Contents - Previous - Next

This is the old United Nations University website. Visit the new site at

Part 1 : The ecological outlook

Ecological prospective for tropical Latin America
Rich and poor ecosystems of amazonia: an approach to management
Archaeological perspectives on the potential of Amazonia for intensive exploitation
Distribution and interannual variability of rainfall in Brazil


Ecological prospective for tropical Latin America

1 Introduction
2 The current condition of the tropical Latin American ecosystems
3 Modelling ecological changes
4 The reference scenario
5 The sustainable scenario
6 Conclusions


Gilberto C. GallopÍn and Manuel Winograd

1 Introduction

The Latin American tropics and subtropics contain many unique ecosystems of high biological diversity, many endemic species, and a great potential in terms of renewable natural resources. The ecosystems have evolved under climatic regimes of relatively low variability, characterized by high temperatures and precipitation, resulting in very complex and intricate ecological interrelationships. On the other hand, the ecosystems exhibit a high degree of fragility in face of human perturbations, particularly those associated with the indiscriminate application of modern technologies originating in the industrialized countries and generated under quite different climates and social settings.

However, indigenous pre-Hispanic cultures and civilizations had reached a rather sophisticated level of technology in the management of complex ecosystems, which proved to be sustainable over long spans of time. Different civilizations such as the Maya, the Inca, and the hydraulic cultures of Brazil, Bolivia, Colombia, Ecuador, and Mexico, maintained sustainable agricultural production for centuries, developing original and efficient solutions for managing the environmental resources without destroying the ecological base of production, at relatively high population densities and often including metropolises of considerable size (GallopÍn, 1985; Gligo and Morello, 1980; Vitale, 1983).

The fragility of the Latin American tropical ecosystems, therefore, must be considered in relation to the technology utilized, and not necessarily as an intrinsic attribute precluding any kind of human intervention.

It is clear, however, as will be discussed later, that the current trends in the region are characterized by very high and accelerated rates of ecological deterioration, expressed as deforestation, desertification, soil erosion and depletion, agricultural, industrial, and domestic pollution, accumulation of wastes, and increased vulnerability to catastrophic landslides, droughts, and floods (CEPAL/PNUMA, 1983a; Damascos et al., 1989; Doureojeanni, 1982; Sancholuz et al., 1989; Sunkel and Gligo, 1980). A large proportion of those changes occur in the tropical and subtropical areas of the region, where the advance of the agricultural frontier is most dynamic.

The problem lies not in the transformation or alteration of the natural ecosystems (transformations that in principle could be positive), but in the actual modality and results of these transformations, implying an accelerated degradation of the ecological basis of production, a veritable impoverishment and destruction of the region's renewable natural resources and vital ecological processes. It should be emphasized that many alterations, such as desertification and soil erosion, are irreversible in practical terms.

The destruction of the Latin American tropical ecosystems, and particularly of the tropical forests, is cause for serious concern. From the regional and local viewpoints, the destruction of forests, besides representing a waste of resources for development, has serious ecological impacts, generating micro- and mesoclimatic changes, through variations in the albedo and residence time of rainwater, increases in surface runoff, reductions in evapotranspiration, increases in maximum temperatures and daily thermal amplitudes, and reductions in precipitation (Salati et al., 1989), as well as soil erosion, floods, and other effects. From the global viewpoint, tropical deforestation is a significant contributor to the greenhouse effect, and might possibly affect the regulation of the planetary atmospheric circulation; tropical deforestation is also considered one of the major current causes of species extinction.

In the face of this situation, it is worth investigating the potential for a sustainable management of the major tropical ecosystems in

Latin America. While a number of studies are available around the world, addressing the issue of sustainability at the micro-level and examining alternative technical solutions for the sustainable use of natural resources by a particular human community, or for a given specific ecosystem, studies at the macro-level (regional or global) are very scarce. This paper presents the results of an investigation of sustainability for the whole of tropical Latin America, centred on an ecologically and technically feasible prospective scenario. This scenario is defined as an alternative to the ecologically degrading trajectory being followed in the region.

2 The current condition of the tropical Latin American ecosystems

Latin America is a region where tropical (including subtropical) ecosystems predominate. They cover about 85 per cent of the total surface area of the region, including dense forests (55 per cent), open forests and savannas (33 per cent), and deserts and semi-deserts (12 per cent).

By 1980, 23.5 per cent of the surface of the tropical ecosystems was exploited for ranching and 7 per cent for crop agriculture; 0.6 per cent was under urban uses and 0.3 per cent under plantations. The original (virgin or semi-virgin) ecosystems represented 46 per cent of the total tropical area; the altered ecosystems accounted for 20.5 per cent, and the wastelands (irreversibly decertified or degraded lands) amounted to 2 per cent (see table 2.1).

The tropical portion of Latin America is relatively well endowed with natural resources. It includes 46 per cent of the tropical forests of the world, containing an estimated minimum of 40 per cent of the tropical plant and animal species. It has important reserves of fresh water and minerals, and the highest untapped hydroelectric potential in the world. About 10 per cent of its lands are suited for intensive agriculture and ranching, and another 32 per cent are suitable for agroforestry, agro-silvo-pastoralism and ranching.

However, owing to the advance of the agricultural frontier, as well as the inappropriate management of pastures and agricultural lands associated with high rates of soil erosion and desertification, the degradation of productive lands, deforestation, and land reconversion are advancing at accelerating rates.

The contribution of Latin America to the emissions of carbon dioxide by 1988 has been estimated (GallopÍn et al., 1991) as roughly 14 per cent of the world total, of which about 2.9 per cent is due to the burning of fossil fuels and the rest is mainly from biomass burning associated with tropical deforestation. By contrast, the developed countries contribute more than 70 per cent of the total carbon emissions (Holdgate et al., 1989).

Table 2.1 Surface area (10³ Km³) under each category for the major life-zones of tropical Latin America in 1980 (percentages of total life-zone area appear in parentheses)

  Natural Agric-ultural Grazing Altered Plan-tations Urban-ized Waste-land Total
T & ST moist forests 5,795 583 683 1,023 20 17 3 8,124
(71.3) (7.2) (8.5) (12.5) (0.2) (0.2) (0.1)  
T & ST montane moist forests 158 178 473 390 6 41 5 1,251
(12.6) (14.2) (37.8) (31.2) (0.5) (3.2) (0.5)  
T & ST dry forests 1,068 377 1,612 1,557 21 20 92 4,747
(22.5) (7.9) (33.9) (32.8) (0.5) (0.5) (1.9)  
Tropical savannas 423 32 485 125 0 1 0 1,066
(39.7) (3.0) (45.5) (11.7) (0) (0.1) (0)  
T& ST mangrove forests and deltas 52 8 42 82 0 2 0 186
(28 0) (4.3) (22.6) (44.0) (0) (1 1) (0)  
Paramo and puna 173 23 422 253 0 6 45 922
(18.8) (2.5) (45.8) (27.4) (0) (0.5) (5.0)  
T & ST deserts and dry scrub 354 79 392 146 0 26 165 1,162
(30.5) (6.8) (33.7) (12.6) (0) (2.2) (14.2)  
Tropical Latin America 8,023 1,280 4,109 3,576 47 113 310 17,458
(46.0) (7.3) (23.5) (20.5) (0.3) (0.6) (1.8)  
Latin America 8,287 1,562 5,476 4,505 58 136 393 20,417
(40.5) (7.6) (26.8) (22.1) (0.3) (0.7) (2.0)  

Source: Winograd, 1989a.

Key: T & ST = tropical and subtropical.

Deforestation is certainly the most pressing ecological problem of the region in terms of land use and the loss of renewable natural resources. In the 1980s,¹ the destruction of tropical dense forests (including both natural and already altered forests) amounted to 4.34 million hectares per year, or 0.59 per cent per year, and the elimination of tropical open forests added another 1.37 million hectares per year.² If only natural (virgin and semivirgin) forests are considered, dense forests are being depleted at a rate of 0.63 per cent per year, and open forests at a rate of 1 per cent per year. Mangrove forests, deltas, and savannas are also diminishing fast. On the other hand, only 0.5 million hectares per year were reforested during the same period, implying an average reforestation to deforestation ratio of 1:11, ranging from 1:7 in the mountain and dry forests to 1:15 in the lowland moist forests (tables 2.2 and 2.3).

Deforestation in tropical Latin America is mainly caused by commercial and subsistence ranching, shifting agriculture, and land speculation. Shifting agriculture accounts for 35 per cent of the deforestation in tropical and subtropical moist forests, and 15 per cent in the tropical dry forests (FAO, 1981; Lanly, 1985; Winograd, 1989a). Peasant and shifting agriculture together generate more than 50 per cent of the agricultural products for final consumption. On the other hand, the expansion of pastures and commercial ranching exhibited sharp increases in the 1960s in Central America and in the 1980s in South America. In the Brazilian Amazon, for instance, this activity is responsible for about 80 per cent of the deforestation in its tropical moist forests. Despite the allocation of important subsidies to this activity, commercial ranching accounted for less than 0.1 per cent of the Brazilian gross internal product in 1981 and employed only 1 per cent of the labour force in the Amazonian agricultural sector (Browder, 1989).

Besides the waste of valuable natural resources, tropical deforestation is a major engine of species extinction. Recent calculations (Lugo, 1988), although lower than other previous figures, suggest that from 30,000 to 100,000 species could disappear irreversibly by the year 2000 in tropical Latin America.

In the dry forests and woodlands, deforestation is also affecting a growing number of people. It is estimated that in 1980 about 26 million persons were suffering from acute fuelwood deficits (Lanly, 1985; Lugo, 1987).

Table 2.2 Deforestation life-zone in tropical Latin America in the 1980s


(10³ Km²)


(10³ Km²)

Deforestation Annual rate


On natural On altered Total
(Km²/yr) (Km²/yr) (Km²/yr)
TF 5,588 752 30,350 4,000 34,350 0.54
STF 207 271 3,300 1,200 4,500 0.94
T&ST F 5,795 1,023 33,650 5,200 38,850 0.57
TSTMF 128 283 3,000 450 3,450 0.84
TLMF 30 108 850 290 1,140 0.83
T&ST MF 158 391 3,850 740 4,590 0.84
TDF 393 632 6,800 1,700 8,500 0.83
TVDF 496 311 2,380 535 2,915 0.36
STDF 179 614 1,885 450 2,335 0.29
T&ST DF 1,068 1,557 11,065 2,685 13,750 0.52
TS 423 125 1,280 200 1,480 0.27
D&M 52 82 450 150 600 0.45
Total 7,496 3,178 50,295 8,975 59,270 0.56

Source: Winograd, 1989a.

Key: TF = tropical moist forests
STF = subtropical moist forest
T&ST F = tropical and subtropical moist forests
TSTMF = tropical amd subtropical montane forests
TLMF = tropical lower montane moist forests
T&ST MF = tropical and subtropical montane moist forests
TDF = tropical dry forests
TVDF = tropical very dry forests and thorn woodlands
STDF = subtropical dry forests
T&ST DF = tropical and subtropical dry forest
TS = tropical savannas
D&M = tropical and subtropical deltas and mangrove forests

Soil erosion and desertification are also a problem in tropical Latin America. In 1980, about 226 million hectares of tropical and semiarid lands were suffering from desertification, affecting primarily productive lands (pastures and rainfed and irrigated croplands), and 50 to 75 per cent of the productive lands in the mountain areas were exposed to soil erosion (Doureojeanni, 1982; Masson, 1987; Winograd, 1989a).

Table 2.3 Forest losses and gains m tropical Latin America in the 1980s

  Forest area¹
(10³ Km²)
tion rate
tion (Km²)
reforestation (Km²)
T&ST F 6,818 0.57 38,850 2,600 1:15
T&ST MF 549 0.84 4,590 700 1:7
Total dense forests 7,367 0.59 43,440 3,300 1 :13
T&ST DF 2,625 0.52 13,750 2,050 1: 7
Total dense and open forests 9,992 0.57 57,190 5,350 1 :11
TS 548 0.23 1,480 0 -
D&M 134 0.45 600 0 -
Tropical Latin America 10,674 0.56 59,270 5,350 1:11

Source: Winograd, 1989a
Key: ¹ Natural + altered
T&ST F = tropical and subtropical moist forests
T&ST MF = tropical and subtropical montane moist forests
T&ST DF = tropical and subtropical dry forests
TS = tropical savannas
D&M = tropical and subtropical deltas and mangrove forests

The human population in the tropical and subtropical areas reached about 280 million (80 per cent of the total for Latin America) in 1980. Per capita agricultural land in 1980 amounted to 0.46 hectares per person. However, this average hides large disparities in the availability of agricultural land, as well as in the destiny of agricultural products. For instance, in the same year, agricultural land amounted to 1.17 ha/person in the tropical and subtropical moist forests, where a large part of the production is directed towards export. On the other hand, in densely populated zones such as the tropical and subtropical mountain moist forests, and the paramo and puna, per capita agricultural land reaches only an average of 0.19 and 0.14 ha/person, respectively. There, subsistence crops (corn, beans, potatoes, etc.) dominate (table 2.4).

In general terms, inappropriate modalities of land use in tropical Latin America translate into low grain production, insufficient har vests of roots and tubers, and the decline and even disappearance of traditional food crops; alternatively, the latter are produced to satisfy external demand (Winograd, 1989a).

Table 2.4 Current (1980) and anticipated (simulated for 2030) per capita availability of agricultural land in the Latin American tropics

  1980 2030
Reference Sustainable
ha/person PIL ha/person PIL ha/person PIL
T&ST F 1.17 L 0.73 L-I 0.96 I
T&ST MF 0.19 L-I 0.13 I 0.084 H
T&ST DF 0.88 L-I 0.46 I 1.1 I-H
TS 1.6 L 0.79 L-I 1.5 I
D&M 0.17 L 0.08 L-I 0.09 I
P&P 0.14 L 0.11 L-I 0.11 I
T&ST 0.11 L 0.06 L-I 0.08 I-H
Total 0.46 L-I 0.27 I 0.32 I-H

Source: Winograd, 1989a.

Key: PIL = predominant input level: L = low; I = intermediate; H = high
T&ST F = tropical and subtropical moist forests
T&ST MF = tropical and subtropical montane moist forest
T&ST DF = tropical and subtropical dry forests
TS = tropical savannas
D&M = tropical and subtropical deltas and mangrove forests
P&P = paramo and puna
T&ST DDS = tropical and subtropical deserts and dry scrub

In 1980, livestock amounted to 270 million animal units (AU) in tropical Latin America; that represents an average of 0.96 AU/person. However, the efficiency of ranching is very low, average meat production being about 45 kg/ha/year. Ranching is essentially extensive, with low animal loads: in ten-year-old pastures they may diminish to 0.2 AU/ha (Hecht et al., 1988).

3 Modelling ecological changes

For the purpose of exploring alternative ecological futures for tropical Latin America it was necessary to choose a land classification system capable of including the ecological characteristics of the region and its potential and limitations, and adaptable to the type and quality of the available information. The life-zone approach (Holdridge,1967) was considered appropriate, subdivided into specific categories (e.g. savannas, mangrove forests) according to the criterion of actual vegetation.

The spatial extent of the tropical zone was adjusted by taking into account the ecological and productive characteristics of the different areas (Brown and Lugo,1980; Winograd, 1989b); as a consequence, some ecological units exceed the geographical limits (23°27' South to 23°27' North) of the strict definition of the tropics. A total of twelve tropical and subtropical life-zones were identified for the region, aggregated into seven major zones for the purposes of the present paper (figs. 2.1 and 2.2).

Simulation models were implemented (GallopÍn and Gross, 1989; Winograd, 1989a) for each of the twelve zones. Each zone is modelled as a set of compartments representing different ecological categories or conditions and with different structural, functional, and productive characteristics. The following seven categories were defined:

¹ "Natural": virgin areas, and areas with past alteration but currently similar to the original ecosystems; ² "Altered": denotes a mosaic of patches of land under production coexisting with patches of original and secondary vegetation, and areas with slight to moderate soil erosion; ³ "Agricultural": annual, permanent, and nontraditional (i.e. coca, marijuana) crop areas, including fallow from permanent agriculture;³ (4) "Grazing": ranching areas in natural or artificial pastures; (5) "Plantations": reforested areas used for forestry and watershed protection; (6) "Wastelands": unproductive lands irreversibly transformed by extreme soil erosion and desertification (natural deserts are not included here); and (7) "Urban": urbanized areas (mainly cities).

Every year, land shifts from one category to others according to the intensity and nature of the human activities (defined by an assumed scenario) and of the natural processes occurring on it (fig. 2.3). Simulations span the period 1980-2030. A simple compartment model was used. Each compartment represents the surface of a land category, for each life-zone, and it changes according to the following equation:

where S = surface of a given land category (Km²); Inflows = surface of land of other categories converted into the considered category in a given year (Km²/year); 0utflows = surface of land of the considered category converted into other categories (including itself) in a given year (Km²/year); Smax = maximum potential surface of the category (Km²); 1= set of all land categories. The scenario yearly defines the process generating the transformations (human activities or natural regeneration; see fig. 2.3) for each category and life-zone, specifying the portion of the category affected by the activity and the rates of conversion to other categories. The scenario is exogenously defined, taking into account the current situation, the assumed rate of growth of the activity, and the availability of land. Models were run under both a reference and a sustainable development scenario (see next section).

Figure 2.1 The major tropical and subtropical life-zones of South America. T&ST F = tropical and subtropical moist forests; T&ST MF = tropical and subtropical montane moist forests; T&ST DF = tropical and subtropical dry forests; TS = tropical savannas; D&M = tropical and subtropical deltas and mangrove forests; P&P = paramo and puna; T&ST DDS = tropical and subtropical deserts and dry scrub.

Figure 2.2 The major tropical and subtropical life-zones of Central America and Mexico. T&ST F = tropical and subtropical moist forests; T&ST MF = tropical and subtropical montane moist forests; T&ST DF = tropical and subtropical dry forests; D&M = tropical and subtropical deltas and mangrove forests; T&ST DDS = tropical and subtropical deserts and dry scrub.

Figure 2.3 Potential transitions of land between ecological categories, simulated for each major life-zone of Latin America Boxes indicate land categories; circles indicate processes generating the transitions; arrows denote the transitions. The thick circle operates in the sustainable scenario only. SA = shifting agriculture; PA = permanent agriculture; FE = forest exploitation; EA = extraction activities; RA = ranching; PL = plantation; RC = reconversion; NR = natural regeneration; RS = restoration; UR = urbanisation.

While the simulation models, in their present state, do not calculate production, but only the surfaces of land under different categories and production systems, the estimates of production are based upon the expected improvements in agricultural yields (compatible with the historic changes within the region and probably underestimating the future increases).


Contents - Previous - Next