The methodology developed to assess biodiversity over the Western Ghats using remote sensing, involves a synthesis of information derived at several spatial scales.

In India we have 320 million hectares of land, and 200 million hectares of exclusive economic zone in the sea, within which are distributed some 120, 000 known and perhaps another 400, 000 as yet undescribed species of microbes, plants and animals In a country with this rich heritage of biological diversity, it is obviously not possible to census the distribution of each and every species based on field studies alone.

A two-level combination of remote sensing and field studies, can be used to derive information on the distribution of large numbers of species.This methodology appears simple - but there are several possible difficulties associated with its use over large areas. Although several species are known to exist in tight association with their habitats, some species are more wide ranging, and may occur in a wider variety of habitats. Certain groups of organisms, such as birds, may require a mixture of several ecotopes, some for foraging and others for breeding, for example. It may not be possible to relate their distribution to the presence of a single ecotope type. The spatial scale at which habitats are differentiated by the remote sensor may also crucially affect the results obtained. For example, herb species may respond to ecotopes at a very fine scale, much finer than the spatial resolution of the remote sensor.

Still, such a two scale coupling of remote sensing and field sampling still shows great promise and, indeed, has been recommended by the Global Biodiversity Assessment.

Study area
Biogeographically, the hill chain of the Western Ghats constitutes the Malabar province of the Oriental realm, running parallel to the west coast of India from 8 ° N to 21 ° N latitudes for around 1600 km. Rising up from a relatively narrow strip of coast at its western border, the hills reach up to a height of 2800 m before they merge to the east with the Deccan plateau at an altitude of 500-600 m. The average width of this mountain range is about 100 km. This bio-region is highly species rich and under constant threat due to human pressure, and is considered one of the 18 biodiversity hot spots of the world. With its complex, heterogeneous landscapes and high levels of biodiversity, it forms an ideal ground for the testing and elaboration of this methodology.

Methods and Results
The methodology developed to assess biodiversity over the Western Ghats using remote sensing, involved a synthesis of information derived at several spatial scales. At the largest scale of mapping, satellite imagery covering the area of the Western Ghats, was collected from IRS 1 B LISS 2 sensors, during the pre-monsoon seasons of either 1992, 1993 or 1994 (depending on availability of cloud free data). The scenes were manually co-registered and pasted together to create a composite image of the study area. Using information derived from the Normalized Difference Vegetation Index (to minimize effects of inter-image variability), according to the methodology described by Nagendra and Gadgil, a relatively coarse, broad scale map of the Western Ghats was created. This map classifies the Western Ghats into 200 patches belonging to eleven ecomosaic types. Each ecomosaic type is a characteristic set of several ecotope types, both manmade such as different types of agricultural crops and plantations, and natural, such as forests in various stages of degradation. The resultant map was compared with pre-existing information on the distribution of forests, agricultural lands, climatic and topographic features, and population, for interpretation.

At the next scale of mapping, twelve landscapes belonging to five of the eleven ecomosaic types were taken up for more detailed investigations. These landscapes range in size from 9 to 54 sq. km in area. Supervised and unsupervised classifications of LISS 2, 4 band data was carried out at this scale, using single date pre-monsoon imagery collected either in 1992, 1993 or 1994. Supervised classification accuracies of these twelve landscapes ranged from 70% to 92%. Unsupervised classification accuracies were uniformly much worse.

From the supervised classification maps, information on the size, shape and inter-patch distance of ecotopes was calculated. This information is believed to affect the presence and distribution of various species within a landscape. In addition, ecotope type richness, and Shannon’s index of ecotope diversity (based on proportion of landscape area occupied by various ecotope types) were computed. Statistical analysis determined that landscapes belonging to specific ecomosaic types tended to be similar in their ecotope characteristics, thus providing us with confidence in the ecomosaic map of the Western Ghats.