Cover Image
close this bookThe Fragile Tropics of Latin America: Sustainable Management of Changing Environments (UNU, 1995)
close this folderPart 4 : The semi-arid north-east
close this folderChanging aspects of drought-deciduous vegetation in the semiarid region of north-east Brazil
View the document(introductory text...)
View the document1 Introduction
View the document2 Study sites and methods
View the document3 Results
View the document4 Discussion and conclusion
View the documentReferences

(introductory text...)

1 Introduction
2 Study sites and methods
3 Results
4 Discussion and conclusion


Ichiroku Hayashi

1 Introduction

White sand soils are amongst the most infertile of the generally infertile tropical soils. They are composed mainly of quartz sand, and support a distinctive vegetation. This, varying from open savanna to closed forest, is characterized, by pronounced sclerophylly, low diversity, and high endemism. Forests on white sand soils are reported from several humid tropical regions of the world and are variously designated: Amazon caatinga or campinarana in Amazonia (Anderson, 1981); wallaba forest or muri-bush in Guyana, and heath forest, kerangas or padang, in Borneo.

As shown in figure 11.1, extensive areas of white sand soils and related vegetation (white sand formations) are present in Amazonia, occurring in the Rio Negro basin, in Serra do Cachimbo on the Pará-Mato Grosso boundary, on the Chapada dos Parecis in Rondônia and along the Atlantic coast near the mouth of the Amazon, as well as in Maranhão. Small patches are present in many other parts of Amazonia.

Amazonian white sand soils occur under diverse geological and geomorphological conditions. They are found on the low uplands (called terra firme), mainly composed of arenaceous sediments; on natural levees in a floodplain (várzea); sand ridges or dunes in a coastal lowland (restinga); and on plateaux (chapadas) of Cretaceous sandstone, or on hill areas of granitic rocks (Whitmore and Prance, 1987).

Figure 11.1 Distribution of the on the white sand (white sand formation) in Brazilian Arnazonia. (After Whitmore and Prance, 1987.)

White sand soils also cover a considerable area of North-East Brazil, or the Nordeste. The objective of this paper is: (1) to clarify the distribution of white sand soils in North-east Brazil;(2) to examine their characteristics and genetic processes; and (3) to consider the influence of deforestation on their formation.

2 Study sites and methods

The study sites are situated between 7° and 8°S and between 32° and 38°E in the Brazilian North-East (fig. 12.1). Mean annual temperature and total annual precipitation are 20°C and 800-1,000mm, with a severe dry season from October to the following February (Nishizawa, 1976; Rizzini and Pinto, 1964). The soils are classified as solodized solonetz and regosols Ministério da Agricultura e Ministério do Interior, 1971).

Figure 12.1 The study area in North-East Brazil. 1. João Pessoa; 2. Campina Grander 3. Patos.

The vegetation varies from place to place, according to land use, which includes ranching, cultivation, firewood collection, and charcoal Production.

The intensive study sites were located in the vicinity of Campina Grande and Patos, where the vegetation is typical for the region. Nine 10m x 10m quadrats were set out at sites dominated by Mimosa tenuiflora (synonym of Mimosa hostilis) and Caesalpinia pyramidalis. In each quadrat, I measured the diameter of tree stems at 130 cm high (DBH) for each species, and counted the number of shrubs less than 150 cm tall for each species.

For selected specimens of Caesalpinia pyramidalis, Mimosa tenuiflora and Aspidosperma pyrifolium, the weight of stems (Ws:kg), branches (Wb:kg), and leaves (Wl:kg) was taken after first measuring tree height (H:m) and stem diameter (D:cm) at 130 cm high.

The nitrogen and carbon contents of stems and leaves and of the surface soil were determined in the laboratory.

3 Results

The floristic composition of the caatinga in Campina Grande is shown in table 12.1. Mimosa tenuiflora and Caesalpinia pyramidalis predominate, with Pithecellobium foliorosum and Aspidosperma pyrifolium also present. In the shrub layer, Croton sonderiana, Sida cordifolia, and Croton campestris were the dominant species. Observation indicated that the bare sites produced by disturbance were covered first by Sida cordifolia and Croton campestris. These were then replaced by Mimosa tenuiflora and Caesalpinia pyramidalis. The final stage of succession of this area appears to be Aspidosperma pyrifolium, be cause the trees are of big stature, with erect stems, emerging above the other tree species.

Table 12.1 Floristic composition of the caatinga at the vicinity of Campina Grande (quadrat area 10 m x 10 m; 1st and 2nd layers above and less than 1 m in height)

Species 1st layer 2nd layer
Mean DBH (cm) No. of trunks No. of individuals
Caesalpinia pyramidalis 4.4 12 7
Pithecellobium foliorosum 4.2 4 0
Mimosa hostilis 4.0 28 3
Croton sonderiana 3.3 4 13
Aspidosperma pyrifolium 2.7 4 3
Encholrium spectabile - - 8
Spondias tuberosum - - 1
Croton species - - 1
Pilosocereus peutedrophorus - - 1
Unidentified species 1 - - 2
Unidentified species 2 - - 2

Table 12.2 Quandiative caracteristics of caatinga stand at the experimental sta tion,

Campina Grade  
Number of species per 100 m²  
Tree layer 7
Herb layer 13
Number of trees per 100 m² 39
Mean DBH (cm) 4
Above-ground biomass (kg per 100 m²)  
Trunks and branches 246
Leaves 29
Total 275
Herbs 1 9
Gross total 294

The Caesalpinia pyramidalis stand included 11 species and 52 individuals per 100 m². The height of the tree layer was about 5 m. The density and mean DBH of the trees were 39-52 individuals per 100 m² and 3.7-4 cm. The aboveground phytomass of the stand was 294 kg per 100 m², including 246 kg of trunks and branches, 29 kg of leaves and 19 kg of herbs and grasses (table 12.2).

The relationship between mean DBH and number of trees per 100 m² (tree density) is shown in figure 12.2, which was obtained by the survey of 9 stands of caatinga. The relationship shown in this figure is described as follows:

N= 217 exp ( - 0.42 D) (1)

where N and D are tree density and mean DBH. This equation suggests that the number of trees per unit area decreased exponentially with increment of mean DBH.

The relationship between DBH (cm) and weight mass of trees (W:kg), trunks and branches (Wt:kg) and leaves (Wl:kg) is shown in figure 12.3. These relationships are as follows:

W = 0.226 D2.274 (2)
Wt = 0.206D2.273 (3)
Wl = 0 .018D2.369 (4)

After measuring the diameter of stem, we are able to estimate the tree weight from equation (2). The tree weights of Caesalpinia pyramidalis and Aspidosperma pyrifolium were 13.3 kg and 22.8 kg in estimated, and 14.5 kg and 23.0 kg in measured values, respectively (Hayashi, 1986).

Figure 12.2 Relationship between mean diameter at breast height of tree stem (DBH) and number of trees per unit area (N) of caatinga stand.

Figure 12.3 Relationship between DBH and weight of leaves, trunk, and branches and whole tree of caatinga species.

Figure 12.4 DBH growth of Pithecellobium foliorosum (Jurema-branca) from 1976 to 1985. Open circles are measured value.

The area of the annual ring formed in the stem of Pithecellobium foliorosum was computed for each year from 1975 to 1985. Based on the area of annual ring (s:cm²), I obtained the stem diameter (d:cm) using the following equation:


The DBH was obtained by adding bark thickness to the stem diameter (d). The growth of DBH obtained above is shown in figure 12.4. for Pithecellobium foliorosum.

The growth curve of DBH of P. foliorosum was described as

D (t) = 9/{1 + 1.04 exp [ - 0.14 t]} (6)

where D (t) is DBH at year t, t is year from 1976 (1976 is 0 year). As shown by this figure, the logistic curve obtained by the least-squares method describes the growth of this tree. According to the equation, the maximum DBH and relative growth rate are 9 cm and 0.14 per year.

In a previous paper (Hayashi, 1986), I reported the same relationship for Caesalpinia pyramidalis and Aspidosperma pyrifolium, which are the dominant species of the caatinga. These equations were as follows:

D (t) = 7.7/{1 + 1.16 exp ( - 0.098 t)} (7)
(Caesalpinia pyramidalis: from 1963 to 1984)

D (t) = 12.3/{1 + 2.33 exp ( - 0.067 t)} (8)
(Aspidosperma pyrifolium: from 1963 to 1984)

Figure 12.5 Predicted changes in DBH histogram of caatinga stand from 1984 to 1989. n and d are number of trees and mean DBH of the stand.

Assuming that the trees in the caatinga grow in accordance with equations (6), (7), and (8), based on a computer simulation, I predicted a change of DBH histogram of a caatinga stand from 1984 to 1989. The DBHs of trees measured in 1984 in Campina Grande were used in the simulation. The changes in the DBH histogram produced by computer simulation are given in figure 12.5. According to these results, the mean DBH changed from 4.1 cm in 1984 (measured value) to 6.4 cm in 1989.

The number of trees per unit area (predicted value) is expected to decrease according to equation (1) from 39 individuals per 100m² in 1984 to 16 individuals in 1989. During this period, the tree mortality in the stand from self-thinning is expected to be 3 from 1984 to 1985, 6 from 1985 to 1986, 7 from 1986 to 1987, 4 from 1987 to 1988 and 3 from 1988 to 1989. A total of 23 trees from the stand are expected to die during the five years.

Table 12.3 Carbon and nitrogen of the plants of the doniinant species in caatinga

Species Carbon (%) Nitrogen (%) C:N
Mimosa hostilis
Wood 46.0 1.0 49.7
Bark 52.1 1.7 31.2
Leaf 43.1 2.4 18.0
Caesalpinia pyramidalis
Wood 47.1 1.0 49.5
Bark 45.5 1.2 37.6
Leaf 46.1 2.5 19.0
Aspidosperma pyrifolium
Wood 47.4 1.1 43.1
Leaf 42.2 2.7 15.7
Panicum trichoides      
Whole plant 44.2 2.8 15.7
Encholrium spectabile      
Whole plant 46.1 1.7 26.5
Herhaceons. plants 41.5 2.5 16.6

The mass of leaves, trunks, and branches from dead trees was estimated, using equations (2), (3), and (4). Based on the DBH shown in figure 12.5 and using the above equation, it is possible to estimate the total mass of litter for each year. Substituting the DBH of dead trees (which are assumed to be the smaller trees within the stand), with equation (3), the mass of trunks and branches from the dead trees were estimated for each year. The mass of fallen leaves at the end of rainy season was also estimated using equation (4). Thus, the total mass of litter supplied to the stand is expected to be 280 kg (164 kg of leaves and 116 kg of trunks and branches) per 100m² during the five years from 1985 to 1989 (Hayashi, 1986, 1988).

The carbon and nitrogen content of the plants are given in table 12.3. Nitrogen content was 2.4-2.8 per cent in the leaves, 1.0-1.1 per cent in the wood and 1.2-1.7 per cent in the bark. The C:N ratios were 43.1-49.7 for the wood and 15.7-19.0 for the leaves, respectively. Uhl et al. (1982) reported that the leaves of tree shoots from the Amazonian caatinga contained 0.78 per cent nitrogen, which is one third of the nitrogen content in the leaves of caatinga trees in North-East Brazil. The plant carbon contents were: wood 46-47 per cent, leaves 42-46 per cent and bark 45-52 per cent. Multiplying 2.4 per cent and 1.05 per cent by total mass of leaf and trunk litter, 5.26 kg/100m² of nitrogen is expected to be supplied to the stand in the form of plant litter during a five-year period. The carbon supplied to ground surface by the litter is expected to be 126.8 kg during the same period, which is equivalent to 466 kg of carbon dioxide (table 12.4). This suggests that the vegetation absorbs the 466 kg per 100m² of carbon dioxide from the atmosphere during the fiveyear period.

Table 12.4 Amount of plant litters and of carbon and nitrogen in the litters produced by the caatinga stand for 5 years (kg/100m²)

  Litter Carbon Nitrogen Carbon dioxide
Leaf 164 71.8 4.10 264
Trunk and branch 116 55.0 1.16 202
Total 280 126.8 5.26 466

Table 12.5 Soil texture of the caatinga stands under different human impacts in the vicinity of Campina Grande

Soil depth (cm) Content in percentage  
Clay Silt Fine sand Coarse sand  
Site 1
0-5 14.0 10.2 71.3 4.5 SL
6-10 19.4 11.4 62.0 6.9 SCL
11 -15 23.7 12.8 57.3 6.2 SCL
16-20 25.1 11.0 58.5 4.9 SC
21-25 23.5 12.1 58.0 6.4 SCL
Site 2
0-5 19.4 9.1 54.1 17.4 SCL
Site 3
0-5 5.0 9.3 71.0 14.7 SL

Site 1: developed stand of caatinga;
site 2: disturbed stand of caatinga; 11 km south from C. Grander
site 3: heavily disturbed stand in the vicinity of C. Grande.

The texture of caatinga soils is given in table 12.5. The fine sand decreases with depth, although the coarse sand content remains constant. By contrast, the clay content increased with soil depth. Surface soils of disturbed and severely disturbed sites contained a larger proportion of coarse sand than that of developed stands.

Soil nitrogen and carbon concentrations are given in table 12.6. The top 20 cm of caatinga soil contained 0.01-0.14 per cent of nitrogen and 0.39-1.65 per cent of carbon. The soil nitrogen is very low comparable to B horizon for soils of well-developed temperate forest. Grove et al. (1986) reported that the surface soil of a Eucalyptus marginata forest in south-western Australia, which is similar to caatinga in physiognomy, contained 0.14 per cent of nitrogen. After burning, the soil nitrogen increased to 0.17 per cent. The nitrogen content before the fire was similar to the caatinga soil.

Table 12.6 Carbon (C) and nitrogen (N) content (%) of the soils of caatingas under different human impacts at the vicinity of Campina Grande

Soil depth (cm) Site 1 Site 2 Site 3
0-5 1.14 0.02 1.65 0.14 0.67 0.05
6-10 0.56 0.02 0.75 0.07 - -
11-15 0.43 0.01 0.46 0.05 - -
16-20 0.39 0.01 0.43 0.04 - -

Site 1: developed stand of caatinga;
Site 2: disturbed stand of caatinga 11 km south from C. Grander
Site 3: heavily disturbed stand of caatinga at the vicinity of C. Grande.

4 Discussion and conclusion

The flora of the semi-arid region of North-East Brazil was described by Luetzelburg (1922-23). According to him, the dominant species in the caatinga were Caesalpinia ferrea, Caesalpinia pyramidalis, Caesalpinia echinata, Aspidosperma pyrifolium, Spondias tuberosa, Magonia pubescens, Hymenaea martiana, Mimosa verrucosa, Mimosa tenuiflora, Cereus squamosus, Cereus jamacaru, and Pilocereus gounellei. Recently, Gomez (1981) studied the vegetation of Cariris Velhos, Paraíba state, and reported that the dominant species in that region were Caesalpinia pyramidalis, Mimosa tenuiflora, Croton sonderianus, Combretum leprosum, Aspidosperma pyrifolium, Jatropha pholiana, Spondias tuberosa, Pithecellobium foliorosum, and Pilocereus gounellei. These floristic compositions are similar to my results in Campina Grande and Patos.

According to Whittaker (1970), the average biomass of shrubland formation is in the range of 2 to 20 kg per m², which includes the value of 2.94 kg m², thus bracketing our value of 2.94 kg m² obtained from stands used for firewood and ranching. The phytomass of the stands varies from site to site according to the intensity of human activities.

The area has been degraded by varying human impacts such as insufficient fallowing period within the shifting cultivation; charcoal production; felling wood for fuel used in brick production; demand for fuel wood by the bakeries in (Campina Grande and Patos; and fence construction to enclose goats (fig. 12.6).

According to Saito et al. (1988), productivity of cultivated crops for 1986 in the município of Patos was 1,050 kg per 9,683 ha, for perennial and annual cotton; 44 kg per 2,299 ha, for maccather bean; 600 kg per 1,962 ha, for maize; and 1,200 kg per 60 ha, for rice. Livestock figures in Patos in 1980 were 14,274 head of cattle, 1,762 head of sheep, and 293 head of horses for the total area of 59,499 ha. Nishizawa and Pinto (1988) reported that during 1981, 16,253 tons of charcoal and 1,569,612m³ of firewood were consumed in the region, including Seridó Paraibano, Piemonte da Borborema, Depressão do Alto Piranhas, Cariris Velhos, and Agreste da Borborema. In 1981, 1,890m³ of timber was produced for house and fence construction. Saito et al. (1988) noted that a single family consumes 100 kg of firewood in a week (five tons per year). Increasing demand for firewood by the bakeries in Patos was also reported by Nishizawa and Pinto (1988).

These human activities have resulted in a lot of bare ground within NorthEast Brazil, thus promoting land degradation. Once the vege tation has been removed, tropical rains create a sandy surface soil and vegetation recovery, including seed germination and seedling establishment, is limited by the severe condition of the soil surface (Hayashi 1981, 1988). We should, with proper use of the land, conserve plant cover to protect the soil against erosion. It is necessary, therefore, that the production of livestock, crops, and fuels be managed within the capabilities of the natural biological potential for the area.


Gomez, M.A.F. 1981. Padrães de caatinga nos Cariris Velhos, Paraíba. Ceará, Fortaleza.

Grove, T.S., A.M. O'Connell, and G.M. Dimmonk. 1986. "Nutrient changes in surface soil after an intense fire in jarrah (Eucalyptus marginata Donn ex Sm.) forest." Australian Journal of Ecology 11: 303-17.

Hayashi, I. 1981. "Plant communities and their environments in the caatinga of Northeast Brazil." Latin American Studies 2: 66-79.

. 1986. "Vegetation and soils of humid and semi-arid regions in Northeast Brazil." Latin American Studies 8: 49-62.

. 1988. "Changing aspect of caatinga vegetation in semi-arid region, Northeast Brazil." Latin American Studies 10: 61-76.

Luetzelburg, P. von. 1922-23. Estudo botônico do Nordeste, vols. 2, 3. Rio de Janeiro.

Ministério da Agricultura e Ministério do Interior. 1971. Mapa exploratorio-reconhe-cimento de solos: Estado da Paraíba (1:500,000). FIBGE, Rio de Janeiro

Nishizawa, T. 1976. "Some characteristics of rainfall in the Northeast of Brazil." Tokyo Geography Papers XX: 53-61.

Nishizawa, T., and M.M.V. Pinto. 1988. "Recent changes in firewood and charcoal production and their affects on deforestation." Latin American Studies 10: 122-31.

Rizzini, C.T., and M.M. Pinto. 1964. "Areas climático-vegetacionais do Brasil, segundo os métodos de Thornthwaite e Mather." Revista Brasileira de Geografia 26 (4): 523-47.

Saito, I., H. Maruyama, and K.D. Muller. 1988. "A comparative study of land use between the Campo Alegre in the sertão and Sitio Açude de Pedra in the agreste, Paraíba, Northeast Brazil." Latin American Studies 10: 78-99.

Uhl, C., C. Jordan, K. Clark, H. Clark, and R. Herrera. 1982. "Ecosystem recovery in Amazon caatinga forest after cutting, cutting and burning, and bulldozer clearing treatments." OIKOS 38: 313-20.

Whittaker, R.H. 1970. Communities and ecosystems. Collier-Macmillan, London.