Analysis of Land Use Change: Theoretical and Modeling Approaches
Helen Briassoulis, Ph.D.


5.1. Theories and Models of Land Use Change: The Main Issues
5.1.1. Summary of the main issues pertaining to theories and models of land use change
5.1.2. Scale issues in the analysis of land use change
5.1.3. The relationship between theories and models of land use change
5.1.4. The use of theories and models in land use decision making
5.2. Future Research Directions: Theoretical, Methodological and Practical Needs

5.1. Theories and Models of Land Use Change: The Main Issues

Land is used to meet a multiplicity and variety of human needs and to serve numerous, diverse purposes. When the users of land decide to employ its resources towards different purposes, land use change occurs producing both desirable and undesirable impacts. The analysis of land use change is essentially the analysis of the relationship between people and land. Why, when, how, and where does land use change happen? To provide answers to these closely interrelated questions, theories have been advanced and models have been built in the last 200 years. This contribution attempted to provide a panorama of theoretical and modeling approaches to the study of land use change as well as to examine broadly how well they reflect the drivers, processes and implications of this change. This section first summarizes the main issues

which pertain to the theories and the models presented in chapters 3 and 4 and, then, focuses on a number of selected broader issues; namely, the importance of scale in the analysis of land use change, the relationship between theories and models, and the use of models in making land use decisions. The whole discussion is set within the context of the broader quest for theories and models of land use change which can offer meaningful and essential guidance in understanding land use change and making decisions for future sustainable uses of land in a variety of real world settings.


5.1.1. Summary of the main issues pertaining to theories and models of land use change

The presentation of theories of land use change made clear one basic point – that each theory focuses on particular aspects of the subject. Each theorization tradition specializes more or less on a given spatial and temporal level which determines, to a considerable extent, the nature of and the emphasis placed on the components of the system studied. At lower levels, theorizing is usually more concrete and gives (or aspires to give) more realistic accounts of the agents, context and mechanisms of change. At higher levels, theorizing is more abstract and getting from theory to the real world is not always simple and straightforward. Similarly, each theorization tradition conceptualizes land and land use change differently; some conceptualizations are more realistic while others are abstract and "space-neutral". Some theories specify the land use patterns that result in the process of change while some others give only vague indications. The former are associated usually with a state of equilibrium while the latter make no such assumption considering land use change as a continuous, rarely equilibrating process.

The role of the theorization tradition is critical with respect to the identification of the drivers of land use change. Some theories emphasize the economic, some others the socio-political, while some others the environmental determinants of land use change. The recent trend is towards more integrated theoretical schemata although the influence of the "mother discipline" remains strong in most cases. Explanation of land use change and direct reference to the mechanisms of change varies also considerably among theories depending on their epistemological basis. This is one of the reasons why, in their present form, very few theories have filtered down to models of land use change.

Lastly, the desirability and possibility of a general theory of land use change is basically an open question. The diversity of real world situations casts doubt on whether a general theory of land use change will be able to provide, besides broad explanatory driving factors, patterns and processes of change, those details which may be critical in explaining land use change in particular contexts and circumstances. At present, it seems that a sensible approach to formulating a general theoretical framework for land use change is to attempt a synthesis of extant theories employing each at the spatio-temporal level of detail for which it is mostly fit.

The models of land use change presented in Chapter 4 constitute a diverse universe in terms of purpose and object of study, level of aggregation, dynamics, underlying theory, functional specification, data requirements, and real world applications. Descriptive, predictive, prescriptive, and impact assessment models of land use change have been built for urban/metropolitan areas, regions, nations, as well as for groups of regions and nations and the globe as a whole. The level of functional and spatial aggregation of the models varies with their purpose and object of study mainly. Models which account for a few (two or three) uses of land have been the norm until now. The level of spatial representation ranges from a few, coarse zones to a great number of detailed zones. Spatially explicit models at fine levels of spatial resolution -- the individual parcel level -- are increasingly being developed as the required computational and technological infrastructure improves continuously and as data at this level are becoming available. The improvement in the functional detail of types and drivers of land use will hopefully follow suit soon.

Although land use change automatically implies the concept of time, dynamic models are rare at a level of spatial and functional resolution which is relevant in most practical situations. The difficulties of building truly dynamic models are not only technical but theoretical as well. Specifying and interpreting the results of a dynamic model requires a corresponding theory of change. But models differ in important respects, in terms of underlying theory. Broadly, there are models which are based on some kind of theory and those which are not. Models based on an explicit theory do not necessarily provide acceptable accounts of change, however. The same is true for models which adopt an instrumental approach utilizing simple theoretical statements to justify their assumptions and functional form.

Despite the diversity of model functional forms, the majority of models adopt simple functional forms – statistical or linear programming models – or rely on heuristic techniques (simulation). The recent emphasis – which runs in parallel with the progress made in GIS – is towards spatially explicit models which make possible the explicit treatment of the spatial incidence of the causes and the resulting changes of the uses of land. However, a major impediment to the full realization of such models – in addition to the lack of appropriate theories – is the availability of data of a given quality and specifications; more specifically, the demand is for compatible, consistent, reliable, timely, updated, transferable, and low-cost data sets.

Lastly, of the variety of models of land use change presented in this contribution, several have remained at the level of the proposal, some have been calibrated with data from various real world settings, while some others have been used in the context of policy analysis. Whether and to what extent the use of models has improved decision making on land use issues is a question which cannot be answered satisfactorily as important information on the "politics" of model use is not usually reported.

5.1.2. Scale issues in the analysis of land use change

The importance of scale – spatial, temporal, institutional, etc. – in the analysis of land use change has been mentioned on several occasions in this contribution. In fact, the issue of scale is a cross-cutting theme for both theories and models and for all disciplines which study spatial phenomena (at least). This section brings together and summarizes the main issues related to scale in this context starting with a definition of the term.

The usage of the term "scale" as well as its connotations are not uniform among scientists and among disciplines. Gibson et al. (1998), in a concise treatment of scaling issues in the Social Sciences, provide a definition of scale as "the spatial, temporal, quantitative or analytical dimensions used by scientists to measure and study objects and processes" (Gibson et al. 1998, 8; the definition first appeared in Turner et al. 1989; see also, Turner and Gardner 1991, 7). Related to this conceptualization of scale, are the terms level, extent, resolution, grain, hierarchy, absolute and relative scale (Definitions of these terms are given in Appendix 5.A). The literature does not comply always with this definition using it frequently interchangeably with the term "level" although this latter term denotes a subdivision on a scale (Gibson et al. 1998).

Based on the above definition, several types of scale can be distinguished. The spatial and the temporal scales are the most widely known and used. However, various other scales are being used, sometimes unknowingly or without properly acknowledging their usage. Buttimer (1998) provides a useful categorization of these scales. "Administratively-defined scales, from local through national to global, through which various social and political functions are normally processed; these are spatially circumscribed and representable as a nested hierarchy of domains …. Secondly, there are functionally-defined scales, such as those within which industrial systems, urban fields of influence and service networks radiate their influences. These are nodally-organized spaces and their dimensions reflect the varying strength of particular functions, production or service-based sectors of the economy…. Perceptually-defined scales of reach vary among people and places, reflecting culturally-diverse traditions and aspirations" (Buttimer 1998, 18-19). "Reach" denotes access to (and responsibility for) resources, information, and decision making (Buttimer 1998).

The above discussion implies that scale, irrespective of type, is not given always in an absolute sense but it is socio-culturally produced and modified and varies with the phenomenon being studied. The distinction between absolute and relative scale underlines this fact. In the discussion of the theory of uneven development, it was mentioned that the capitalist mode of production produces different scales as a way of coping with its need for a spatial fix to its internal contradictions. The famous problem of the ecological fallacy in the Social Sciences as well as of the more general modifiable areal unit problem (MAUP) in geography are among the most salient manifestations of the significance of (spatial) scale. In landscape ecology models, scale is the fist and foremost concern in analyzing land use change. In the present context, the issue of scale is crucial because of its important implications for the analysis of land use change. These are discussed in the following under the headings of: (a) definition and classification of land use types, (b) measurement/assessment of land use change, (c) explanation of land use change, (d) assessment and evaluation of the impacts of land use change, and (e) decision making in the context of land management and land use planning.

Land use classification systems are tied usually to particular spatial scales but they reflect also functionally-defined scales such as the scale of agricultural or industrial organization. At the world level, the FAO distinguishes four or five major land use types. At the level of nations, the number of land use types increases to around ten. At this level, other scales enter the classification system to differentiate further the land use typology – agricultural land is further subdivided according to the type of product (annual, perennial), pastures are distinguished according to ownership status into public or private, etc. At the level of a parcel of land, land use classification becomes very detailed capturing local environmental, socio-cultural, demographic, economic and other details. In general, at lower levels of the spatial scale, land use types are defined along additional scales which reflect modes of economic, social, and institutional organization. The analysis of land use change is essentially performed at the level of detail of the land use classification system adopted which, in its turn, reflects a certain combination of scales. It is, thus, important to examine the extent of agreement between the intended level of analysis and the actual level at which the analysis is eventually performed to provide consistent interpretations of the results obtained.

The first step in any study of land use change concerns the measurement and assessment of land use change involved. This is influenced significantly again by the level of scale at which the measurement is conducted – spatial and temporal scale primarily but also social, economic, institutional and cultural scale. For short time intervals at the level of the globe, no land use change may be discernible while at the level of a field measurable change may be recorded. Longer time intervals reveal significant changes in the uses of land. The case may be, however, that, depending on the socio-political and geographic context, very long time intervals may conceal the true "history" of land use change. Similar observations apply to the level of the spatial scale used. For a given time interval, land use change may not be discernible at higher spatial levels while at lower levels – e.g. at the level of a settlement – very large changes may be measured. The role of the land use classification system is critical in this context also as the measurement of change involves the land use types which this system includes. Hence, regardless of the intended level of analysis, the results obtained and the ensuing description of land use change refer to land use change at the level of the spatial, temporal and socio-economic scale to which the land use classification system refers. Prediction of land use change, similarly, refers to the scale of the land use classification system used.

Explanation of land use change – answering the question of "why" of the "what" measured – is inextricably related to the scale at which the analysis is pitched. Meaningful explanation of land use change (and of its impacts) at a given level of any scale, requires that the relevant explanatory factors are identified at the levels of the particular (and of other relevant, perhaps) scale at which they operate in reality. The critical point is that "the relevant explanatory factors" – with the exception of the bio-physical determinants – are associated with particular individual and collective actors – agents involved directly or indirectly in the process of land use change. Essential explanation focuses on these agents, their differing resource endowments to influence land use change, and their actions through which land use change is effected. For example, to explain land use change and its impacts at the farm or the parcel level, relevant explanatory factors may include soil type, slope, water availability, local climate, and the characteristics of the household or of the head of the family; in a few words, factors which operate at the same level of the spatial and perhaps of the temporal scale. However, other relevant factors which influence land use change at the farm or the parcel level operate at other – higher and/or lower – levels of the spatial and temporal as well as of economic, organizational, and institutional scales such as financial assistance, agricultural policies, product prices, climate change, past types of land use, past policies, etc. Hence, the need to employ a nested set of scales for a comprehensive explanation of land use change in concrete settings (Blaikie and Brookfield 1987, Veldkamp and Fresco 1996b). In other words, the drivers of land use change as well as the determinants of the resulting impacts have to be sought at a variety of scales and levels of these scales.

Focusing on just one scale and on a particular level of that scale, leads to biased explanation as either the full set of factors is left unaccounted or a wrong or irrelevant set of determinants is taken into account. Turner et al. (1995) discuss this issue in the context of global vs. local level studies of environmental change in general. At higher spatial levels, the PAT variables – Population, Affluence, Technology – are found to have important statistical associations with environmental change implying that these may be the underlying drivers of change. However, local level studies do not reveal such associations; on the contrary, the most important associations relate to such factors as institutions, policy, social organization. A related problem arises when the assumptions made for the purposes of analysis at a particular level of a scale are transferred to another level at which they are most probable not to be valid and produce biased explanation. For example, the assumption of homogeneous household preferences (by group) may be valid at the regional level but may be inappropriate and misleading at lower spatial level or in the longer run.

In a similar vein, the impacts of land use change and the degree of severity attached to them are influenced by the scale of the analysis. Land use change at the level of a settlement may have a number of local, direct, short-term impacts – environmental (e.g. air pollution), economic (e.g. changes in land values, tax base), social (e.g. disruption of social cohesion). However, these same changes may produce supralocal, indirect and longer term impacts; i.e. impacts at different levels of the spatial, temporal and social scale. These may include negative environmental impacts on agricultural production in neighboring areas, increased demand for exurban space (caused by increases in urban land values), population and jobs migration. At the regional level, land use changes impact not only on regional climate, economy and social structure in the short run but also on individual land owners (e.g. farmers), or on the larger national level both in the short and in the longer run. Deforestation due to over-exploitation of forest resources or other natural and/or anthropogenic causes impacts on the hydrologic balance, soil quality, regional income, population migration. But it also impacts on local lives; households face a different environment within which they have to make a living, the family structure changes (fewer and perhaps older members, lower reproduction), household income and quality of life changes. The impacts of deforestation may show up also on higher spatial levels as changes in productivity, changes in the balance of trade, regional imbalances. The indirect impacts set in motion by deforestation may also show up in the longer term as further population migration, further land use change, institution of state policies to assist declining regional economies. The list is endless. What is important is that all these impacts are not confined to only one scale or to one level of the spatial, temporal or organizational scales but they diffuse to higher and lower levels and to different scales.

Less visible but of no lesser importance is the buildup of small impacts at lower levels of the spatial and temporal scales to generate impacts on higher levels of these scales; this is the case of cumulative impacts which are caused by incremental impacts at the individual level and are felt usually after some period of time at the regional or even the national level. Water abstraction by a single farm may not disturb the water table but a dramatic drop may be caused if all farmers follow this practice. Temporally, cumulative impacts appear as lagged effects of land use change. Salinization , acidification , desertification are caused usually by the buildup of smaller, short run changes in the use of land. Lastly, another class of impacts are the distant impacts; those caused in a place by land use changes in another place (or places). The development of exurban space results from changes of land use, among others, in urban areas. Water shortages in a region may be due to excessive water abstraction to serve a fast growing tourist area. The issue of scale is implicated in all these and similar instances and makes imperative the use of "scale-sensitive" analytical approaches.

The evaluation of impacts caused by land use change is influenced also by the scale of analysis. Land use change may have important impacts (beneficial or detrimental in an economic, social or any other sense) at the farm level but it may be of no importance at the level of the rural settlement or the larger region. Moreover, the same change may be important in the short run but it may loose its importance in the longer run. The myriads of land use changes which occur continuously on the earth’s surface are evaluated differently at different scales. This is not surprising as evaluation implies an "evaluator". Different interests are involved at each scale with different criteria and priorities attaching, consequently, different importance to the various impacts of land use change. The importance of scale in the evaluation of impacts is closely related to the last, and perhaps the most important, consideration to be discussed, the role of scale in decision making in the context of land management and land use planning.

Land use change results from direct or indirect decisions to alter the current uses of land at the level of an individual land owner, of a regional or national authority, of an international body, or of any other land-related interest. Whatever the form these decisions take, the important point is that they involve decision making units and decision making processes at particular levels of one or more scales. In other words, the analysis of land use change necessarily asks "who decides to change the use of land, where, when, and why". The factors which are taken into account in the analysis relate to the particular decision making units and processes as well as to those influences which impinge on the range of choices open to the decision making units (see, Blaikie and Brookfield 1987 for a discussion of this topic in the context of land degradation). The assessment of land use change, the assessment and evaluation of the resulting impacts as well as the decision to act are all related to the pertinent decision making units and processes. Land management and land use planning in response to land use change or with the purpose of effecting desirable land use change are tied to decision making units at various scales. The meaningful and useful analysis of land use change in support of these functions should, therefore, pay due attention to the different scales involved and to their relationships.

5.1.3. The relationship between theories and models of land use change

The linkages between theories and models of land use change have not been strong over time in general. Early theories and models tended to be interrelated (e.g. von Thunen’s and Alonso’s theories and models) but the level of abstraction on which they operated and the limited number of real world situations which they could successfully approximate did not lead to any widely known useful operational tools. Urban economic theory has guided model building and has provided theoretical support (the instrumental approach to theory ) to modeling efforts but the scope of this theory is limited judged in the light of the socio-culturally and geographically variegated nature of land use change. At several instances, theories and models developed independently; hence, neither theories led to models nor models were based on theories. Most frequently, however, especially in contemporary practice, models fall in the gray area mentioned to in Chapter 4; in other words, they attempt to include in their design determinants of land use change revealed by theory – this is the case, for example, of several (although not all) statistical models of land use change which choose the independent (explanatory) variables on the basis of broad theoretical considerations as well as of simulation models which attempt to approximate the working of the urban or regional system being modeled.

Risking a rough comparison of theories and operational models of land use change, it seems that, overall, the latter are more developed than the former. Although this is not the place to delve into the reasons for this gap in development between theories and models, two broad groups of reasons are suggested for further elaboration: substantive and practical. Substantive reasons pertain to the difficulties associated with building theories for such complex phenomena as land use change and with trying to disentangle the interactions among their diverse determinants. Important (and socio-politically sensitive) among those difficulties are those associated with identifying the role and contribution of institutional and cultural factors especially at micro-spatial levels where they may be more influential and decisive of the direction and quality of land use change. Moreover, an elaborate and detailed theory will, most of the time, result in a rather involved and contrived model whose use will be questionable. Hence, it is comparatively easier to evade these difficulties by adopting an instrumental approach to theory , simplifying the observed relationships, and reducing them to manageable quantitative (or, simple qualitative) mathematical expressions which can be easily understood and manipulated.

Practical reasons which may account for the priority given to models over theory include both the availability of resources of various kinds (money, time, personnel, know-how, effort, administrative support) as well as the demands of the decision making "clientele". The former are critical and they are usually directed to activities which will bring "visible" and "operational" results within reasonable time; i.e. to activities with high value-for-money. Theories compared to models may have lower value-for-money, at least in western societies. The decision making "clientele", on the other hand, irrespective of their attitude towards theories, may be placing higher priority to operational and easy-to-use tools in decision making where decisions may have to be made in relatively short time, controversial issues are to be avoided, and the results have to be easily translated into concrete actions. Theories may be rating low in this respect too although one should keep in mind the controversial nature of several models also. However, models seem to sell better than theories as they "say it with numbers" and, thus, appeal to a much broader and diverse decision making "clientele". The modern trend to produce visual versions of model results with the use of GIS are even more appealing as they use "the seductive power of the visual medium " (Longley and Batty 1996, 9) to enhance the market potential of models. Naturally, these are very sketchy explanations of the gap between theories and models and further analysis is needed to give them more concrete form and substance. Whatever, the outcomes of this analysis may be, however, the true explanation should always lie in the synthesis of substantive and practical reasons.

It must be admitted that the linkage between theories and models is not an easy one to achieve. Theory is indispensable in meaningful model building as "the role of theory is to explain experimental findings and to predict new results" (LUCC, 199, 89). But one of the reasons why satisfactory models of land use change have not appeared yet is "the lack of a comprehensive and integrative theory of human-environment relationships. Land use/cover change research is embedded within human-environment relationships. To date, these relationships have proven difficult to conceptualize in a meta-theoretical framework (Blaikie and Brookfield 1987; Turner 1997)" (p.89). However, what is important for the development of models appropriate for concrete spatio-temporal contexts and decision settings is the synthesis of elements from the variety of available theories to help explain the dynamics and interactions between land use and the drivers of its change in the particular situation being studied. LUCC (1999) considers the "thorough understanding and modeling of these complex interactions … a prerequisite to generate realistic projections of land cover changes … into the future" (LUCC 1999, 89).

Another difficulty surrounding the linkage between theory and models is that theories usually place heavy demands for operationalization especially when important land use change drivers are qualitative and there is no consensus on how best to express and measure them. Scale considerations complicate the operationalization issue further. Models at the micro-level of the parcel require a theory of how individual and higher level factors combine to produce the land use changes observed as well as how to aggregate micro-level changes into higher level changes in land use patterns. Models which attempt to provide for more realism in the representation of the modeled socio-economic and physical entity may become burdensome and, ultimately, unusable. Hence, trade-offs between theoretical rigor and practical usefulness are inevitable. The nature of the linkage between theories and models is a matter of the trade-offs chosen in particular applications.

The issue of communication between theory and model builders should not be discounted. Disciplinary fragmentation and compartmentalization frequently impede the smooth communication between those developing theories of land use change and those building models. Frequently, these individuals reside in different scientific compartments which make the mutual exchange of ideas, knowledge and tools problematic. True interdisciplinary research, a basic prerequisite for the development of theoretically informed and sound models of land use change, is rarely practiced. In the absence of information about the availability of the variety of theories which deal with the multiple dimensions of land use change, then, model builders rely inevitably on the most widely publicized and easy to access theoretical frameworks. This is the case with economic theory and its wide use in supporting models of land use change at all spatial scales.

5.1.4. The use of theories and models in land use decision making

The above discussion leads to a more essential aspect of the whole enterprise which concerns the contribution of theories and models of land use change to improved, informed, and rational decision making on land use issues. The question is whether and how all this available stock of knowledge and analytical capability can assist in making decisions that improve the quality of human life by averting undesirable and promoting desirable forms of land use change; i.e. that lead to sustainable land use. The issue of theory and model use is particularly important nowadays also as there is a proliferation of spatial decision support systems (SDSS) in many disciplinary quarters spurred by developments in computer technology and GIS. Their stated purpose is to offer support and improve the quality of decisions – on land use issues in the present case. The following discussion examines the use of theories and models in land use decision making focusing on the central concern of the issue – the users of theories and models. Two questions are addressed in this respect. First, who are the actual users i.e. who are the people interested in using theories and models of land use change when making land use decisions. Second, what kinds of users theories, models, and SDSS assume; i.e. what is the underlying stereotype of users they purport to serve.

The first question of the actual users of theories and models does not have a unique answer. Interest in using theories and models in making land use decisions, and effective demand thereof for them, depends, among others, on the socio-economic, cultural and institutional context as well as on the decision making tradition on land use issues at particular spatial levels and over time. In other words, it is a matter of whether a "culture" of using science, in general, in decision making in the public or in the private domain exists. The literature on this issue is not very rich but it seems that demand for decision support tools is higher in the developed compared with the rest of the world. Most of the applications of the models examined concern western countries; the applications in other countries reported are usually led by professionals educated in western institutions. In a great number of countries, however, land use decision making follows different patterns and logic which do without the use of formal theories and models.

Even in those cases where real world applications are reported, however, the critical question is why theories and models are used; i.e. for what purpose. Is there a genuine purpose of improving the quality of decisions, as theory and model developers usually assume? Are they used for merely symbolic reasons? Are they used to justify decisions already made? All answers are possible! The politics of theory and model use in decision and policy making, in general, has received considerable attention in the related literature but its discussion is beyond the scope of this contribution (see, for example, Weiss 1972, D. Lee 1973, Greenberger et al. 1976, House and McLeod 1977, Rothenberg-Pack 1978, Walker 1978a, 1978b, 1981, Wildavsky 1979, Mann 1981, Szanton 1981).

Assuming that theory and model users have a genuine interest in getting assistance in making land use decisions, the critical question arises whether they are capable of using the suggestions of theories and the results of models. At this point, Machiavelli’s advice to the Prince is worth remembering: "A prince not himself wise cannot be well advised" (cited in Szanton 1981). However elaborate, sensitive, carefully designed, and technically perfect theories and models may be, their ultimate contribution to informed decision making depends on how "wise" their users are. "Wisdom" in this case has to be construed in the particular sense of users comprehending the theories and models used, being aware of the range of their applicability, and being able to judge whether they are appropriate for the problem at hand. This means that sensible and correct use of theories and models requires that users are aware of and understand the assumptions underlying theories or models, recognize their possibilities and limitations, and use them for the uses for which they are designed and not as hammers which make everything look like a nail, a panacea for addressing all land use ills. The question that arises, then, is whether theory and model users satisfy these requirements.

The theories and models presented in this work leave no doubt that a certain level of education, not to say of specialized education in several cases, is needed in order to comprehend most of them and, consequently, to use them sensibly and appropriately. The higher the level of sophistication of a theory or model, the greater the demand for scientific (broadly conceived) competence on the part of the user. This is particularly true for the contemporary generation of models which utilize diverse visual and other devices to present their results but provide very little guidance or indications of the limits of the validity of these results. A user unaware of a model’s assumptions and caveats is prone to succumb to the seduction of the mode of presentation and rely on the results offered regardless of whether they make sense for the problem at hand. The crucial implication of this discussion is the inevitable fact that the wealth of knowledge and information provided by theories and models of land use change is essentially accessible to an elite of educated users. Matters of proprietary information or information withholding aside, the greatest barrier to using any theory or model is general and special education. But is it possible that all interested users of theories and models possess the requisite education? Probably not, especially in the present information age when the production and dissemination of information move at unprecedented speed and a whole set of issues arise with respect to the social and economic consequences of the so-called information technologies (see, for example, Castells 1998). At this conjunction, the question of ethics is unavoidable. Given all the constraints surrounding the proper use of theories and models, the elite who possess the requisite education and skills has an ethical obligation to guide the users of theories and models, i.e. the actual decision makers, to making wise use of them in the sense already discussed previously. It rests, therefore, with those individuals who "control" the available information to assure its sensible and appropriate use.

Turning now to the stereotype of the user underlying theories, models, and the fashionable spatial decision support systems (SDSS), it seems that, in most cases, their developers have evaded addressing the tough question of the real model users and their needs. What they assume, instead, implicitly at least, are various types of users such as "intelligent individuals", "educated specialists", and the like. Or, they rely frequently on a vague and abstract notion of a "policy maker" or "decision maker" who usually asks questions the theory or the model is designed to answer! Overall, the underlying stereotype of the user agrees, not unjustifiably of course, with the elite of educated users referenced above. The various SDSS being developed recently include in their stated purposes the intent to provide easy access and "user friendliness" to a variety of users. However, several reservations can be expressed in this respect which have nothing to do with the good intentions of their developers.

The term SDSS is used rather vaguely in most cases and without adequate explanation of its content. In fact, this is a term which means different things to different people in different contexts. It is noted that of the variety of available SDSS, the present discussion refers to those related to land use decision support (as opposed to those designed for facility location or network design support). The question is basically what is the meaning of the terms "spatial", "decision", and "support" in the context of a given SDSS, for what kinds of users and related needs it is designed, and how it is used. Any SDSS will necessarily be designed to answer certain types of questions and to serve certain purposes (e.g. description, prediction, prescription, scenario and impact analysis). It will refer to a certain range of spatial and temporal resolution. It will adopt some kind of theory about reality; i.e. particular views of land, land use, land use change, the users of the land, the drivers and the impacts of land use change. It will utilize some kind of modeling device – either a formal or a heuristic model – to provide answers to questions. It will utilize particular types of data (of varying degrees of quality and accuracy) as input to the calculations it performs. It may also require input from the user if it is used in an interactive mode – a common trend recently to render SDSS more sensitive to the actual decision environment (e.g. preferences, priorities, constraints). It will utilize some types of presentation aides to communicate the answers to the questions asked. These are critical aspects which should be made clear to users to help them decide whether the SDSS can address adequately and consistently their particular decision support needs. Evidently, intelligent use of a SDSS is not a matter of easy, low-cost access and user friendliness but of the real ability of the user to comprehend it. The availability of high-tech decision support aides makes even more timely the demand for "wise princes".

5.2. Future research directions: theoretical, methodological and practical needs

Future research needs on the subject of theories and models of land use change have been mentioned or indicated on several occasions in the previous chapters. This last section outlines more general, "higher level" research needs which address a central research requirement; that of integrating the various pieces of knowledge and producing coherent theories and methodologies to guide future land use change towards sustainable paths. Integration is needed at the theoretical, the methodological, and the practical levels. At the theoretical level, despite efforts to bring together the diverse and interacting drivers of land use change, there is still a long way ahead in terms of true integration of theoretical analyses at various spatial levels and within particular time frames. At the regional and higher levels, theory has not addressed adequately yet the close relationships between the dynamics of urban and regional development and the demands placed on ex-urban space as well as the uses of land within and between regions. For example, land use change (and its environmental and socio-economic implications) in rural areas cannot be explained solely by looking at these areas in isolation but by embedding them in the broader framework of urban-rural or regional, in general, dynamics. Demand for agricultural, industrial, recreation and other types of land use depends on the food, mobility, recreation and other needs of urban populations but also, and perhaps more critically, on the demands of capital (production activities). An important need for integration arises exactly at the interface between theories of production and theories of consumption in all three theorization traditions . More specifically, theories which concentrate on the patterns and dynamics of the location of firms need to be tightly coupled to theories which study the land use patterns of residential, commercial, recreational and other extensive land using activities. At lower levels, where the actual land users reside, make decisions and implement them, theory needs to integrate the psychological, socio-economic (e.g. land tenure and ownership, income, family structure), and other aspects of individual decision making with the institutional, cultural (value systems), economic and other higher level forces framing and directing individual as well as collective land use decision making.

At the methodological level, similar developments are expected. The trend is definitely towards integrated modeling but integration is insufficient in several respects. The urban and the exurban dimensions have not merged yet well in modeling. Regional and higher level models mostly ignore the representation of the workings of the urban components of a region. The same applies to urban models which operate as if the surrounding territory in uniform, undifferentiated and unrelated with what goes on within the urban region. It is important to note that integrated models with essential explanatory power are contingent upon the development of proper theoretical structures. Another significant development in modeling land use change is the more realistic representation of land use types, a trend already begun but in need of wider adoption and application. IIASA’s modeling projects integrating FAO’s land evaluation methodology (the AEZ methodology) with land use analysis is an indication of a more general research direction towards integration. Progress in remote sensing technology is expected also to assist in injecting more detail in the frequently used coarse land use representations (e.g. developed/undeveloped) which will make them more useful to a broader array of land use decisions.

At the practical level, the main issue which future research should continue to address is perhaps the availability of appropriate data sets for integrated analysis of land use change. Data are provided at various levels and degrees of spatial and temporal resolution and from various sources partly reflecting the traditions and needs of particular disciplines (e.g. environmental sciences, demography, economics) and partly reflecting technical, technological and organizational constraints. Standardization of the various dimensions of data is a much needed development which will contribute to the consistent analysis of land use change across space and over time. Another, equally important, practical research theme is the monitoring and analysis of model adoption and use. Although applications and use of several models in actual decision contexts have been reported, for the majority of models not enough is known as regards the level and quality of their utilization except for isolated cases of particular models and specific issues. This research direction is not meant to satisfy academic curiosity but to analyze experience with model use and employ it to the design of models and perhaps to the education of the rich diversity of model users.