Overview
While pre-1972 global vegetation patterns can only be estimated using a variety of his torical sources, post-1972 global vegetation patterns can be directly determined using satellite remote sensing. Our ability to assess and evaluate future changes in these patterns has thus changed dramatically through the application of satellite remote sensing. These data sources (satellite remote sensing and in situ or field studies) have been used to make an up-to-date estimate of the impacts of both natural and anthropogenic changes to vegetation cover in selected regions since the 1970s. For example, recent studies of tropical forest and savanna vegetation in Brazil have documented changes in an area of 7,000,000 square kilometers (km2). While estimates of reported vegetation cover change derived using different methodologies vary, there is agreement that a substantial alter ation of vegetation cover in Brazil has occurred since 1972. The extension of these techniques to other regions of the tropics (i.e., non-Brazilian South America, Southeast Asia, and tropical Africa) was discussed at this Aspen Global Change Institute (AGCI) session, and regional/historical differences between these areas and the Brazilian Amazon were identified.
The various tools for satellite-based assessment were also discussed at this AGCI session. These tools include optical remote sensing, and passive and active (radar) microwave remote sensing, all coupled with "ground truth" validation and numerical model simulations. The model simulations are necessary to evaluate and assess the impact of global vegetation changes on environmental parameters and processes, as ecological, bio-geochemical, and climate models are all needed to completely evaluate the potential impacts of these vegetation changes on the global environment.
Introduction
Changes to global vegetation cover resulting from human activities have profound implications for ecosystem functioning, biogeochemical cycles, and climatic stability. It is generally assumed that human activity has dramatically altered the natural vegetative cover of our planet, especially in the past 300 years (Turner et al. 1990). Upon closer analysis, one discovers not only a considerable ignorance of the present global distribution of terrestrial land cover types, but no systematic, reliable, and comprehensive compendium of human-caused changes to the natural vegetation cover on a global scale (Townshend et al., 1991). In an effort to improve this state of knowledge, the Aspen Global Change Institute devoted its first of three 1995 summer sessions to understanding changes to global vegetation patterns and their relationship to human activity.
A review of existing estimates of global land cover reveals a high degree of disagreement (see figure i.1, Townshend et al., 1991). This is also true for global estimates of agricultural land and forested land, two categories of land cover which can be easily mapped. Substantial disagreements in extent of various land cover types globally result from the fact that estimates rely on reconciling numerous separate sources employing widely-varying criteria. It is thus not surprising that such a high degree of disagreement is present. Furthermore, not only does the total area of the various classes vary substantially among authorities, but the specific spatial distributions of ten vary widely even when the total global estimates of a cover type are similar.
Major efforts have been made to synthesize current global land cover knowledge to generate global digital data bases (Matthews, 1983 and Henderson-Sellers et al., 1986). Although these represent improvements on previous knowledge, they suffer from unavoidable errors inherent in the primary data upon which they are based.
Not only are accurate estimates of global land cover of the utmost importance for under standing the coupled Earth-climate system, the extent of land cover type fragmentation is also important. For example, if Earth's climate does become warmer, the ability for terrestrial ecosystems to adapt by "ecosystem migration" could be compromised if natural ecosystems are broken up into non-contiguous pieces. In addition, biological diversity concerns are directly related to habitat destruction and fragmentation.
Regional Examples or Case Studies of Vegetation Changes
Satellite remote sensing has been suggested as a possible means to overcome the tremendous disagreement obvious in figure i.1 (Tucker et al., 1985; Justice et al., 1985; Goward et al., 1985; Townshend and Justice, 1986; Townshend et al., 1987, 1989, 1991; Defries and Townshend, 1995; Loveland et al., 1995). Two different satellite-based approaches have been used to produce estimates of land cover extent:
Landsat Multi-Spectral Scanner (MSS) and /or Thematic Mapper (TM) data, coupled with a geographic information system (Skole and Tucker, 1993) has been used to map deforestation and surface cover alteration in tropical areas. The use of Landsat data to map land cover is especially important because it can detect many human-caused surface changes which are not detectable using AVHRR data.
Two very different approaches are used to assemble the "seamless" data sets required to study areas millions of km2 in size. With AVHRR data, the daily satellite data are processed into time series composite images which are in turn analyzed to produce land cover type estimates. Thus seamless satellite data are produced and then analyzed. By comparison, for Landsat data hundreds of Landsat scenes are individually analyzed to arrive at the same scene attributes; these determined attributes are subsequently incorporated into a geographic information system, edge-matched, and a seamless data set of those attributes results. The remote sensing determination of tropical deforestation and also the conversion of cerrado or savanna to agricultural use in Brazil were used to illustrate the AGCI discussions of human modification of the natural environment as detected through the use of satellite data.
AVHRR Data: Continental-Scale Estimates of Land Cover
AVHRR data are useful for large-scale land-cover mapping because of their multi-temporal nature. Daily AVHRR data are formed into normalized difference vegetation index (NDVI) composites which are estimates of vegetation photosynthetic capacity. This follows from the basic relationship of the NDVI to intercepted photosynthetically active radiation (PAR), and this relationship to photosynthesis. Summed over the growing season, PAR and NDVI are highly related to total photosynthesis, and hence to total dry biomass production (figure i.2). Thus, multi-temporal AVHRR NDVI composites allow classification of the surface based upon photosynthetic capacity and how this varies in time.
AVHRR data have been formed into time series of the NDVI and used to investigate the possibility of continental land cover determination(s) (Townshend et al., 1985 and 1987; Tucker et al., 1985; Tucker et al., 1991; Tateishi and Kajiwara, 1992; Eastman and Fulk, 1993). The assumption in all of the AVHRR continental land cover studies is that the various large aggregations of land cover have different NDVI responses through time. These different magnitudes and time variations are the means to classify or discriminate between cover types (desert, semi-arid steppe, savanna, forest, etc.).
Problems arise when more and more land cover aggregations are desired, or when larger and larger areas are studied. Problems also occur when multiple continents are studied together. These problems stem from a lack of clear boundaries between different vegetation aggregations and make exact boundary delineation impossible in many situations.
An example of where this has been overcome has been reported by Tucker et al. (1991 and 1994) and involves determining the area of the Sahara Desert from 1980 to the 1990s. In these studies, precipitation data were correlated with coincident NDVI data for 42 Sahelian station locations (figure i.3). This provided the means to use a specific NDVI value as the threshold corresponding to a specific precipitation amount. Tucker et al. , (1991 and 1994) used the 200 mm/year precipitation isoline as the boundary between desert and non-desert, on the south side of the Sahara. This approach was then used from 1980 through the present to document the expansion and contraction of the Sahara to the south. Other land cover determinations must use specified thresholds for assigning areas studied to specific land cover aggregations. Without such thresholds, it is difficult to assign specific boundaries between the various cover types which have meaning.
Figures i.2 and i.3 represent two methods for determining the accuracy of NOAA AVHRR satellite NDVI data in terms of coincident ground observations from specified areas. Other relationships from different locales have been reported by Nicholson and her students (Davenport and Nicholson, 1993; Malo and Nicholson, 1990; and Nicholson et al., 1990).
Landsat Data/GIS Tropical Forest and Savanna Example
At the present time, a substantial alteration of natural vegetation is occurring through out the tropics, in both tropical forests and tropical savannas. Recently researchers have suggested that tropical deforestation plays a major role in the global carbon cycle and has profound implications for biological diversity. Tropical deforestation has also been the subject of intense popular media coverage the past decade. It is thus instructive to review our knowledge of the extent and rate of tropical deforestation, as this is directly pertinent to the theme of this Aspen Global Change Institute session.
Tropical forests (including moist evergreen and seasonal forests) once covered ~24,500,000 km2 of Earth's terrestrial surface (Whittaker and Likens, 1975) and are now estimated to cover only ~10,000,000 km2 (Wilson, 1986). This dramatic decrease has generated widespread concern and calls for tropical forest conservation. These concerns can be traced to the fact that approximately 8,000,000 km2 of tropical land have been converted to agriculture, ~3,000,000 km2 are under shifting cultivation, and ~3,500,000 km2 have been converted to pasture (Salati and Vose, 1983 and 1984). The Amazon Basin contains ~60% of our planet's remaining tropical forest with Brazil containing ~40% of the global total (Mares, 1986).
Tropical forests are home to the greatest diversity of plant and animal life known on Earth, containing over half of our planet's plant and animal species. They are estimated to contain far in excess of 5 million plant and animal species (Prance, 1982). Wilson (1986) notes that ~1.7 million species have been described, including ~750,000 insects, ~47,000 vertebrates, and ~250,000 plants. Some researchers believe the total number of species could be higher than 30 million (Erwin, 1986). Taxonomists have identified only a small percentage of the several million tropical species. One principal adverse effect of tropical deforestation would be mass extinctions comparable to that which last occurred ~60-70 million years ago at the end of the Mesozoic era when the dinosaurs, among many other species, became extinct. These mass extinctions would be expected to result primarily from tropical forest habitat destruction.
Brazil contains the majority of the tropical forests within the Amazon Basin. The so-called "Legal Amazon" covers ~5,000,000 km2, of which 70% is occupied by the non-flooded ( terra firme) forest. The rest includes swamp forest, flooded grassland and cerrado (bush land, grassland and savanna). The difficulties of monitoring changes in this large area have resulted in substantial disagreements on the extent of deforestation. For example, the World Bank published figures stating that as of 1988, ~600,000 km2 (~12%) of the Legal Amazon was "cleared" (Mahar, 1989) while Brazilian sources stated that only ~280,000 km2 (~5%) was cleared in 1988 (Tardin and da Cunha, 1990). These two reports, both for 1988, illustrate the late 1980's disagreement over actual figures for deforestation in the Amazon Basin of Brazil (see also Washington Post, 1989) and prompted David Skole at the University of New Hampshire and Compton Tucker at NASA Goddard Space Flight Center to investigate this matter.
Skole and Tucker (1993) have reported 230,000 km2 of deforestation in the Legal Amazon as of 1988, as well as calculating the total areas of isolated forest and deforestation/forest edges. When considering the effect of tropical deforestation upon biological diversity, three components must be considered simultaneously: 1) the deforestation per se (i. e., habitat destruction), 2) areas of isolated forest fragments, and 3) the total area of intact forest in direct contact with areas of deforestation (i.e., the "edge" effect).
Skole and Tucker (1993), drawing upon previous deforestation remote sensing research (Kaufman et al. , 1990; Malingreau and Tucker, 1988; Nelson and Holben, 1986; Nelson et al., 1987a and 1987b; Setzer, 1988; Tardin et al., 1980; Tucker et al., 1984; Woodwell et al., 1983 and 1987), developed a geographic information system approach to analyze tropical deforestation using Landsat TM data as the primary data source (Skole, 1992; Skole and Tucker, 1993). This technique is student-labor intensive but much less expensive than any other means for studying deforestation of large areas while providing a high degree of accuracy. Skole (1992) has reported the accuracies of various analytical approaches (table i.1).
Once the deforestation data are incorporated into the geographic information system, they can be combined with other information such as vegetation distributions, elevation, hydrology, etc., and then used for a variety of purposes. The geographic information system can then be used to determine land cover spatial descriptions such as patch size, edge length, distance between patches, etc. Most importantly, the geographic information system also functions as a data management system. This is an important consideration when dealing with assembling several hundred Landsat images into a "seamless" data set.
Ecological Consequences of Fragmentation
The habitat which remains after parts have been deforested is broken up into fragments which can be isolated to varying degrees (Lovejoy et al., 1984, 1986; Wilcove et al., 1986). The connectivity with other similar fragments, time since fragmentation, and distance between fragments are all important in determining the biological effects of fragmentation (Miller, 1978; Wilcox, 1980; Harris, 1984; Saunders et al., 1991). In order to study this question, the landscape must be specified in terms of fragment size, fragment shape, and location in the landscape (Franklin and Forman, 1987; Ripple et al., 1991). One way to study large-scale habitat fragmentation is to employ a geographic information system using Landsat TM data to provide the required information about surface conditions.
Savanna Alteration in Brazil Since 1970
While Brazil contains the largest expanse of continuous tropical rain forest on the planet in its Amazon Basin, it also contains a very large area of tropical savanna. The ~5,000,000 km2 of Brazilian Legal Amazon is comprised of 80% tropical forest, 5% open water, and 15% savanna or cerrado. To the south and east of the Legal Amazon is an expansive mosaic of cerrado comprised of evergreen woodland, savanna, and grasslands. This mosaic is often referred to as the Brazilian Cerrado and totals ~2,000,000 km2 of which ~1,400,000 km2 lies outside of the Legal Ama zon (Nepstad et al., 1995).
Cerrado vegetation ranges from open grassland to closed-canopy woodland which differ in proportions of herbaceous and woody vegetation. The herbaceous cover is very active during the rainy season which usually lasts from October to March; ~90% of the cerrado receives between 1,000 to 2,000 mm of precipitation per year. During the dry season precipitation is very low, daytime temperatures high, relative humidity low, and evaporation very high. Consequently, the incidence of fire is very high and the accumulation of dead herbaceous vegetation facilitates fire occurrence (Adamoli et al., 1986).
The Brazilian Cerrado has experienced and continues to experience large-scale land use conversion, from formerly natural ecosystems to various types of human-managed activities. Because much of the exploitation of these areas has occurred since the early 1970s, this is an ideal area to evaluate the use of Landsat and other satellite data to identify conversion of natural ecosystems to areas of human activity.
It is much more difficult to use Landsat satellite data to quantify land-use changes in areas of cerrado or savanna than in closed-canopy forest. This results from the Brazilian Cerrado being a seasonal savanna with a continuous layer of herbaceous species at peak periods of growth, with scattered trees and bushes that can sometimes form continuous canopies. In addition, corridors of gallery evergreen forest occur along rivers (Klink et al., 1993). The difficulty in using Landsat satellite data to distinguish undisturbed cerrado from cerrado converted to agricultural use is that differences between these are more subtle compared to those in closed-canopy forested areas. It is perhaps most difficult to distinguish undisturbed cerrado from cerrado converted to agricultural use in the dry season when cloud cover is at a minimum. The only way to overcome this dry season limitation is to use Landsat data from more than one date. Complicating this further, there is disagreement regarding the "natural" cerrado vegetation was it open grassland savanna or was it closed canopy woodland that vanished through the increased use of fire since Colombian times?
Until 1965, the cerrado of Brazil was used primarily for cattle grazing. Since 1965, the Brazilian cerrado has been increasingly transformed into cultivated pastures, field crops, dams, urban areas, and degraded areas. It is estimated that ~40% of the natural vegetation of the cerrado has been converted from the natural state (Dias, 1994). Total population in the cerrado area grew from 6.5 million people in 1970 to 12.6 million in 1991. Principal uses of the converted cerrado areas are agricultural and include cultivated pas tures, soybeans, corn, rice, and other crops (Nepstad et al., 1995).
As Nepstad et al. (1995) point out, the lack of information on land-use in the Brazilian Cerrado (and by extension, to all tropical savannas) is remarkable when compared to the attention given to closed canopy tropical forest. The need for information regarding land-use changes in natural non-forested areas can be extended to include not only tropical savannas but temperate grasslands or steppes. Satellite data are the only economically viable means of determining land-use change in savannas and grasslands.
References
Adamoli, J. et al., 1986. Caracterizacao da regiao dos cerrados. In: W. J. Goedert. Solos dos Cerrados. Sao Paulo, EMBRAPA and Nobel Press, pp. 33-74.
Adams, J. B., Smith, M. O., and Johnson, P. E., 1986. Spectral mixing modeling: A new analysis of rock and soil types at the Viking Lander 1 site. J. Geophys. Res. 91:8098-8122.
Adams, J. B., Smith, M. O., and Gillespie, A. R., 1989. Simple models for coples natural surfaces: A strategy for the hyperspectral era of remote sensing, in Proceedings of the IEEE Geoscience Remote Sensing Symposium '89 , I:16-21.
Cale, P. and Hobbs, R. J., 1991. Condition of road side vegetation in relation to nutrient status. pp. 353-362 in D. A. Saunders and R. J. Hobbs, eds., Nature Conservation and the Role of Corridors. Surrey, Beaty, and Sons, Chipping Norton, Australia.
Davenport, M. L. and Nicholson, S. E., 1993. On the relationship between rainfall and the normalized difference vegetation index for diverse vegetation types in East Africa. Int. J. Remote Sens. (in press).
Dias, B. F., 1994. "Conservacao de natureza no cerrado," in Cerrado: Caracterizacao, Occupacao, e Perspectivas. Brazilia Univ. Press, Brazilia, pp. 583-640.
Eastman, R. R. and Fulk, M., 1993. Long sequence time series evaluation using standard principal components. Photogram. Eng. Remote Sens. 59:991-996.
Erwin, T. L. 1986. News and Comment. Science , 234:14-15.
Fearnside, P. M., 1986. Spatial concentration of deforestation in the Amazon Basin. Ambio 15:74-81.
Franklin, J. F. and Forman, R. T. T., 1987. Creat ing landscape patterns by forest cutting: Ecological consequences and principles. Landscape Ecology 1:5-18.
Goward, S. N., Tucker, C. J., and Dye, D. G., 1985. North American vegetation patterns observed with the NOAA-7 advanced very high resolution radi ometer. Vegetatio 64:3-14.
Harris, L. D., 1984. The Fragmented Forest: Is land biogeographic theory and the preservation of biotic diversity. University of Chicago Press, Chicago, Illinois.
Harris, L. D., 1988. Edge effects and conservation of biotic diversity. Conservation Biology, 2:330-332.
Henderson-Sellers, A., Wilson, M. F., Thomas, G., and Dickinson, R. E., 1986. Current Global Land Surface Data Sets for Use in Climate-Related Studies. NCAR Technical Note 3 272-STR, National Center for Atmospheric Research, Boulder, Colorado.
Janzen, D. H., 1983. No park is an island: increase in interference from outside as park size decreases. Oikos 41:402-410.
Justice, C. O., Townshend, J. R. G., Holben, B. N., and Tucker, C. J., 1985. Analysis of the phenology of global vegetation using meteorological satellite data. Int. J. Remote Sens. 6:1271-1318.
Kaufman, Y. J., C. J. Tucker, and I. Y. Fung, 1990. Remote sensing of biomass burning in the tropics. J. Geophysical Res. 95(D7) :9927-9939.
Klink, C. A., Moreira, A. G., and Solbrig, O. T., 1993. "Ecological impacts of agricultural development in the Brazilian cerrados," in The World's Savannas: Economic Driving Forces, Ecological Constraints and Policy Options for Sustainable Land Use . Young, M. D. and Solbrig, O. T. eds., pp 259-283, Man in the Biosphere Series 12, Parthenon Publishing, London.
Lovejoy, T. E. et al., 1984. "Ecosystem decay of Amazon forest fragments," pp. 295-325 in Ex inctions, M. H. Niteki, ed., University of Chicago Press, Chicago, Illinois.
Lovejoy, T. E. et al., 1988. "Edge and other effects of isolation on Amazon forest fragments," pp. 257-285 in Conservation Biology, M. E. Soule, ed. The Science of Scarcity and Diversity. Sinauer Associates, Sunderland, Massachusetts.
Mahar, D. J., 1989. Government Policies and De forestation in Brazil's Amazon Region. World Bank, Washington, D. C., 56 pp.
Malingreau, J. P. and C. J. Tucker, 1988. Large-scale deforestation in the southern Amazon Basin of Brazil. Ambio 17:49-55.
Malo, A. R, and Nicholson, S. N., 1990. A study of rainfall and vegetation dynamics in the African Sahel using normalized difference vegetation index. J. Arid Environ. 19:1-24.
Mares, M. A., 1986. Conservation in South America: Problems, consequences, and solutions. Science 233:734-739.
Matthews, E. 1983. Global vegetation and land use: new high resolution data bases for climate studies. J. Appl. Meteorol. 22:474-487.
Miller, R. L.,1978. Applying island biogeographic theory to an East African reserve. Environmental Conservation 5:191-195.
Nelson, B. W., 1992. "Natural Forest Disturbance and Change in the Brazilian Amazon." Presented at the World Forest Watch Meeting, Sao Jose dos Campos, SP, Brazil, 6 pp.
Nelson, R. F. and B. N. Holben. 1986. Identifying deforestation in Brazil using multiresolution satellite data. Int. J. of Remote Sensing 7(3):429-448.
Nelson, R. F., D. Case, N. Horning, V. Anderson, and S. Pillai. 1987a. Continental Land Cover Assessment Using Landsat MSS Data. Remote Sensing of Environment 21:61-81.
Nelson, R. F., N. Horning, and T. A. Stone. 1987b. Determining the rate of forest conversion in Mato Grosso, Brazil, using Landsat MSS and AVHRR data. Int. J. Remote Sensing 8(12):1767-1784.
Nepstad, D. C. et al., 1995. Land-use in Amazonia and the cerrado. Ciencia e Cultura (in press).
Nicholson, S. E., Davenport, M. L., and Malo, A. R., 1990. A comparison of the vegetation response to rainfall in the Sahel and East Africa, using normalized difference vegetation index from NOAA AVHRR. Climate Change 17:209-242.
Panetta, D. and Hopkins, A. J. M., 1991, "Weeds in corridors: invasion and management," pp. 341 -351, in Nature Conservation the Role of Corridors , D. A. Saunders and R. J. Hobbs, eds. Surrey, Beaty, and Sons, Chipping Norton, Australia.
Prance, G. T., 1982. Extinction is Forever , Columbia Univ. Press, New York.
Ripple, W. J., Bradshaw, G. A., and Spies, T. A., 1991. Measuring forest landscape patterns in the Cascade Range of Oregon, USA. Biological Conservation 57:73-88.
Sabol, D. E. Jr., Adams, J. B., and Smith, M. O., 1992, Quantitative subpixel spectral detection of targets in multi-spectral images. Journal of Geophysical Research 97:2659-2672.
Salati, E. and P. B. Vose, 1983. Depletion of tropical rainforests. Ambio 12:67-71.
Salati, E. and P. B. Vose, 1984. Amazon Basin: A system in equilibrium. Science 225:129-138.
Saunders, D. A., Hobbs, R. J., and Margules, C. R., 1991. Biological consequences of ecosystem fragmentation: A review. Conservation Biology 5:18-32.
Setzer, A. W., 1988. Relatorio de atividades do projecto IBDF-INPE "SEQE"Ano 1987, Relatoria INPE-4534-RPE/565 (Instituto Nacional de Pesquisas Espaciais, Sao Jose dos Campos).
Skole, D. L., 1992. Comparison of SPOT, thematic mapper, and AVHRR for quantifying deforestation in Rondonia. Ph. D. Dissertation, Univ. of New Hampshire, Durham.
Skole, D. L. and Tucker, C. J., 1993, Tropical deforestation and habitat fragmentation in the Brazilian Amazon: satellite data from 1978 to 1988. Science 260:1905-1910.
Tardin, A. T. et al., 1979. Levantamento de areas de desmatamento na Amazonia Legal atraves de imagens do satellite Landsat. Relatoria INPE 411 -NTE/142 (Instituto Nacional de Pesquisas Espaciais, San Jose dos Campos).
Tardin, A. T. et al. , 1980. Relatoria INPE 1649 -RPE/103 (Instituto de Pesquisas Espaciais, Sao Jose dos Campos).
Tardin, A. T. and R. P. da Cunha, 1990. Evaluation of deforestation in the Legal Amazon using Landsat TM images. INPE-5015-RPE/609 (Instituto de Pesquisas Espaciais, Sao Jose dos Campos).
Tardin, A. T. et al., 1992. Evaluation of deforestation in the Legal Amazon using Landsat TM images. INPE-5015-RPE/609 (Instituto de Pesquisas Espaciais, Sao Jose dos Campos).
Tateishi, R. and Kajiwara, K., 1992. Global land cover monitoring by AVHRR NDVI data. Earth Environ . 7:4-14.
Terborgh, J., 1976, Island biogeography and conservation: strategy and limitations. Science 193:1029-1030.
Townshend, J. R. G., Goff, T. E., and Tucker, C. J., 1985. Multitemporal dimensionality of images of normalized difference vegetation index at continental scales. IEEE Trans. Geosci. Remote Sens. GE-23:888-895.
Townshend, J. R. G. and Justice, C. O., 1986. Analysis of the dynamics of African vegetation using the normalized difference vegetation index. Int. J. Remote Sens. 7:1189-1207.
Townshend, J. R. G., Justice, C. O., and Kalb, V. T., 1987. Characterization and classification of South American land cover types using satellite data. Int. J. Remote Sens. 8:1189-1207.
Townshend, J. R. G., Justice, C. O., Li, W., Gurney, C., and McManus, J., 1991. Global land cover classification by remote sensing: present capabilities and future possibilities. Remote Sens. Environ. 35:243-255.
Tucker, C. J., B. N. Holben, and T. E. Goff, 1984. Intensive forest clearing in Rondonia, Brazil as detected by satellite remote sensing. Remote Sensing of Environment 15:255-261.
Tucker, C. J., Townshend, J. R. G., and Goff, T. A., 1985. African land cover classification using satellite data. Science 227:369-375.
Tucker, C. J., Dregne, H. E., and Newcomb, W. W., 1991. Expansion and contraction of the Sahara Desert from 1980 to 1990. Science 253:299-301.
Turner, B. L., Clark, W. C., Kates, R. W., Richards, J. F., Mathews, J. T., and Meyer, W. B. (editors), 1990. The Earth as Transformed by Human Action. Cambridge Univ. Press, Cambridge UK, 713 p.
Washington Post , 1989. "Brazil angrily unveils plan for the Amazon." April 7, 1989, pp. 1 and 18.
Whittaker, R. H. and G. E. Likens, 1975. "Primary production: The biosphere and man," in Primary Production of the Biosphere, eds. Lieth, H. and Whittaker, R. H.; Springer Verlag, New York.
Wilcove, D. S., McLellan, C. H., and Dobson, A. P., 1986. "Habitat fragmentation in the temperate zone." pp. 273-256 in Conservation Biology. The Science of Scarcity and Diversity, M. E. Soule, ed. Sinauer Associates, Sunderland, Massachusetts.
Wilcox, B. A., 1980. "Insular ecology and conservation." pp. 95-117 in Conservation Biology an Evolutionary-Ecological Perspective, M. E. Soule and B. A. Wilcox, eds., Sinauer, Sunderland, Mas sachusetts.
Wilson, E. O., 1986. News and Comment, Science 234:14-15.
Woodwell, G. M. et al. , 1983. Global deforestation: contribution to atmospheric carbon dioxide. Science 222:1081-1086.
Woodwell, G. M., R. A. Houghton, T. A. Stone, R. F. Nelson, and W. Kovalick. 1987. Deforestation in the Tropics: New Measurements in the Amazon Basin Using Landsat and NOAA Advanced Very High Resolution Radiometer Imagery. J. of Geophysical Research 92(D2):2157-2163.