Ecological Research in the Large Scale Biosphere Atmosphere Experiment in Amazonia (lba): a discussion of Early Results

Download 99.32 Kb.
Size99.32 Kb.
Ecological Research in the Large Scale Biosphere Atmosphere Experiment in

Amazonia (LBA): A Discussion of Early Results
Michael Keller, Ane Alencar, Gregory P. Asner, Bobby Braswell, Mercedes Bustamante, Eric Davidson, Ted Feldpausch, Erick Fernandes, Michael Goulden, Pavel Kabat, Bart Kruijt, Flavio Luizão, Scott Miller, Daniel Markewitz, Antonio Nobre, Carlos Nobre, Nicolau Priante Filho, Humberto Rocha, Pedro Silva Dias, Celso von Randow, George Louis Vourlitis

(Keywords: LBA, Amazon, tropical forest, cerrado, Land use and land cover change, carbon, nutrients, trace gases)


The Large Scale Biosphere Atmosphere Experiment in Amazonia (LBA) is a multinational, interdisciplinary research program led by Brazil. Ecological studies in LBA focus on how tropical forest conversion, regrowth, and selective logging influence carbon storage, nutrient dynamics, trace gas fluxes and the prospect for sustainable land use in the Amazon region. Early results from ecological studies within LBA emphasize the variability within the vast Amazon region and the profound effects that land use and land cover changes are having on that landscape. The predominant land cover of the Amazon region is evergreen forest, but nonetheless LBA studies have observed strong seasonal patterns in gross primary production, ecosystem respiration, net ecosystem exchange as well as phenology and tree growth. The seasonal patterns vary spatially and inter-annually, and evidence suggests that these patterns are driven not only by variations in weather but also by innate biological rhythms of the forest species. Rapid rates of deforestation have marked the forests of the Amazon region over the past 3 decades. Evidence from ground based surveys and remote sensing analyses show that substantial areas of forest are currently being degraded by logging activities and through the collapse of forest edges. Because forest edges and logged forests are susceptible to fire, positive feedback cycles of forest degradation may be initiated by land use change events. LBA studies indicate that cleared lands in the Amazon once released from cultivation or pasture usage regenerate rapidly. However, the pace of biomass accumulation is generally dependent upon past land use and particularly by the depletion of nutrients by unsustainable land management practices. The challenge for ongoing research within LBA is to integrate the recognition of diverse patterns and processes into general models for prediction of regional ecosystem function.


The Large Scale Biosphere Atmosphere Experiment in Amazonia (LBA) is a multinational, interdisciplinary research program led by Brazil. LBA is not formally an experiment; instead it links many researchers whose goal is to understand how Amazonia currently function as a regional entity in the earth system. Research within LBA is guided by the recognition that Amazonia is changing rapidly through development. Therefore, LBA researchers seek to understand how changes in land use and climate will affect the biological, chemical and physical functions of Amazonia, including the sustainability of development in the region and the influence of Amazonia on global climate. LBA scientific activities cover seven themes: (1)Land use and land cover change; (2) physical climate; (3) carbon dynamics; (4) biogeochemistry; (5) atmospheric chemistry; (6) land surface hydrology and aquatic chemistry, and (7) human dimensions.

Studies of the effects of land use and land cover changes on Amazonian ecosystems developed as a result of a deliberate planning process engaging Brazilian and international scientist that produced a document informally know as the Manaus Workshop Report (Cerri et al. 1995). The key question guiding ecological studies within LBA was defined as,

“How do tropical forest conversion, regrowth, and selective logging influence carbon storage, nutrient dynamics, trace gas fluxes and the prospect for sustainable land use in the Amazon region?”

“Forest conversion” refers to forest clearing and conversion to agricultural uses, especially cattle pasture, and “forest re–growth” refers to vegetation succession following the abandonment of agricultural lands. The question above calls for an explicit consideration of the effects of these land–cover and land–use changes on terrestrial carbon and nutrient budgets, the fluxes of trace gases between the land and the atmosphere, and the exchange of materials between the land and river systems. Implicitly, the question also calls for an understanding of these budgets, fluxes, and exchanges in “primary” or pre–disturbance forest ecosystems. Scientists participating in ecological studies within LBA have elaborated on this initial question to develop a set of more detailed questions to guide our research in four them areas, land use and land cover change, carbon dynamics, nutrient dynamics and surface water chemistry, and trace gas and aerosol fluxes (Box 1).

Biophysical and Biogeochemical Characteristics of the Amazon Region

The Amazon Basin covers 5.8 million km2 and contains the world’s largest river with an annual discharge of nearly 6 x 1012 m3 y-1 (Salati and Vose, 1984). The natural land cover of the Amazon Basin is mainly closed canopy tropical forest although a substantial portion is covered by savanna, known in Brazil as cerrado. Energy and moisture exchanges in the Amazon region play a significant role in global atmospheric circulation. Between 63 to 73% of the annual rainfall of about 2150 mm y-1 is evaporated or transpired at the surface (Costa and Foley 1999, Marengo et al. 2001). Regional scale deforestation could have significant effects on regional and global climate (Nobre et al. 1991; Marengo and Nobre 2001; Werth and Avissar 2002). Recent studies suggest that deforestation in portions of the region can lead to locally increased precipitation (Baidya Roy and Avissar 2002). This raises the hypothesis that there is a threshold of deforestation amount and distribution that leads to a precipitation decline (Avissar et al. 2002).

The Amazon forest vegetation in Brazil alone contains about 70 Pg of carbon (C), which amounts to between 10 and 15% of global biomass (Houghton et al 2001). The biomass density of Amazon forests is poorly known mainly because of the scarcity of plot based data (Houghton, et al. 2001; Keller et al. 2001). Forests in the Amazon region are mostly evergreen and highly productive. Deep rooting allows Amazon forests to maintain productivity through dry seasons that extend up to 5-6 months (Figure 1) (Nepstad et al. 1994). The productivity of the forests is sustained despite the infertility of the highly weathered soils common to the region (Irion 1978). Although about 70% of Amazon soils are dystrophic Oxisols and Ultisols, more fertile soils still cover substantial areas (Richter and Babbar 1991). While nutrients such as phosphorus (P) and base cations ( K+, Ca++, and Mg++) are relatively scarce or only slowly available in most heavily weathered Amazon soils, under forests nitrogen is often abundant. The rapid cycling of nitrogen support large emissions of the greenhouse gas nitrous oxide (N2O). Emissions of N2O from the forests of the Amazon Basin account for about 0.8 to 1.3 Tg N y-1 or nearly 10 to 15% of the natural emissions of that gas (Melillo et al. 2001).

Land Use Change and Development in the Brazilian Amazon Forests

Over the past three decades the Amazon region has been undergoing a burst of development. This development takes place on a backdrop of forest that has recorded the imprint of human habitation and use extending back more than 10,000 years ago (Roosevelt ...). The region, particularly the heavily populated river corridors, suffered depopulation following the influx of European explorers and settlers in the sixteenth century (ref). By the middle of the nineteenth century, large numbers of migrants colonized the Amazon region mainly to extract rubber. The rubber tappers probably had a considerable, if largely undocumented, influence on the forest ecosystem. {Example?}. The population in the Amazon region of Brazil grew from slightly more than 100,000 in 1840 to 1.2 million in 1912 (Weinstein, 1983). Regional population then stagnated through 1940 (1.5 million) but grew exponentially thereafter reaching nearly 18 million today (IBGE), more than 50% of which is urban defined as living in towns and cities of greater than 5000 population (Browder and Godfrey, 1997).

Agricultural colonization and development schemes have a long history in the Amazon beginning with the Jesuit missions (ref). The history of recent agricultural exploitation began during the rubber boom. During the years 1875 to1900, settlements were encouraged along a railroad line from from Belém to Bragança in Pará; lands along the route were cleared to help feed the growing trading metropolis of Belém (Weinstein, 1983). Forest clearance and agricultural development accelerated enormously in the 1970's and 1980's with the construction of roads such as the Trans-Amazon Highway and BR-364 in Rondônia. Rates of forest clearance have averaged about 20,000 km2 y-1 during the past decade (INPE 2003; Houghton et al. 2001).

Recent trends in land use indicate consolidation of the old frontiers, a new phase of experimentation in land management, and a heightened level of governance (Carvalho et al. 2002). Whereas previous development depended largely on a mixture of logging, cattle ranching, and subsistence cropping, current trends suggest a move toward more intensive management including mechanized production of grains, dairy cattle, or agro-forestry products (refs). Logging remains an important industry and canopy opening following selective timber harvests makes forests more susceptible to fires (Nepstad et al. 1999). Most logging might still be considered predatory or timber mining where valuable species are removed and little or no attention The potential for fire to spread from deforested areas into fragmented forests, represents a threat to long-term ecosystem health and sustainability (Nepstad et al. 1999; Cochrane et al. 1999; Cochrane and Laurance, 2002).

LBA Study Design

Design and planning of LBA began in 1993 (Avissar and Nobre 2002). LBA science and the LBA study sites have been chosen to reflect the history of Brazilian led research in the Amazon region. LBA owes a debt to Anglo-Brazilian Climatic Observation Study (ABRACOS) (Gash et al. 1996) and the Amazon Boundary Layer Experiments (ABLE) sponsored by INPE and NASA (Harriss et al. 1988 and 1990) in the 1980s and early 1990s. Planning for ecological work within LBA was summarized in a document informally named the Manaus Workshop Report led by Carlos Cerri and Jerry Melillo (Cerri et al. 1995). This report established a design to include two notional transects to incorporate the main climatic variability within the Amazon region, especially total rainfall and dry season length (Figure 1). The northern transect traverses more highly weathered and infertile soils compared to the southern transect. The two transects both cover a range in the extent, intensity, and character of land use change.

Study site selection was guided by the transect design. Prince and Steininger (1999) suggested a biophysical classification of the Amazon region to further guide study design. These ideas were considered along with practical considerations, especially site logistics and research history, that strongly influenced the selection of LBA study sites. For example, the long term studies supported by Brazil’s National Institute for Amazon Research, (INPA) in the vicinity of Manaus, made that area a leading candidate for study. Specific opportunities to study land use changes also influenced site selection. The presence of the only Brazilian government managed logging concession in the Amazon region in the Tapajós National Forest outside of Santarem led to the inclusion of sites there for the study of selective logging.

In this paper, we review the results from 19 studies in LBA that consider the science themes of physical climate, carbon dynamics, nutrient dynamics, trace gas fluxes, and land use and cover change. These 19 studies either deal with the Amazon region generally or were conducted at one or more of 13 sites that are listed in Table 1. Early results from LBA focusing mainly on physical climate and atmospheric chemistry recently have been collected in a special issue of the Journal of Geophysical Research (see Andreae et al. 2002; Avissar and Nobre 2002; Avissar et al. 2002; and Silva-Dias et al. 2002 for summaries). Recent work on land cover and land use change has been compiled in a special issue of Remote Sensing of the Environment (Roberts et al. in press?). The papers collected in this issue do not cover all current ecological work within LBA. In this summary paper we attempt to compare the results presented in this volume and to place this work within the context of other recently published studies within LBA and within the broader scope of studies of tropical ecosystem studies without attempting an exhaustive review. We conclude by identifying emerging trends and challenges for future ecological research in LBA.

Physical Climate

LBA studies of physical climate extend from the global and continental scale to the micro-scale. The same range of scales is represented in this issue. Goncalves et al. (this issue) discuss the importance of incorporating land cover heterogeneity in weather prediction models for South America. Rocha et al. (this issue) and Quesada et al. (this issue) quantify water and energy budgets for a forested and a savanna site respectively. The differences in the water budgets between the campo sujo savanna at the IBGE Reserve and the dense forest at the Tapajós National Forest is impressive. Dry season evapotranspiration in the savanna averaged 1.6 mm d-1 versus 4.9 mm d-1 for the forest. Both ecosystems depend upon deep rooting to sustain evapotranspiration during the dry season. Rocha et al. (this issue) also observed that hydraulic lift recharged the forest upper soil profiles each night. At Tapajós, the forest showed no signs of drought stress allowing uniformly high carbon uptake throughout the dry season (July - December 2000) (Rocha et al. and Goulden et al. this issue).

Carbon Dynamics

Results published in this special issue emphasize the seasonal variation of carbon fluxes in Amazon forests and savannas. While the forests of the Amazon basin are mostly evergreen, patterns of seasonal variability are apparent in stem growth (Rice et al., this issue), litter fall (Goulden et al. this issue), soil-atmosphere carbon dioxide flux (Goulden et al. this issue; Chambers et al. this issue), and also in the net ecosystem exchange (NEE) measured by eddy covariance (Goulden et al. this issue; Vourlitis et al. this issue). The seasonality of litterfall in evergreen Amazon forests is a well-known phenomenon (e.g. Luizão 1989) and litterfall generally peaks during the dry season. Leaf shedding may represent an adaptation to water stress, but the common pattern of dry season flushing of new leaves also suggests that leaf phenology has been strongly influenced by natural selection related to herbivory and other biological factors (van Shaik et al.1993). Innate seasonal rhythms may also account for the pulse of stem growth observed prior to the initiation of the wet season in Santarem in 2001 (Goulden et al., this issue).

Soil atmosphere flux of carbon dioxide was highly seasonal although contrasting patterns were observed near Manaus and Santarem. Whereas Goulden et al. (this issue ) found that low moisture content in litter and soil constrained soil respiration in the Tapajós National Forest during the dry season, Chambers et al. (this issue) found that excess moisture appeared to inhibit soil respiration in the Cueiras Reserve during the wet season. In part, the difference between these results may be explained by the contrast in forest habitats investigated near Manaus and Santarem. The Tapajós site is flat and well drained and the Cueiras site contains rolling topography with poorly drained valleys whose soils saturated during the wet season. The importance of the valley sites to the overall NEE of the Cueiras reserve has been emphasized by Araujo et al. (2002).

Net ecosystem exchange (NEE) also varied seasonally. The first eddy covariance study in the Amazon basin that included both wet season (44 days) and dry season (11 days) measurements, noted strong differences in the NEE between the seasons (Grace et al. 1995). Working in the Jaru Biological Reserve near Ji-Parana, Rondônia, Grace et al. (1995) measured NEE of -0.09 mol C m-2 d-1 during the dry season and -0.05 mol C m-2 d-1 during the wet season.

Contrasting patterns of NEE, ecosystem respiration (Reco), and gross primary production (GPP) have been observed in different sites (Figure 2). Most of these data covered only one year of measurements so inter-annual variation may be as important as spatial variation for differences among the seasonal patterns. Nonetheless, these patterns raise some interesting questions. In two sites, Jaru and Sinop, net carbon uptake (the most negative NEE) clearly occurred during the rainy season. At Manaus, NEE was nearly constant across the year while in Santarem, NEE was most negative during the dry season. The latter pattern appears to be driven by the strong decrease in Reco during the dry season at Santarem without a comparable decrease in GPP. Goulden et al. (this issue) found that the vegetation did not show evidence of drought stress during the dry season.

In contrast to Santarem, Manaus, Jaru and Sinop displayed greater GPP during the wet season as compared to the dry season. What controls seasonal differences in GPP across sites? Access to deep soil water may vary across sites. Additionally, innate phenological controls may be play an important role in the regulation of seasonal carbon uptake.

NEE is a relatively small quantity that represents the difference between two large quantities GPP and Reco. As the difference between two large numbers, NEE is difficult both to measure and model accurately. The errors and biases related to the calculation of annual sums of NEE from eddy covariance data is a main subject of two papers in this issue, Kruijt et al. and Miller et al. As noted previously by Araujo et al. (2002), both Kruijt et al. and Miller et al. conclude that interpretation of nocturnal fluxes is the largest single source of error for sites with strong nocturnal stability, a typical situation in tropical moist forests. Not all LBA sites appear to suffer equally from this problem. For example, Kruit et al. present data from the Jaru Reserve near Ji-Parana, Rondônia that show no relation between measured nocturnal NEE and u*, a measure of turbulence.

Difficulties in analyzing NEE from tropical forest sites reinforce the need to use complementary methods to constrain biological fluxes. As Kruijt et al. (this issue) point out, biometric methods and eddy covariance methods provide independent approaches to measurements of NEE. Chambers et al, Rice et al. and Miller et al. (all this issue) make biometric measurements that can be compared to eddy covariance results. Rice et al. and Miller et al. measured aboveground biomass changes and concluded that stands studied at the Tapajós National Forest are either roughly in carbon balance or losing a moderate amount of carbon annually. Miller et al. found these measurements of change in above-ground biomass were consistent with their own measures of NEE using eddy covariance techniques. Chambers et al. made extensive measurements of four components of ecosystem respiration (leaf, stem, CWD, and soil) and scaled these measurements to annual values. They concluded that biometric and nocturnal eddy covariance results for sustained high turbulence conditions were indistinguishable within the errors of the methods. A challenge in both eddy covariance and biometric studies is to analyze and minimize those errors.

Biometric studies provide insights into controlling mechanisms that complement mechanistic inferences available from eddy covariance studies. For example, Chambers et al. (this issue) conclude that the tropical moist forest at the Cueiras reserve has a low carbon use efficiency compared to temperate forests but similar to other tropical forests. Rice et al (this issue) found that stand structure and the abundance of coarse woody debris (CWD) at their site in the Tapajós National Forest indicates that it suffered a recent disturbance that they attribute to severe ENSO related drought in the 1990s. The respiration fluxes from CWD are surprisingly large. Understanding this carbon pool and its site to site variation will be critical to accurate estimation of NEE. The CWD pool at the km 67, Tapajós National Forest Site (48 Mg C ha-1), is 2 to 4 times greater than the standing stock of CWD measured in forests near Manaus (Nascimento and Laurance, this issue; Chambers et al. 2000). Similarly, estimates of the annual CWD respiration are 5.7 Mg C ha-1 y-1 for the km 67 site at Tapajós and a maximum of 1.8 Mg C ha-1 y-1 for forest outside Manaus (Rice et al. this issue; Chambers et al. 2000).

Analysis of NEE from eddy covariance data spanning at least one year of measurements from Sinop, Santarem, and Manaus (k34) as well as biometric data from Santarem indicate that forest NEEs were relatively small 0±2 Mg C ha-1 y-1 (Vourlitis et al., Miller et al., Araujo et al. 2002; Rice et al. this issue). Araujo et al. 2002 found a slightly greater uptake for the C14 tower in the Cueiras reserve even after filling nighttime fluxes for u* < 0.2 m s-1. Interestingly, fluxes of magnitude of only 1 Mg C ha-1 y-1 extrapolated over with the large extent of old growth forest in the Amazon Basin (~5 x 106 km2) lead to globally significant amounts of carbon (0.5 Pg C y-1). Because of the locations of observing stations, even large fluxes are currently below the resolution of global atmospheric inversion models for the Amazon region (Rayner et al.1999 ; Bousquet et al. 2000; Gurney et al. 2002) but not below the threshold for measurement using regional airborne sampling. Using a budget method, a re-analysis of CO2 concentration data from the 1987 ABLE 2B flights showed that the central Amazon region had a near zero carbon flux (-0.03 ± 0.2 µmol m-2 s-1) during the wet season (Chou et al. 2002).

The accuracy of global atmospheric transport model inversions could be greatly improved if high precision CO2 concentration data from the Amazon region were to become available (Rayner et al. 1999). Weekly sampling of tropospheric air to 3 km is planned for coastal and interior sites in LBA. Even so, interpretation of the net fluxes derived from these models will remain ambiguous until we accurately quantify both the seasonality of biological exchanges and the magnitude of emissions caused by extensive annual burning (Alencar et al. this issue; Potter et al. 2002).

Nutrient Dynamics

Efficient nutrient conservation mechanisms allow mature tropical forests to thrive even on acid and infertile soils. Base cations (e.g. K+, Ca++, and Mg++) and phosphorus (P) are generally tightly cycled in tropical forests and thus they are considered possibly limiting factors for forest productivity (Vitousek and Sanford, 1986). In this issue, Markewitz et al. confirmed prior studies showing significant losses of nutrients, carbon (C), nitrogen (N), and P from cleared and burned sites (McGrath et al. 2002). Markewitz et al showed through a budget analysis that base cations derived from forest clearing and burning are tightly retained in the surface soils of “secondary lands” (secondary forest, degraded pasture, and active pasture) at their Paragominas, Pará study site after more than 20 years following the land clearing fires. It is generally accepted that C and N are lost from the ecosystem to the atmosphere during the fires used to clear land and through subsequent mineralization of organic matter. Unlike C and N , P does not have a long-lived volatile phase in the atmosphere (Schlesinger, 1997). The selective loss of P as opposed to other non-volatile elements such as K, Ca, and Mg remains unexplained.

Conversion of forest to pasture was the most common land use change in the exploitation of the Amazon region during the 1970s and 1980s. Pasture is still the most common land use in deforested areas although considerable areas of pasture have been abandoned to secondary vegetation (Alves et al.?). The intensity of prior land use, the distance to seed sources, and the presence or absence of fire are all important factors regulating the pace of secondary succession in abandoned pastures (Uhl, Nepstad refs.). In this issue, Davidson et al. Feldpausch et al. and Markewitz et al. all argue that the scarcity of key nutrients may limit the pace of secondary succession. Davidson et al. (this issue) conducted a fertilization experiment in a 6 year old secondary forest in Paragominas, Pará. They found that additions of N or N together with P increased the rate of above-ground biomass increment by woody vegetation. In contrast, additions of P only favored the growth of herbs and grasses. Davidson et al. concluded that forest biomass increment was limited by N at their site. Markewitz et al. who also worked in Paragominas came to a similar conclusion by inference from the relative rates of accumulation of N and P. The secondary forest they studied had only accumulated 33.5 Mg C ha-1 above ground over 19 years (accumulation rate of ~1.8 Mg C ha-1 y-1). In contrast, Feldpausch et al. observed an above-ground biomass accumulation rate of 5.5 Mg C ha-1 y-1 for a chronosequence of secondary forests (up to 14 years old) developed on abandoned pastures. They found that surface soil to 45 cm depth was accruing N and P. They concluded that it was likely that P or possibly Ca might limit growth and that these nutrients were actively extracted from the sub-soil towards the surface layers.

Rates of biomass accumulation in secondary forests are highly variable. Successful models capture regional behavior of secondary forest regrowth are based on soil texture (a proxy for water and nutrient availability) and growing season limitations (wet versus dry months) (Johnson et al. 2000). But, there are other important sources of variability. Figure 3 shows the range of carbon:nutrient ratios encountered in two studies in this issue. It is not surprising that differences in nutrient availability appear to influence the rate of succession. Factors affecting nutrient stocks include the history of land use and management particularly the use of fire, a practice that impoverishes system N stocks. As shown by Davidson et al., secondary succession on pastures that have been repeatedly burned can be N-limited. Secondary vegetation on pastures that were heavily grazed and frequently burned may accumulate biomass more slowly compared to vegetation in areas that suffered a less intensive use. Future models of secondary forest regrowth should consider prior land management, especially the frequency of fire, in order to accurately predict biomass accumulation.

The importance of fire as a control of biogeochemical dynamics in Amazon ecosystems is difficult to overstate. The Brazilian cerrado (savanna ecosystem) is highly diverse (ref) and currently undergoing far greater relative rates of land use change than the forests of the Amazon (Nepstad et al. 1997; other). The cerrado receives annual rainfall from about 1 to 2 m per year but it is always marked by a strong dry season and fire is frequent. The return time for fire in the cerrado is 2-40 years while it appears to be hundreds of years or more in the Amazon forests (Coutinho 1990; Vicentini 1999; Sanford et al. 1985). Analysis of 15N:14N ratios in cerrado vegetation shows that this vegetation shares wide ranges in these ratios characteristic of N-limited ecosystems ( Bustamante et al., this issue). Frequent fire limits ecosystem N accumulation and N concentration; 15N contents reflected fire frequency even among different vegetation formations within the cerrado.

Trace Gas Fluxes

Estimation of the fluxes of trace gases of the Amazon ecosystems to the atmosphere for both long- and short-lived trace gases is an essential component of LBA. This issue includes four examples of studies of gas fluxes for both long-lived radiatively important gases such as nitrous oxide (N2O) and carbon dioxide (CO2) (Garcia-Montiel, et al. this issue; Varella et al. this issue; Davidson et al. this issue) or short-lived gases such as nitric oxide (NO) (Varella et al., this issue; Davidson et al. this issue) and short-chain aldehydes (Rottenberger et al., this issue) that regulate the production and destruction of atmospheric oxidants.

Garcia-Montiel et al. (this issue) present a new approach for regional estimation of N2O emissions based on field investigations in Rondônia. Soil N2O emission is scaled linearly to the soil emission of CO2; the latter is estimated based upon the TEM model (McGuire et al. 1992). This approach parallels a previous effort by Garcia-Montiel and her colleagues (Melillo et al. 2001) whereby regional N2O emission was estimated from modeled N-mineralization using TEM.

Tropical forests release substantial quantities of volatile organic compounds (VOCs) to the atmosphere (Guenther et al. 1995). Possibly VOC release is important to the ecosystem carbon balance (Crutzen et al. 1999). However, as Rottenberger et al. (this issue) found in their study of small-chain aldehydes, vegetation can be sink as well as a source of VOCs. A full accounting of the influence of VOCs on ecosystem carbon budgets must consider both sources and sinks for these compounds and the reaction products of VOCs including atmospheric particulates.

Land Use and Land Cover Change

Current trends in land use in the Amazon region have caused significant fragmentation of the forest (Skole and Tucker 1993). Increasing fragmentation leads to an increasing length of forest edge and an increasing area of edge habitat. Living on the edge, whether for people or for trees, has its drawbacks. Nascimento and Laurance (this issue) quantify the biomass and necromass in forest edges and interior forests. They found that large tree mortality was accelerated in edge habitats compared to forest interiors.

The acceleration of fire risk on the Amazon landscape has the potential to greatly alter ecosystem structure and function (Nepstad et al. 1999; Cochrane et al. 1999). Cochrane and Laurance (2002) recently demonstrated that fire risk is greatest within 2-3 km of existing forest edges. Alencar et al. (this issue) model fire probabilities based on settlement patterns, infrastructure, and economic activities such as charcoal manufacture. Their work showed radically different probabilities of fire in El Niño versus non-El Niño years. Forest degradation through logging or fire strongly predicted fire occurrence. Alencar et al. (this issue) supplemented remote sensing interpretation with 6 months of field interviews to classify degraded forest. New approaches, such as the automated Monte-Carlo unmixing developed by Asner et al. (this issue; Asner and Lobel 2000) used with LANDSAT imagery show great promise for quantitative measurement of forest degradation [FIGURE 4]. Asner et al. precisely measured canopy opening using remote sensing data for forests that were recently logged and and forests recovering from logging. This measurement is valuable because canopy opening is associated with increased fuel loads in logging gaps and altered microclimates that make the forests more susceptible to fire.

Future Challenges for Ecological Studies in LBA

Studies in LBA are advancing our understanding of the functions of managed and unmanaged ecosystems in the Amazon. The overall challenge for ecological research in LBA is the unification of the results of site-based studies into a regional synthetic framework. Results presented in this issue raise many questions and indicate some directions forward.

While most of the Amazon region is covered by evergreen forest, strong seasonality in rainfall leaves its imprint on the cycling of carbon and nutrients, the fluxes of trace gases and the patterns of land management. Understanding how seasonal patterns are driven by variations in weather as well as by innate seasonal rhythms will be critical for development of reliable models of ecosystem function.

Development in the Amazon region has been accompanied by increasing forest fragmentation and poorly managed logging activities. Both fragmentation and logging increase the likelihood of fire escaping from managed systems into forest, especially during dry years often associated with El Niño. Fire may represent the single greatest threat to the forest ecosystem yet the extent of burned forest in the region is poorly quantified and the conditions leading to forest fires are only beginning to be understood. The spatial extent and magnitude of forest degradation in the Amazon region has not been comprehensively quantified. Nonetheless, the combined use of newly developed remote sensing techniques coupled with intensive ground studies shows great promise for quantifying forest degradation and recovery over vast areas.

While recent development activities have led to extensive changes across the Amazon landscape, the ecosystems of the region, or at least their component species, show an ability to react to change. Secondary forest succession now covers extensive areas and secondary forests can rapidly achieve certain functions of primary forest such as the recycling of water in evapotranspiration (Nepstad..., Brown and Lugo 1990). On the other hand, the vigor of secondary succession may be limited by the shortage of plant available nutrients left behind as a legacy of past land use practices such as overgrazing and repeated burning.

Geologists and other natural scientists depend on the uniformitarian principle of James Hutton summarized by the motto “the present is the key to the past.” Can ecologists and social scientists depend on the present and past as keys to the future? Frontier expansion in the Amazon region is confronting the modern world of rapid communications and globalization. The regional population is urbanized and the regional economy no longer depends strictly on a mixture of extractive industries, extensive ranching and subsistence agriculture. Will future land cover and land use continue to follow past patterns in simply a greater extension or will the future development of the Amazon region follow a different, and perhaps more sustainable, path? The answer to that question depends on choices made the people of the Amazon region countries and the governments that they elect. We believe that LBA is developing new knowledge to guide the decision making for a more sustainable future.

Box 1. Questions for ecological research within LBA.

Land Use and Cover Change
LC–Q1 What are the rates and mechanisms of forest conversion to agricultural land uses, and what is the relative importance of these land uses?
LC–Q2 At what rate are converted lands abandoned to secondary forests; what is the fate of these converted lands, and what are the overall dynamic patterns of land conversion and abandonment?
LC–Q3 What is the area of forest that is affected by selective logging each year? How does the intensity of selective logging influence forest ecosystem function, thus altering forest regrowth and flammability?
LC–Q4 What are plausible scenarios for future land–cover change in Amazonia?

Carbon Dynamics

CD–Q1 What is the (climatically driven) seasonal and interannual variability of the CO2 flux between the atmosphere and different land cover/use types?

CD–Q2 How do biological processes such as mortality and recruitment or succession following land use change influence the net annual C balance for different land–cover/land use types?

CD–Q3 What are the relative contributions of fluxes from natural and disturbed ecosystems to the net Amazonia–wide flux? This question can be approached through a number of subsidiary questions:

CD–Q3a How do pools and fluxes of C and nutrients (in soils) of pasture/cropland change over time and what factors determine C gain or loss?

CD–Q3b How does selective logging change the storage and cycling of C in forests?

CD–Q3c What factors (biologically mediated, land–use history, soil properties, etc.) control the rate of C sequestration in biomass and soils of regrowing forest?

CD–Q3d What portion of the Amazonia–wide C flux is from fire? How do ecosystems recover from fire? What are the relations between land management and fire occurrence/frequency?

Nutrient Dynamics and Surface Water Chemistry

ND–Q1 How do stocks, cycling rates and budgets of carbon and important elements N, P, K, Ca, Mg, and Al change under different land covers and land uses?

ND–Q2 Are nutrients major factors that control the rates of re–growth and carbon accumulation in abandoned pastures and re–growing secondary forests?

ND-Q3. What are the processes and consequences of atmospheric horizontal transport of nutrients (wind) on the nutrient stocks and cycles of ecosystems within the Amazon basin at various spatial and temporal scales?


- Saharan dust inputs

- Losses and redistribution due to fire

- Links between physical climate models and nutrient cycling?
ND-Q4. How do changes in land-use and climate alter the stocks, processes and fluxes of dissolved and particulate organic matter, nutrients, and trace gases from the uplands across the riparian zones and floodplains and down the channels of river corridors?


  How will the composition and quantity of nutrients and organic matter entering and being processed within streams be altered under different land-use change scenarios?


  Are there unique signatures that can be traced downstream?


  To what extent do intact riparian zones buffer streams against changes due to anthroprogenic activities in surrounding uplands?

ND-Q5. What is the importance of periodically wet environments (from moist soils to standing and flowing waters) for the land and atmospheric balances of nutrients, CO2, trace gases, and water and energy on multiple scales?

Trace Gas and Aerosol Flux

TG Q1. How are fluxes of trace gases and aerosols between ecosystems (both upland & wetland) and the atmosphere of Amazonia affected by land cover and land use change?

TG Q2. What is the (climatically driven) seasonal and inter annual variability of trace gas and aerosol fluxes between the atmosphere and different land use/land cover types?

TG Q3. Are losses and gains of carbon from Amazonian ecosystems in forms other than CO2 (e.g. CO, CH4, VOC, organic aerosol) of sufficient magnitude to influence ecosystem carbon balance?

Table 1. LBA study sites discussed in this special issue.

Full Site Name

Short Name

South Latitude

West Longitude


Rerserva do IBGE


Fazenda Cauaxi



Fazenda Nova Vida

Fazenda Nova Vida

Cattle pasture

Reserva Biologica Jaru


Mature forest tower

Manaus BDFF

Forest fragmentation

Manaus C14

Mature forest tower


Secondary forest

Manaus K34

Mature forest tower

Fazenda Victoria


Pastures and secondary forest



Floresta Nacional do Tapajos, km 67

Santarem, km 67

Mature forest tower

Floresta Nacional do Tapajos, km 83

Santarem, km 83

Mature forest tower (later logged in 2001)


Mature forest tower


Funding was provided by the National Geographic Society and the National Science Foundation, Division of International Programs. Additional support from California State University, San Marcos (CSUSM), the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq), Universidade Federal de Mato Grosso (UFMT), the Fundação de Amparo à Pesquisa do Estado de Mato Grosso (FAPEMAT), the Sindicato das Indústrias Madeireiras do Norte de Mato Grosso (SINDUSMAD), and Fazenda Maracaí is gratefully appreciated.


Coutinho (Coutinho, L.M. 1990. Fire in the ecology of the Brazilian Cerrado.

In: Fire in the tropical biota   Ecosystem process and global challenges. J/G.

Goldammer (ed.) Springer Verlag. Ecological Studies, 84:82 105)

2. Vicentini (Vicentini, K.R.F. História do fogo no Cerrado: uma análise

palinológica. PhD Thesis, Universidade de Brasília. 1999, 208pp)

Share with your friends:

The database is protected by copyright © 2020
send message

    Main page