El1virol1mel1ta[ Impact
Land-use in Amazonia and the Cerrado of Brazil
J 2 4 DANIEL C. NEPSTAD , CARLOS A. KUNK , CHRISTOPHER UHL3, IMA C. VIEIRA , PAUL 2 5 6 7 LEFLBVRE , MARCOS PEDLOWSKI , ERALDO MATRICARDI , GUSTAVO NEGREIROS , IRVING F. x JO JJ BROWN , EUFRAN AMARAL'>,ALFREDO HOMMA , ROBERT WALKER
/ Woods Hole Research Center.Woods Hole. MA 02543. USA and lnstintto de Pesquisa Ambienta! da Amazonia, Universidade Federal do Pará, Bclem, PA M075-90(), Brasil, -'Depar/amento de I:'cologia, Instituto de Ciéncias Biológicas,Universidade de Brasilia. Brasilia. DF 70919-970, Brasil, 'Thc Pennsvlvania Stat«Universitv. Deparuneut o(Biol()g\~ State College, PA I ó80 I, USA and Instituto do l lomem e do Meio Ambiente da Arnazonia, Bclent, PA ÓÓ()()O-900, Brasil. "Museu Paraense Emilio Goeldi. Beléni. PA ÓÓ017-970' Brasil.
.< Virginia Polvtechnic tnstitute. Blacksburg. VA 240ó3-0559, USA, "Secretaria de Desenvolvimento Ambiental, Porto Velho, RO, Brasil, 'Co//ege o] Forest Resources. Universitv 01 Washington, Seattle. WA 98 I 95, "/;Vt)()dl'Hole Research Center. ~jlo()ds Holc,MA 02543, USA and Departamento de Gcociéncias. Universidade Fedeml Flunrincnse. Niteroi, RJ 24200-900, Brasil, "PIlI'íIUe Zoobotánico, Universidade Federal do Acre, Rio Branco. AC ó99 I 5-900. Brasil. '''f:'MBRAPA-CPATU, Belém, PA M095-2óO, Brasil and
/I Florida State Universitv. Tallahassee, Florida 323()ó-2050, USA
The total area and annual rate of native vegetation clearing is greatest in the Cerrado region followed by the Brazilian states of Pará, Mato Grosso, Maranhào and Rondônia. Amazonian forest clearing proceeds most quickly where abundant natural resources (wood or land) are accessible by roads and close to markets. These regions are concentrated along the eastern and southern f1anks of Amazonia, particularly in eastern Pará, Cuiabá and Rondônia. There are still large discrepancies in estimates of annual deforestation; Landsat (Thematic Mapper-based) mapping of deforestation in the closed-canopy forests of Amazonia has not include non-Brazil- ian countries and is incomplete for the cerrado biome.Amazonian deforestation was last mapped in 1994. Current estimates of Amazonian forest clearing do not include most ofthe forests that are affected by logging each year, which is an area (about 7,000 km-yr ") more than halfthe size of the area of annual deforestation. Logging changes forest structure and increases forest flarn- mability. The intensity of logging ranges from 1- to 100-species harvest, and averages 20 m3 of wood harvested per hectare. Logging may increase dramatically in the coming years. Fire affects large, but difficult to measure, areas of pastureland, logged forests, secondary forests and pri- mary forests. Forest ground fires are particularly difficult to map from satellite data. Fire is more frequent where forest clearing is taking place, and where seasonal drought is most severe. The destiny of Amazonian forest land cleared for crops and cattle pasture is complex, and highly variable regionally.Areal estimates are needed for managed pasture, degraded pasture, cropland and secondary forests, for these ecosystems are functionally distinct. Most forest clearing is for
Corrcspondcncc to: Danicl C. Ncpsrad, Woods Hole Rescarch Ccnter,PO Box 296, Woods Hole, MA 02543, USA
Volume 49(1/2)· January/ApriJ1997 Ciência e Cultura Journal of the Brazilian Association for the Advancement of Science» 73 · ...... Environmenta[ Impact pasture establishment, followed by shifting cuItivation. Cattle pasture is the logicalland-use for both small-scale and large-scale rural Amazonians because cattle are easily sold or traded, and they maintain their value during inflation. Cattle pastures help secure land c1aims and increase land value.ln the Cerrado, there has been a shift from extensive cattle grazing ofnatural savannas to pastures planted with African forage grasses; mechanized soy bean production is the second most extensive land-use. Pastures are the most important land-cover for the LBA (Large-Scale Biosphere - Atmosphere experiment in Amazonia) science campaign. Brazilian Amazonia expe- riences reduced rainfall during ENSO events. ENSO-related drought is most severe in eastern Amazonia. A basin-wide reduction in rainfall would have its greatest affect on vegetation near the border between savanna and c1osed-canopy forest in Rondônia, Mato Grosso, Pará and Tocantins. The LBA campaign should be conducted in a variety of rurallandscapes to capture the muItiplicity ofhuman effects on native ecosystems, as well as the range of c1imatic and edaphic conditions under which these ecosystems have evolved. It should address the current (ENSO) and predicted variations in climate, and should be designed to recommend those land-uses that best reconcile the maintenance of ecosystem processes with socially equitable economic growth.
() dcsnunamcnto da \'cgcfaçclo nativa no Ccrrud» (' na Ainazonia t ctn nuiadas. Na Amazoniu, o destino dos alltigos áreas dcflorcst a» (/uc sc sido ;11(,/,(:'/I/CIII(/(/O JJOl" vctorc» tais como cstra das (' pcl« proxintida d« transtormaram ctn pastos ou arcas cultiváveis (; complexo c dC/Jclldcll- de centro: consumidorvs. Ainda cxist cn: discrcpuncias nus cst intat ivas /c do rcgit/o, Uni invcnt ario anuul de cada um destes scgmcnto-: de uso da arei: dcsmatad« obti das pelo l.andsut TM C/II [lorcsias icchadu» du dc solo iornas« ncccssurio, unia I'C~ (/IIC cada ccossistcm« [uncionu .tmazonia. hCIlI COIJIO nos cerrados, sendo (1[((' () último 1ll(ljJUlIIlCII!O dc modo distinto. .'1 niaior part« das arcas dcsmatadas t ciu sido usada foi rm/b"lo em 1994. Fstinunivas atuaix dc dC.\/IIllIaIlJeI1!o na Anur.oni« paro /}(/stugc!J/, IUIIO n)~ (/uc o gado (; lUJ/produto de alta liquides, alem nào /('\,UlJ/ cut considcraçào () corte seletivo, () quul poclc: atingir unia árcu do (/llC uiiula a niant cr a ocupu\"i/o do solo c SClI va!or cnmcrcial. ;\:'0 estimada com scnd«: 111(,/(/(/cdauucl a de toda a área dcsrnatnd«. A/h" do Ccrriulo, teu. se obscrvculo utuu mudanç« lUIS /hlslagcl/s. ant es CO/J/- mais, () cor/c seletivo. (fHC oscila entre I (/{(; IO() c:';jJc;cic:,;, atingindo uma postr, 1'0,. S(/I"(IIIUS nat nruis. ogol'o pOI' I"cgc/o\'(/o c"r()/icu atricantt. A media dc ]() ni' de madeira rctiradu /lor hectare, pod« alterar a cstruturu produção niccani zadu de so]« c; a segunda maior atividadc q uant o ao e cunncnt« iI inilcnnubilidade da [lorcstu. As qucinnulas afctcnn sobrenia- uso do solo nu rcgii/u, Est a nta tri : CUIJI/l/c"ra co mpos:« /NJr varios ncira iainda (/IfC diticil de .""(TCII}mcnsuradus) os áreas de /hlstagem, .110- ccossislc/J/us COIJl liSOS, [ttncioncnncnto e /Je'!"s/}('clil'as t ào ditcrcntc» resta, prinunias c sccundárius. 0.l(Jgo rasteiro c diiici! de ser detectado deve SCI' a bus« dc pl'ogrumas vol t ados pura () cst udo dcst« rcgit/o, por satclite-: scnd«: I}wisfi"c(/I"icn/c P'?!' ocusii/o do dcsniatc. bem COIllOna lvuscando assim a nta nutcnçá«: dos proc('"~".'...os nuturais associadas (/0 epocc! da seca, muito embora cst a» rc/a\"/)cs ai ndu não tcuh am sido cs- crcscilJ/cll!o economico c igualdade social.
eoPle are altering thc native eco.systems ofAn.lazonia sources by replacing high, species-rich forests with African and thc Cerrado region throught their efforts to de- forage grasses, by clearing forest for agriculture, through rive sustcnancc and wealth frorn agriculture, ranch- the harvest of forest fruits, nuts anel gamc, through tirnbcr Ping and timber harvest. These land-use activitics highgrading, placcr mining for gold, industrial mining, and provide an importam tlow of food, fiber and othcr products traditional and industrial fishing (1-1 I). Each of these prac- to human society, but not without a pricc. Agriculturc, ranch- tices exerts a unique intluence on the region 's biota. Ever- ing and logging disrupt the storage anel tlow of cncrgy, ear- green deeply-rooted forests are replaced by seasonally-dor- bon, water and mineral elerncnts in native ccosystcms, anel rnant Iorage grasses or short -I ived food crops (3.4), agro- may therefore damage those ecological processes that sus- extractivists deplete some game and plant species while tim- tain life in these regions. lt is thc goal of thc Large-Scale ber highgraders open the humid understories ofburn-resist- Biosphcrc-Atmosphcrc Expcrimcnt in Amazonia (LBA) to ant forcsts to the ravages of firc (1,2.4,5,6, II ). Placer min- conduct scicntific investigations that determine the ccologi- ing rclcascs scdimcnts and mercury into streams (5,7,8,9) cal effects of land-use activities in Amazonia and thc ad- while industrial mining spurs spontaneous mining-towns joining cerrado "savanna" biome. Thcse investigations bcyond the fences 01' its development sites ( 12); geleiro ice should allow for the identifícation 01' those forms of land- boats sweep seasonal varzea lakes of the físh that once sup- use that provide the max irnum flow of goods to human soci- ported traditional ribeirinho populations ( 10). ety with the minimurn disruption of life-sustaining eeologi- Just to the south and east of Arnazonia, thc giant Bra- cal processes. zilian cerrado mosaic 01' evergreen woodland, spccies-rich ln this paper, we sumrnarizc current knowlcdgc ofland- savanna, and grassland ecosystems is undcrgoing a diffcr- use in Amazonia and the Cerrado regions anel idcntify some ent se! 01' transformations, Extensivc grazing of natural signifícant gaps in this knowledge. The biggest ehallenge grasslands is giving way to cattle production on planted in this revicw is prescnted by the great diversity 01' land-use pastures, while the well-structured oxisols of the region are activities that can bc íound in the region. The managcrs of ti Ileel and fertilized for the production of soybcans, rice and rural Amazonia 's soils, Iorcsts and rivers utilize thcsc re- other annual crops. Along with thc scttlcmcnt ofthe cerrado
74· Ciência e Cultura Journal of the Brazilian Association for the Advancement of Science Volume 49(1/2)· January/April 1997 · ...... El1VirOl1mel1ta[ Impact
has come an increase in the frequency of fires that may have Vegetation Cleared as of 1991 changed the structure and function of cerrado vegetation in 600 20 ways that are difficult to measure (13,14,15,16). 18 1---- Annual Clearing 1 500 It is within the context of this complex array of Iand- 16 uses that the LBA science pIan is being formulated. The 14 integrated research programs will study the impacts of 122; o Amazonian and cerrado land-use on carbon storage, trace o 10 ",- gas fluxes, and nutrient dynamics. A synthesis of the large .'o" amount of unpublished and recently-published information 8 i on land-use is needed to inform this planning process, and 6 to improve the quality of the proposals and science to fol- 4 100 low. ln this document, we review information on the major 2 forms of land-use, their geographical distribution, and their o o Q) .", probable trends over the next decade. We also discuss these t; o, « '"E land-uses as they relate to the occurrence of EI Nifío/South- « ern Oscillation rainfall anomalies, and projected reductions in rainfall. The lion's share of research on land-use in Amazonia Figure I. The total area of natural vegetation cleared as of 1991 and lhe has focused on region's of closed-canopy forests, which com- average annual rale ofclearing from 1988 Ia 1991 for lhe two-mitlion-km' prise about 80% of the Brazilian Legal Amazon. This re- Cerrado region and lhe states of Brazilian Aniazonia. Data for lhe cerrado are from Klink et ai (95), and are based 011 land-cover trends registered in view is divided into a large section concerning closed-canopy lhe IBGE agricultural censusfrom 1970-1985. DataforAmazonian states forests of Amazonia, followed by a smaller section concern- are from INPE (18), and include areas of "old "(pre-1960) deforestation. ing the savanna-like, cerrado vegetation. The total area of Aniazonianforest clearing as of 1991 lI'as 426,000 kll/
Closed-canopy forests
Spatial and temporal patterns ofdeforestation The conversion of closed-canopy forests to pasture and cropland can be mapped using imagery ofthe Landsat The- matic Mapper satellite. Unlike the cerrado biome of Brazil, the closed-canopy forests of Brazilian Amazonia have been mapped entirely by different research groups, using TM irnagery for several time periods (17,18,19).APan-Amazo- nian map of deforestation has been prepared using the rela- tively coarse-resolution (16 krn pixels) NOAA-AVHRR im- agery (20). The clearing of closed-canopy, Amazonian forest has been concentrated in the east and south of the region, in the Brazilian states of Pará, Maranhão, Mato Grosso and LJ~~.~. -'--_--' Rondônia, in decreasing order (Figs. 1 and 2) (Table I). The two recent efforts to map deforestation in Brazilian Amazonia using prints of Landsat TM images have yielded remarkably different estimates of the total area cleared in each state. For 1988, the estimates of total area deforested j Manaus of Skole and Tucker (17) are lower than those of INPE (18) AMAZONAS PARA by 20%(Rondônia), 30%(Mato Grosso), 33% (Acre), 40% M""~ "(
(Amazonas) and 50% (Tocantins), but are quite close for Sant. Aragua~ ( ( ACRE
Pará and Maranhão. The lower estimates ofSkole and Tucker TOCANTINSi\... (17) may be a result of the coarser resolution employed by this group (1 :500,000) compared to the resolution employed by INPE(1:250,000) or the fact that Skole and Tucker (17) used a single spectral channel compared to three employed by INPE.Whatever the cause, there continue to be large uncertainties associated with this most basic parameter of Figure 2. Clearing of closed-canopy forests in Brazilian Amazonia (top land-use change in Amazonia. panel) based 011 /988, digitized tracings of clearings as interpreted from Landsat Thematic Mapperpaperprints ai 1:500,000 (from Skole and Tucker The rate of clearing of closed-canopy, Amazonian for- (/7)), and state boundaries and cities of Brazilian Amazonia and adjacent ests was highest from 1978 to 1988 (21,000 km2 per year), states of the Cerrado region (lower pane!).
Volume 49(1/2)· January/April1997 Ciência e Cultura Journal of the Brazilian Association for the Advancement of Science· 75 ·...... Environmenta[ Impact .
thcse logging sears disappear frorn TM Table 1 - Estimares of lhe arca of savanna vegeration (cerrado biome) and closcd-canopy forest (Bra- zilian Amazonia) clcared as of 1988, 1991 and 1994. seenes within three years time as logging roads and loading zones are grown over 10110'.1ofs" quare kilometers with recovering vegetation (23). 19118 1991 1994
Source: (I) (2) (3) (I) (3) (3) Fire scars Cerrado 590 620 The area of forested or agricultural land Brazil Amazonia: Total 230 280 (378)" 329 (426)" 358 (470)" that burns each year is one of the most im-
Brazil Amazonia: By statc portant paramcters nccded ror the LBA sci- ence carnpaign, since burning influences Pará 95 92 (131)" 108 (148)" 121 (160)" of'the Maranhào 32 33 (91)" 36 (94)" 38 (96)" virtually ali ecological processes that Mato Grosso 4~ 72 86 104 the campaign will address. IPE's hot Rondônia 24 30 35 42 pixel mapping of tires has provided im- Tocanrins 11 22 23 24 portant data on this issue, dcmonstrating Amazonas 12 20 23 25 Acre 6 9 11 12 that hundreds 01' thousands of individual Roraima 2 3 4 5 Iirc cvents occurred in Arnazonia in 1987, Amapá 200,000 krn' (11). Howcver, these rnaps of AVHRR U Valucs in parcnthesis include areas of"old dcforestation ", approximately prc-1960. pixcls that exceed thermal thresholds are diffieult to translate into precise arcas of firc scar by vegetation cover type because declincd until 1990191 (11,000 km ' per year) and then in- of problcrns of pixel saturation by srnall tires. the prolonged crcased during the period from 1992 to 1994 (15,000 krn- smoldering of woody biomass, and the coarse resolution 01' per ycar)( 17,18,19) (Tablc I). This temporal scquence of lhe data (I x 1 km pixels) (24,25). Forest ground tires are forest clearing estimates was madc using Landsat TM im- particularly difficult to detect frorn space. agery, and is the worlds best rccord of deforestation in a Pereira and Setzer (26) have also demonstrated that largc tropical forest region. Forest clearing trends beyond Landsat TM imagery can be uscd to detect fire scars in 1994 arc not yet published. However, the largest number of Arnazonia on the basis of spectral signatures. Using this Arnazon fires dctected by lhe NOAA-AVHRR satellitc sincc approach, Nelson (27) inspected TM prints for ali of Brazil- 1987 were registered in 1995 (21), suggesting that forest ian Amazonia with ground-truthing at scveral locations, and clearing may have increascd furthcr since the last INPE map measured >500 krn' of burned prirnary forest that had un- of deforestation was produccd in 1994. dergone crown death and another 1,000 km2 01' primary for- est that had undergone delcafing frorn ground fire during lhe 1983 drought. He predicts that these values greatly un- Whatforest alterations are missed by TM-based analysis? deresrimate the arca of prirnary forest that burned during The dichotornous, forcst vs nonforest approach to mapping this scvere, EI Nino-rclated drought. land-use activities neglccts important alterations of thc for- Using visual interpretation 01' a 1988 TM image, a one- est that are difTicult to mcasurc reliably using Landsat TM thousand km2 tire scar was mapped in an area of mostly irnagcry. In this seetion, we discuss the problern of quanti- loggcd forcsr north of Paragominas (28). This scar was prob- fying the arca of forcst affccted by seleetive logging and ably not included in the regional deforestation estimares, tire, and the problem 01' understanding the fate of c1eared since it appeared to have substantial forest rernaining. How- lands. ever, a site visit revealed that the forest rernnants visible in the satellite image were actually thc crowns ofremnant trees pcrsisting in an impovcrished forest. Locallandholdcrs stated Logging that portions of this forest had burned three times (29). The More than 4,000 km ' 01' forcst is altered through logging area of this single scar represented -5% of thc 1988 esti- each ycar in Pará state alone (see Logging and mining: Thc mate of forest clearing. In another forest ground tire ob- case of Pará statc, below). This is roughly equivalent to the served in Paragorninas, 45% ofall trees >10 em diameter at area of forcst that is cleared each year in this state (see Fig. breast height were killed (30). Forest ground fires are diffi- I), and about half of the total area c1eared annually in the cult to map using Landsat TM imagery, but excrt a pro- early 90s (18,19)(sce Fig. I). Logged forests frequently burn found influence on foresr structure and funetion. because ofthc large fine fuelloads and drier microclirnate of The tire scar near Paragominas underscores the impor- the litter laycr in logged vs prirnary forest (6.22). Estimates tant interaction between logging and forest flammability. 01' deforestation based on visual inspection ofTM prints may By opening the forest canopy and incrcasing the tine fuel capture some small portion of the very pronounced logging load on the forest floor, logging greatly increases the risk scars of the most intensive forrns 01' selective logging, such that forests will experience ground tires that are initiated in as those found within a few km of Paragorninas. But even adjacent agrieultural lands (622).
76· Ciência e Cultura Journaf of the Brazilian Association for the Advancement of Science Volume 49(1/2)·January/April 1997 Environmenta[ Impact
The írequency offire rnay inerease in the future through Land-cover change in Pará, Amazonia a positivo Ieedback with regional climate change. The re- duction in rainfall predicted by most modelling cxperiments Bragantina, 90-year·old, Small-holder Frontier 1984 1991 done thus far (e.g., References 31,32 and 33, but not prc- dicted by Polchcr and Laval (34) may exacerbate drought- induced forest burning which, in turn, may lead to less evapotranspiration, less rainfall and more firc (4).
•Forest • ore See. Forest ~ Yng See. Forest Altamira, 25-year-old, Small-holder Frontier Thefate ofcleared land O Pasture 1985 1991 By ] 994. about 470,000 krn- 01' Brazi Iian Amazonian fores! O Crops had been cleared and replaced with managed (disk-tilled, ~Other fertilized and replanted) pasturcs of African forage grasscs, "dcgraded" pastures that have largc arnounts of shrubs and nonforagc hcrbs, fields of annual crops, pcrcnnial crops and regrowing íorests on abandoned pasturc- and cropland ( 18),
KnowJcdge of thc areal extent 01' each of thesc vegetation Paragominas, 25-year-old, Ranching Frontier typcs is important because they vary greatly in the amount 1988 1991 af carbon they store or scquester from the atrnospherc, in •Forest • Sec, Forest their patterns of runoff and cvapotranspiration, and in their O Pasture & Crops patterns of trace gas tlux ( I ,3,4,35-49). ~Other In thc absencc of Amazonia-wide analyses of the areal covcrage ofdiffcrcnt anthropogenic ccosysterns, we can turn to case studics for an understanding of thc eurrent distribu- tion and temporal trends of thcse land-cover types. The FiJ(lI1'e 3. l.cnul-cover chang« in three [rontiers ofAmazonia showins; lhe I3ragantinc Zone, east of Belérn, is the oldest agricultura I lurg« arca (~r"lI('/jFe" pasture. cropland. andvoung secondarvforests itt lhe frontier in Amazonia, and providcs a glirnpse into the fu- IJragllllfil1C' zune. lhe hllge arca offon-st f/1(I/ is stillfound in lhe vounger [rontiers otAltamira and Patugontinas: and thc apparcnt tendcncv toward ture of Arnazonian landscapes that are currently dominated land lIhlIlU/O/IIJU!/J1 in thes« !5-yeul'-o/d [nmtiers. The [ores: clIleg(J/:" ill- by small holder farmcrs, and close to markets. Although cludcs arcas ofsclcctive logging. The Bragantine ZOf1e stud, used TA1 iin- this region is characterizcd as "old deforestation" by the agcrvfro,» 1984 III/d 19C)I,./ór thc countics or!'eixe Boi, Nova Timboteua and Capuncm«. Total arca \1'11,\' !3 7,835 ha.Classiiicmio» condurtcd in Amazonia-widc e lforts to rnap deforestation (17,18,19), IDR/S'! "'"S tcstcd l/si/lg a Iarnicr \ 11111J1of'35 ugriculíural plots dnring lhe analysis of Landsat TM scencs and fíeld studies in three dares (i( 111<'ill/(/gc>s, and addiüonal Iield visits (/00). The Altannra ,\11/(11· I3ragantinc counties rcveal that land-usc in this region is ('I11J1/()I'ed TM imagesfron: 1985 and 1991 in a studv arca 0(233,466 hu. quite dynarnic. Aftcr nearly a century of agriculture, pri- Classiíicotíon Il'lIS tcsted throughfield work, and through a more thcoreticul cvaluution olthe six-band spcctralputtern ofeach land-cover elas» (50,54). mary forest had been reduced to 15% of the landscape, and Tlie Pu/'{/go1l/illas studv covered 347,500 hectares on /988 cuul /991, The was concentrated along streams on sites where agriculture classification was conducted in !f)R/S/ and tcsted in thcfick! t l tt l ). was inviable. Old secondary forests (more than 20 years) cornpriscd only 23'1'0 ofthe arca, Thosc portions ofthe land- scapc that are still in use by farmcrs, including young sec- tion of sawrnills in Amazonia. The reduction in area 01' ac- ondary "fallow" forests (29%). pastures ( 13%) and agricul- tive pasture and cropland in this region may be a result of tural crops (9%), made up halfofthe land area (49) (Fig. 3), the intensifícation of cattle production on smallcr areas of In a similar changc detection study conducted ncar lhe pasturc. town 01' Altamira (50), a 25-year-old fronticr that was initi- In Paragominas and Altamira "forcst" was the most ated as a colonization project, two thirds 01' the forest was common type ofland-eover aftcr 25 ycars of agricultural ser- remaining and onc fourth of the landscape was covered by t\cment, but had been redueed 10 the margins 01' streams af .. pasturcs, crop ficlds and young secondary forests (5 I) (Fig. ter 90 ycars of settlement in the Bragantine region. The "for- 3). In this region, the total area in pastures and croplands est" category includes those forests that had becn loggcd 01' declined slightly during the 6-year study pcriod, but the por- that had experienced ground tires, which could be 1110st of tion of this arca in young secondary forcsts incrcased by thc forest in the case of Paragorninas. Sixty-six percent of 8%, This region does not have good highway connections thc land within a 25-km radius of Paragominas still supports with major urban centers, and may be undergoing a process forest covcr, but only six percent 01' this land supports forest of agricultural stagnation. that has bcen neither logged nor burned (52), In Altamira Ouring a three-year period in Paragorninas, one of the and Paragominas, there appears to be a trend toward aban- centers 01' the large, government-subsidized ranches im- donrnent of pasture and cropland followed by sccond forest planted during the I 970s, the area of forest decIined by 4% regrowth such that pastures and croplands occupied less than and the arca 01' pasture and cropland declincd by 6% (see half ofthe land originally clcared for agriculture. In Altarnira, Fig. 3), Paragorninas, contrary to Altarnira, has exccllent where small-scale farmers are prevalent, much of the young highway links with Belérn, and has the highest concentra- secondary forest may be agricultural fallows and, hence, not
Volume 49( 1/2) • January/ April 1997 Ciência e Cultura Journal of lhe Brazilian Associalion for lhe Advancemenl of Science· 77 ·...... El1virol1mel1ta[ Impact .
truly abandoned. In Paragominas, where only 4% ofthe land eredit program for Amazonian farrners and ranchers, pro- is oeeupied by small holders, secondary forests are mostly vides funds for agricultural recuperarion of lands that are abandoned pastures that also may come baek into produe- already cleared, but not for new clearing. tion through pasture reforrnation. Forest clearing also occurs where it is not planned. These and other studies (1,36,54-60) show that the phe- Expanding agricultural frontiers ean be found around the nomenon of land abandonrnent followed by forest regrowth Manaus' free tradc zone, near Santarérn, Pará and Cuiabá. in Amazonia appears to occur following ali but the most in- The railroad corridor of the Carajás iron mine has fostered tensive forms of pasture managcmcnt. such as those that in- forest clearing and charcoal produetion that was not antiei- volve mechanization and/or numerous cyeles of burning, pated in thc mines development plan (68,69). The poten- herbieide application and overgrazing. Where pasture man- tial for unplanned, spontaneous forest elcaring may increase agement has depleted soi I seed banks of tree species, ki Iled greatly in the future as industrial mines provide new foci of the residual tree roots that can give rise to new shoots through ernployrnent and forest access roads to Brazil's land-seek- sprouting. and introduced populations of seed- and seedling- ing citizenry. For examplc, a S 160 million loan was recently consuming ants and rodents, the reestablishment oftree cover approved by the World Bank (lnternational Finance Corpo- following pasture abandonment ean take >5 ycars, and longer ration) to finanee a kaolin mine in northeastern Pará that if these sites burn periodieally (53,59). would provide 7% of the world's supply of this elay min- One particularly promising approach to the mapping eral, and that will include aeeess roads, a 180-km pipclinc, of sccondary vegetation involves the modelling of spectral and cmployrnent for 5,000 laborers (12). Dozens of similar endmembers (62,63). This approach rnay also provide a tech- mining operations are nearing the end ofthe licensing proc- nique for mapping logging and fire scars that are otherwise css. lnforrnation on the spatial distribution of new indus- difficult to see in satellite images. lt is unclear if this ap- trial mines allows us to identify those regions that should proach would allow for thc satellite-based classification of have the greatest ehanges in access to forested regions and secondary forcsts by age, biornass or spccies composition wherc opportunities for ernployrncnr and land acquisition over large areas of Arnazonia, since the floristic composi- will arise for the landless poor. Most new mining opera- tion of secondary forests is highly variable frorn one rcgion tions are currently being planned for westcrn and southern to the next. Pará, northern Amazonas and Rondônia statcs, but the avail- ability 01' electricity may encourage more mining in south- ern Pará (see also Logging and Mining: The case of Pará The macro causes ofdeforestation state, below), The spatial pattcrn of forest clearing in Brazilian Amazonia has closely tracked thc construction ofhighways. The Bclém- Brasília highway (BR-O 10) in eastern Pará and Maranhào, The proximate causes of deforestation the BR-364 in Mato Grosso, Rondônia and Acre, the People clear Amazonian forests as the firsr step in pasture Transarnazon highway in central Pará and Amazonas, and formation and erop field preparation. The biornass of the the PA-150 in east-central Pará (see Fig. 2) have ali becomc felled forest is allowed to dry and is then burned, providing eorridors for human settlement and forest conversion in ru- an input of nutrient-rieh ash to the soil surface. This ash ral Arnazonia. In tandem with highway construction, gov- promotes the growth of pasture forage grasses and crop ernrnent programs have provided land or fiscal incentives plants, but this fertilizing effeet declines with time, espe- for agriculture, leading to focal points of settlemcnt and for- eially in crop fields where weed invasion and dcclining lev- est conversion. For examplc, much ofthe "old deforestation" eis of soil fertility restrict the productive period of reeently east of Bclém and in Maranhào occurred near the turn of the cleared land to 1-3 years. It is thc rapid decline in erop pro- century through agricultural clearing by immigrants from ductivity that forces small-scale, "slash and burn" farmers Brazil's drought-stricken Northcast who were provided land to clear more forest. in the region through a government setrlement program (62). Although srnall-scale farmers are numerous, and are ln the 1970s, the Brazilian government's prograrn to pro- dependent upon frequent c1caring of foresr for their agricul- mote pasture establishment resulted in clearing along the tural producrion, it is the largc-scale landholders that ap- Belém-Brasília highway in castern Pará, giving rise to towns pear to be responsible for most forest clearing in Amazonia. such as Paragominas (63,64).Numerous colonization pro- Fearnside (70) has analyzed the relative contribution to de- grams in ali statcs of the region, but particularly in Mato forestation of small holders (lots < 100 ha) vs large-scale Grosso, Pará and Rondônia led to clearing by landless immi- landholders by conducting regression analysis of ccnsus data grants that were pulled to the region by thcse programs and and deforestation estimares for thc nine Amazonian states. pushed from their land in southern and northeastern Brazil He estimates that -30% of forest elearing in 1991 can be by industrial agriculture and drought (56,64,65,66,67), ln attributed to small holders, and 70% to large-scale recent years, government subsidy programs to promete the landholders. This is not to say that small holdcrs do not colonization of areas of pristine forest have declined. Land become large-scale landholders as they accumulate land and is still being distributed by govcrnment agencies, but direct cattle, and as their household labor pool expands. financial incentives to clear forest have dried up. FNO (Fundo The logic that leads landholders to form cattle pastures Constitucional do Norte), which is the major govcrnment may have changed in rceent years. In the 1970s and carly
78 • Ciência e Cultura Journal of the Brazilian Association for the Advancement of Science Volume 49(1/2)· January/April 1997 · ...... El1virol1mel1tal Impact .
80s. large-scale landholders formcd extensive cattle pastures with cattle pastures acting as a form of "savings account". to establish ownership over land and to "capture" the esca- The trajecrory of new ciearing is tied to the geographical lating land values and government subsidies (64,71,72,73). distribution of logging, gold mining, and industrial rnin- As governrnent subsidies for pasrure formation have disap- ing, as we illustrate here with a review of logging, ranching pearcd, and as new cattle pasture managernent systerns have and mining in the state of Pará. cmergcd, the logic behind cattle pasture formation appears to have evolved. Cattle pastures are now the best mcans by Logging and mining: The case of Pará state whieh rural Amazonians can accumulate the wealth gencr- The state of Pará (see Fig. 2) is the source of eighty percent atcd by logging. artesanal mining, or agriculture - the "sub- of the timber harvested from Brazilian Amazonia. Hence, sidies frorn nature" that drive todays Amazonian land-use most of the Amazonian timber industry can bc understood systerns (72). For both large- and small-scalc landholders through analysis of logging in this castern Amazonian state. (74), cattle pastures provide only a small profit margin (less Today most timber is extracted at low intensity over than $IO/ha/yr) using the traditional extensive model of large areas of forest. This "exrensive" system is charactcr- pasture rnanagernent (75), but investrnents in pasture are ized by extraetion ofonly high-value speeies with little plan- highly liquid cornpared to investrnents in perennial erops or ning of logging roads and skid trails. The depletion of for- forest management for tirnber, The priee of beef is more est timber stocks is seen in parts ofthe estuary region of the stable than other agricultural products and the market for Amazon river where logging has been going on for centu- cattle is ubiquitous, such that the bovine has beco me a form ries (81). of alternative currency on the agricultural frontier. Cattle More recently there has been a trend toward more in- pasturcs rhercfore represent an inflation-free, liquid invest- tensive logging in old frontier areas along the Belém-Brasilia ment that strengthens legal claims to land. In addition, the highway. Here 100 or more speeies are harvested and sig- current rate of pasture formation may reflect demographic nificant damage is inflicted on the canopy and ground sur- phenomena interacting with econornic conditions at the face through tree felling and log extraction. Forest Icaf canopy houschold levei, promoting widespread shifts to investment cover deereases 50%, and up to half of the forest tree basal crops from production bascd on annual crops (76,77). arca is lost through direct harvesting, toppling or damage An important faetor making Amazon ranching an ar- (80).Unfortunately, this more aggressive harvesting has not tractive land-usc is the low price of land compared to the been aeeompanied by measures to guarantee future harvests. traditional ranching rcgions in Central and South Brazil. Unless forest management methods are implemented, tim- Land prices have increased dramatically in Central and South ber cxtraction in Pará may have a short life span. Forest Brazil duc to the increascd cultivation of agricultural ex- managernent is not utilized in most areas of Pará. port crops (c.g., soy beans). The mean land price for agri- It is tempting to describe lhe wood industry in the Bra- cultural/pastureland in the states ofSão Paulo, Minas Gerais zilian Amazon using summary statistics and averages, but and Goiás (place of origin ofthe majority of Pará ranchers), this approaeh provides only a limited amount of informa- was 3.6 times greatcr than in Pará in the early I990s, $1,541 tion about this sector. There are very fcw central tendencies vs 5)427. Low land prices and, therefore, lower costs of capi- in the industry, but there are some patterns that ean be de- tal, explain why eastcrn Amazonia ranchers can compete in tected. Five faetors influence the characteristics ofthe wood the national market for rncat, even though they are far from industry: 1) Local stand composition (especially the prcs- consumer markets (75). ence of high-value species); 2) transport options (fluvial vs The logic behind pasture forrnation continues to evolve. terrestrial); 3) availability of investment capital; 4) market The future of Amazonian cattle ranching may be found in options (domestie vs export); and 5) local socioeultural tra- regions such as Paragorninas, a ranching and logging center ditions. The role of these factors is evident in Pará state of eastern Amazonia, where serni-intensive cattle pasturc where most logging activities in the Brazilian Arnazon are systems involving disking, fcrtilization and ncw types of eoncentrated. We have idcntified five modcls of logging in forage grasscs (espccially Brachiaria brizanthav, generate this state - two in "várzea" (floodplain) forests and three in average annual profits of $50/ha (78). In this sarne region, the "terra firme"(upland) forests (80,81,82,83). 111 consid- even greatcr profits are generated by small-scale dairy pro- ering lhe impact ollogging on carbon stocks, nutrient cv- duetion (78). This intcnsification of bccf and dairy produc- cling and so forth, it will he important to design research tion systerns does not generally involvc new forest ciearing that incorporates lhe different intensities ollogging that in the Paragominas region. are found in Amazonia. Other land cover dynamics are also in evidence. AI- Amazonian logging is emerging as a major growth in- though substantial concern has focused on pasture forma- dustry. At present thc logging industry in Pará harvests some tion (given ecological impacts), some deforestation leads to 8.3 million cubic meters ofroundwood/year (lMAZON, data the adoption of pcrennial crops, which is a form of agricul- base). Considering that, on average, 20 rn' are extracted/ha, tural intensificarion. This is an important process from the some 4,150 km? of forest are logged in Pará state cach year. perspective of aehieving sustainable development (79). But the wood sector is not statie. The Brazilian economy is Hence, in the place of government loans and subsidies, projected to grow at 7% in the period marking the transi- thc major sources of capital driving expansion of the agri- tion from the 20th to the 21 st eentury. If this oceurs, and if cultural frontier today may be timber and minerais (65,78,80) the Amazon wood industry experiences this same levei of
Volume 49(1/2)· January/April1997 Ciência e Cultura Journal of the Brazilian Associatíon for the Advancement of science- 79 ...... Environmenta[ Impact .
growth, wood production will double 10 16 million rrr'/yr in ing. The opening ofnew mines in southern Pará, whcrc clcc- len years. In this case, lhe area of forest logged each year tricity is available and licenses 10 mine havc been solicited, would be approximately 8,000 km", Assuming a cutting cy- could lcad to deforestation zones siIIIilar to those found along ele of 70 years in lhe absence of forest management (84), lhe Carajás railway corridor. Mines that are locatcd on riv- the total area of lhe state of Pará nccdcd 10 double present- ers and that transport their prcduct by bargc or pipcline have day roundwood production (i.e., supply 16 million cubic lhe potcntial to incrcase the sediment load and to intrcducc meters of hardwood per year) on a sustainablc basis would toxins into these rivers. be about 560.000 krrr' (8,000 km2 x 70 year cutting cycle) or Gold mining is far more pcrvasivc than industrial rnin- half lhe area of thc stare. Of coursc, if production were 10 ing at lhe present, and may be influencing more river sys- double over thc next ten ycars, it is unlikely that it would terns, Although the rclcasc of mercury into river systerns then suddenly stabilize. If lhe projccted 7% annual growth through gold mining has bccn fairly well documented, lhe in timber proeluction continued for additional ten years, we effects of mining on stream sediment loads, stream chernis- calculate that the annual proeluction would have exceeded a try, fishcrics, and the forests that surround rnining carnps levei beyond what coulel be sustained given the availablc havc not been described in detail, An understanding oflanel- area. Therefore, although the impacts ofthis activity are use effects on Amazonian surfacc waters wil! require field relatively small now. increasingly ir is lhe drive for wood studics of this widespread form 01' mining. that wil! be a catalyst determining how A mazonia is occu- pied and lhe associated impacts. Future trends There are two additional logging-related topics that Thc present spatial distribution of resource extraction ac- merit special attention. The first is vines. The long-term tivitics determines, in large part, the future patterns of hu- sustainability of a logging econorny in Amazonia may be man occupation. In the case of Pará, the rnaps providcd in severely jeopardized by vines, This plant group is favored Figure 4 show that farming, ranching, and logging activi- by gaps, debris and stem cutting (vigorous sprouters) - ties are linked to established road networks. The areas bor- lhe very disturbances thal logging prometes. Vine cut- dering roads to a distance of 5 to 50 krn have been cleared ting eonducted in conj unction with forest management and are used rnainly for ranching. ln the proximity oftowns, might have an effect on regional rainfall. Vines often have roadside land is frequently dcvorcd to agriculturc. Logging root systcms that extend deep into the soil (10 III or more) activities are practiced behind lhe ranches to a distance of and therefore vincs might be responsible for a significant 50 to I00 km. Where wood of exceptionally high vai ue (e.g., fraction of forest evapotranspiration during dry periods mahogany) is involved, loggers may go as far as 500 krn (85).The removal of vines frorn lhe forests of eastern frorn lhe mil! in search of logs (82). In Ihese cases they ac- Amazonia mighr reduce evapotranspiration with concorni- tually build the roads themselves. tant effects on regional rainfall. At present only about 10% 01' the state has been af- Finally, just as forest rnanagcmcnt eould affect lhe re- fcctcd by these resource-use activities, but this is changing gional landscape, 50, 100, changes in the regionallandscape rapidly. Expansion Irends for the next several dccadcs will might affect foresrry. Particularly worrisome for sustainable likely rcsult in thc occupation of 40-50 80· Ciência e Cultura Journal of the Brazilian Associarion for the Advancement of Science Volume 49(1/2)· January/April1997 Environmenta[ Impact creasing pressure frorn logging, agriculturc and mining ac- soils. However, only one fourth of the state is suitable for tivities. Logging and mining are likely to be the dominant agriculture, and many of the subsequent farmer colonists activities and deservc special mcntion. aehieved only low levels of agrieultural produetivity com- Logging currently provides the second largest ineome- pared to the first wave of settlcrs. The massive influx of generating activity for the state. Tropical timber rcsources colonists led to very high rates of deforestation.Frorn 1978 continue to enjoy a warm reception on both the national and to 1993, the area of eleared forest in Rondônia went frorn 2 world markets. Philippine and Malaysian buyers now regu- 4,000 krn- to 40,000 km (- 17% ofthe state territory) as newly- larly visit Belérn and other wood processing centers such as arri ved farrners c1eared and burned the forest to prepare the Paragominas and Marabá. Stocks of valuable mahogany in soil for agrieulture (17,18,19,87) (see Table 1). southern Pará are likely to be depleted before the end ofthe The plans for the rural sector of Rondônia developed present century.At the sarnc time, lcsser-value timber will in lhe 70s included governmental incentives for the plant- increasingly bc extracted to feed thc many sawrnills already ing of annual and perennial crops. However, the trajeetory establishcd along the roadways and in urban ccnters through- ofRondônia agrieulture has now converged with that of other out southern Pará and lhe Transarnazon highway. There is agricultura] frontiers in Amazonia in that cattle pasture is no sign of sustainable logging practiccs in the region.Moreo- the 1110st cornmon use 01' cleared land (88) (Table 2). ver, if unchecked, the secondary effects of logging whieh Rondônia farmers turn to cattle pasture for the reasons cited inelude spontaneous eolonization along logging roads and above (see "Proximate causes of deforestation"), but may forcst tire, eould forcclose the possibility of future cuts. also have an added incentive in lhe substantial infrastruc- Mining has bccn and will continue to be a key activity ture for the sale and processing of beef and milk. In 1994, in Pará over the next several decades. Between 1975 and the state of Rondônia had 13 dairy units and 5 slaughter- 1987, mineral production grcw frorn 0.1 % to 7.0% 01' lhe houses (89). The arrival in 1994 of a single milk produeer state's intcrnal production. 11currcntly reprcsents the larg- in Ouro Preto do Oeste (Parmalat) inereased the price paid est incomc-generating activity for the statc, and continues to farrners for 11ilk1 by 50%. to grow in the I990s. Today, nearly 2,800 perrnits to search Hencc, after a little more than two decades Rondônia for mineral resources hold c1aim to 16% ofthe state (120,000 has becorne a clear cxample of the potential for poorly- km-), but only a small fraction of that area is being actively planned settlement schcmes in tropical forest regions to pro- exploited. The major limitation to expanding activities in mote rapid and extensive losses of tropical forest while pro- formal mineral extraction is the activity's depcndenee on viding little irnprovcrnent in the lives of those people that c1eetricity. In southeastern Pará.blesscd with electrical powcr lhe schcmes were designed to help. 111an effort to mitigate supplies and roads, mineral extraetion activities will con- some of these ill effects of Rondônia 's settlcment scherncs, tinue to expand in both the industrial and the gold mining the state government of Rondônia, with support from the scctors. Southwcstcrn Pará is less devcloped today, but plans World Bank, has now launched a new program of agro-eeo- cxist to build roads and provide elcctricity in the near fu- logical zoning of the state in an attempt to eontrol lhe un- ture. Although much mineral prospccting has bcen eonductcd regulated expansion of agriculture and logging in the re- in northern Pará, thc lack of power lines makes extensive gion. This prograrn, ealled "PLANAFLORO", is also in- mineral extraction activities unlikely in the near future. tended to improve the state's rural infrastructurc, to provide As Pará's natural resources beeome exhausted, urban credit for small farmers, to demarcate and ereate extractive migration, poverty and ccological stress increase. In 1991 reserves, and to cornmunities of Amerindians. The perform- Pará's urban population accounted for more than 50% of ance of this arnbitious zoning program should provide valu- the statc's approximate 5 million people (86). This is the able lessons abour the potential for this approach to rural first time that urban populations are larger than rural land-use planning to direct the course of rural development populations in the state. As a reflection of this growth, in Amazonia. populations in urban centers in Pará have more than dou- bled in the past 15 years. Emergingfrontiers: The case of Acre state Acre statc, in western Arnazonia, is symbolie of regions of Government colonization projects: "incipient" frontier expansion. While Acre was linked to Tire case of Rondõnia state the rest ofBrazil only by an unpaved two-lane highway (BR- Land-use in the state of Rondônia (sce Fig. 2) has been domi- nated by large, government-sponsored projeets designed to Table 2 - Major land-use activities in three differcnce areas in Rondônia eneourage farrncrs to settle in forested regions and, more (49). Mean arca per property ofthree land-usc practices. rcccntly, designed to fix the problcms that arose frorn these settlement schemes. Colonization programs in Rondônia Citv Average area Annual Perennial Pasture started in the early 70s, and suecessfully attracted thousands ofproperties (lia) crops (lia) crops (ha) (ha} offamilies, mostly from southern Brazil (56). The first wave Rolim de Moura 80.0 5.4 3.7 26.7 of settlers was given foresr plots situated on good soils, and Ouro Prelo 73.0 6.4 6.0 23.0 the news quickly spread that Rondônia was a land of rich Alto Paraíso 80.0 6.1 12.9 13.9 Volume 49(1/2) • January/April 1997 Ciência e Cultura Journal of the Brazilian Association for the Advancement of Science» 81 El1virol1mel1ta[ Impact 364), the influx of settlers to the region and the associated mable and are invaded by grasses (4,31). The recent, fre- forest c1earing were slow. As of 1994, only 12,000 km2 (8% quent occurrence of ENSO events, and the predicted reduc- of the state territory) had been cleared (18,19). Acre has tion in rainfall for Amazonia (31,33,94) point to the need to now been selected, however, as the "gateway to the Pacific"; understand forest responses to seasonal and interannual an all-weather road may connect Brazil to the Peruvian sea- drought. coast via Acre in the coming years.Hence, Acre represents an opportunity to accompany the process of rural coloniza- tion and frontier expansion at an early stage. Recommendation Acre also contains some of Amazonia's most impor- In sum, TM-based, Amazonia-wide analyses offorest c1ear- tant experiments in land-use categories designed to recon- ing are needed as a first order estimate of the general trends cile the rural economic development and forest conserva- in Amazonian land-use. A fuller understanding of the tra- tion. Two of the largest extractive reserves in Amazonia, jectory of Amazonian land-use, however, will require infor- the Chico Mendes (one million ha (I)) and Upper Jurua mation from other types of studies.Land-uses such as selec- (about 500,000 ha) are located in Acre. These reserves al- tive logging, the escape of accidental fires into logged and low the communities of rubber tappers and Brazil-nut prirnary forests, and secondary succession on abandoned land harvestors that live within their boundaries to gather forest can alter the carbon storage, nutrient cycling, hydrology and products and practice shifting cultivation, while placing the flora and fauna of Arnazonian forests, but are inad- constraints on the amount of forest that can be converted to equately deJected by Thernatic Mapper-based monitoring. agriculture. Some of the key information needs that could be addressed by the LBA science campaign for Arnazonia include the fol- lowing. The climatic stratification ofresearch in Amazonia Virtually alI ofthe ecosystem processes that will be studied • Mapping of deforestation across Brazilian Amazonia, by the LBA campaign are affected by rainfall regime. Sea- is conducted at 2-3 year intervals by INPE and should be sonal and interannual drought influences net primary pro- continued. It should be extended to non-Brazilian Amazonia ductivity, evapotranspiration, runoff and streamflow, nutri- and to the Cerrado region, it has never been conducted for ent export, and soil microbial processes. The research de- ali of Amazonia using Landsat TM. sign of the carnpaign should be stratified to include gradi- •Mapping of the expanding frontier of selective log- ents of rainfall seasonality and interannual variability. ging, classified according to logging intensity. Such an ef- Both total rainfall and rainfall seasonality vary from a fort would require information on the spatial distribution of general gradient that runs frorn northwestern Arnazonia (high rain and seasonal) to southeastern Arnazonia (low rain and seasonal). In São Gabriel de Cachoeira (northwestem Amazonas), annual rainfall is -4000 111mand >100 mm each month; in Santana de Araguaia, at the transition between c1osed-canopy forests of Àmazonia to savanna ecosystems of the cerrado, annual rainfall is -1550 mrn and there are 5 months each year with <50 mrn of rain (Fig. 5). Forest c1ear- ing has been heavily skewed toward these areas of seasonal drought; the two million km2 of closed-canopy forest that have at least 3 months each year with <50 mrn of rainfall contain 75% of deforested lands (3). The major source of interannual rainfall vananon in Amazonia is the EI Nifío/Southem Oscillation phenomenon (ENSO) (90,91).ENSO leads to greatly reduced rainfall (>30% below average) in eastern Amazonia (92). Based on a model that tracks plant-available soil water (PAW), pre- cipitation, and the minirnum rooting depth necessary to sup- ply evapotranspiration, we estimate that the evergreen for- ests in this region deplete PAW to soil depths of >5 m dur- ing ENSO events (93). Field measurernents during the 1992 ENSO event found PAW was depleted to >8 m soil depth at Figure 5. Map of Brazilian Amazonia (Legal Amazon) showing lhe gradi- enl ofdry season intensity from lhe northwest to lhe southeast, At the bor- one site near Paragorninas (3). der between closed-canopy, evergreen forests {ligh! green) and savanna A second focus of research on the effects of drought on (yellow), there is an average of 0.5 111mofrain per day during lhe driest forest function should be the zone of transition between three-month period of lhe year.Amazonia S evergreen, closed canopy for- closed-canopy forest and savanna vegetation. A general re- ests that are subjected 10 intense drought during lhe dry season rely 011 water uptake from deep in lhe soil profile 10 mainrain evapotranspiration duction in rainfall could lead to a shift in the position ofthis (3). Most of the area of human settlement in Amazonia has been in regions transition, as drought-stressed forests become more flam- of intense seasonal drought, From Nepstad et 01 (3). 82 • Ciência e Cultura Journal of the Brazilian Association for the Advancement of Science Volume 49(1/2)·January/April1 997 ...... El1virol1mel1ta[ Impact . sawmills, their production rates, and their harvest intensities, (cerradão) that has disappeared through the increased fre- and might draw on radar and high resolution, aircraft- quency of tire or was it primarily open savanna (cerrado rnounted sensors. stricto sensu)? There has, nonetheless, been progress in • Mapping of fire scars. Such an effort may require rnapping the land-cover of the region.EMBRAPA-CPAC high resolution mapping (pcrhaps from aircraft) with ground has rnapped the land-cover (native vegetation, pastures, truthing over sarnple landscapes. Forest ground fires are cropland) of Goiás (which surrounds Brasília), and south- particularly difficult to map, and may be affecting a large ern Maranhão using Landsat TM imagery and ground- area of logged and prirnary forests. Moreover, a mechanis- truthing. Cerrado regions in Mato Grosso, Tocantins and tic understanding of the role of drought and land-use prac- Minas Gerais have not been mapped. The estimates of na- tices is needed to better project the future importance offor- tive vegetation clearing presented in Figure 6 and Table I est tires. were based on agricultural census data (95). Further rnap- •Modelling ofsecondary forest dynamics. Information ping of land-cover in the Cerrado region is needed. is needed on the rates at which agricultural land is aban- The lack ofTM-based maps ofland-use patterns in the doned to secondary forest regrowth, and the arnount of tirne cerrado region has perhaps masked the considerable impor- that passes before these secondary forests are brought back tance ofthis region for the questions posed by the LBA carn- into prcduction through new clearing. This information will paign. The areal extent of pasture- and cropland in the probably require a combination of remote sensing studies in cerrado is greater than that of Arnazonia (Figs. 1 and 6), combination with ground surveys that characterize the land- with large but poorly understood implications for the tlux of use strategies adopted by fanners and ranchers.Measure- methane, nitric and nitrous oxide, COz and water vapor to rnents of secondary forest biornass from space using radar, the atmosphere (96,97,98). for exarnple, would greatly facilitate this research effort. Such Cerrado has one of the richest savanna tloras in the studies should be conducted by mulridisciplinary tearns that world. According to Heringer et ai (99), of the 774 woody can couple ernpirical measures of land-cover change with species, 429 compose the proper tloristic stock ofthe cerrado, socioeconomic studies that reveal the logic ofthe land man- 300 species belong to forest forrnations and the rernaining agers that are causing this change. 45 to other vegetation types. This number ofthe 429 unique • The design of the LBA carnpaign should take into woody savanna species is not matched by any other savanna consideration the spatial distribution of rainfall with and tlora in the world. The nurnber of herbaceous species is still without ENSO events. unknown. Cerrado vegetation Cerrado occupation Cerrado exploitation started with the Portuguese searching The Legal Arnazon of Brazil contains 850,000 km2 of cerrado for precious mineraIs and lndians for enslavernent in the vegetation concentrated in the states of Mato Grosso (43%) 16th century. The first perrnanent settlements were estab- and Tocantins (29%) (17) (see Fig. 2). This vegetation is lished in the early 18th century and were associated with conspicuously absent from the efforts to map deforestation gold mining. Some farrning developed among the mining in Arnazonia, even though the Tocantins and Xingu rivers find their headwaters in this biornc. The cffects of land-use on these large rivers have not been studied, even though the 100,------, areal extent ofnative vegetation clearing appears to be larger • Crops in thcse drainage basins then anywhere else in Legal !l!ICultivated Amazonia. 80 pasture Taken as a wholc, the cerrado biome is the second larg- <õ [j Clearedbut s: not in use est of Brazil, after the Amazon rain forest, representing 22% Õ IJ) 60 ofthe country, or approximately 2 million km-. It is a tropi- t: cal seasonal savanna, with a continuous layer of herbaceous .2 species (mainly C, grasses) at the peak of the vegetation :§. 40 tO growth, with scattered shrubs and trees that sometirnes forrn Q) a continuous canopy. lt has a characteristic tlora, which dis- ~ tinguishes it frorn other Brazilian biornes, such as the Ama- 20 zonian and coastal rain forests or the northeastern dryland "caatingas". Savanna transforrnation through land-use is difficult o,_,_,_,_,_,_,~,_,_,_,_,_,_,_,_2= to quantify using a dichotomous, TM-based approach be- cause the spectral differences between natural and agricul- Figure 6. Projections ofland-cover change in lhe cerrado biome hused ()J1 tural vegetation are more subtle than in closed-forest re- government census statistics through /985. The arca ofcultivoted (planted) pasture in /985 (31J.IJIJO.OOIJ hu} is approxinuuelv equivalent 10 lhe total gions. Moreover, there is some debate regarding the "origi- area of forest cleared in lhe closed-canopvforests ofAniazonia as o( /988 nal" cerrado vegetation; was it closed-canopy woodland (95) (TaNe I) Volume 49(1/2)·January/April1997 Ciência e Cultura Journal of the Brazilian Association for the Advancement of Science· 83 · ...... Environmenta[ Impact communities but cconomic activity shifted to cattlc raising tares to 38 million in the same period. Thc delay in the re- with the exhaustion 01' the mines. A new wavc of explora- lease of census information makes it difficult to determine tion began in the middle 01' the 19th century aftcr the Para- cattle growth in recent years. guay war and during the 1920-1930 period, stimulated by Projections of the evolution of the main crops, culti- the coffcc boom in São Paulo statc ( 13). vated pastures. and opcncd arcas. indicate the futurc of land- Two main factors wcrc responsible for the modern oc- use in the cerrado. In 1985. 51 million hectares of Iam! had cupation of cerrado. The construction of Brasília, Brazils been cleared in thc rcgion. It is estimatcd that the area 01' new capital. in the late 50s. and the adoption 01' develop- cleared Iam! in 1994 was 70 million hectares. equivalem to rncnt policies and investments in infrastrueture between 1968 39%) of thc total arca of the region. Considering appropriate and 1980. The construction of highways allowed the occu- conditions for agricultura! cxpansion, good market condi- pation 01' space and thc cxpansion of commercial agricul- tions, and changes in economic policies. the total area of ture in the Cerrado. clearcd land in the cerrado will be 88 hectares in 2000. rep- Thc policies with the highest irnpact on agricultural resenting 49% of cerrado Iam! surface. expansion wcre subsidized credito tax excmption on agri- The transformation of natural landscapes in the cerrado cultural activitics, minimum prices for crops, subsidies to is causing water pollution, soil crosion, fragmentation of fucl, and the devclopment of farrning technology suitablc to vcgctation, extinction ofwildlife, invasion ofexotic species, the region 's climatic and soil conditions (95). Subsidies have and modification of the tire regime ( 13). favored commercial crops at the expense of staplc crops and the environment. Since concession loans wcrc available in proportion to crop arca (not production), they had a dra- Recommendations matic impact on the nature of dcvclopment and the area clcarcd ( 13). The lack of information on land-use in the cerrado biorne is As a consequence 01' the expansion of agriculturc, total rernarkable given the amount of attention that thc scientific cerrado population grew from 6.5 million in 1970 to 12.6 cornmunity has given to the closed-canopy forest biome of million in 1991. Throughout the region there lias been a Amazonia. Thc LBA science campaign should complete TM- strong trend towards urbanization and decline of the rural based mapping of land-cover change in that area 01' the population. Almost a fourth of the population concentrates cerrado biomc that is within, or closc to. the Legal Amazon. around the metropolitan arcas ofBrasília and Goiânia, Goiás Similar to the recommendations for the closed-canopy for- state capital. ests of Arnazonia. information is needed to predict the fu- ture geographic trends in cerrado vegetation convergion, and the logic that is adoptcd by the farmcrs and landholders who Cerrado land-use managc this vegetation. The scicncc campaign represents Until 30 years ago thc region was used prirnarily for exten- an excellent opportunity to analyze the changes that have sive cattlc raising. Today it is estimated that 37% of its natu- taken place in this vast biorne, much ofwhich drains through ral vegetation has becn transformed into cultivated pastures, the Amazon cstuary, and most of which has already been crop ficlds, dams, urban scttlcmcnts, and degraded areas ( 15). severely altcrcd by land-use practices. • The most signifieant forrns of land-use are cultivatcd pastures and comrnercial crops, mainly soybcans, corno ricc, coffee, beans and manihot (scc Tablc 3). Cerrado produc- References and notes tion of soybeans in 1994 was 8.8 million tons (a quarter of the national production). Cerrado com comprises 16'% of I. Brown Ir. D Ncpstad, 10 rires. L Luz. A Alcchandre 1992 Carbon the national production, rice 13%l. co ffe e 8%). beans I 1% siorage anel land-use in extractive reserves. Acre. Brazil. Environ and manihot S'Yrl. Conserv 19: 307-315 The cattle hcrd also displayed substantial growth be- 2. Ncpstad De. IF Brown. L Luz. A Alcchandre, V Virgilio 1992 Biotic impovcrishmcnt 01' Amazon forests by rubber tappcrs, loggers and cat- tween 1970 and 1985: Frorn 16.6 million to 38 million ani- tlc ranchers. p 1-14. /11 Non-timbcrforcst productsfron: tropical trccs: maIs. Cultivated pastures have grown from 8.7 million hec- Evutnation ofa conscrvution and developtncnt stratcg): De Ncpstad. S Schwartzman cds. Thc New York Botanical Gardcn. NY _'. Nepstad DC, CR de Carvalho. FA Davidson, r Jipp, r Lcfcbvrc, Gil Table 3 - Agricultural production in the Cerrado and in Brazil in 1990. Negreiros. ED da Silva. T Stone, S Trumborc, S Vicira 1994 Thc role IBGE 1990. citcd in Klink ct ai (98). ofdccp roots in thc hydrologieal and carbon cyclcs ofAmazonian for- ests anel pastures. Nature 372: 666-669 Crop Production in thc Production in Production 4. Nepstad DC, r Jipp, r Moutinho, G Negreiros. S Vieira 1995 lorcst Cerrado Brazil Cerrado: rccovcry fol lowing pasture abandonrnent in Amazonia: Canopy aooo tons) (I ()()()tons) Brazil scasonality, tire rcsistancc and ants, P 333-349. 111 Evaluatin.: 011<1 inoni- lol'illg lhe hcalth oflarge-scal« ecosvstenis. D Rapport cd, NATO ASI Soybeans 5.050 19.890 0.25 Series. Springer- Vcriag. NY Com 3.400 21.340 0.16 5. Uhl C. () Bezerra, A Martini 1993 An ccosysrcm perspective on threats Riee no 7.420 0.13 to biodivcrsity in castem Amazonia, Pará state, p 213-231.lll Pcrspcc- Beans 240 2.230 0.11 tives Ofl biodivcrsitv: Case snuiic» o(gcnclic rcsource ('()I7SCITOfiOI1 Coffee 250 2.930 0.08 and dcvclopmcnt. C Pottcr, [) Janczcwski cds, AAAS Press, NY Manioc 1.260 24.310 0.05 A. lJhl C..IB Kauffman 1990 Deforestation cffccts on fire susceptibility 84· Ciência e Cultura Journal of the Brazilian Association for the Advancement of Science Volume 49(1/2) • January/April 1997 El1virol1mel1ta[ Impact alld lhe pph,.'llli:!I rc-,pOIlSL· pr,rL'L' :--p"..'I..:iL·-, \l) lirc in thc r.un J(lI'L'Sh 01'1111...' .~7 lillitn:lr:-Ie:-. \\'\'\. P[\,I lcaru ....idc l{jl.)-t 1/'01"-,<":'1'01111" bionui«, of '('('- l';l:-,H.:rll :\Ill:\/Pll. F,·%.'.!.,\" 71:.-.L~7-...J-l1.) IJlld(l/:l·/O/"('.\/ \·('.!.!.l'!tlfioll;1I u/!ullt!oll('d li(l\flfl't'\ ;/1 UnciliclJl IlJ/tCfJIÚ" - I'li:ilh'r \\ (',l.I) 1,:lCL'rlb I 'IXX \krclIr\ inpur- il110IhL' ,\Ill:l/OI1 rc- -I'11]lllhli,hcd ~illll. Ih;lIil. r'II1'inJII J~·(1J 1.('11 I): 32:'-.~30 3~, 11<'1I;!h10IlR,\,.1 I 'nruh. I' l.cicl» rc I'I'I.~ CUrrL'111l.nnl-cov cr 111IhL' X, \l:ihl1(I, \\( l'kirkr,C'\I\1 SOIl/:I, RRC'lIlhcr I'I'I() \krclIr"polhlliol1 t1'npíL':--:!nd il:-- po1l'lltial ruI' ;-..~qlll'"tl'rillg curbon. (;J(JhClI NiIJ.'.!.I'O'-!ll'lIl dlh.: IIl12ídd lllil1i!l~ in til\..' 1\·1;\(lcir:t Rrx cr B""in. (h~I/i1 .. IJIlh;') 1l): 11-1:' (:1'('(('.\ 7: .''fJ:'-.'.20 'I, vt.minclhI ,\, .I1{ I c'rrL'i1':,I1 ll{ h1r,hL'rt', RI \'icl"ri" I'IXX '\krl'lln Y). Sl1roL'lkr J>..1 \Yilljllll \i){);=i :\ •.....I...>,:--ingIh'l/il·~ rurbun hlld~l't. I. BillllC puo!s. /-iJ/"('Sf /:"ell( \Itlllll,':..!. t:Olll<1lllit1:\lillll in ihc .\111:1/011:;1gdld ru....h .l'(lll"iL'qIlCIlCL', uubi«, 17: 2:'2- c.ubon 75: 77-'i':dl 2:,-l -111,I'"hcr \1.1,1\1 R'I\', \1.\ '\,'ar/;l, CI': l.a-cano. .11S:lII/, R.I 1111'"1;", I{ 111,\1<:(ir:llh D, I, C'c"lro,(' lurcnuna. Il ,'\111:1""1,I.(',,!ahri:l 1'1'1.1lishcr- \ l.'r:\ 19l)-~ I.k'l'I'-rUulL'd gra ...•L'..-,.•;-.;tOI\c:arhou in SOlllli .\tllL'riL',Ul -,oil-; ic -, .uid thc L'\ 0111i1ou 01" I"L'SOUrCL' 1lI:lIla~L·tlll·llt 011 th,: 10\\ cr AIl1 \'(/l'{//III(/S. !:1'(}!/lIlIli(' drtl·lll,'..!, /o/,('('s. ('('ulu,t!..i('ul ('oll\/nlil//\' (/lIi/f'u/i('_, :I1Pllg l'lll\.'''iI-l\l-p;\SllII'L' l'hrullllsl'ljllL'lll'L'S in thc \\l'..;kl'lll-h:l/ili~l1l i\111~\- opt i o n» IUI' s/l\I(flI/U/l/(' l.nul-u.«: 1)Í\;1 'Últlll!:!, O'l Splhrig l'(j-.;. \1.\H /(lll Ilil'"1. 1),',ol(/,.;i" ')'1: I 12-1 17 "LTil' -, 1:2.ParlllL'IH1ll l'uhl: ..h..ill. g. l.ondon -I~ l'i,'c"I<1 \IC ,( \clll, CC CcTri 1')'1-1:-':,'1uiuoucn uuncrulivatiou '"Ie1 1-\, l)i;l •...1~1 1\)l)2 (·(,!"I·lIdo. ('1/111 ClII'(/I'!('ri:o\,t/rI. lI/t'rIllI/;I'IIS d(' 111..'1 11ilrilil';Itillll d('\('//I'(/!l'ÚIl('Il!O do ('('/nulo \1({fI('jo t ' C(lI/,\('/"\'({(/U do\' ('('cllI'StI_' SI,il Ih2: (,1-7il 1/(/1/1/'(11\ 1'<'/I0\,(;I','i\, I' 11-2,', ])1" H r\,\TI'I{,\,IIL\\1 \, IklSili" -1(" '\L'ill c. \IC' l'icu,ln, 1',\ Srcudlcr..1\1 \:klilln, Ill'c'i,:1. ('C Cl'rri I ')'I~ I ~, I )í:l:-- I{t' II.)\)-l ('llllSL'1'\;1\.,;;11'lI:1 n.uurcv.: no CL'IT:Hlp. ~X_~-(l-l-O. 111 \itrtlgL'11 LI: n.uuic ..•in soil-, ot linl..'-,i and acli\ c p:I:--lurL's in lhe \\L' ....tem ( crr.ul«: t·C/l·tlchT/:U{"t[o. (/("II/hl(ÚO(' 1'(·I"\I"'dil·u.\. \P\ Pinto c.L Ilr",iliill1 ,\II1a/OI1 Jl:l'"1. .\oilllioll!io'{,"1I1 27: 111l7-117_' Ih;\ -,íli:\ l 'lIi\ l'r~il~ Pres:-;. Br'l:-.ília -17, I'L'it'1 11.1.,1\1 \klilln, CC ('L'ITi I')l)~ Ch:ll1t'c, in ihc on,:il1 .md rhc lju;\l il: I ;...{liIllr~:l1lil' Jll:llh.'r :lnd ir-, :--i/L' rral'l i011:-a-ücr dclorc-a.niun lor I", (Illl'dl.'rt \\".1 I\):-(, \-bll:l.;.!l'lIll'1I1llrthL' t,xrr:ldp ",li...; nfBr;l/i1: ,11\.'\ IL'\\. Ir .I \,,,il Sei 3-1: -I{)~--I2X pa-,IUl'L' introduvtion in Rondóni» (I h:I/i1). Nio'..!.('()(-!/('Illi.\"//T 17-:--: 21<~() 17, Sknk I). (' IlIl:kl'r 11)1).' Trupll·allkrlll'L':-.tatioJ) :lIhl h.rbii.u fr:!~lllL'll- -I~, I lodncu \ I. \ I (h :Imn, ,I'1',"'1;1,,'1];(,\ dc'() \ larquc hlho I 'I'ill Com- r.n ion iu 111..' 1:.\Illa/oll :-;:IICllik d~II:\,'rnlll 1')7X lu I')XX .. \'(';C'Jli.·,, 2(,()' 1)~lri""olb llrtollg-Icrm soil \\ :11('1' ...;1(lr:l!;Lh' L'h:l\ ior undcr pnxturc rj,oll).t!,l'cllll FII.'.!.;IJ /«'/JI .\('11\ (){): 5-t 1-551 111/11/1:',',,1 c I: 1-21 21 Sc'I/,'r \, \ I I'CIL'lr:l - I 11]'lIhli,lll,,1 ,];(1:1 '2, .vlcnc.n ,\,\, IlC '\L'phl:lcLI \klldnl)/,1.I) Rilll:l, li, 1\1'\1\\\1,I' I.ckll\ re lhl ( I{ 111I,c:hh,"'hcr I<)~'\ di,llIrhillt' "IlL'rt'i,," hL'I\\L'Cl1c':lllk Il)l)7 ;'().'.!.u IIlfIJJIlI::()lIilf, ('11/("0 ("fllt/OV /h' ('I (\'/), \\"\)rld Ibllk. 1'11\11 1':111L'l1ill1h:u2rllillg pral"lil'l'''i ,lllll sL'k'l'li\'l' lI'L'L'11<11'\L:slill~ ill1hL'!-.:ISlL'rll jlro~r~11l1lix 1he ( ·011...1..'1'\':11iUll ()r IhL' I~I';ll.i Ii"lll Ri! i'1 h )I'l''';\'Ih:1SíIi;\ .\111:\/1111. IJiO/l'U/Jil'iI I :2h~-.2h~ ',;, \L']1';I"d DC, (1,lhl. (\ Perelr", ,1\1( d:l Sih" 1')')(,\ u'"1]1:1r"ll\c' I L·ji,.'l)\I'e P. T SII)lh.' 1\)l)-t \1(}lIil(jl'illt~ ,\t)C('/il'(' (1I,'..!.,'...!.ill,'..!.;/1 l'({\/t'/"II :'lLIdy urtrL'L' L'slahlishllll'1l1 111;dl;\lldlllll'd p:l"iIUrl' alld lllillul'l.' r(lI·I..'S!()j' /1,'o:lli(/1I 1/1/({:()/Iia lI,ill~~ 1I11I/li-/('III/)01·1//1.01/(/S(// 1l1,'IIIi/li('\/If/J!Jl'l' L'~hILTll :\ma/unia. ()dos 7h: 2~-Y) ~-I, \Iall,cll', y \\u, Y 1.1.1 \lur:lI1, L Ill\,"di/i" 1'1'1,1\]'L'(I"III.I,'''llli'':I- flll",""/'I', ISI'I 1'1'\)11hi\l1 !l1ihS hllrllill~. ('1illllllil ('1/llll'.!.,' l-L ~-t,~-2h:2 (;<'()('clf"/() IfI! K: I-li ,~, SI,"1c'Ti\, (iM\\","hc'c'll. I{A 11'"I,!lII"\1 I')K'I I r"plc:l1 ,kl,'rL''':lI'''11 111 2:'. I{I lhi 11....•1)11.1IIF) I 111'1r..n' H1I "iP:ll'l" 1;!.lllha11I rI..'1..'\ ;dU:l! lU 11usill~ i!l rr:1n,~'d I'elllnle ...1.... ·Il:--iIlS.11//'/1(('/11 _\('fI\ l~: .'-~-l 1""':I,lh,,/iI:;\l1al\',i,\\llh 1,,\'\I)S,\I ,,,,d ,h 111.-1I il",,~illt' 1<;ld:l'-,\,]1 I\TL'ira \ 1< .\\\" Sel/cr Il)').')Spcl.:lr:t1 cllar:u.:k'ri:--lic ...•llr rire :--L'ar:--ill 192-19.':;;. 1'l'oc<'(·(IiIl.'.!,s f(;. IN,\"\" ,\ ol".\Ill:I/I\1I dL'Ii.\rL....I:lliol1. \'U/IIFt' .~-t2: -t 11--lI." :\Ill:lIolli;lJIland:-;c;ljlL': 11.11\.'r:L'-.1l'O\ L'r: :llld ;\~ricl1ltllral rL':'({lralioll. IIII"io 1'"kIIL'r,I," 1:1,:11 1'1'1-\'I hL'illl]':lCI ,,1',\ !'riC:I11,,"d ,\IIIa/Olli,," de](,r- 20: 2-lx-2~:' l·....I:ltinn 011Il'tlpil.::d c1illl;lk . .Illidr(J115~: .~:\l)_-t():' (,O, \d""".II\. \' I':lp""\10 -;Illilll. R \llIlci,i;, l'ilh",\R (idk,piL', 11..\ S,d,llll;'lIl I{ P. 1)( '\CpSl:HI. 1< ViL'Ir" Il)97 Hit11ll;h ...a• C estoqllc (k RI)hl..'rt-.. II}l)() .\ Ill'\\- 1_;!lhl..a..l. \ lL'\\ o" \;!Ild·u ....c .ill .\IIl;:v{llli;1. 111/ In" 1.::1rhplh)lk !lurcsLI ...lr. ppic:\is pril1l;Ú'Ii1l' Sl~L·1I1hl;'ln;1.111prc"is. 11/ /-lnn'_\/u I'flu(u'.!.,.""l U('1I1 '\('IlS .2X: 177-1:-:5 III/,/:U'I/('" /);II(;lIllj'lI, I't·'.!.I'lIt'!·UVIIIJ (' 1I/(II!<'jtJ, ( (l:l:--l'Oll. 111( (,I ..\d:llll' ,lIl, I) S:lb,,1.\ ":11"'" R ,\IIllc'id:l I ill"" 1)\ ]{"hL'Ih,\ j( I\lIlilh, \I{ (lilk...;píL' I \)I)~ (!:h"ilíl':11hiJ) \\i'llllll1hj1L'L'I1';,1 illl:l!..'.('...h. :bL'd PI1I'r 1\;)liI1 Cienc;a E' Cultura ,}oflrni1/ of the Brazilian Assocía/íon fo!' the Adv8ncefnent of SClencc· 85 El1virol1mel1ta[ Impact 62. Vieira IC RI' Salomão. N Rosa. D Nepstad, J Roma 1996 Rcnascimcmo cial, cconomic, and ccological conscqucnccs ofsclccrivc logging in an da floresta no rastro da agricultura. Ci l Ioie 20: 3X-45 Amazon frontier: the case ofTailândia. Forcst Eco! !'v!lIIwg 46: 243- 63. I3rowder JO 19X5 Public policy and dcforcstation in the I3razilian Ama- 273 zon, p 247-2'!7. 111Public policv anel lhe ntisus« II/fim'sl rcsourccs, R X4. Barreto P, C Uhl..I0 Yared 1993 Potencial de produção sustentável de Rcpctto, M Cíillis cds, Cambridge Univ Prcss, Cambridge madeira em Paragominas, Amazônia Oriental: considerações ecológicas 64. Hccht S, R Norgaard, (j Possio 19X5 Thc cconornics ofeattlc ranching c econômicas. Anuis do J o. Congresso Floresta] Pananicricano. P 3R7- in eastern Amazonia.llllerciellcia 13: 233-240 293. Curitiba 65. Schrnink M. C Wood 1992 Contcstcdfronticrs, Columbia University 85. Restem T 1996 Contribuiçáo dos cipiÍs a cvapotrunxpiracito de Prcss. NY florestu sccundáriu 1111Amazouia oriental. 8e!<:II/.Brctzil, Ms Thcsis, 66. Moran F 19X I f)el'ellllli/,g lhe 1111/1I~oll.Indiana Univ Prcss. Universidade Federal do Parir, Bclcm Bloomington 86. FIBGE (Fundação Instituto I3rasileiro de Geografia e Estatística) 1993 67. Fearnside pM 19X6 111111/(/11CW'ITillgcupacitv II{II/(' Brazilian rain- .tnuàrio est atistico do Brasil, Rio de Janeiro [orest. Columbia Univcrsity Prcss, NY X7. Stone TA, IF Brown. CiM Woodwell 1991 Estimares. by remete scns- 6X. Andcrson AI3 1990 Smokcstacks in the rainforcst: Industrial dcvclop- ing, 01' deforestation in central Rondónia. Brazil. Forcst Eco! !'vIII11 aI'. ment and dcforcstation in thc Amazon Basin. Work] Dcvclop I X: I 191- 38: 291-304 1205 88. Browdcr JO 1993 A sitnaçiio [undiariu c agricolu 110 estado d« 69. Fcamside PM 19X9 Thc charcoal of Carajás: a thrcat to the forests 01' Rondoniu: Resultados preliminares de 1111/ projeto de PCS(f"i.\'II- Un- Bravils custem Amazon region. Ambio IS: 141-143 publishcd 70. Fcarnside pM 1993 Dcforcstation in I3razilian Arnazon: The cffcct of g<). Matrieardi EAT 1994 Pecuaria in Rondônia. Consultant RcportIor population and land tenure. Antbio 22: 537-545 the Virginiu Polvtcchnic lnstitute and Stut« Lnivcrsitv, Virginia 71 Hccht S 1993 Thc logic of livcstock and dcforcstation in Amazonia. 90. Ropclewski CT MS Halpert 1987 Global and regional seale precipita- Bi"Sci('//ce 43: 6X7-695 tion patterns associatcd with the EI Niúo/Southcrn Oscillation. Monthlv 72. Hccht S. AB Andcrson, I' May 19SX Thc subsidy frorn naturc: shitiing Wcather Rei' I 15 1606-1626 cultivation. successional palm Iorcsts, and rural (Jevelopment. HIIII/ 91. Diaz HF. e; Kiladas 1993 Atmosphcric tcleeonnections associatcd with Organi: 47: 25-35 cxtrcmc phascs ofthc Southcru Oscillation. p 7-2X. !I/ U Niiio: His- 73. I3rowder JO 19XX Thc social cost 01' rain forcst dcstruetion: the ham- torical and paleocliniatic aspcct« ofthe Southern Oscillation, IIF Diaz, burger debate. lntcrcicncia 13: 115-120 V Margrafeds, Cambridgc 74. Small-scalc ranchcrs. thosc with lcss than 200 animais. as a group, pos- 92. Nobre CA, AS Oliveira I 9~6 Prccipitation and eirculation anomalies scsscd 2R'I<, ofthc total caule hcrd in Pará state in 19XO. In 19X5, their in South Arncrica and the 1982-83 EI Niúo/southcm oseillation cpi- panicipation grew to 3Y~;l.Thc numbcr of animais in this group grew sode. /I l nt ern at i o n a l Co niercn ce o n South ern Hemisph erc twicc as much as thc statcs hcrd in general (53% vcrsus 27%) during thc vtctcrologv. Arn Meterological Soe, Ncw Zcland same pcriod 93. Negreiros G, D Nepstad 1994 The mininnun depth ofrooting ofAma- 75. Arima F, C Uhl Ranching oiBrccilian Amazon in thc national context: zoniuniorcsts. ISpRS- Ri094, RJ econotnics. policv III/(!pructicc» - Unpublishcd 94. Vietoria RL, LA Martinelli,.I Mortatti,.I Riehey 1991 Mcchanisms 01' 76. Walker R. AKO Homma 1996 Land use and land eover dynamics in the water recycling in thc Amazon basin: lsotopic insights. Ambio 20: 384- Brazilian Amazon: an ovcrvicw. Ecol Fcon I X: 67-XO 387 77. Homma A, i\ Rocha, A Santos. A Conto, C Rodrigues, C Ferrcira, P 95. Klink CA. RH Maccdo, CC Mucllcr 1995 De grilo em grào, o cerrado Oliveira, R Walkcr, R Carvalho 1994 trin.imica dos sistema» de perde espaÇ'l). Cerrado: lmpuctos do processo de ocupaçào, LS Mar- produção 1111Transamaconica. EMHRApA/CpATlJ, Bclórn tins, C.lR Alho cds, WWF and pRÓ-CER, Brasília n. Mattos M, C Uhl 1994 Economic and ecological pcrspcctivcs on ranch- 96. Davidson E 1991 Fluxes ofnitrous oxide and nitrie oxide from terres- ing in the eastern Amavon. TYorldDcvelop 22: 145-158 trial ecosystems. 111Microbial production and consuniptinn ofgrcen- 79. Hornma A, R Walkcr, F Scatcna, A Conto. R Carvalho, A Rocha, C house gases: mcthune. nitrogen oxides, undhalomcthnncs . .lE Rogcrs, Fcrreira, A Santos 1993 Dvnamic-; 0/ dctorcstatiou and burning in WH Whitman cds, Am Soe Microbiol, Washington, DC Amazonia: A microcconomic analvsis. Rural Development Forestry Net- 97. Potter C. L Verchot. E Davidson Methane fluxcs frorn terrestrial eco- work Paper 16c. Overseas Dcvclopmcnt Institutc, London systcms - Unpublishcd XO. Vcrissirno A, P Barrete. M Mattos. R Tarifa, C Uhl 1992 Logging im- 98. Meir P,.I Grace, AC Miranda.} Lloyd 1996 Soil respiration in a rain- pacts and prospccts for sustainablc forcst managcmcnr in an old Ama70- forest in Amazônia and in cerrado in central Brazil, p 319-330.11/ /I/11a- nian fronticr: thc case ofParagorninas. Forest Eco! Manag 72: 39-60 ~OIli1111deforestcnion and cliniate. Gash et ai eds, Wiley & Sons, NY 81. Barros A, C Uhl 1995 Logging along thc Amazon river and cstuary: <)9. Heringer Ep, CiM Barroso, JA Rizzo, CT Rizzini 1977 A tlora do Pattcrns, problcms and potential. Forcst Eco! Ma)/Ilg 77: X7-1 05 Cerrado, p 211.232. IV Simpásio sobre o Cerrado, São Paulo. MG X2. Veríssimo A. P Barrcto. R Tarifa, C Uhl 1995 l.xtraction ofa high-valuc Ferri cd, Editora da Universidade de São Paulo, LDUSp, SI' natural rcsourcc frorn Amazonia: thc case of mahogany. Forcst Eco! 100. Vicira L. D Ncpstad, A Alcncar, I' Lefcbvre - Unpublishcd data Manag 72:39-60 I () I. Lefebvre I' Vegetation ehange dctection in an Amazonian frontier. 83. Uhl C. A Vcríssimo. MM Martes. Z Brandino, ICe; Vieira 1991 So- Paragominas from 19X8-91-Unpublished 86· Ciência e Cultura Journal of the Brazilian Association for the Advancement of Science Volume 49( 1/2) • January/ April 1997