Soil Bid. Biochem. Vol. 29, No. 36, pp. 771-774, 1997 (r7 1997 Elsevier Science Ltd. All riehts reserved Pergamon

PII: SOO38-0717(%)00226 Printed in &eat Britain

0038-0717/97 $17.00 + 0.00

BIOLOGICAL NITROGEN FIXATION IN THE TROPICS: SOCIAL AND ECONOMIC CONTRIBUTIONS

JOHANNA DBEREINER EMBRAPA - Centro Nacional de Pesquisa de Agrobiologia (EMBRAPA-CNPAB), Km 47,

Seropdica, Itagua, 23851-970, RJ, BRAZIL

(Accepted 4 July 1996)

Summary-Brazil has become the world leader in replacing N fertilisers by biological NZ fixation (BNF). Even though agriculture in Brazil is one of the main export-producing activities, it has become the country which uses the lowest N applications (mean 10 kg ha-). Soybeans selected with zero N applications have become the countrys largest export product. Cereals, also selected with N supplies much below their need, unknowingly were selected for associations with diazo- trophic bacteria and can obtain up to 30% of their N from BNF when fertilised with ample PK and minor elements. The largest effect in this group was obtained with sugar cane, which can obtain up to 150 kg N ha- from BNF. This has become the key to the success of the Brazilian Bio-fuel programme where because of increases in the energy balance when the cane is planted without any N fertiliser and without buming the leaves, the energy produced is more than 5 times that required to produce it. Recently possibilities of expanding this biofuel programme to diesel oil are arising from the isolation of probably new diazotrophs which colonise the sterns of oil palms, of the Brazilian palm genotypes which usually are not fertilised with N. 0 1997 Elsevier Scieice Ltd

INTRODLKXION

Agriculture in the tropics might be expected to be more dependent of N fertilisers than that in temper- ate regions, because heavy rains and more rapid or- ganic matter decomposition lead to leaching and rapid losses of the N fertilisers applied. In addition to ground water pollution, other problems due to high N fertiliser applications occur in the tropics. High N fertiliser prices, especially in Brazil where they are not subsidised, can make agriculture unprofitable. Due to this, plant genotypes have been selected for high yields with low N fertiliser applications thus leading to crops which produce relatively high yields with low or no N fertilisers. Table 1 gives an overview of the fertiliser appli- cations in various representative countries in the World.

It can be seen that among al1 countries in the world, Brazil applies on average the lowest N fer- tiliser doses. This is not, however, due to fertile soils, because most soils in Brazil under agriculture are very N and P deficient as are most tropical soils. It also is not due to the large areas available for agriculture, but mainly to the high prices of these fertilisers here. Low fertilizer application has not only made agriculture more economically viable and competitive in the world, but it has also reduced groundwater pollution problems turn- ing Brazil into one of the least polluted countries in the world.

THE BRAZILIAN LEGUME CROPS

When soybeans were introduced into the country in the 196Os, adaptation and selection of genotypes of this crop were made with relatively high P, K and minor element applications, but with zero nitrogen added. Consequently Brazil and Argentina, where our soybean varieties are used, became the only countries in the world to obtain high yields of soybeans with absolutely no N applications. Mean yields in Brazil are now more than 2 t ha- and the total yield on the 12.5 million ha planted with this crop make up 25 million tons per year worth 5.7 billion US$. With an N content of 6%, most of which comes from N2 of the atmosphere, this is equivalent to 150 million t of nitrogen valued at 1.6 billion US$. Since nitrogen fertilisers have to be applied at rates twice the amount plants take up, the economy of the Brazilian soybean crop amounts to 3.2 billion US!! N fertiliser economised.

Beans (Phaseolus vulgaris) the main protein source of the Brazilian people, yield on average 600 kg ha- with inoculation used only in part of the country. Selection of more efficient and com- petitive Rhizobium strains and bean genotypes could increase annual yields to 1500 kg ha- with 4% N equivalent to 375 million US$ (Hungria and Neves, 1988).

Many other legume crops have potential in Brazil, such as peanuts, and forage legumes, which contribute to the economy and are partly respon-

771

772 Johanna Dbel-emer

Table 1. Fertiliser applicatmns m the World (kg ha-) (Solos e Adubos. 1988)

Country

Europe China USA Mexico India Brazil

NPK N

335 172 142 x7 105 58 63 36 46 27 51 10

NP

13 114 5. 7 h

7 0 IJ

sible for the low N fertiliser inputs in Brazihan agri- culture. The centra1 highland savannahs of Brazil. called Cerrados, have been partly transformed into highly productive agricultural areas with the cultiva- tion of soybeans, using inoculants specifically selected for these regions (Botelho (12 ul., 1988; Scotti et al., 1988). If the whole cerrado area (210 million ha) were planted with soybeans, enough protein could be produced to feed one third of the worlds population at U.S. standards (100 g protein dd).

New projects of reforestation with legume trees inoculated with species of selected rhizobia and mycorrhizal fungi, and grown before transplantion in to the field in bags containing 50% rock phos- phate, apply an additional very important technol- ogy to enrich eroded soils with N and organic matter (see paper of Franco CJI cl/.. this symposium).

NITROGEN FIXATION IN CEREALS AND FORAGE GRASSES

Since nitrogen fertilisers are not subsidised in Brazil, most genotypes, and subsequently commer- cial varieties, of cereals have been selected for high yields with N fertiliser rates much below their real requirement. This unconsciously favoured varieties which obtain part of their nitrogen from associ- ations with diazotrophs. Initially this was attributed to rhizosphere bacteria, but it seemed ditlcult to explain that as much as 20 to 4006 of plant N could be supplied by such associations. During the last decade, however, it became clear that such gen- otypes are colonised endophyticalfy by various dia- zotrophs (Dbereiner, 1992). some of which have not yet been identified. Herhaspirillum spp, Burkholderia spp and certain strains of Azospirillum spp have been shown to colonise roots, sterns and leaves of maize and, Pennisetum and even of Cs plants such as rite (Boddey et al., 1995; Baldani et al., this symposium) and wheat (Baldani et ul., 1986, 1987). Programmes of selection of maize and rite genotypes for high contributions of these diazo- trophs promise good chances of success (Blow and Dbereiner, 1975A; Toledo. unpub. Ph.D. thesis 1996).

ENERGY CROPS

Progress made in al1 areas of biomass energy has been much greater per unit expenditure than pro-

gress achieved in the pursuit of nuclear fusion (Rosillo-Calle et al., 1994). These authors suggest that if half of the money spent in the world on nuclear energy had been applied to the study of bio-energy alternatives, large amounts of renewable energy sources would have been developed.

The elimination of N fertilisers for bio-energy crops represents a major key to high energy bal- ances. The Brazilian ethanol programme is the best example of this (Dbereiner, 1994). Sugar cane grown in the country for centuries never received high N applications and therefore the genotypes developed and grown today obtain the highest BNF contributions among al1 non-legumes. When grown with ample P fertiliser and foliar applications of molybdenum. this crop can obtain up to 150 kg N ha- year from BNF (Urquiaga et al., 1992). Sugar cane is now planted on 4.2 million ha in Brazil, 8% of the land under agriculture. With mean yields of 64 t haar, in addition to sugar, 14.2 billion litres of ethanol are produced per year, equivalent to 260000 barrels of petrol per day. Four million cars run on 80% ethanol, and al1 gasoline sold in the country contains 20% absolute ethanol. Even though petrol prices over the world are cur- rently low. the government is convinced of the social and ecological impacts of the biofuel pro- gramme and plans to support it further. The key to this program is the high energy balance which is shown in Table 2.

Due to the high N contributions certain sugar cane genotypes receive from BNF, we are now recommending that farmers plant this crop without any N fertiliser and use the funds otherwise used for N fertiliser for increased phosphate applications, foliar spraying of Mo solutions (0.5 kg Mo ha-) and irrigation. Elimination of leaf burning before harvest, in addition. increases soil fertility and the soil cover from plant residues reduces needs for irri- gation. The higher labour need for cutting unburned sugar cane provides more jobs and costs are compensated by further increased yields (Oliveira er ai.. 1994). The Pro-alcool Programme has already created more than one million jobs and

Table 2. Energy balance of ethanol production from sugar cane m Brazil (Boddey, 1993)

Mean crop yields Mean ethanol ylelds Energy produced Ethanol Bagasse Total Energy expended Agriculture Factory Energy gain Overall energy balance Energy balance assuming all factory power derived from bagasse Eoergy balance assuming zero N fertiliser applicatlon

65 t ha- yr- 3564 L ha- yr-

18.747 Mcal 17.500 Mcal 36.297 Mcal

4.138 Mcal 10.814 Mcal 2 1.345 Mcal

2.43

4.53

5.79

Biological nitrogen fixation in the tropics 113

could be further increased, thus helping to solve one of the major problems of our country, the over-population in the main cities. Elimination of cane buming also wil1 further reduce air pollution; use of fuel alcohol has already reduced the lead content in the air of the large cities by 75%. Motor cars running on alcohol also emit 57% less CO, 64% Iess hydrocarbons and 13% less NO, than cars running on gasoline.

POSSIBILITIES FOR THE REPLACEMENT OF DIESEL FUEL

In addition to the alcohol programme, possibili- ties are being studied by Petrobras, the Brazilian state-owned oil company, to replace 20% of diesel oil by palm oil. Oil palms in Brazil are grown in the poorest North East area and in the Amazon regions, but Brazil is one of the tropical countries which produces relatively little palm oil, only 0.6% of world production. There are large areas available where oil palms could be planted, and it has been estimated that 18% of this already-deforested area could produce sufficient oil to replace al1 diesel oil used in the country (460000 barrels d-) (Oliveira, 1985; Boddey, 1993). Planting costs are relatively high, but harvest costs are very low; harvesting is by hand and mules are used for transport. These areas could be reforested with legume trees mixed with oil palms, and this would not only help to re- store the Amazon forest but also would create im- portant income for the poor Amazon population without causing environmental problems. Oil palms can be harvested continuously throughout the year, and the fuel can be processed by simple pressing. This leaves much less residue than other biofuel production systems. Among al1 liquid biofuel alternatives oil palms produce the highest amount of energy ha- followed by bio-ethanol from sugar cane (Table 3). Among the different oil-producing crops, the oil palms (Elaeis guianensis and Bactris gasipaes) produce the highest oil yields ha-], 10 times more than soybeans or rape (Table 4).

Investigations on growing palm trees without N fertilisers, showed insignificant responses to N ferti- liser by genotypes planted in Brazil (Chepate et al., 1988). This is apparently because oil palms have always been planted in the poor Northeast and Amazon regions where no N fertilisers are used. Based on this, we recently started to search for

Table 3. Energy yields of the most productive biofuel plants (Dbereiner et al., 1981)

Yields (t ha- yr-) Fuel L ha- yr- Energy TOE

Oil palm 35-38 3780 3.7 Sugar cane 60-90 4020 2.1 Mank 13-40 2340 1.2 Sweet sorghum 35-50 1925 1.0

Equal to 10 Kcal ha- yr-.

Table 4. Oil and energy yields of different oil crops (Purseglove, 1968)

Plant Growth period

(days) (t ha-) Oil Energy TOEa

Oil palm 365 4.0-8.4 3.1-7.8 Pejibaye 365 4.8 5.7 Coconut 165 1.5 1.8 Rape 150 0.7 0.8 Soybeans 120 0.6 0.7

Equal to 10 Kcal ha- yr-.

their source of N and found about 106 diazotrophs g- fresh weight of roots, sterns and leaves of both oil palms and pejibaye. Herbaspirillum, Azospirillum amazonense and some apparently new diazotrophs were isolated from these trees, and even the seeds of the palm trees contained the bacteria within the endosperm, suggesting how they are transmitted (unpublished data from our laboratory).

These findings fumish an excellent basis for exploring the use of palm oil for partial or even complete replacement of diesel oil. In the poor northern regions of Brazil, large areas of oil palms could be planted in mixed forests with legume trees and thus rejuvenate eroded deforested areas with good profits for the local populations. This practice could then be extended to other tropical countries, primarily Africa, where the cultivation of oil palms has recently been decreasing.

REFERENCES

Baldani V. L. D., Baldani J. 1. and Dbereiner J. (1986) Effect of inoculation of Azospiriflum spp on the nitrogen assimilation of field grown wheat. Plant and Soil 95, 109-121.

Baldani V. L. D., Baldani J. 1. and Dbereiner J. (1987) Inoculation of field grown wheat (Triticum aestivum) with Azospirillum spp in Brazil. Biology and Fertility oj Soils 4, 31-40.

Boddey R. M. (1993) Green energy from sugar cane. Chemistry and Industry (London) 11, 355-358.

Boddey R. M. (1995) Biological nitrogen fixation in sugar cane: A key to energetically viable biofuel production. Critical Reviews in Plant Sciences 4, 263-219.

Boddey R. M., de Olivares 0. C., Urquiaga S., Reis V. M., de Olivares F. L., Baldani V. L. D. and Dbereiner J. (1995) Biological nitrogen fixation associated with sugar cane and rite: Contributions and prospects for improvement. Plant and Soill74, 195-209.

Botelho G. R., Peixoto R. C., Ramos L. R. G. and Dbereiner J. (1988) Adantation of an efficient Bradyrhizobium japonikm strain for survival in cerrado soils. Anais da Academia Brasileira de Cincia 60, 379.

Biilow J. and Dbereiner J. (1975) Potential for nitrogen fixation in maize genotypes in Brazil. Proceedings of the National Academy of Sciences (USA) Vol. 12, pp. 2389-2393.

Chepate R. E., Valle R. R. and de Santana C. J. L. (1988) Resposta do dendezeiro a adubao mineral. Revista Brasileira de Cincia do Solo 22, 251-262.

Dbereiner J. (1992) Recent changes in concepts of plant bacteria interactions: Endophytic N2 fixing bacteria. Cinciu e Culturu 44, 310-313.

774 Johanna Dbereiner

Dbereiner J. (1994) Alternatives for biofuel production: Rosillo-Calle F., Hall D. O., Arora A. L. and Carioca A. The Brazilian example. PontiJitiae Academiae (1994) Bio-ethanol production: economie and social Cienciarum Scripta Varia (In press). considerations in failures and successes. In

Dbereiner J., Silva Araujo Neto J. and Arkcoll D. Biotechnology: Economie and Social Aspects. Issues of B. (1981) Energy alternatives from agriculture. Developing countries. (E. J. Da Silva, C. Radledge and Ponificiae Academiae Cienciarum Scripta Varia 46, 432- A. Sassen, Eds.), pp 23-53. Cambridge University 458. Press, Cambridge.

Hungria M. and Neves M. C. P. (1988) Interao entre cultivares de Phaseolus vulgaris and estirpes de Rhizobium na fixao e transporte de nitrognio. Pesquisa Agropecuaria Brasileira 21, 127-140.

Oliveira H. P. (1985) Utilizao do dende coma alimento e energetica. In Alternativas de Energia. (E. L. La River and M. Robert, Eds.), pp. 323-369.

Ohveira 0. C., Urquiaga S. and Boddey R. M. (1994) Burning cane: The long term effects. Inrernational Sugar Journol%, 2722275.

Scotti M. R., Sa N. M. H., Vargas M. A. T. and Dbereiner J. (1988) Avaliao da resistncia a tetraci- clina, kanamycina e cloramfenicol de estirpes de Rhkobium japonicum, Bradyrhizobium spp e Stylosanthes isolados de solos de cerrados. Revista de Microbiologia 19, 60-66.

Solos e Adubos Boletim Informativo (1988) So Paulo 21, 141.

Purseglove J. V. (1968) Tropical crops: Dicotyledons and Monocotyledons. Longmans, London.

Urquiaga S.. Cruz K. H. S. and Boddey R. M. (1992) Contribution of nitrogen fixation to sugar cane: Nitrogen- 15 and nitrogen balance estimates. Soil Science Society of America JournalS6, 105-114.