Version classiqueVersion mobile

Forest landscapes of the southern western Ghats, India

B.R. Ramesh
Rajan Gurukkal

Part I. Approaching The Landscape

1. Identification and Mapping Landscape Units of Kerala

Gérard Bourgeon, Danny Lo Seen et B.R. Ramesh

Texte intégral

1A landscape is a heterogeneous land area composed of a cluster of interacting ecosystems. Under this general definition of a landscape, however, lies a wide range of spatial scales that depend on the object under study. Here, our primary objective is to develop planning for biodiversity conservation and its sustainable utilization by stakeholders at landscape level. In this context we have to define a landscape as a heterogeneous system composed of a mosaic of landscape elements (patches, corridors) such as forests, grassland, plantations, settlements etc. The structure (type, pattern, arrangements within a mosaic) of these elements is primarily determined by physical factors (bioclimate, soil, topography, drainage) and is modified by human activities, which are generally driven by socioeconomic and cultural factors among others. As a result, complex environments develop in a landscape due to the adaptations and transformations.

2The key components of a landscape, as generally referred to in the scientific community, are the matrix (the most connected portion of the landscape, in our case, the vegetation type -natural or not- that is most contiguous), the patches (areas of vegetation that are relatively homogenous) and corridors (elements that connect similar patches). Landscapes are organized in hierarchy across scales such that a landscape patch at a given scale can become the landscape matrix at a finer scale.

3In more practical terms, however, the extent of a given landscape unit must encompass an area that is homogeneous enough to be considered as a forest management unit. It should correspond to a potentially homogenous forest type (e.g., the Cullenia exarillata – Mesua ferrea – Pallaquium ellipticum type of forest of the Periyar region, (Ramesh et al, 1997) or to a transformation of that forest type by anthropogenic activities (e.g., tea plantations in Peermade area, and cardamom in Elamalai Hills). In other words, the landscape unit contains elements that have evolved differently from the same originally homogeneous natural vegetation type.


General principles

4In books dealing with landscape ecology (e.g., Forman and Godron, 1986; Burel and Baudry, 2006) precise guidelines to delineate landscape units are not provided; on the contrary, authors generally consider that the method and the units, which are identified, must be adapted to the objectives. Nevertheless, some concepts appear to be important in the different examples given in their books, for instance those of ‘levels of organization’ and ’hierarchy’.

5Burel and Baudry (ibid) assume that “landscapes are complex systems in which an entire series of ecological phenomena occur, each having its own spatio-temporal scale”. The idea of spatio-temporal scale is not a recent one and was proposed more than fifty years ago by Cailleux and Tricart (Tricart, 1965) to classify the phenomena studied by the earth sciences.

6More recently, the hierarchy theory of Allen and Starr (1982) has offered a conceptual framework to treat all phenomena, which are dependent on scale. Further, the theory suggests that landscapes can be analyzed at different levels, using different criteria, depending on the context of the analysis (O’Neill et al, 1986).

7For an ecologist, the Western Ghats is one of the biogeographical regions of Peninsular India and is the topmost level of organization to be considered here. For a physical geographer (Gunnell, 2001), the Western Ghats region corresponds to a passive margin and has evolved according to a general model described by Summerfield (1991). This evolution has given rise to three major physical regions (Widdowson, 2001): the escarpment (the Ghats itself), its piedmont (the coastal areas) and its back-slope (the interior plateaus). We will hereafter divide these major units to identify and delineate other landscape units at a scale more appropriate to our objective, which is the ecological management of Kerala forests.

Criteria used for identifying and delineating landscape units at appropriate scale

8The identification of landscape units follows a “top-down” approach, that is, in our case, starting from peninsular India; according to different criteria, the region is successively subdivided into smaller tracts. A Geographical Information Systems (GIS) is a good and modern tool to retrieve and extract information from thematic maps (layers) and delineate the landscape units at different scales, a work that was previously time consuming and difficult to achieve for large territories.

  • The first criterion combines climate, geology, and topography to delineate major geographical units as defined by Singh (1977) in his book ‘India - a regional geography’.
  • The second criterion used is a combination of topography and soil. The previous geographical units are further subdivided according to their morpho-pedological characteristics. That is, the resulting units are characterized by similar landforms (slope, flat valleys, convex hills, etc.) and association of soil types.
  • The third criterion is the potential extent of natural vegetation types, knowing that for a given area, the natural vegetation type (wet evergreen, dry deciduous forests, grassland) is the visible expression of various characteristics of that area. These characteristics mainly result from abiotic conditions such as climate, geomorphology and soil.


Mapping of landscape units

9The results presented here concern the application of the three sets of criteria to organize in hierarchy the levels of organization of space as set by the abiotic environment (climate, landform and soil). Although results are presented only for Kerala state, the analysis had to take into account physical features that naturally extend into neighbouring Tamil Nadu and Karnataka states.

Geographical regions

10For the first level of the analysis, as already mentioned, we have used the regional partitioning of Kerala made by R.L. Singh (ibid). The three main reasons for which R.L. Singh’s regional partitioning has been retained here are:

  1. These regions correspond to geographical entities that are widely accepted by all, even outside the scientific community; their names have been for long used in common language. This can only facilitate acceptance at local level during implementation of the scheme;
  2. These geographical regions generally result from particular evolutions of the earth crust during relatively long geological periods during which their characteristics have been cast. In particular, the regions have different structures corresponding to morphological traits that can legitimately be considered in the hierarchy of landscape analysis;
  3. It is very unlikely that a landscape unit defined at a finer scale would not conform to the hierarchy of larger regions (for example if it overlaps two neighbouring larger regions). But even if it was the case, a landscape unit that cut across two regions could be considered as one for operational reasons.

11In Singh’s Regional Geography, India is divided into four Macro-level regions: Great Plains, Himalayan Mountain regions, Peninsular Uplands and Indian Coasts and Islands (Figure 1.1). These are subdivided into 28 Mesolevel regions, out of which only three concern Kerala including adjacent part of neighbouring states: (XXII) Karnataka Plateau, (XXIV) Tamilnadu uplands and South Sahyadri and, (XXVI) West Coast region. These three regions are further subdivided into 9 first-order regions, 27 second-order and 54 third-order regions. Among these, 15 regions are entirely or partly situated in Kerala state (Figure 1.2). The boundaries given by R.L. Singh have been refined using GIS data (mainly topography / contour lines) extracted from 1: 250,000 toposheets.

Figure 1.1 Hierarchical illustration of organization of India into landscape units and corresponding potential vegetation types and their transformations

Figure 1.2 Fifteen geographical regions, entirely or partly located in Kerala

Landforms and soils

12The following step was the introduction of landform and soil information to differentiate ‘morpho-pedological units’ (often called ‘soilscapes’) that correspond to a close relationship between geomorphological units (scarps, plateaus etc.) and the soil types. Landform analysis was carried out through interpretation of topographical maps at 1: 500,000 scale to get a general view of the pattern of main landform units. The same type of analysis was previously preformed for Western Karnataka, Goa and North Kerala (Bourgeon, 1989). For the piedmont and the backslope of the Western Ghats escarpment, this analysis is based on the following assumptions:

  • The relief had evolved under planation processes.
  • The planation surfaces resulting from this evolution are recognizable and they display different stages of conservation (vs degradation / dissection) as shown in Figure 1.3.

Figure 1.3 Different types of relief used to describe landforms

13After the landform analysis, soil data were extracted from the Soil Resource Map of Kerala prepared by NBSS & LUP (National Bureau of Soil Survey and Land Use Planning; Krishnan et al, 1996). The soil classification was updated to fit the 1998 Keys for Soil Taxonomy.

14The combined analysis of landform and soil information results in a map comprising 32 morpho-pedological units (Figure 1.4).

Figure 1.4 Thirty two morpho-pedological units of Kerala derived in the initial step
(Reffer "morpho-pedological units" in the CD for the definition of the units)

Figure 1.5 Profile 1 (8° 45’ N) – Transect through Travancore and Agastyamalai hills;
Profile 2 (9° 45’ N) – Transect from Kottayam to Varushanad hills

Figure 1.6 Profile 3 (11°N) - Transect from Kozhikode region Milgiris: Profile 4 (11° 40° N) Transect from Camanore region to Wayanad

15The legend describing these units with information on soil association present in each unit is given in the CD, which has been provided in the back cover of this book. In order to better illustrate the importance of such units in landscape analysis, four cross-sections at representative locations are drawn across Kerala state. Then for each section, a topographic profile (shown on the left side) and a diagram indicating the distribution of the morphopedolological units are shown (on the right side) (Figures 1.5 and 1.6).

16The morpho-pedogical units thus derived using different scales of information were found to be too detailed, with disparate size (last column of the legend) and shapes, to be used as distinct forest units landscape for management. Without much compromise on the principle used, the 32 units were grouped by combining geomorphologically closer units and using potential boundaries of forest types. Subsequently, the number of units in the state has been brought down to 19 units, and for practical purposes, these are considered as landscape units for the Kerala state (Figure 1.7). The combination of physical characteristics of these units in terms of geography, geomorphology and soil is given in Annex 1.1. In the same Annex the way the original morpho-pedological units have been grouped is also indicated.

Landscape units and potential area of forest types

17In Kerala, physiography plays a major role on local variability in bioclimate. This in turn is expressed in various kinds of biota of that region. Potential areas of 12 main forest types and two facies or subgroups have been identified apart from coastal vegetation (Figure 1.8) based on floristic composition in relation to bioclimate (Ramesh et al, 1997; Ramesh et al, 2002). These types are closely correlated with rainfall regimes, temperature and length of dry season (Annex 1.2). Wet evergreen, moist deciduous and dry deciduous forests are clearly distinguished by the gradient in mean annual rainfall, due to orographic effect of the Ghats, whereas low, medium and high elevation wet evergreen forests are distinguished by the decrease in minimum temperature with increase in altitude, and increase in dry season with increase in latitude.

Figure 1.7 Nineteen landscape units of Kerala after regrouping nearly identical morpho-pedological units
(for the definition of the landscape units, see Annex 1.1)

Figure 1.8 Potential area of primary forest types in Kerala
(for the definition of the floristie types, see Annex 1.2)

18Although the boundaries of landscape units and potential area of forest types may not exactly match, major area of certain forest types falls within a unit or group of geomorphologically or geographically closer units. For example potential area of ‘Dipterocarp’ forest types of low elevation largely come under dissected hilly hinterlands of North and South Malabar, and strongly dissected scarps and residual hills of South Sahyadris and western slope of the Wayanad region; ‘Cullenia’ forest types of medium elevation were mostly found in partially dissected to strongly dissected surface of Agatyamalai hills, Anamalai, Elamalai and Palghat hills and, scarp and residual hills of the Wayanad; high elevation and montane forest types were generally found on upper slopes and rounded crests of western slopes and highlands of North Nilgiri region, and partially or strongly dissected high surface of Anamalai region.

19Primary moist deciduous forest mostly confined to hilly hinterlands of South Malabar, low residual and broad valleys in Palghat region, and multi convexo-concave reliefs of Wayanad. Primary dry deciduous types, characteristically found in north and south slopes of Palnis and Konganad upland regions and the eastern part of the Wayanad region.


20Landscapes are generally defined according to the objectives of the study. As the scope of the present study is related to forest management, the landscape units are identified using the well known geographical regions, topographic features and for certain extent considering vegetation types (as an indicator of ecology). The derived units although differ substantially on the typical forms and shapes they nevertheless forms the unique units identifiable on the ground. The understanding of the similarities and dissimilarities of the these units in terms of characteristics of landscape elements and their functioning would help the forest managers to rationalize the existing management units, which are currently based on administrative criteria.

21Although the landscapes defined is confined to Kerala, the mapping method and the analysis carried out can be easily extended to the neighbouring states without major changes as the same contour maps, soils maps, geological maps and regional analysis of Singh (ibid) are available.

Table des illustrations

Légende Figure 1.1 Hierarchical illustration of organization of India into landscape units and corresponding potential vegetation types and their transformations
Fichier image/jpeg, 62k
Légende Figure 1.2 Fifteen geographical regions, entirely or partly located in Kerala
Fichier image/jpeg, 78k
Légende Figure 1.3 Different types of relief used to describe landforms
Fichier image/jpeg, 18k
Légende Figure 1.4 Thirty two morpho-pedological units of Kerala derived in the initial step(Reffer "morpho-pedological units" in the CD for the definition of the units)
Fichier image/jpeg, 79k
Légende Figure 1.5 Profile 1 (8° 45’ N) – Transect through Travancore and Agastyamalai hills;Profile 2 (9° 45’ N) – Transect from Kottayam to Varushanad hills
Fichier image/jpeg, 50k
Légende Figure 1.6 Profile 3 (11°N) - Transect from Kozhikode region Milgiris: Profile 4 (11° 40° N) Transect from Camanore region to Wayanad
Fichier image/jpeg, 57k
Légende Figure 1.7 Nineteen landscape units of Kerala after regrouping nearly identical morpho-pedological units(for the definition of the landscape units, see Annex 1.1)
Fichier image/jpeg, 73k
Légende Figure 1.8 Potential area of primary forest types in Kerala(for the definition of the floristie types, see Annex 1.2)
Fichier image/jpeg, 54k

© Institut Français de Pondichéry, 2007

Conditions d’utilisation :


Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search