Contents - Previous - Next

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

Pasture restoration

Given that cattle pastures are likely to remain an important fixture of the agricultural economy of Amazonia, technologies and management practices are needed to improve their productivity. Improvement of existing pastures and livestock, rather than trying to "re-educate" people not to eat beef, is also likely to be more fruitful in Central America (Nations and Komer 1983). With proper management, artificial pastures in Amazonia can be reasonably productive and sustainable (Falesi 1976; Falesi, Baena, and Dutra 1980).

Periodic upgrading with a range of technologies, such as improved germ plasm, is a key part of pasture management in Amazonia. The trend towards improved management of pastures in the region strengthened considerably in the 1980s (Serrão 1989), and remained strong in the early 1990s.

The move to restore pastures has been spurred by a number of factors. First, more technologies, particularly improved pasture grasses, are available than in the past. More productive, pest-resistant grasses have been developed through collaborative research between CIAT (Centro Internacional de Agricultura Tropical) in Cali, Colombia, and national agricultural research systems, such as Brazil's EMBRAPA (Empresa Brasileira de Pesquisa Agropecuária). Second, with a slow-down in road construction in the 1980s, less "virgin" land was available for opening new pastures. Third, land prices have been rising in many parts of Amazonia because of the slow-down in road construction and population increases (Serrão 1989).

How quickly a pasture degrades depends mainly on its management history and to a lesser extent on inherent soil fertility. Some first-generation pastures of guinea grass on alfisols (terra roxa) derived from weathered basalt in the Altamira area of the Transamazon Highway are still reasonably productive after 20 years without fertilizer, provided that they are periodically weeded and fenced to allow rotation. Some guinea grass pastures on relatively infertile ultisols and inceptisols near Maraca, Roraima, were still in good shape after a decade of grazing; pastures tend to degrade when stocking rates exceed 1 head/ha (Eden, McGregor, and Vieira 1990). A guinea grass pasture at Fazenda Piquia, some 60 km south of Paragominas, has been kept in production for 20 years in spite of the acid, infertile sandy loam on which it has formed. Fazenda Piquia has kept its guinea grass pasture going with periodic weedings and fertilization with rock phosphate. Transported from central Brazil, rock phosphate is broadcast at the rate of 35-50 kg P2O5/ha.

On the other hand, a guinea grass pasture approximately two decades old at km 30 of the Marabá-Altamira stretch of the Transamazon Highway was in extremely poor shape during the 1992 dry season; overstocking and poor weed management had contributed to compaction and in some cases erosion of the relatively infertile oxisol. A regularly weeded and fenced P. maximum pasture on yellow oxisol near Paragominas was replaced with braquiarao after 12 years. Even with relatively careful management, the rancher considered the inputs required to keep the guinea grass pasture going would be better spent on "upgrading" to a newer grass, after mechanically loosening the soil and fertilizing with phosphorus.

At the São Judas Tadeu ranch 19 km from São Miguel do Guamá in the Bragantina zone, a 21-year-old pasture of Brachiaria decumbens is still grazed. This pasture was cleared in second growth and has never been fertilized, even though the soil is an oxisol. It is still productive because it is fenced, rainfall is plentiful in the 2,5003,000 mm range, and the land has been weeded periodically.

The cost of upgrading pastures varies according to the degree of treatment and the location (table 7.2). At the high end, pasture restoration involves bulldozing weeds and logs into windrows (fig. 7.5), discing the cleared soil, fertilizing with phosphorus, sowing improved pasture grasses, erecting fences so that pastures can be rotated, and, in some cases, intercropping with forage legumes. In the mid-1980s, an investment of approximately US$150-200/ha was necessary to rehabilitate degraded pastures in areas with good road connections to major markets, such as around Paragominas (Serrão and Toledo 1988; Uhl and Vieira 1989). By the early 1990s, the cost of restoring pastures with the "full treatment" in eastern Pará, had climbed to the US$230350/ha range. In the vicinity of Manaus, increased costs for fertilizer, manual labour, and equipment rental drive up the price for upgrading pastures to the US$350-500/ha range. In Acre, restoration costs average some US$400/ha, but in one case reached as high as US$800/ha.

Table 7.2 Costs of restoring degraded pasture en upland areas of the Brazilian Amazon, 1991-1992

Location Cost (US$/ha) Year Observations
Nr. Manaus, Amazonas 460 1991 Mechanized, P fertilizer
and intercropping with
Desmodium ovalifoliuma
Fazenda Vitoria, 350 1992 Mechanized, P fertilizer
Paragominas, Pará,     and fencing
Fazenda Boi Branco,

Paragominas, Pará,

300 1991 Mechanized, P fertilizer
(rock and super
phosphate) at 200 kg/ha
Fazenda Paraiso, Rio 234 1991 Mechanized, P fertilizer
Gurupi, Pará,     and fencing.
Fazenda Modelo, km 12 PA 116 1992 Mechanized, fencing, but
150 Morada Nova,     no fertilizer
Marabá, Pará,      

a. This herbaceous legume was originally introduced to the Manaus area in 1978 by EMBRAPA as a ground cover for experimental jacarandá (Dalbergia sp.) plantations. b. For a breakdown of cost per item see table 7.3.

Larger ranches may employ aircraft to accelerate seeding of their pastures. One ranch near Rio Branco, Fazenda Cipoal, for example, has employed aircraft to seed pastures with B. brizantha, a practice that is economically feasible only with target areas in excess of 50 ha. Private companies are the main source of seed for ranchers and farmers in Amazonia. General farming supply stores, such as Agroboi in Altamira, stock seed of several pasture grasses.

The price differential between rehabilitating existing pastures and clearing forest is important when considering land-use changes in Amazonia. In the case of the Rio Branco area of Acre, the cost of felling forest was approximately US$120/ha in November 1991. The higher the cost of pasture restoration, the more likely it is that ranchers will open forest. Research on ways to reduce the cost of pasture improvement could thus help deflect the axe from forest stands.

In spite of the relatively high cost, some 700,000 ha of artificial pasture had been rehabilitated in the Brazilian Amazon by the late 1980s, reaching approximately 1.5 million ha by 1993 (Serrão and Homma 1989, 1993). The income generated from selling timber in the Paragominas area has helped ranchers restore their pastures. Indeed, the financial incentives to clear forest for pasture have decreased in areas with sawmills and a relatively good road system; in such places, the wood is worth more than the ash fertilizer from forest burns (Buschbacher, Uhl, and Serrão 1987).

Encouragingly, the trend to upgrading pastures is widespread in the Brazilian Amazon. Artificial pastures in Amazonia are being upgraded with a variety of techniques. Fertilizers, particularly phosphorus, are usually essential. More productive and pest-resistant grasses are usually deployed, and legumes, such as Calopogonium mucunoides, Pueraria phaseoloides, and species of Centrosema, are sometimes interplanted for improved ground cover and to fix nitrogen. If the pasture is not bulldozed, weeds are cut and burned, but in heavily compacted sites the soil must be mechanically disced or raked to restore productivity (Serrão 1986a). As in Central America, fencing and rotation of pastures planted in former rain-forest areas can substantially increase the yield and sustainability of cattle operations (Parsons 1989).

The amount of money that ranchers and farmers invest in upgrading their pastures depends on the degree of degradation and available resources. At the lower end, small-scale farmers, such as colonists along the Transamazon Highway with 100 ha lots, simply cut and burn weeds in pasture and seed with braquiarão. Larger operators often opt for hiring a bulldozer, and purchase fertilizer and barbed wire for fences.

Table 7.3 Breakdown of the costs of restoring pasture on two ranches in the Brazilian Amazon

Item Observations Ranch A
Ranch B
Bulldozing debris into
Discing with tractor
Ranch A bulldozed in rainy season when rental rates are lower
Two discings are typically needed
at 1 hater




Phosphate fertilizer
Brwantão seed
Seed and sowing costs; ranch A
application rate was 16 kg/ha,
ranch B application rate was 12




Clean-up and burning
Moving large logs, firing windrows
Enclosing 50 ha plots at cost of
US$1,400 per km of fence



TOTALS   234 116

A: Fazenda Paraiso, Rio Gurupi, Pará, 1991

B: Fazenda Modelo, km 12, PA 150 Morada Nova-Belém, 1992

Phosphorus deficiency is common in Amazonian soils and managing P levels in soils is essential to maintaining the productivity of pastures. When available phosphorus declines in soils, weeds usually proliferate (Falesi 1992). Although Brazil is virtually self-sufficient in phosphate, costs of the mineral fertilizer remain high in the Brazilian Amazon. Phosphate fertilizer is one of the major items in the cost of upgrading pastures (table 7.3). In the vicinity of Belém which has among the cheapest prices for fertilizer in Amazonia, superphosphate fertilizer cost US$180/ton in May 1990. Simple superphosphate is usually applied to pastures at the rate of 1 ton per 5 ha. Large distances between factories in central and southern Brazil elevate the cost of fertilizers for ranchers and farmers in the North region. In spite of these high costs, however, ranchers increasingly employ fertilizers to upgrade the productivity of their pastures.

Earlier predictions that ranchers in the Brazilian Amazon would eschew phosphate fertilizers because of their high cost (Fearnside 1980b) have not come to pass. Between a fifth and a quarter of the firstcycle pastures in the Brazilian Amazon have been upgraded, many with phosphate fertilizer. To prolong the effect of phosphate fertilizer, some ranchers use a combination of superphosphate and rock phosphate. Substantial deposits of phosphate at Maicuru near Santarém estimated at 100 million tons, may eventually reduce the cost of this important nutrient for ranchers and farmers. However, these phosphate deposits are deep and in a thin band, which will drive up extraction costs.

The restoration treadmill

Restored pastures provide an immediate productivity boost. How much time elapses between each major upgrading of a pasture depends on a host of factors, including the state of finances of the ranch and grazing pressure. Weeds are usually cut and burned every one to four years, but some five to ten years typically elapse before fertilizers need to be applied.

Upgraded pastures in the Paragominas area appear to be profitable, even though they often require fertilizer every five years (Mattos, Uhl, and Gonçalves in press; Nepstad, Uhl, and Serrão 1991). On Fazenda Sinuelo at km 23 of the Rio Branco-Porto Velho Highway, a sevenyear-old B. brizantha pasture has been weeded only once, shortly after the grass seed germinated, and has not needed any fertilizer. On infertile oxisols in the Paragominas area, several ranches have not had to fertilize their brizantão pastures after five years.

With adequate care, pastures of B. brizantha are reasonably productive. Well-managed brizantão pastures on moderately fertile ultisols near Rio Branco, Acre, permit cattle to reach 500 kg within three years. At Fazenda Boi Branco near Paragominas, superior Nelore stock can attain 350 kg within 18 months when grazing on wellmanaged brizantão pasture and supplemented by chopped elephant grass mixed with mineral nutrients (fig. 7.6). At Fazenda Paraiso on the Rio Gurupi, Pará, cattle typically gain 144 kg per year on brizantão pastures without supplemental feed.

Restoration of pastures is not a one-shot affair. Inevitably, rehabilitated pastures require more fertilization as well as new forage species or varieties. For all its favourable attributes, for example, brizantão is no panacea for improving pastures in Amazonia. When planted on poor soils, for example, productivity is likely to fall off without periodic fertilization with phosphorus. Furthermore, Brachiaria brizantha may be allelopathic, thus diminishing the chances of successful interplanting with forage legumes.

Another factor mitigating against a long tenure for Brachiaria brizantha var. Marandu in many areas of the Amazon basin is that pest populations are likely to change. For example, one small patch of Brachiaria brizantha pasture on Fazenda Rio Branco near Ariquemes, Rondônia. appeared to be suffering from spittlebugs in February 1992. If such proves to be the case, then spittlebugs have evolved a new race, thus putting at risk the large areas planted to B. brizantha. As in the case of silviculture and field crops, pasture ecosystems in Amazonia are highly dynamic, thus requiring a continued research effort to maintain their long-term productivity.

Scale and pace of rehabilitation

Small- to medium-scale ranchers appear to be most actively upgrading artificial pastures throughout Amazonia. In 1991, ranches with restored pastures were observed along the Manaus-Itacoatiara road; the Transamazon Highway from the Altamira area to km 80 of the Altamira-Itaituba stretch; the highway linking Tomé-Açu and Paragominas (PA 256); the Belém-Brasília Highway from Paragominas to Belém the road from the Belém-Brasília Highway to Capitão Poço, Pará, and the Rio Branco-Porto Velho Highway in Acre. In 1993 and 1994, rehabilitated pastures were observed along BR 364 and associated side-roads in Rondônia. and along the Obidos-Oriximina Highway. At least a quarter of the ranches observed along the above roads had all or some of their pastures under improved management. Ranchers in northern Mato Grosso are also actively upgrading their pastures.

How fast the restoration of pastures in Amazonia proceeds hinges largely on market signals and fiscal policies, since technologies for pasture improvement are no constraint. Relatively low market prices for beef in 1989 and 1990 discouraged investment in pasture improvement. In addition, hyperinflation in Brazil, which reached 60 per cent a month by February 1990, dampened the investment needed to improve the productive capacity of businesses and industry.

Prior to Collor de Melo's presidency in March 1990, the federal government was offering financial instruments (Letras Financeiras do Tesouro) that paid rates of return greatly in excess of inflation to help finance the burgeoning deficit. Working with commercial banks, investors with as little as US$100 were obtaining an unsecured rate of return of about 3 per cent every 24 hours in February 1990 through this financial instrument. Over US$50 billion was tied up in "overnight" deposits in early 1990, funds that could be used more productively to improve pastures and other enterprises. Fiscal measures were taken after March 1990 to dampen inflation and eliminate the "overnight" option, but investors switched to 30-day notes.

Inflation dropped dramatically with the introduction of a new currency, the real, in July 1994. The architect of the piano real, Fernando Henrique Cardoso, was elected President of Brazil in October 1994, and the investment climate will likely improve for upgrading pastures and other agricultural enterprises.

Modest fiscal incentives are available for pasture improvement from the Amazon Regional Development Bank (BASA - Banco da Amazonia, S.A.). BASA taps funds from the Fundo Constitucional de Financiamentos do Norte (FNO) for this purpose. To qualify, ranchers must intercrop pastures with nitrogen-fixing Pueraria in an effort to improve their sustainability. Few ranchers appear to have taken advantage of BASA loans for this purpose, partly because Pueraria does not withstand much trampling by cattle. Stocking rates have to be carefully controlled when Pueraria is intercropped, and not all ranches have established expensive fencing to rotate pastures carefully.

Further research, fresh germ plasm, and new management techniques will be required to keep pastures sustainable over the long term. As in the case of agro-forestry systems, a better inventory of existing practices by farmers and ranchers might uncover promising avenues for further development. The owner of Fazenda Baixa Verde, located at km 46 of the Rio Branco-Porto Velho stretch of BR 364, is restoring 2,500 ha of pasture in stages using maize as an intercrop to reduce costs. At Fazenda Rio Doce, 46 km east of Belém on the BR 316 highway, Brazil nut (Bertholletia excelsa) shells from factories in Belém are recycled on to pastures after they have been mulched for several months with cattle manure (fig. 7.7). Other ranchers in eastern Amazonia are experimenting with grain crops as a phase in pasture rehabilitation (Veiga 1986).

Intercropping pasture with perennials

The notion of planting perennial trees in pastures, or along their margins, is attractive on many counts. First, it would diversify income for ranchers. Second, a more ecologically diverse landscape would ensue, with possible benefits for some wildlife. Third, nutrients and water would be sequestered from soil layers out of reach of grasses, thereby improving the productivity of the land and increasing evapotranspiration, particularly in the dry season. Fourth, depending on the species employed, soils could be replenished with nitrogen. Fifth, cattle would benefit from shade, and would concentrate dung around the trees, thereby fertilizing them.

Although the benefits of intercropping pastures with fruit, nut, or timber trees, and of encircling them with living fences, are well recognized, few ranches have implemented such practices. Annual crops are often planted before a pasture is established, or during the initial stages of restoration, but they are phased out as the grass becomes established. Jari discontinued its pasture/Caribbean pine operation to concentrate on pulp production.

Ranchers are reluctant to try intercropping pasture with perennials because of management concerns. Management intensity would necessarily increase as the operation diversifies. Most ranchers lack the knowledge base to work with an assemblage of perennials. The cost of fencing off tree seedlings so that they are not damaged by cattle is a concern. Fire, a common management tool to reduce weeds, would also damage tree seedlings.

Small farms are likely to be the first successful testing ground for intercropping pastures with trees and bushes. Small farmers often have experience with both livestock and a mixture of crops. Land tenure is an important dimension here: only if the farmers own the land are they likely to invest in perennials and other long-term conservation measures (Vosti 1991).

For the most part, ranchers in the Brazilian Amazon are not currently interested in agro-forestry systems that incorporate pasture grasses with perennial crops. A few medium-scale ranchers intercrop rubber with pasture, such as in the vicinity of Tomé-Açu but ranchers generally prefer to concentrate on pasture or pasture-legume intercrops. At Fazenda Aruanã near Itacoatiara an attempt was made to intercrop Brazil nut trees with pasture: Brazil nut trees were initially planted at 20 metre intervals, but the pasture was already degraded when the Brazil nut seedlings were planted and the effort to combine beef and Brazil nut production on the same land was discontinued because of weed problems. A decision was made to double the planting density of Brazil nut trees instead of trying to combine them with cattle grazing. The rancher is still in the business of producing beef and Brazil nuts, but on different parcels of his property. Approximately 4,000 ha of Brazil nut trees have been planted at Fazenda Aruanã at 10 metre intervals. Most of the Brazil nut trees have been grafted in an effort to ensure higher yields, but 12year-old trees are still not producing adequately, possibly because of pollination or soil fertility problems.

For larger ranches, the first "breakthrough" in diversifying the pasture landscape is likely to come with living fences. Fence posts have to be replaced every four to five years, mainly because of termite damage. If suitable species can be found that do not require expensive labour for pruning, ranchers are likely to adopt living fences, as in Costa Rica, where at least 57 plants are employed as living fences (Sauer 1979).

Most of the living fence species in Costa Rica are native to Central America and were already adapted to agricultural conditions since they are also volunteers in open sites. The practice of using living fences pre-dates the Spanish Conquest in Costa Rica. Suitable candidates for living fences in Amazonia might come from the native second-growth flora, from indigenous groups, or from other tropical regions.

On-farm and on-ranch experiments are ideal ways to introduce new technologies. In the vicinity of Paragominas, the Brazilian agricultural research service (EMBRAPA) and Woods Hole Research Center in Massachusetts are collaborating with a rancher to test the suitability of rehabilitating pastures with various intercropped fruit trees.

Pasture intercropping raises a host of research issues. Allelopathic interactions can thwart some intercrops. Perennials are long-term investments, but market conditions can shift unpredictably. For example, low latex prices have forced the owner of Fazenda Rio Branco near Ariquemes, Rondônia. to cut down rubber intercropped on 750 ha of his 2,000 ha of pasture. The mix of perennials suitable for any given area will depend on soil conditions, allelopathic behaviour, rainfall patterns, and marketing infrastructure, among other factors.

Intercropping with leguminous ground cover

Although a number of ranchers introduce some herbaceous legumes to their pastures, particularly when they are being upgraded, this practice is not as well developed as it could be. The advantages of intercropping pasture with a leguminous ground cover are well known increasing soil nitrogen levels, provision of a nutritious forage, and, in areas with a pronounced dry season, increased food for cattle - but few artificial pastures in Amazonia are systematically planted with legumes. Less than 5 per cent of the sown pastures in Amazonia include planted herbaceous legumes (Serrão 1986a).

Two main reasons appear to account for the slow adoption of intercropped legumes in pasture. First, the introduction of a leguminous ground cover implies new management practices. Weeds are frequently suppressed in pastures by fire, so legumes can be destroyed when pastures are "cleaned." Another management issue with forage legumes is that they can become weedy and interfere with crops. Pueraria and mucuna preta (Mucuna aterrima) can spread quickly; Pueraria is known as the notorious kudzu vine in Florida and Georgia. Also, while legumes are becoming established, careful attention to grazing pressure is usually needed. Second, farmers may not have used the appropriate germ plasm for their soils and climate.

The push to experiment with leguminous ground cover in Amazonian pastures is nevertheless well under way. In addition to trials conducted by EMBRAPA in collaboration with CIAT, many ranchers are experimenting with legumes on their own. The manager of Sementes Garanhão in Marabá, for example, reports an increasing demand for seeds of Pueraria, mucuna preta, and soja perene (Glycine wightii) from local ranchers (Borba, pers. comm.). Sementes Garanhão has seed production facilities in Mato Grosso do Sul and also buys seed from other producers in the south-central state. The owner of Fazenda Nova Aurora on the outskirts of Quatro Bocas further north in Pará, has established experimental plots intercropping brachiarão with soja perene and mucuna preta.

Research carried out by EMBRAPA's cerrado research centre (CPAC - Centro de Pesquisa Agropecuaria do Cerrado) in conjunction with CIAT scientists has identified a number of different strategies for improving pastures, some of which may soon pay off for ranchers in Amazonia. For drier areas of Amazonia, Stylosanthes guianensis holds promise because it withstands drought well and provides valuable forage when pasture grasses are suffering from moisture stress. This waxy-leafed legume is a native of the cerrado and parts of the Amazon; it is at home in areas with long dry seasons and is adapted to poor, acid soils. Poor seed production will need to be overcome and trials in various parts of Amazonia will have to be carried out before S. guianensis can be recommended for planting in the North region. Another promising candidate for Amazonian pastures is Arachis pinto). This relative of the groundnut quickly establishes a lush green carpet, and is thus particularly appropriate for bunch grasses, such as Panicum maximum. A further advantage of the legume is that it produces abundant seed, tolerates drought and shade, and is eaten eagerly by cattle (J. Spain, pers. comm.).


In upland areas, extensive campos in Roraima and Amapá and smaller patches of native grassland in Rondônia. Amazonas, Pará, and Amapá cover at least 20 million ha and have supported lowdensity cattle populations for centuries (Eden 1990: 3; Moura 1989: 92; Pires-O'Brien 1992; Serrão 1986b). The impressive extent of savannas in some areas is due in part to human activities, both at present and in remote times. Edaphic factors, such as poor drainage caused by a clay hardpan near the surface, create a core of natural grassland much of which has expanded after burning.

In the distant past, hunters and gatherers ignited savannas to facilitate hunting, and some indigenous groups still follow this practice (Smith in press). Cattle were first introduced to the savannas of Roraima in 1787, and by 1930 some 300,000 head of cattle roamed the annually torched grassland (Rivière 1972). Ranchers ignite the campos in the dry season to promote more succulent and nutritious growth when the rains come. Only fire-hardy tree species, such as Curatella americana, survive in grasslands that are annually torched. The carrying capacity of such rangeland is low, but could be improved by new management techniques, such as interplanting with more nutritious forage species (Braun and Ramos 1959). Flood-plain pastures along the Amazon are much more productive, but they are only seasonally available. Some 50 million ha of savanna and floodplain meadows are currently underutilized by livestock (Serrão 1989), an area two-and-a-half times greater than planted pastures in deforested areas of the uplands. Stocking rates on "natural" or non-planted grasslands could increase from the current 6 million head of cattle and water buffalo to at least 30 million without cutting down any trees (Serrão 1990). The potential and pitfalls of ranching on the flood plain are explored in more detail in the following chapter.

Contents - Previous - Next