Geoderma 110 (2002) 1 – 17 www.elsevier.com/locate/geoderma
Pedogenesis and pre-Colombian land use of ‘‘Terra Preta Anthrosols’’ (‘‘Indian black earth’’) of Western Amazonia Hedinaldo N. Lima a, Carlos E.R. Schaefer b,*, Jaime W.V. Mello b, Robert J. Gilkes c, Joa˜o C. Ker b b
a Universidade Federal do Amazonas, 69077-000 Manaus, Amazonas, Brazil Departamento de Solos, Universidade Federal de Vicßosa, 36571-000 Vicßosa-MG, Brazil c Soil Science and Plant Nutrition, University of Western Australia, Nedlands, Australia
Received 12 December 2001; received in revised form 25 March 2002; accepted 12 April 2002
Abstract The ‘‘Terra Preta de I´ndio’’ (Indian black earth) or Terra Preta of Western Amazonia is a thick, dark-coloured, anthropic epipedon, usually rich in nutrients. It occurs mostly at the fringes of the Terra Firme, along the Amazon river banks, overlying deep strongly weathered soils. We studied selected chemical, physical and mineralogical properties of seven soils, ranging from the Tertiary Plateau down to the Amazon river floodplain in the Iranduba district, near Manaus, Amazonas, Brazil. Three Terra Preta soils were classified as anthropogenic (Anthropic Xanthic Kandiudult, Anthropic Xanthic Kandiudox and Anthropic Dystropepts). Chemical, mineralogical and micropedological attributes, such as high total and available P and mica flakes in pottery remains found in the Terra Preta, indicate that the origin of soil materials of these anthrosols is closely associated with neighbouring floodplain (va´rzea) soils and sediments. Amazon floodplain soils were the source of soil material for pottery, since 2:1 clay minerals are not found in the Tertiary Plateau (Terra Firme) sediments. Total and available P contents of Terra Preta are associated with microfragments of bone apatite with high P and Ca values. In the anthrosols under cultivation, these values are less, with increasing Al release suggesting acidification and losses of nutrients. Large amounts of Mn and Zn occur in the anthrosols and in high-fertility floodplain soils. It is unlikely that well-drained Tertiary Plateau (Terra Firme) area far away from lowland Amazon floodplain soils could develop highfertility Terra Preta on the top of nutrient-poor Oxisols (Latosols). The suggested model of Terra Preta formation between the Tertiary Plateau and nutrient-rich Amazon floodplain does not extend to other nutrient-poor, smaller, floodplains draining the deep-weathered interfluves of the Brazilian Uplands. This raises reservations about estimates of precolonial human population densities for the
*
Corresponding author. Fax: +55-31-899-2648. E-mail address:
[email protected] (C.E.R. Schaefer).
0016-7061/02/$ - see front matter D 2002 Elsevier Science B.V. All rights reserved. PII: S 0 0 1 6 - 7 0 6 1 ( 0 2 ) 0 0 1 4 1 - 6
2
H.N. Lima et al. / Geoderma 110 (2002) 1–17
Amazon basin as a whole, assuming the widespread occurrence of such anthrosols farther inland. D 2002 Elsevier Science B.V. All rights reserved. Keywords: Amazonia; Terra Preta de I´ndio; Anthrosols; Human carrying capacity; Soil formation; Organic phosphorus
1. Introduction The Terra Preta Anthrosols of Amazonia (Indian black earth) are mainly Oxisols, Ultisols and Inceptisols with an anthropic A horizon. They have been described by Katzer (1933), Gourou (1949), Sombroek (1966), Ranzani et al. (1970), Eden et al. (1984) and Andrade (1986), yet many aspects of their origin remain obscure. Detailed studies of these precolonial anthropogenic soils can help answer questions about population distribution, soil carrying capacity, settlement pattern and land uses of ancient Amazonian peoples. According to Roosevelt (1997), the history and ecology of Amazonian habitats are of theoretical and practical relevance to the conservation of the vast tropical rainforest, and yet they are poorly documented. According to radiocarbon dating of Terra Preta sites (Hilbert, 1968), these pre-Colombian societies inhabited the Amazon valley and its main tributaries between 2400 F 75 and 1525 F 58 years BP.
Fig. 1. The distribution of known anthrosols in the Amazon and the location of the present study.
H.N. Lima et al. / Geoderma 110 (2002) 1–17
3
Terra Preta Anthrosols in the lower Amazon valley and lower Tapajo´s have been chemically studied by Kern (1988), Kern and Kampf (1989), Zech et al. (1990), Pabst (1991), Kern and Costa (1997) and Glaser (1999) among others, but little is known about the Terra Preta sites of the middle Amazon valley. Near Manaus, there are many discontinuous patches of well-drained Tertiary Plateau, where Xanthic Oxisols (Latosols) are overlain by black earth deposits pedogenically transformed, containing archaeological artifacts such as pottery fragments, weathered bones and organic remains. The extent of these patches of high fertility epipedons is of local and regional importance, and has been considered as an indication of former sustainable land use (Smith, 1980; Glaser, 1999). Even today, anthrosols are intensively cultivated by local population (the‘‘caboclos’’), highlighting its importance to the Amazonian social and ecological landscape (Fig. 1). The aim of this paper is to relate selected chemical, physical and mineralogical attributes of Terra Preta Anthrosols (TPA) with aspects of precolonial land use and the origin of TPA. Also, TPA attributes were compared with neighbouring nonanthropogenic soils, ranging from the Tertiary Plateau to the floodplain of middle Amazon, emphasizing the pedogeomorphological relationships of their occurrence and implications for precolonial human societies.
2. Material and methods We studied seven soils along a toposequence ranging from the Tertiary Plateau (regionally called Terra Firme) down to the Amazon river floodplain (regionally called va´rzea), in the Iranduba district, near Manaus, Amazonas, Brazil (Fig. 2). The soils were classified in the Brazilian System of Soil Classification (EMBRAPA, 1999) and Soil Taxonomy (USDA, 1999), respectively, as Anthropic Yellow Podzolic (Anthropic Xanthic Kandiudult) (P1), Anthropic Yellow Latosol (Anthropic Xanthic Kandiudox) (P2), Anthropic Cambisol (Anthropic Dystropepts) (P3), Yellow Latosol (Xanthic Kandiudox)
Fig. 2. Location of the soils along of the transect Tertiary Plateau (terra firme) – floodplain (va´rzea) near Iranduba, Western Amazonia.
4
H.N. Lima et al. / Geoderma 110 (2002) 1–17
(P4), Plinthic Yellow Latosol (Typic Plinthudox) (P5), Low Humic Gley (Typic Fluvaquent) (P6) and Alluvial (Typic Udifluvent) (P7). The pits were dug to a depth of 1.50 m, and soil profiles described according to EMBRAPA (1999). Undisturbed samples were collected for micropedological analyses. The samples from genetic soil horizons were subjected to chemical, physical and mineralogical analyses. Particle size distribution was determined by wet sieving and ultrasonic dispersion, adapted from EMBRAPA (1997). The clay ( < 0.002 mm) and silt (0.002 –0.053 mm) fractions were separated by sedimentation. Subsamples of clay for X-ray diffraction analysis were flocculated in 5 M NaCl solution. The mineralogy of the clay fraction was determined for all horizons by X-ray diffraction analysis (XRD) using monochromatic CuKa radiation on oriented clay samples. The diffractograms were interpreted following Brindley and Brown (1980). Available P, pH, Ca, Mg, K, exchangeable Al3 + and H + Al were measured by standard procedures (EMBRAPA, 1997). Soil humic substances were chemically fractioned as recommended by Swift (1996) and total organic carbon was determined according to Yeomans and Bremner (1988). Total P was determined by the method of Kuo (1996), and citric acid extractable P according to the procedures of USDA (1996). For each horizon, 1 g of fine earth was digested by 20 ml 1:1 H2SO4, and the relative proportions of SiO2, Al2O3, Fe2O3 and TiO2 were measured by atomic absorption spectrometry (EMBRAPA, 1997), to provide data required for the Brazilian System of Soil Classification (EMBRAPA, 1999). The total microelements in 100 mg clay were extracted by HF + HNO3 + HCl (USDA, 1996), and the elements were determined by atomic absorption spectrometry. Undisturbed soil samples were impregnated with crystic resin under vaccum, following recommendations of FitzPatrick (1993). The micromorphology of these selected horizons was studied in thin sections at 30 or greater magnifications. Pottery fragments and pedological features such as structural units, porosity, pedofeatures (nodules, concrections) and clay coatings were described according to Bullock et al. (1985) and Fitzpatrick (1993). Selected thin sections were polished down to 1 Am with diamond paste, carbon-coated and submitted to SEM/EDS analyses, using a JEOL 6400 fitted with EDS, in order to identify and analyse the P forms.
3. Results and discussion 3.1. Chemical and physical characteristics The five soils from the Tertiary Plateau and its border (P1 to P5; Fig. 2) are generally dystric in subsurface, and dominated by kaolinite in the clay fraction, similar to those developed from preweathered sediments on the Tertiary Plateau elsewhere in Brazil (Resende et al., 1995) (Tables 1 and 2). The two nonanthropogenic soils, P4 and P5, have high levels of Al3 + in the exchange complex. At the surface, the anthropogenic A horizons of P1, P2 and P3 show a distinct eutrophic character, with elevated base saturation and very high ‘‘available’’ P content (Table 1). In the soil under continuous
H.N. Lima et al. / Geoderma 110 (2002) 1–17
5
Table 1 Chemical and physical characteristics of the soils Horizon/depth (cm)
pH (H2O)
Pa
K
mg kg
1
Ca2 +
Mg2 +
cmolc kg
Al3 +
1
Clay
360 300 200 180
240 180 140 100
70 50 20 20
320 460 630 700
0.00 0.00 0.00 0.00 0.00
6.37 6.21 4.31 3.83 3.36
71.0 70.2 69.7 65.2 58.5
390 310 290 280 280
130 140 120 120 110
130 190 180 140 120
350 360 410 460 490
1.04 0.44 0.86 0.06 0.08
0.00 0.00 0.00 0.00 0.00
7.00 5.78 3.52 1.82 1.50
52.7 51.6 48.4 30.0 31.1
450 380 410 430 480
110 130 100 140 110
140 130 90 90 90
300 360 400 340 320
0.01 0.01 0.01 0.01 0.01
0.03 0.01 0.01 0.01 0.01
1.06 1.34 0.86 0.77 0.77
6.84 5.10 3.52 2.88 2.72
1.1 0.6 0.8 0.9 0.9
340 290 270 200 390
220 220 180 180 220
50 30 50 40 60
390 460 500 580 330
16 11 2 2
0.01 0.01 0.01 0.01
0.03 0.03 0.01 0.01
1.63 1.15 0.99 0.67
6.89 5.73 3.52 3.36
1.2 1.2 0.7 0.7
330 270 230 180
200 210 190 120
50 40 50 40
420 480 520 560
69 34 33 33
46 39 30 44
9.86 12.45 11.92 13.01
3.21 4.99 5.33 7.37
2.50 0.48 0.35 0.08
6.37 3.44 2.57 2.57
67.4 83.6 87.1 88.9
0 0 0 0
30 60 60 0
700 650 660 580
270 290 280 420
25 108 45
79 38 44
10.62 10.88 11.17
2.52 2.42 3.44
0.51 0.10 0.42
5.53 3.20 3.20
70.7 80.7 82.1
0 0 0
480 440 140
370 380 590
150 180 270
0.19 0.45 0.19 0.13
Anthropic Xanthic Kandiudox A1 0 – 30 6.2 1991 A2 30 – 60 6.2 2935 A3 60 – 100 6.4 3921 AB 100 – 130 6.5 3537 Bw 130 – 150 6.5 1567
55 49 53 44 27
14.13 13.98 9.34 6.69 4.37
1.32 0.53 0.44 0.36 0.30
Anthropic Dystropepts A1 0 – 15 6.3 A2 15 – 40 6.4 A3 40 – 55 6.3 AB 55 – 110 6.4 Bi 110 – 180 6.0
1332 2032 816 115 92
70 44 36 24 18
6.59 5.60 2.35 0.66 0.55
Xanthic Kandiudox A 0 – 18 4.6 AB 18 – 40 4.3 Bw1 40 – 64 4.4 Bw2 64 – 90 4.4 Bw3 90 – 150 4.4
1 1 1 1 1
15 4 3 2 2
Typic Plinthudox A 0 – 20 4.7 AB 20 – 40 4.7 Bw 40 – 95 4.7 Bwf 95 – 150 4.8
2 1 1 1
Typic Fluvaquent Ag 0 – 13 4.8 ACg 13 – 35 5.8 Cg 35 – 62 5.9 2Cg 62 – 100 6.5 Typic Udifluvent A 0–5 5.4 2C2 24 – 34 5.8 5C5 50 – 150 5.6
d
Silt
1
41.6 30.3 26.1 30.0
0.63 0.28 0.18 0.12
c
F.S.d
6.32 6.16 4.47 3.36
3.85 2.38 1.39 1.31
P-Mehlich-3. Base saturation. Coarse sand. Fine sand.
C.S.c g kg
12 6 5 4
b
BSb %
Anthropic Xanthic Kandiudult A1 0 – 23 5.2 173 A2 23 – 42 5.2 136 Bt1 42 – 73 5.2 257 Bt2 73 – 130 5.1 145
a
H + Al
6
H.N. Lima et al. / Geoderma 110 (2002) 1–17
Table 2 Clay mineralogy, Fe contents extracted by dithionite-citrate (Fed) and ammonium-oxalate (Feo) and Feo/Fed ratios of fine-earth ( < 2 mm) of the soils Horizon
Feo
Fed
Feo/Fed
Clay mineralogy
Anthropic Xanthic Kandiudult A 2.6 AB 2.2 Bt1 1.4 Bt2 1.0
35.7 47.1 48.4 51.6
0.073 0.047 0.029 0.019
Kt, Gt, Hm, Ana
Anthropic Xandic Kandiudox A1 5.5 A2 4.8 A3 3.6 AB 2.9 Bw 1.7
35.2 40.7 41.4 46.4 52.8
0.156 0.118 0.087 0.063 0.032
Kt, Gt, Hm, Ana
Anthropic Dystropepts A1 4.4 A2 3.1 A3 1.2 AB 0.4 Bi 0.1
43.6 59.2 71.9 49.1 34.2
0.101 0.052 0.017 0.008 0.003
Kt, Gt, Hm, Ana
Xanthic Kandiudox A AB BA Bw1 Bw2
4.5 1.8 1.0 0.5 0.3
33.7 37.4 37.8 47.8 50.2
0.134 0.048 0.026 0.010 0.006
Kt, Gt, Hm, Ana
Typic Plinthudox A AB Bw Bwc
2.7 3.3 0.7 0.4
35.9 39.0 40.2 63.0
0.075 0.085 0.017 0.006
Kt, Gt, Hm, Ana
Typic Fluvaquent A ACg Cg 2Cg
15.0 13.8 12.2 10.5
25.7 23.8 24.2 23.9
0.584 0.580 0.504 0.439
Kt, Smec, Ill, Cl, Bi, Gt, Hm
Typic Udifluvent A 2C2 5C5
11.8 10.3 12.5
23.6 22.1 25.2
0.500 0.466 0.496
g kg
1
Kt, Gt, Hm, Ana
Kt, Gt, Hm, Ana
Kt, Gt, Hm, Ana
Kt, Gt, Hm, Ana
Kt, Gt, Hm, Ana
Kt, Smec, Ill, Cl, Bi, Gt, Hm
Kt, Smec, Ill, Cl, Bi, Gt, Hm Kt, Smec, Ill, Cl, Bi, Gt, Hm
Kt—Kaolinite; Gt—Goethite; Hm—Hematite; An—Anatase; Ill—Illite; Smec—Smectite; Bi—Biotite; Cl— Clorite.
H.N. Lima et al. / Geoderma 110 (2002) 1–17
7
cultivation (P1), however, there is more extractable Al, similar in amount to the recently reported incipient Al release in eutrophic soils under shifting cultivation in Amazonia (Vale, 1999). The high silt content and considerable textural variation with depth of the floodplain soils (P6 and P7) reflect the complex sedimentary history of this environment, and the very weak degree of pedogenesis compared with the Tertiary Plateau upslope (Schaefer et al., 2000). Clay in the floodplain soils is typically dominated by high activity and some horizons contain large amounts of exchangeable Al (Table 1) associated with 2:1 expanding clays (Table 2). This suggests weathering of the smectite under present conditions. The widespread occurrence of petroplinthite in the transition segment of the Tertiary Plateau, as observed in P5, is attributable to lateral Fe-flux from upland sources and its precipitation along the escarpment edge above the floodplain (va´rzea). The distribution of Fe-dithionite between anthropogenic and nonanthropogenic profiles is not very different (Table 2), but there is more Fe-oxalate in surface horizons of the anthrosols, a feature possibly related to the greater organic matter content. The higher Feoxalate/Fe-dithionite ratios (0.47 and 0.58) in the soils of the floodplain (P6 and P7; Table 2) indicate the dominance of less crystalline forms of Fe-oxides, and confirm the relative immaturity of the soils associated with aquic regimes, compared with those at higher levels. In the Anthropic Xanthic Kandiudox, total P2O5 reached the exceptionally high value of 13,870 mg kg 1 P2O5 in the anthropic epipedon, whereas the distribution of total P in the Anthropic Xanthic Kandiudult is relatively uniform, probably because of repeated cultivation and mixing of this site (Table 3). In noncultivated Terra Preta, the surface values are much enhanced in relation to the underlying soil horizons. The high levels of P and their association with organic matter are being further investigated in a broader project currently underway. Scanning electron microscopical (SEM) observations suggest that most of the phosphorus in the anthropic epipedon is in amorphous/low crystalline forms, associated with bone apatite from fish middens (Lima, 2001). Two examples are illustrated in Fig. 3, showing the chemical composition by EDSmicroprobe. Table 3 Total, Mehlich-1, citric acid P2O5 contents and phosphate maximum adsorption capacity (PMAC) of the anthropogenic soils Horizon
Total mg kg
Mehlich-1
Citric acid
PMAC
1
Anthropic Xanthic Kandiudult A1 3070 Bt2 3460
396 332
345 323
0.78 0.77
Anthropic Xanthic Kandiudox A1 13,870 Bw 7180
4559 3588
4548 3172
0.50 0.17
Anthropic Dystropepts A1 Bi
3050 211
3066 124
0.31 0.07
8800 960
8
H.N. Lima et al. / Geoderma 110 (2002) 1–17
Fig. 3. SEM photomicrographs of some P forms in Terra Preta anthrosols: (A) bone-apatite microfragment with high P/Ca values on EDS chemical mapping (CaO 35.45%; P2O5 16.45%); (B) secondary P concentration along a biological channel, revealing a high Fe/Al composition (CaO 3.35%; P2O5 11.29%; Fe2O3 12.62%; Al2O3 20.23%).
3.2. Organic carbon and humic fractions The soils developed on Tertiary sediments (P4 and P5) generally contain more organic carbon than those on the floodplain (Table 4), because of the dystrophy, and therefore less favorable conditions for mineralization. Also, the organic carbon values in the anthrosols are greater than in the nonanthropogenic soils. The floodplain soil contains the least organic carbon, suggesting rapid mineralization or burial by floods. All the anthrosols are dominated by strongly humified fractions (humins and humic
H.N. Lima et al. / Geoderma 110 (2002) 1–17
9
Table 4 Total organic carbon, soil humic substances and HAF/FAF ratios in superficial soil horizons Horizon
FAFa
TOC g kg
HAFb
Humc
HAF/FAF
1
Anthropic Xanthic Kandiudult A 18.3
1.7
7.3
7.2
4.29
Anthropic Xandic Kandiudox A 34.6
3.3
11.8
19.5
3.55
Anthropic Dystropepts A 35.3
0.7
10.5
20.1
14.22
Xanthic Kandiudox A
14.8
3.5
3.0
8.6
0.86
Typic Plinthudox A
13.6
4.1
1.5
7.4
0.36
Typic Fluvaquent A
8.3
1.5
0.4
5.6
0.28
Typic Udifluvent A
9.7
1.3
1.1
6.8
0.82
a b c
Fulvic acid fraction. Humic acid fraction. Humin.
acid—HAF) with smaller amounts of the more soluble and mobile fulvic acid fraction (FAF), as indicated by the high HAF/FAF ratios (Table 4). These results agree with those Zech et al. (1990), who reported small amounts of mobile humic substances and rapid humification in Terra Preta of the lower Tapajo´s area. It is possible that most humins and humic acids are formed by progressive polymerization, with decreasing aliphaticity and slow turnover (Duchaufour, 1998), but a marked contribution of refractory, ‘‘inherited’’ humin as charcoal (black carbon), through burning and pedobiological mixing has been reported by Glaser et al. (2000) and Lima et al. (2001) in Terra Preta soils. These results highlight the importance of this sequestered organic carbon pool in Terra Preta Anthrosols due to long-term charcoal formation, enhancing the content of aromatic carbon, only partly oxidized. The greater humification of the Terra Preta soils compared with the others analysed may also be related to the larger amounts of Ca in the exchange complex (Table 1), which favors earthworm activity and renders the organic matter less soluble by forming more stable aggregates. The Terra Preta A horizon would therefore be equivalent to the ‘‘Ca-rich eutrophic mull’’ (Duchaufour, 1998). Its complex crumb structure consists of very stable microaggregates < 50 Am across with larger macroaggregates >250 Am (Guggenberger et al., 1995) possibly created by bioturbation (Schaefer, 2001), attributable to the earthworm and termite activity, indicated by abundant burrows and channels.
10
H.N. Lima et al. / Geoderma 110 (2002) 1–17
3.3. Microelements The clay fraction of the A horizons of the anthrosols contains more Mn and Zn than either the B horizons or the A and B horizons of nonanthropogenic oxisols (Table 5). The floodplain soils also have large amounts of Mn and Zn. On the other hand, there are little differences in Cu, Cd, Ni and Cr between Terra Preta and floodplain soils. High Mn and Zn concentrations in the Terra Preta Anthrosols of the lower Amazon have also been reported by Kern and Kampf (1989) and Kern and Costa (1997). 3.4. Micromorphological features Human activity in the Terra Preta soils is also indicated by the micromorphological features of the Anthropic Xanthic Kandiudult (A horizon). The crumb structure (FitzPatrick, 1993) is typical of soils with mollic epipedons (Fig. 4a), and indicates a progressive downwards mixing of organic aggregates (Fig. 4a and c). The overall impression is of efficient pedobiological mixing of the organic-rich A horizons with the underlying Bw/Bt
Table 5 Microelements concentration in clay fraction of the soils Horizon
Fe g kg
Mn 1
mg kg
Cu
Zn
Cd
Ni
Cr
1
Anthropic Xanthic Kandiudult A 79.4 Bt 90.2
627 71
104 62
150 45
21 22
78 84
172 146
Anthropic Xandic Kandiudox A 57.4 Bw 68.5
387 84
90 42
245 97
20 19
97 94
17 1
Anthropic Dystropepts A 59.3 Bi 54.2
289 97
69 49
248 73
18 21
88 97
3 0
Xanthic Kandiudox A 53.7 Bw 76.8
84 81
81 36
41 44
21 21
86 89
92 138
Typic Plinthudox A 67.2 Bw 71.0
87 102
123 101
41 50
21 20
91 84
98 89
Typic Fluvaquent A 72.3 C 57.7
778 379
148 106
176 168
22 18
108 99
92 80
Typic Udifluvent A 74.9 5C5 64.1
755 502
121 82
156 158
17 16
103 91
0 54
H.N. Lima et al. / Geoderma 110 (2002) 1–17 11
Fig. 4. Photomicrographs of A (a, b), AB (c) and B (d) horizons of the Anthropic Xanthic Kandiudult.
12
H.N. Lima et al. / Geoderma 110 (2002) 1–17
horizons, as indicated by earthworm channels filled with black material in the B horizons and with B horizons material in the A horizons. In the Bt horizon of the Anthropic Xanthic Kandiudult, there is a coalesced pattern of oxidic microaggregates, associated with features of clay illuviation (argillans) (Fig. 4d) along burrows, channels and aggregates. There are also abundant black charcoal fragments documenting long-lasting human settlement, followed by intense biological mixing. The pottery fragments contain many mica flakes (Fig. 4c) and XRD analysis of the pottery fragments indicated an abundance of illite, but no kaolinite; this suggests that the pottery clay came from the gley soils from the floodplain, as mica (illite) is absent from the Tertiary sediments, uplands (Table 2). In the Bw horizon of the Anthropic Xanthic Kandiudox, the granular microstructure is typical of the Oxisols from elsewhere in Brazil (Schaefer, 2001), and fragments of laterite or petroplinthite are common (Fig. 5b). The A horizon shows the typical crumb structure of mollic epipedon (Fig. 4a), reflecting intense earthworm activity. The floodplain soils contain features indicating weak pedogenesis only. Wetting and drying (Fig. 5d) have resulted in incomplete, incipient development of peds, and redox processes have resulted in Fe precipitation in voids/channels, or between sand and silt particles (Fig. 5c). 3.5. Terra Preta Anthrosols and human carrying capacity In the Western Amazon, as elsewhere in Amazonia, the Terra Preta Anthrosols are closely associated with flat-tops of escarpments of the well-drained Tertiary Plateau (Terra Firme), where they form patches resulting from ancient middens (waste deposits). It is unlikely that any well-drained Tertiary Plateau located far from the rich Amazon floodplain environment could ever attain such high concentrations of P, Ca and K such as those reported here, without an anthropic influence. Terra Firme soils, especially Oxisols and Ultisols, are extremely poor in nutrients (Schaefer et al., 2000). It is also unlikely that the Amazon as whole could sustain a high population density, if one considers the occurrence of ‘‘Terra Preta’’ as the basis for calculation. This is because the Amazon floodplain covers less than 5% of the Amazon basin, and represents the only chemically enriched enviroment, apart from the Terra Preta (Schaefer et al., 2000). The abundance of total and exchangeable nutrients is probably a result of the sustained high primary productivity of the neighbour floodplain (va´rzea), the only area capable of providing high-nutrient middens (Schaefer et al., 2000). It seems that the only environment in the Amazon lowlands suitable for prolonged cultivation is the floodplain, where nutrients removed by leaching and harvesting are replaced annually by flooding. In the floodplain, maize can be grown at lower levels, reaching maturity in 3 months, whereas cassava requires at least 6 months, and would be better adapted to the higher transition zone to the Tertiary Plateau, due to soil drainage requirements. At the time of European contact in the Amazon, there were reports of up to six harvests of maize per year (Meggers, 1971) and widespread gathering of wild rice. Indian villages along the escarpment of the Tertiary Plateau were said to have as many as a thousand inhabitants in 1542 (De Carvajal, 1894), but their presence dates back to 2400– 1500 years BP. Some cultural sophistication and hierarchical division can be further
H.N. Lima et al. / Geoderma 110 (2002) 1–17
Fig. 5. Photomicrographs of A (a) and B (b) horizons of the Xanthic Kandiudox and C (c, d) horizon of the Typic Fluvaquent. 13
14 H.N. Lima et al. / Geoderma 110 (2002) 1–17 Fig. 6. General aspects of ‘‘Terra Preta’’ sites in precolonial times along the fringes of Terra Firme with seasonal inputs of fish, bones, pottery and other mineral/organic materials from the floodplain. In this scenario, the floodplain was unsuitable for permanent settlement.
H.N. Lima et al. / Geoderma 110 (2002) 1–17
15
implied from the Terra Preta evidence of these sites (Meggers, 1995), based on pottery and artifact studies. According to our preliminary model (Fig. 6), the Terra Preta Anthrosols result from diversification in adaptative strategies to face natural adversities in the ‘‘human floodplain’’ ecological adaptation. When people settled on the fringes of the Tertiary Plateau, there were considerable constraints to permanent settlement in the floodplain. A possible scenario is a combination of seasonal high watertable level or flood intensity and limiting land for housing. Thus, prehistoric settlement along the Amazon developed towards a Tertiary Plateau – floodplain complement, as recently postulated by Denevan (1996). Consequently, estimates of precolonial population density for the whole Amazon assuming similar soil qualities between the rich floodplain and the nutrient-poor well-drained soils can be erroneous. Seasonal fishing of the nutrient-rich Amazon and its associated system of lakes and channels (Fig. 6) was the main source of protein for these pre-Colombian societies. The floodplain had the natural resources to support large permanent settlements and higher social complexity than could be sustained in the nutrient-depleted Tertiary Plateau, but was apparently unsuitable for permanent settlement. On this basis, it seems unlikely that there was ever a substantial settlement on the active floodplain.
4. Conclusions The investigated chemical, mineralogical and micropedological attributes, such as high available P and mica flakes in pottery remains of the Terra Preta, indicate that the allochthonous materials in the Terra Preta Anthrosols have their source in the neighbouring floodplain soils and from the Amazon river. Amazon floodplain soils were the only source of soil material for pottery, since 2:1 minerals are not found in upslope Tertiary Plateau soils. The total and available P contents of noncultivated Terra Preta soils (Anthropic Xanthic Kandiudox) are greater than values reported in the literature. In the Terra Preta under cultivation (Anthropic Xanthic Kandiudult), these values are less, but still large. The levels of Mn and Zn in the anthrosols are similar to the floodplain soils, corroborating a floodplain source. It is unlikely that any large area of interfluvial Tertiary Plateau, far away from lowland Amazon va´rzea soils, could develop widespread Terra Preta Anthrosols on the top of nutrient-poor Oxisols. The formation of large areas of Terra Preta Anthrosols containing high P contents in areas of rivers draining deep-weathered terrains of the Brazilian Plateau is unlikely. This raises reservations about estimates of precolonial population densities for the Amazon basin as a whole, assuming a widespread occurrence of nutrient-rich anthrosols away from the Amazon floodplain.
Acknowledgements The authors are grateful for the careful reviews and comments by Prof. J. Catt and Dr. G. Guggenberger on earlier version of this paper. This work has been partially
16
H.N. Lima et al. / Geoderma 110 (2002) 1–17
supported by CAPES-Brazil and CNPq, during Prof. Schaefer’s sabbatical in Western Australia. References Andrade, A., 1986. Investigation arqueologica de los anthrosolos de Araracuara. Banco de la Republica, Bogota. Brindley, G.W., Brown, G., 1980. Crystal Structure of Clay Minerals and Their XRD Identification. Mineralogical Society, London. De Carvajal, G., 1894, (1542). Descubrimiento del Rio de Orellana. Descubrimiento del Rio de Las Amazonas, compiled by Jose Torribio Medina, Ed. Rasco, Sevilla, pp. 1 – 83. Denevan, W.M., 1996. A bluff model of riverine settlement in prehistoric Amazonia. Annals of the American Geographers 86 (4), 654 – 681. Duchaufour, P., 1998. Handbook of Pedology—Soils, Vegetation and Environment. A.A. Balkema, Rotterdan. Eden, M.J., Bray, W., Herrera, L., McEwan, C., 1984. ‘‘Terra Preta’’ soils and their archaeological context in the Caqueta Basin of southeast Colombia. American Antiquity 49, 125 – 140. EMBRAPA. Centro Nacional de Pesquisa de Solos, 1997. Manual de me´todos de ana´lise de solo. EMBRAPACNPS, Rio de Janeiro. EMBRAPA. Centro Nacional de Pesquisa de Solos, 1999. Sistema Brasileiro de Classificacß a˜o de Solos. EMBRAPA-CNPS, Rio de Janeiro. FitzPatrick, E.A., 1993. Soil Microscopy and Micromorphology. Wiley, New York. Glaser, B., Balashov, E., Haumaier, L., Guggenberger, G., Zech, W., 2000. Black carbon density fractions of anthropogenic soils of the Brazilian Amazon region. Organic Chemistry 31, 669 – 678. Gourou, P., 1949. Observacßo˜es geogra´ficas na Amazoˆnia. Revista Brasileira de Geografia 11, 354 – 408. Guggenberger, G., Zech, W., Haumaier, L., Christensen, B.T., 1995. Land-use effects on the composition of organic matter in particle-size separates of soils: II. CPMAS and solution 13C NMR analysis. European Journal of Soil Science 46, 147 – 158. Hilbert, P.P., 1968. Archa¨ologische Untersuchungen am Mittleren Amazonas. Marburger Studien zur Vo¨lkerkunde 1. Dietrich Reimer, Berlin. Katzer, F., 1933. Geologia do Estado do Para´. Boletim do Museu Paraense Emilio Goeldi de Historia Natural e Etnografia. Oficinas Gra´ficas do Instituto ‘‘D. Macedo Costa’’, Bele´m. Kern, D.C., 1988. Caracterizacßa˜o pedolo´gica de solos com Terra Preta Arqueolo´gica na regia˜o de Oriximina´, Para´. Porto Alegre, UFRGS, Faculdade de Agronomia (MSc Thesis). Kern, D.C., Costa, M.L., 1997. Composicß a˜o quı´mica de solos antropogeˆnicos desenvolvidos em Latossolo Amarelo derivados de lateritos. Geocieˆncias 16 (1), 141 – 156. Kern, D.C., Kampf, N., 1989. Antigos assentamentos indı´genas na formacß a˜o de solos com terra preta arqueolo´gica na regia˜o de Oriximina´, Para´. Revista Brasileira de Cieˆncia do Solo 13, 219 – 225. Lima, H.N., 2001. Genesis, Chemistry, Mineralogy and Micromorphology of Soils from Western Amazonia. PhD Thesis, 176 pp. (in Portuguese). Lima, H.N., Benites, V.M., Schaefer, C.E.G.R., Mello, J.W.V., Ker, J.C., 2001. Caracterizacßa˜o de a´cidos hu´micos extraı´dos de terra preta de ´ındio. Annals of Brazilian Society of Humic Substances, III Congress. Meggers, B.J., 1971. Amazonia: Man and Culture in a Counterfeit Paradise. Aldine, Chicago. Meggers, B.J., 1995. Amazonia in the event of European contact: ethnohistorical, ecological and anthropological perspective. Revista de Arqueologia Americana 8, 91 – 115. Pabst, E., 1991. Crite´rios de distincßa˜o entre Terra Preta e Latossolo na regia˜o de Belterra e os seus significados para a discussa˜o pedogene´tica. Boletim do Museu Paraense Emı´lio Goeldi. Se´rie Antropologia 7 (1), 5 – 19. Ranzani, G., Kinjo, T., Freire, O., 1970. Ocorreˆncias de ‘‘Plaggen Epipedon’’ no Brasil. Noticia Geomorfologica 10, 55 – 62. Resende, M., Curi, N., Rezende, S.B., Correˆa, G.F., 1995. Pedologia: bases para distincßa˜o de ambientes. NEPUT, Vicßosa. Roosevelt, A.C., 1997. The influence of geology on soils, biota and the human occupation of Amazonia. In: Costa, M.L., Ange´lica, R.S. (Eds.), Contribuicßo˜es a` Geologia da Amazoˆnia. Contribuicßo˜es a` Geologia da Amazoˆnia, pp. 1 – 13.
H.N. Lima et al. / Geoderma 110 (2002) 1–17
17
Schaefer, C.E.G.R., 2001. Brazilian latosols and their B horizon microstructure as long-term biotic constructs. Australian Journal of Soil Research 39 (5), 909 – 926. Schaefer, C.E.G.R., Lima, H.N., Vale Jr., J.F., Mello, J.W.V., 2000. Uso dos solos e alteracß o˜es da paisagem na Amazoˆnia: cena´rios e reflexo˜es. Boletim do Museu Paraense Emı´lio Goeldi, Se´rie Cieˆncias da Terra 12, 63 – 104. Sombroek, W.G., 1966. Amazon Soils: A Reconnaissance of the Soils of the Brazilian Amazon Region. Centre for Agricultural Publications and Documentation, Wageningen. Swift, R.S, 1996. Organic matter characterization. In: Sparks, D.L. (Ed.), Methods of Soils Analysis: Chemical Methods. Soil Science Society of America, Madison, pp. 1011 – 1069. USDA, 1996. Soil Survey Laboratory Methods Manual. Soil Survey Investigations Report No. 42, Version 3.0. Washington, USDA Office of Communications. USDA. Soil Survey Staff, 1999. Soil Taxonomy, 2nd ed. United States Department of Agriculture-Natural Resource Conservation Service, Washington, DC. Vale Jr., J.F., 1999. Pedogenesis and management effects of shifting cultivation in soils developed from acid and mafic volcanic in Roraima, Amazonia. Universidade Federal de Vicßosa (DSc Thesis). Zech, W., Haumaier, L., Hempfling, R., 1990. Ecological aspects of soil organic matter in tropical land use. In: McCarthy, P., Clapp, C.E., Malcolm, R.L., Bloom, P.R. (Eds.), Humic Substances in Soil and Crop Sciences: Selected Readings. American Society of Agronomy, Madison, pp. 187 – 202.
