Cover Image
close this bookConservation and Development in Northern Thailand. Proceedings of a Programmatic Workshop on Agro-forestry and Highland-Lowland Interactive Systems, Held at Chiang Mai, Thailand, 13-17 November 1978 (UNU, 1980, 114 pages)
View the documentEditorial note
View the documentPreface
close this folderIntroduction
View the documentProgramme on the use and management of natural resources
View the documentHighland-lowland interactive systems in the humid tropics and subtropics: The need for a conceptual basis for an applied research programme
View the documentNorthern Thailand: The problem
close this folderResearch reports
View the documentImplications of socio-economic, demographic. and cultural changes for regional development in northern Thailand
View the documentAgricultural intensification and the role of forestry in northern Thailand
View the documentProblems of land use and recent settlement in Thailand's highland-lowland transition zone
View the documentCartographie de la dynamique des paysages dans les hautes et basses terres du nord de la Thaïlande
View the documentMapping landscape dynamics in the highlands and lowlands of northern Thailand
View the documentClimatological, pedological, and geomorphological processes in tropical mountain ecosystems
View the documentLocal climatological differences between highlands and lowlands in Thailand
View the documentLand use and its relationship to agriculture in Pangsa, Chiang Rai: A case study
View the documentIncreasing farm production in the highlands of northern Thailand
close this folderResearch and training requirements
View the documentResearch and training in southeast Asia in relation to priority areas of the programme on the use and management of natural resources
close this folderOn the significance of the watershed management approach in studying highland-lowland interactive systems
View the document(introductory text...)
View the documentReview of current research on agriculture in the highlands of northern Thailand
close this folderSubcommittee reports, plenary discussion, and recommendations
View the documentReport by the subcommittee on documentation, training, and personnel
View the documentReport by the subcommittee on mapping and cover-type data requirements
View the documentReport by the subcommittee on climate, soil. and soil erosion data requirements
View the documentReport by the subcommittee on socio-cultural, demographic, and economic aspects
View the documentPlenary session
View the documentRecommendations
close this folderHuai Thung Choa highland project
View the documentThe Huai Thung Choa highland project: Status and opportunities
close this folderAppendix
View the documentWorkshop participants, observers, and staff
View the documentOther UNU publications

Mapping landscape dynamics in the highlands and lowlands of northern Thailand

Michel Bruneau


This study is based on the experience of six years of mapping landscape dynamics (five maps published and two others in the process of being published) drawn to different scales, based on a chronological series of aerial photographs and Landsat images of Northern Thailand. This approach is particularly effective for the analysis of occupied landscape and their dynamics, demonstrating the impact of man upon the natural environment.

After having defined the notion of landscape, we attempt to show its cartographic translation at different scales (large, intermediate, and small) and its application in the study of the highlands and lowlands in Northern Thailand. Finally, we try to demonstrate the importance of such a mapmaking process for regional planning and for study of the land development problems of this region.

The Landscape and Its Dynamic in Northern Thailand

The term ''landscape'' is used here in a purely physiognomic sense. It is the visible, observable space in its entirety As the visible form of contact between the biosphere, lithosphere, and hydrosphere, landscape is characterized by natural vegetation that is more or less degraded, and by human use and occupation of the soil in a given geomorphological framework. The study of landscape, or of units of landscape,' requires a synthetic approach. beginning from the actual. considered as a whole. taking each of its elements not separately but in their interrelationships and combinations.

The aerial photograph and, more recently, the satellite image have given to this notion of landscape its extreme richness. They are the faithful reflection of landscape according to the various horizontal directions with real relative surfaces and at comparable scales. Geographic documents par excellence. they reveal landscapes in their completeness: natural settings that are more or less transformed by the human groups that occupy or have occupied them.

Each (mappable) landscape or landscape unit is a combination of features belonging to three different types of phenomena the physical-chemical setting (relief. climate, bodies of water), the biological setting (the living plant and animal communities), and the human use and occupation of these settings (production methods and socio-economic organization).'These different phenomena evolve according to very dissimilar chronological rhythms. Geological and geomorphological elements evolve over a very long period of time. Soil and plant cover are transformed in the course of shorter periods. As for the action of social groups, they play a role on an even shorter timespan and conform to a logic that is radically different from that of the first two orders of phenomena. Seen from the historic perspective of a period of twenty or thirty years, a view that one has when in possession of the available photographic documents of the space under consideration, it is the last order of phenomena (anthropic action) which is here analysed. In our study the dynamic of mapped landscapes concerns the periods 1953 1972 and 1976 1977. These periods are rich in socioeconomic transformations (passage from a precapitalistic economic system to a merchant economic system) which have brought about important changes in the landscapes themselves

The first aerial photographic coverage of Thailand dates from 1953 1954 (approximate scale of 1:50,000) and corresponds to a period when the methods of economic production were still largely traditional and when demographic growth had not as yet made its impact felt. The landscapes of that period are thus still strongly influenced by the old form of society and civilization. In 1968 1970 (second aerial photographic coverage of Thailand at a scale of about 1:60,000), and even more in 1972 (coverage of some regions at a scale of 1:20,000) and more recently (satellite data beginning in 1973), modernization, penetration of the merchant economic system, and demographic explosion mark and at times even disrupt the landscapes. The dynamic of landscapes thus partly reflects the dynamic of socio-economic developments.

Landscape can be divided into different units according to the scale at which it is observed: large scale (1:20,000), small scale (1:1,000,000), and intermediate scale (1 :250,000). The limits of each is the reflection of geographic separations proper to its particular scale. For each of them we have a sectioning of space into landscape units of a certain type which do not necessarily interlock with each other. The significance of each unit must be determined and the type of representation that is to be given by each of the scales must be ascertained in order to show how they can be combined so as to produce a deeper understanding of a whole region

Large-Scale Cartography of the Dynamic of Landscapes

For rural landscapes analysed here, the large scale is generally considered to be 1 :25,000, but it can go as far as 1 :50,000. The landscapes presented are either homogeneous physiognomic units belonging to a vegetation type or to its different forms of degradation, or homogeneous types of soil utilization (for example. paddy-fields).

In the colour maps produced from this scale of photography, the shades used conform to the principles of H. Gaussen for the natural vegetation, from indigo for the thick evergreen forests of the mountains to orange for the clear forest of the piedmont. and employs greens or yellow-greens for the mixed or semi-dense forests. Blue represents the paddyfields.

Landscape transformations are indicated by stripes or white bands alternating with colour bands. These are. for example. the rai extracted by clearing the forest on the pioneer front (Map of Si Satchanalai) or the anthropic savannah (former opium rai) in the Doi Inthanon area. Recently developed paddy fields have horizontal or slanted bands.

In the Karen zone of the Doi Inthanon area it has been possible to distinguish. in addition to the swiddens of the year shown in yellow, three stages of recovery of the vegetation on fallow and (aerial photographs at a scale of 1 :20,000 taken in 1972): a recent fallow strip (1 to 2 years) in a state of low thickets, an old one (3 to 5 years) in the thicket state. and a much older fallow strip (6 to 10 years) in a pre-forest state. Each is represented with a range of colours extending from yellow-green to dark green. the green darkens as the strip approaches the forest state. It is therefore possible to perceive the cycle of the Karen fallow strips which cover a period of 5 to 10 years, depending upon the site.

On the transect of Phrom Phiram the colours used on the two preceding maps with an ecological meaning are replaced by hachures running in various directions. The horizontals correspond to the wettest environments. flooded the longest (swampy meadow). Rotation of the hachures in a counterclockwise direction indicates drier and drier settings (swampy thicket, swampy forest, semi dense forest). The more closely the 1968 physiognomy (aerial photographs taken to a 1 :20,000 scale) approximates a forest state, the darker the zone is. Increasingly pronounced human impact is depicted by progressively lighter tones. Intermediate stages (a discontinuous thicket in the course of development) are represented by bands that are alternately dark and light, the light band indicating the transformation of the setting (in 1968) with relation to its former state (in 1953)

Such principles enable one to procure a monochrome cartography of landscapes and of their transformations on a relatively large scale (from 1 :20,000 to 1 :50,000). This representation of landscape units facilitates a rather detailed analysis of ecological settings and of land use along particularly significant transects.

The cartography of the dynamic of landscapes drawn to a large scale, supplemented with maps of slopes. soils. and soil erosion. has a positive value for the study of agroforestry systems in the highlands and lowlands. The map of Doi Inthanon (1 :25,000), for example, enables one to analyse two of these systems: the Karen system, where shifting cultivation with fallow strips on the flanks and irrigated paddyfields in the valley bottoms increases towards intensive use but without upsetting the basic ecological balance; and the Hmong system. where the opium poppy and corn growing on the same parcels several yeses in succession hinder the reconstitution of a forest plant cover, which instead is replaced by a savannah.

The retreat of the dense or semi-dense forest of the mountains does not have the same significance in each of these systems. A transition zone (intermediate forest or wooded savannah) is found along the points of contact between these two systems. The distribution of population and inhabited dwellings shows that the Karen system can handle greater population densities than can the Hmong system. However, recent growth of the latter (development of paddy-fields and peach orchards) can orient it in the future towards an intensification of agricultural techniques, as in the Karen case if the forest zones to be cleared disappear and if the population stabilizes.

On the map of Sisatchanalai, two clearly contrasting zones can be seen. The lowlands, consisting of valley bottoms and rolling low hills, where a pioneer front of commercial crops is advancing very rapidly at the expense of the dense and semidense forest, stand out from the highlands, where abandonment of shifting cultivation or forest exploitation favours forest regeneration. Between 1953 and 1968 the farmed area had doubled. destroying the forest in over 13.5 per cent of the area mapped.

The transect of Phrom Phiram shows the progress of landdevelopment on the Mae Nan flood plain between 1953 and 1968 and facilitates the derivation of the dynamic from it without mapping the entire area.

Such large-scale mapping done with chronological series of aerial photographs requires long and detailed work of interpretation, knowledge of design and of the terrain involved. It can only be considered for areas particularly representative of much larger entities. The main advantage of the transect approach is that a maximum number of different environments car, be covered while the mapped space is reduced to a minimum. These large-scale studies are necessary in order to intensify the analysis of the landscape dynamics and to understand their mechanisms by referring to surveys taken at the level of the rural community or of a few communities. They can also be used for a perimeter development which is at the same time a test zone (Huai Thung Choa. for example)

Intermediate - Scale Cartography

On the intermediate scale. Iandscape units are more complex. Each is defined by a combination of traits, some belonging to the natural environment. others to modifications and improvements brought about by human action. They have been delimited either in relation to one or two dominant traits or in relation to the combination of three or four traits. These landscapes-some simple and uniform in structure, others complex and mosaic-have their own dynamic, which is governed by human action. They often correspond to geomorphological units: valleys, pediments, river banks, karst, and so on.

The map ''Soil Utilization and Landscapes in the North of Thailand" gives a picture of the distribution of the large types of landscapes on mountains. hills, and on the alluvial plains of the Chiang Mai and Lampang basins in 1970. These landscapes evolved mainly in the mountains and on the terraces, but this evolution during the course of a period of fifteen years is translated better with a 1:20,000 or 1:50,000 scale map than with one of 1:250,000. The smaller scale does not lead to a very significant shifting of limits. We thus contented ourselves with a cartography at a specific date (the aerial photographs of 1968 to 1970). without directly translating the dynamic.

The natural or developed landscapes which have been little affected by recent transformations appear in flat tones corresponding to the ecological mapping principles of H. Gaussen (the wettest settings shown with the darkest colours. the driest with light shades). Human action in the form of rai clearance is shown with slanted bands: alternation of a colour band of the setting (type of vegetation) and of a white band the width of which is proportional to the degree of human action. The anthropic savannah of the highlands, which corresponds to the opium poppy growing zone, is indicated with green broken lines of the same colour as that of the mountain forest. which is the original formation. This map shows the distribution of the major ecological systems and the degree of human impact on the mountain environment the terraces, and pediments along the borders of the basins or valleys.

The representation of the dynamic on this scale poses special problems because of the great complexity of the landscape units. Sometimes the whole of the landscape has been transformed, at other times one or two units among others have changed during the 15-year period under consideration (map of the dynamic of landscapes of the Sukhothai plain drawn to a scale of 1:250,000). The dynamic is indicated with the absence or presence, the orientation, and spacing of bands. Flat tones correspond to landscapes which either have not changed or have changed very little. Width and spacing of the bands expresses the degree of change. Slanted bands refer to the clearings in the highlands for the cultivation of dry-weather crops, horizontal bands imply that there has been a land development programme for the introduction of paddy-fields between the two dates considered (1953 and 1968). We have been able to distinguish landscapes that have been slightly, partially, or totally, transformed during the course of the period studied. The number of bands joined for the same landscape unit is in relation to the degree of complexity of this unit. Complex landscapes assure the transition between two and at times three simple landscapes. These are the types of transition that unite the traits of several simple contiguous units.

The map of Sukhothai, with slightly complex figures indicating a wide diversity of situations, clearly shows the rapid progress of land development in the course of these fifteen years. This includes the occupation of zones subject to flooding and the older alluvions (uplands) between Mae Yom and Mae Nan and the development of the pediments or hills along the borders of the plain, especially for dry-weather crops or, less frequently. for paddy-fields

A comparison of the large-scale maps and those described above shows that the legend has been simplified by about a third (for the same zone, there are, on an intermediate scale map, three times fewer the number of units represented than on a large scale map). This intermediate-scale cartography enables the acquisition of a view of the whole province or small region, of the occupation of the land and of its recent dynamic. Over a wide space, it simplifies and generalizes the analytic data of a large-scale map while at the same time retaining a sufficient precision as regards the localization and contours of the landscape units.

Landsat Remote-Sensing Data and Small-Scale Cartography

On Landsat images, landscape units drawn to a small scale are more extended and more uniform. In general. a trait of plant cover or of land use dominates the unit and confers its homogeneity to it. There is often a connection between these landscapes and large morphological-structural units.

Landsat images are of great value for studies conducted at a small scale. Band 7 is useful for indicating the morphologicalstructural units. the wettest zones and bush-burnt areas as well as the swidden. Band 5 provides an overall view of plantcovered landscapes and of certain aspects of land use. Colour composite images, combining bands 4, 5 and 7, are richer and more detailed with better contrasts. For the mapping of landscapes in general, they are better than band 5 (twelve landscape units instead of nine on the map of Northern Thailand). but less clear than band 7 for swampy areas or wet surfaces.

A comparison between images taken at the beginning (December. January) and at the end (March. April) of the dry season is possible and very instructive for understanding Northern Thailand. Evergreen forests of the mountains and anthropic savannahs can only be identified on the images of the beginning of the dry season. while forest and clear thicket cannot be easily distinguished from the semi-dense mixed forest except at the end of the dry season (the images of March or April). Delimitation of the zone of irrigated farming and of the stubble of dry paddy-fields is possible during both periods, but certain phenomena of the use of the environment by man (swiddens or dry crops in an advanced state of growth) appear only at one period (April) (Bruneau and Le Toan 1979). Rainy season images cannot be used because of the high cloud cover and uniformity of the plant cover (green over the entire area). They can, on the other hand, be useful for the study of floods and rice growing.

Very simple maps. plotted in black and white, were drawn after an interpretation of three Landsat images that cover a large part of Northern Thailand (Bruneau 1976). For this purpose. these images were enlarged to a scale of 1:500,000. Enlargement to 1:250.000 is also possible for cartographic purposes. But these Landsat images can only be interpreted correctly if based upon a good knowledge of the terrain and in conjunction with aerial photograph (1:20,000 to 1 :50,000) analysis of representative areas.

These images provide an overview of large areas far more effectively than that obtained from a mosaic of aerial photographs. It is then possible to delimit spatial units of the ''ecological regions" type which have a certain homogeneity with regards to morphologic structure. phytogeography. Iand utilization, and human occupation. Ecological variables that are preponderant on this scale, generally of a geomorphologic nature, provide a unity to a group of characters of a varied nature (vegetation. Iand use, soils, communication routes, dwellings. ethnic groups) and enables the perception of the broad integrity of the landscape (Bruneau 1976).

The repetition (theoretically every 18 days) of Landsat images and of their sequences (since 1973 with Landsats 1, 2, and 3) has the great advantage over aerial photographs of allowing a comparison between seasons and years, thus facilitating the study of the dynamic of phenomena. The increase in the data thus acquired will require more and more automatic processing and consequently a systematic recourse to digital data. But this problem is still in a study stage. The work of supervised classification and automated mapmaking can generally only be performed on a small portion of the image.

An attempt was made to produce a supervised classification for a part of the Doi Inthanon mountain region using four bands of the Landsat tape of 3 March 1975 (Bruneau and Le Toan 1979). The result is disappointing because of the shadow cast, which falsified the classification of the forests. Only the paddy fields in a stubble stage and the stretches of barren soil could be identified without any ambiguity To overcome this problem, ancillary data must be introduced through application of a more sophisticated software, as L.D Miller, K Nualchawee, and C. Tom (1978) have done for a small mountainous area in the vicinity of Chiang Dao. This technique is not yet operational for the cartography of larger areas

Results, however, are far better in an irrigated plain region (the Chiang Mai basin in the surroundings of San Patong). A supervised classification using bands 5 and 7 on two different dates (27 July and 3 March 1975) provides a precise cartography of the types of land use (6 in all) and of their changes between the two dates. These studies will be continued in 1979 and 1980 with the University of Chiang Mai, the objective being to produce an inventory of land use and of its changes, with the aid of three tapes of Landsat digital data (beginning and end of the dry season, five-year interval), for the whole of the Chiang Mai basin

Although they represent future solutions, these techniques cannot, for the moment. be applied efficiently except to restricted plains areas and thus are of use only for studies made with large scale maps where they can complement the analyses derived from aerial photographs.

Basic Documents for Rural Land Improvement and Development

The purpose of a cartography of the dynamic of landscapes is to locate and measure the degree and intensity of human impact on the environment. In Northern Thailand it brings into focus the transformations of the past 10 or 20 years These landscape transformations are related to the increase in population pressure and to penetration by the merchant economic system. which has caused an expansion of dry commercial crops resulting in degradation of natural vegetation types.

Whether it is in colour or monochrome, this cartography always has an ecological basis which shows the main vegetation types and distinguishes rice growing from dry crops, with the topography being expressed by contour lines. It is really a basic map, more concise than a map of vegetation. For more particular points, ecological conditions can be detailed in the insets: hypsometry, bioclimates. geomorphology and slopes, and types of vegetation. The main objective of the map-the human landscapes and their dynamic-is treated in greater detail. The major land-use categories have been delimited and situated chronologically in relation to each other; they often express the degree of agriculture which is associated with them.

Between the natural landscapes (principal vegetation types) and the human landscape (major land use), we have shown what can be called mixed landscapes in which the human element is only partial. It is in this zone where the changes that have occurred in the course of the past twenty years are most pronounced. One sees in them either forms of degradation: natural plant formations evolving towards their regrowth after abandonment, or more frequently, a progress of land clearing accompanied by land development and dissappearance of the natural vegetation.

Thanks to the topographical and ecological basis of these maps, a permanent correlation can be established visually between the natural environment and the use that man has made of it Thus an overview of the exploitation of ecological settings by human groups is provided The dynamic perspective of this representation of human landscapes is fundamental. The stages of the implementation of changes and the direction taken in the growth phases, including those that will appear in the near future. emerge from it.

The approach using different scales is necessary, especially for the small and large scales. For the first. Landsat imagery can be used profitably only by complementing it with a sampling of aerial photographs and with observations made in the field. The Landsat images are of interest both before undertaking a regional study and once it is completed. When a region is poorly known, they enable one to acquire more rapidly than from aerial photographs an overview of the terrain that is. however, still imprecise This approach also enables one to locate the areas that present problems, especially if the region is rather vast (more than 5,000 km²) One must then go to the large scale along a transect chosen as particularly representative, or to small areas on which a development action is exerted. The use of large scale aerial photographs and enquiries in the field are then required. The Landsat images can then be re-examined for generalization of the results thus obtained for a larger area.

Intermediate scales (between 1:60,000 and 1:250,000) are useful for the improvement of small regions exceeding the framework of a small watershed or of a small ecological or social unit (community or group of communities) which belong to the large scale (between 1:5,000 and 1:30.000)

Aerial photographs at scales close to 1:50,000, generally available for several periods (1954 and 1970), for they are employed to draw topographic maps, are easily used for this purpose The districts (Amphoe) of Mae Taeng, Chiang Dao, and Pail which form a group of highlands and lowlands in the vicinity of the perimeter of Huai Thung Chao, can thus be mapped at 1:100,000. This would enable the land improvement project under consideration to be situated in its immediate regional setting. At a smaller scale (1 :500,000), the watershed of the Mae Ping, to the North of Chiang Mai, and that of the Salween River basin in the same latitudes and extending as far as the Burmese border. would furnish. with the aid of Landsat images, a larger regional setting. This is considered necessary in order to provide an overview of the phenomena in all their complexity (savannahization of the highlands, main forestry ecosystems. highland-lowland interactive systems).


In a project of integrated rural development, the cartography of inhabited and man-created landscapes and of their dynamic is not a goal in itself but a necessary intermediate stage. Made at different scales, it furnishes basic documents that are at the same time analytic (large scales) and synthetic (small scales), which constitute a kind of study support. These documents play a role both at the beginning of such a research project, for clarifying objectives and for posing problems, and towards the end, as spatial reference in order to locate systems and structures, both ecological and socioeconomic. They contribute to establishing a link between natural and human phenomena.


Bruneau. M.1963. ''Dynamique des paysages et organisation de l'espace rural dans la plaine de Sukhothai (Thaïlande).'' L'Espace Geographique. 3: pp. 207-223. (With English summary.)

_________1976 ''Analyse des paysages et grands espaces en Thaïlande septentrionale a partir d'images du satellite Landsat 1.'' In Télédétection et Environnement Tropical, pp.125 - 149. Travaux et Documents de Geographie Tropicale, No. 25. Bordeaux: CEGET. (With English summary.)

Bruneau, M.. and G. Cabaussel. 1973. La dynamique des paysages en zone tropicale. Travaux et Documents de Geographie Tropicale, No. 9. Bordeaux: CEGET. 73 pages. (With English summary.)

_________1974. ''Paysages. dynamique et organisation de l'espace rural a Phrom Phiram (Thaïlande).'' Etudes Rurales, 53-56: pp. 169 - 191.

Bruneau, M., and T. Le Toan. 1979. ''An Interpretation of Northern Thailand Swiddening and Multiple Cropping Systems Using Multidate Landsat Images and Computer Compatible Tapes.'' In Proceedings of the 12th International Symposium on Remote Sensing of Environment, 20 26 April 1978, Manila, pp. 1883 1894. ERIM. Vok 3. Ann Arbor, Mich.. USA.

Miller. L.D., K. Nualchawee, and C. Tom. 1978. ''Analysis of the Dynamics of Shifting Cultivation in the Tropical Forests of Northern Thailand Using Landscape Modeling and Classification of Landsat Imagery'' Technical Memorandum 79545. NASA, Goddard Space Flight Center.