization type (LUT) is classified: 'prescriptive land suitability classification'. During this second step, for each LUT-LU combination, the input-land quality-output combination is selected which places the land unit in the highest possible land suitability class: a kind of optimization process.
The use of simulation models of specific land use processes and mechanisms holds much promise for land-use systems analysis and is therefore likely to increase, particularly in situations where the physical and/or socio-economic conditions seriously limit a satisfactory matching between land qualities and land requirements.
Such models will probably relate primarily to specific partial land-use problems, e.g. drainage, soil tillage, the behaviour of nitrogen or chemical fertilizers, and to potential yield. In the immediate future the use of mathematical models solely for simulating all input-ouput relations influencing the performance of a land use system will probably remain too complex to satisfy practical land evaluation entirely. Thus land evaluation must compromise between scientific ideals and the limitations posed by data availability, data reliability, and the possibilities for data handling.
Meanwhile land resources inventories should aim increasingly towards the collection of data that explicitly characterize the fundamental environmental regimes (i.e. land qualities) influencing the physiological and agricultural mechanisms and processes, to improve the possibilities for land-use simulation and the prediction of landuse performance.