7.021 Comparisons of land use effects on soil bulk density in the Amazon Basin ‘Wenceslau G.TEIZEIRA \, Gotz SCHROTH ®), Gladys F. de SOUSAM), Bernd HUWE ® (0 Embrapa Amazinia Ocidental, Caixa Postal 319 — 6901-970 - Manaus — AM, Brazil, lax@cpaa embrapa br, ) INPA — Biological Dynamics of Forest Project, C.P. 478 - 69011-970- ‘Manaus, © Soil Physics Group, University of Bayreuth, Bayreuth, Germany Understanding the processes that lead to degradation is an essential element to define suitable land use systems far these soils, Declining productivity of tropical soils under Continuous cultivation even with supplementary fertilisation, is well documented Nowadays, with the development of models for simulating plant growth or transport of solutes in soils, the concept of a static and qualitative description of coil ctructure ic changing into the concept of finding ways of adjusting and optimising structue to achieve maximum yields or reduced leaching rates. Bulk density (p) measurement is proposed as @ rough indicator of the ileal structure for a land use system, and a simple statistical approach is proposed for comparing different scenarios of soil structure, using average weighted values of p as indicator parameter. “The study was carried out on the research station of Embrapa Amazinia Ocideatal near Manaus. The land use systems mvestigated were descrihed in Table 1. Undisturbed soil sarnples ‘were collected at ss 40 cm from the tank of the respective species as well as under the cover crop (Table 1), The experimental design for statistical analysis wat a randomised block design (RBD) with three subsamples within each treatment. The estimation of the weights of the spot areas are based on the area of mfluence of a specific specie, proportional to the entire of the land use system, The domain of influence of a specific plant is a “disc” centred at the stem. These discs were calculated based on’ the crown area of the plants at that stage of growth and ‘management practices These specific areas of influence of each species were then multiplied by the umber of individuals, The remaining areas predominant eflect of the cover crop were calculated by subtraction. After estimating the radkus of muence of these spots around a stern and calculating their relative means, plant species and whole land use systems can be compared ‘using contrast analyses, and hygotheses of practical interest can be formulated in termas of the p- weighted values, ‘An exploratory analysis of the data is shown in Table 1. Table 2 shows differences between the p evaluated near the different plant species (a < 0.05) and does not indicate a veriation between the blocks. The computation of the subsample variance, with three samples per position, shows that the sampling error contributes with 0,003 to the error variance of 0.010 ‘The analysis of subsampling was computed as an indication of the adequacy of the sampling scheme adopted in this study. It seems to be appropriate, because, if the number of samples ‘were doubled, the sampling ersor contribution would be halved Therefore, the main source of the variance is not from a presumed inadequate number of samples, and the reduction of the variance with increasing sampling (in this example, as well in many other situaticns) may not he ‘worthwhile. The mean values of p observed in Table 1 lie in the small range of 0.82— 1.01 Mg ‘m°, with 2 mean of 0.91 Mg m, This range is not typical, considering thet this Ferralsol has clay cantent at the surface of more than 610 g kg? and the organic carbon content is not very high The reason for this is that the clay particles in thece soils are locculated in maicroaggregates (intra-aggregate) and a secondary pore system (inter-aggregates) was established as a result from an intense biolagical activity and fissures. Tae only significant difference in Table 1 is between the p near bacabas in the primary forest and the position between the cupuagus in monoculture covered by grasses and puerania. The higher values of p found between the cupuacus growing m the monacultwe are die to inadequate growth of the cover crop in this land use system, resulting, presumably, in compaction of the soil structure owing to direct impact of the wopical rain drops on the soil, and the more intensive cycles of heating and drying, In the monoculture of peach palm for fruit and palm heart, the sample positions between the plants tend to show higher p values than the values near the trunks (Table1) This agrees with observation of ponded water, for short periods, after intense rainfalls in those postions. The high aumber of reots (with a low density) per volene maintains the soil p values low. ‘The results and discussion of comparison within and between the whole land use systems was divided in the three land use systems evaluated and some results concerning the agroforestry system are presented below The complete smdy is presented by Teoceira, 2001 However, within the agroforeswy system the bulk density (f) near the peach palms and under the pocraria tends to chow lower values of p cone significant difference between the siz plants growing within the agroforestry system were found (Table 1). The effect of pueraria to fecuperate the porosity of the soil and to maintain the soil structure quality in © good level confirms the results found m carly experiments cared out in the Ferralsols in Manaus, Practical hypotheses and its statishcal formulation concerning the agroforestry systema as a whole system are show in Equations 1, 2 and 3, Is there a significant difference in p between on the agroforestry and the primary forest? Hy 0104 jg POLL pe + 0.18 p+ O02 py +0.03 pip 40.62 py OSD pny 0.50 pus = 0 oy Is there a difference in p between on the agroforestry and the secondary forest? Hy:0.04 f+ O11 p+ O18 p+ OO jy +0108 pp +082 fy — Las =O (2 Isthere a difference inp between on the agroforestry and the monoculture of cupuags? Hy:0.04 jy #011 e+ 0.18 pe + O02 4+ 0.03 py +062 fy - 0.07 y-093 ite = 0 (3) Here, ji are mean values of p and the coefficients are the estimated weight factars for ea mean in relation to the whole system Table 3 shows a significant difference (a” » 0.91) between the agroforestry system and the cupuagy monoculture, The other agncultural land use systems do not show significant differences st a> 0.05 in relation to the agroforestry. It should be emphasized that the agroforestry system had been installed in the feld only 3 to 4 vears before these measurements Thus, at did not yet represent a stabilived land use system, Therefore, much free space exasted between the plants, At the time of evaluation, the space hetween the trees was covered by pueranie, which m the agroforestry system wes gowng well and this 1s one reason fr the relative good performance of the agroforestry system analysed as a whole system. Simulations of two hypothetical scenanos were calculated and statithcally compared. Tn the first stlation, the remaining areas beoween the wees m the agroforestry system, instead of being covered by pueraria, were covered by grasses (the mean and variance of the p fir crasses, in thts simulation ‘were those found in the positions beween the capuagu im monoculture), whereas the other ‘means and the weight factors remained the same, An opposite hypothetical sinuation, where 2 better structure was conceptualised, was the second scenario. Inthis theoretical land. ure system, the sil of the monoculture of cupuagy was covered by puerana mstead of beng covered by vas The results in Table 3 show that for the first scenano the agroforestry system shows a significant difference from the primary forest cauced by the increased values of p» when the soil isnot well covered. Inthe second scenario, an opposite effect was found with the improvement of the average sail structure in a hypothetical monocubare of cupuagu system well covered by pucraria, In thi scenario, the monoculture of cupsagu would not show significant difference from the primary forest. The bullc densty is an appropriate physical indicator far the quality of the soil structure, or a mosaic of different soil structures within an agrofcrestry system, Plant growth, yields and leaching rates may be used as indicators of the efficiency of the ideal structure for the suitatility of a land use system. A simple and useful tool to compare different combinations of plants or temporal scenarios are statistical models that combine contrast analysis wath weighted means that represent the diferent spots within a land use system, This technique may be incorporated in lané use planning as an additional tool to define a suitable combmetion of plants and management practicesTuble| Explontcry enelysis of bulk density (Mgan3] evalused on diferent land use aystem. and use system and specie Mean |S Exo: 0? Miamum | Masmum ‘Agroforestry System Capuagn — Thasbroma grandiflonam 093 ab] 205 ost Ta Ancatto — Bia orsllans 090 eb) 247 ‘ova 059 | Bran nut- Berthoiletia exelsa 09% eb| 221 oss 107 Peach palm Bact: gasipacs Fats 058 ab 2? os 100 Peach palm Palm hestt Om ab 355 oe? | ose Fuse Os ab] 119 ‘ogd | 094 Tonemaltere __ Capea iose [196 | Cover crop in Cupaaga To) eb) 84r ust 15 Paach palm ~ Fouts — aesrplanie 08 eb 300 om | ose Peachpaln— Fruits -betwoen palms 098 eb) 212 86 108. Peach palm Palm kesct— near palme 31 a7 107 es - c w ~Frimary forest Metdmata -Richeweleirasp [087 eh Frimay forest —Beesbs Oeracarpus taceba 0220 | 308 ‘Avetige a) 7a ‘Means followed by the same letters within the column, do not differ oy Tukey's test (0 < 0.05). Teble 2 Analisis of variance using complete randomised block design with subeampling for balk density Sewrce SunofSquares df Meso-Square rato P Landure andphst 0.241 14 024 24650020 Brock ome 2 oon 1394 0.65 Enor oar 28 aio Subsanpting ome 80 003 suboanling p00} a} 0.004 1200 0.283 v.land uve and plant [0120] 23] .004 iam 020 vi block [oni] io 0.003 358 0557 Ezror of subeampling [0189 ] (58) 0.003 oo 134 Table Analysis of varincefor the hypothesis conceming agroforestry system. ‘Source of Variation ‘Sum ofequares Degrees of ‘Mean Square | F F freedom Bpainiiny we pamay fora 00100 T Tod es TE ‘Agroforecuy ve aecondury forest <0 0000 1 0.0000 | 0.0021 09540 ‘Agoferesty va. cupuagy o.oras 1 0245 | 7.4278 M109 ‘Agioforestry ve. veach pala 0.0109 1 O.o109 | 3.3184 0.0702 Ear Doo2e 3 canes Aull a rac hppa oivenig ppc uid wes anne av. pamay fest 0.3 058s | 7 7518 WOOT Cupurulvspnmarg forest 1 8002 10 i THe OT] OMS 9ST Error 0.0932 = 0033 The adjusted level @) SeoSoklR. Rand RolF J Homewise New Vouk, Fresman, 1953 S7p Referencie Teixeira, W. G. Land use effect on soil physical and hydraulic propesties of a clayey Ferralsol an the Contest Amazon, Bayreuther Bodinicunde Berichte, Bayreuth Band 72 255.