Southern Madagascar

Madagascar is one of the world's most "threatened hotspots" for terrestrial biodiversity and a priority for conservation action. Green and Sussman (1990) used maps generated from aerial photographs and satellite images to study the history of deforestation of the rain forests of eastern Madagascar. They were able to show that, in 1985, only 34% of the estimated original rain forest remained and that deforestation was most rapid in areas of low topographic relief and high population density. Observations of the dry and deciduous forests of southeastern and southern Madagascar suggest that they are being reduced at a similar or even faster rate than the tropical wet forests in the east (Sussman et al. 1994, 1995).

The dry forests of the south and west are unique and are inhabited by many plants and animals found nowhere else on Earth. Although much attention has been directed to deforestation in the East, the unique semi-arid habitat of southern Madagascar, containing more endemic plant families than elsewhere on the island (Figure 1 shows Alluaudia procera trees of the endemic Didiereaceae family), is even more vulnerable to human disturbance. Furthermore, these forests mostly cover areas of low topographic relief and, unlike the remaining eastern rain forest that are afforded some inherent protection by steep slopes, the lack of topographic relief in the south leaves this unique habitat more vulnerable to clearing.

Concerning Madagascar, Smith (1997, p.441) states: "seasonally dry...forests have declined more severely, and are under greater threat, than tropical wet forests...leaving western dry forests relatively ignored and underprotected." Protected areas cover only around 2% of the total remaining forest in the dry forest areas of Madagascar compared with approximately 5.3% for lowland evergreen forest (Du Puy and Moat 1996). Recently, the spiny forest of southern Madagascar has been listed as one of the 200 most important ecological regions of the world (Olson and Dinerstein 1998). Thus, there is great urgency to document the deforestation as well as to determine the rate and patterns of habitat loss in southern and southwestern Madagascar.

We employ scenes from six contiguous Landsat footprint locations acquired at three different dates (and satellites): 1973 (from Landsat 1), 1985 (from Landsat 5), and 1999/2000 (from Landsat 7). A total of 20 scenes were purchased and geometrically registered to the 1:100,000 scale topographic maps. Topographic maps generated from photo-interpretation of aerial photos acquired around 1950 are also examined.

Multi-temporal color composite map products of scale 1:100,000 were produced from Landsat images from 1973, 1985, and 1999/2000. The multi-temporal image composites were constructed using the red wavelength bands from three different dates: band 2 from the 1973 MultiSpectral Scanner (MSS) scenes, band 3 from the 1985 Thematic Mapper (TM) scenes, and band 3 from the 1999/2000 Enhanced Thematic Mapper Plus (ETM+) scenes. For the multi-temporal Landsat derived map products the 1973 image brightness was set to the color blue, the 1985 image brightness was set to green, and the 1999/2000 image brightness was set to red. color of the composite can be used to detect land cover change between the three dates.

In these multi-temporal products land cover that does not change appreciably between the three dates appear black, white or a shade of gray, depending on its inherent brightness at red wavelengths (0.6 to 0.7 micrometers). At these wavelengths forested terrain is generally dark, while savannah, cleared land and riverbank sands are bright. Thus, stable forest (between 1973 and 1999/2000) is depicted as black or dark gray in the multi-temporal composites, and unchanging savannah is generally bright gray or white. On the other hand, areas which experience land cover change between the three dates exhibit different image brightness in the 1973, 1985 or 1999/2000 scenes and this imparts a color to the multi-temporal composite map product. Thus, areas that experience land cover change can be identified in these products by their color. The particular color gives the particular sequence of those changes.

Deforestation associated with this commercial charcoaling is shown in Figure 2. The red areas in this multi-temporal color composite show those areas of forest cleared between 1985 and 2000. Dark grey or black areas on the image are predominantly forested and have experienced little change. The yellow/green swathe in Figure 2 (oriented southwest to northeast) running from the middle of the left side of the figure to the savannah edge, parallels Route Nationale #7. These areas were cleared between 1973 and 1985 for commercial charcoal production. Charcoal is produced by local inhabitants where the trees are cut, then transported by ox-cart to a location along the highway accessible by large truck. Along the road this bagged charcoal is then sold to distributors (Figure 3).

The charcoal is then transported to Toliara where it is sold in the public markets. Currently most residents of Toliara cook all their food (mostly rice) with charcoal. The practice of cutting forest to produce charcoal for Toliara has resulted in a large and expanding zone devoid of many woody plants to the northwest of the city. Much formally forested land present along Route Nationale #7 until the late 1960s is now characterized by abundant bare limestone rock exposures (Figure 4). Observations of land cleared for charcoal prior to 1987 reveal little re-growth of woody plants in the subsequent 14 years.

 

 

Madagascar Links

• Madagascar Wildlife Conservation

Much of this material is drawn from Sussman et al. (see below) accepted in Primate Conservation.

Related Publications:

Robert W. Sussman, Glen M. Green, Ingrid Porton, Ony L. Andrianasolondraibe, and Joelisoa Ratsirarson, 2003, A Survey of the Habitat of Lemur catta in Southwestern and Southern Madagascar, Primate Conservation, n.19, p.32-57.

Endre A. Nyerges, and Glen M. Green, 2000, The Ethnography of Landscape: GIS and Remote Sensing in the Study of Forest Change in West African Guinea Savanna, American Anthropologist, v.102, n.2, p.271-289.

Charles M. Schweik, and Glen M. Green, 1999, The Use of Spectral Mixture Analysis to Study Human Incentives, Actions and Environmental Outcomes, Social Science Computer Review, v.17, n.1, p. 40-63.

Sussman, R.W., Green, G.M., and Sussman, L.K., 1995, The Use of Satellite Imagery and Anthropology to Assess the Causes of Deforestation in Madagascar, in Tropical Deforestation: The Human Dimension, Sponsel, Headland and Bailey, editors, Columbia University Press, New York.

Sussman, R.W., Green, G.M., and Sussman, L.K., 1994, Satellite Imagery, Human Ecology, Anthropology, and Deforestation in Madagascar, Human Ecology, v. 22, p. 333-354.

Green, G., and Sussman, R., 1990, Deforestation History of the Eastern Rain Forests of Madagascar from Satellite Images, Science, v. 248, p. 212-215.

408 North Indiana Avenue, Bloomington, IN 47408-3799
Phone: (812) 855-2230
TDD: (812) 855-7654
Fax: (812) 855-2634

Last Updated: May 11, 2005
Comments: cipec@indiana.edu 
Copyright 2005, The Trustees of Indiana University.