Productivity of aboveground coarse wood biomass and stand age related to soil hydrology of Amazonian forests in the Purus-Madeira interfluvial area

Introduction Conclusions References


Conclusions
References

Introduction
The increased number of permanent plots in the recent years has contributed to improve the understanding of regional variation in forest productivity across the Amazon Basin (Baker et al., 2004;Malhi et al., 2004;Giardin et al., 2010).The use of adjusted allometric models for different regions (Chave et al., 2005;Feldpausch et al., 2012) contributed to more reliable estimates on biomass storage and productivity in Amazonian forests.However, the existing studies do not cover the full range of forest ecosystems across the Amazonia.Most of the studies on wood biomass production focus on nonflooded terra firme forests.Few studies focus on wooded wetlands seasonally flooded by large rivers, but other types of wooded wetlands remain mostly overlooked.In the Amazonia, wetlands constitute about 30 % of the total area (Junk et al., 2011), of which old fluvial terraces located in interfluvial areas not flooded by large rivers cover a large portion.Little is known about environmental factors such as soil conditions, climate and hydrology controlling wood biomass productivity in these interfluvial wetlands.
It has been shown that in non-flooded terra firme forests wood biomass productivity negatively responds to severe droughts such as in 2005 and 2010 (Phillips et al., 2009;Corlett et al., 2011;Lewis et al., 2011).In seasonally flooded forests (v árzeas and igap ós), on the other hand, an enhanced tree growth can be attributed to El Ni ñoinduced droughts, since the flooded period in El Ni ño years is shorter than in other years (Sch öngart et al., 2004;Sch öngart and Junk, 2007).There is also evidence that variations in productivity are related to soil fertility in both terra firme and floodplain forests (Malhi et al., 2004;Sch öngart et al., 2005) which consequently results in different stand ages (Sch öngart et al., 2010) and biomass stocks (Malhi et al., 2004(Malhi et al., , 2006)).However, interfluvial wetlands present conditions different both from the terra firme and the large river floodplains.As terra firme forests are not homogeneous over the Amazon basin, wetland forests also show considerable variation in structure and functioning.Introduction

Conclusions References
Tables Figures

Back Close
Full Interfluvial wetlands are found along vast regions in the central and western Amazonia (Sombroek, 2000).In these regions the water table is shallow and the soils are poorly drained and seasonally waterlogged, presenting a patchwork of seasonally flooded, saturated or non flooded areas.Flooding or saturation may occur by pooling of water on the lower poorly drained sites every year during the rainy season, while adjacent sites may be well drained (Rosseti et al., 2005).This probably happens because interfluvial areas are lowland plains (Sombroek, 2000).In addition, the fluctuations of the underground water may generate a firm layer of varying depth by deposition of iron and other nutrients (Quesada et al., 2010), generating a complex hydrology that may influence ecosystem processes in ways much different both from the terra firme forests and the large rivers floodplain forests.The complex effects of soil, climate and hydrology on productivity of seasonally waterlogged interfluvial wetland forests are still poorly understood.
It must be considered that the Amazon basin will undergo a severe transition towards a disturbance-dominated regime, mostly due to changes in land-use and climate (Malhi et al., 2008;Cook et al., 2012;Davidson et al., 2012).Hence, the severe impacts caused by continuous deforestation and degradation in the Amazon Basin, mainly associated with the large infrastructural program of the Brazilian Federal Government, generate demands for rapid assessment of information to generate databases that help discriminate areas for forest management and conservation, and therefore contribute for a sustainable development of the Amazon region.In this context, information on tree growth, tree ages and forest productivity are important key data to establish criteria to define areas of conservation priorities and sustainable policies for sustainable management.
The most common field method to estimate wood productivity in the tropics is monitoring tree growth.However, only after many years of repeated diameter measurements in permanent plots, the estimates of diameter increment rates are reliable (Clark et al., 2001).Alternatively to monitoring tree growth, tree-ring analysis has been applied for reliable estimates of tree ages and mean diameter growth rates which are necessary Introduction

Conclusions References
Tables Figures

Back Close
Full to predict the woody biomass production.Recently, tree-ring data have successfully been used to estimate wood biomass productivity in different forest types of Central Amazonia (Stadtler, 2007;Oliveira, 2010;Sch öngart et al., 2010) and also in the Pantanal wetlands (Sch öngart et al., 2011).In the Amazon, annual tree rings occur in the non-flooded terra firme forests as a consequence of the rainfall seasonality with one dry and one rainy season during a year (Vetter, 1995;Worbes, 1989Worbes, , 1999;;D ünisch et al., 2003;Brienen and Zuidema, 2005;Zuidema et al., 2012).The rainfall seasonality in large catchment areas of the Amazon River and its large tributaries results in a monomodal flood-pulse also leading to the formation of annual tree ring in the wood of tree species in the floodplain forests (Worbes, 1989;Sch öngart et al., 2002Sch öngart et al., , 2004Sch öngart et al., , 2005)).
The presented study is the first field-based estimates for tree ages and wood biomass productivity in the vast interfluvial region between the Purus and Madeira rivers.We estimate aboveground wood biomass productivity in eight 1 ha plots distributed over 600 km along the interfluvial region between the Purus and Madeira Rivers.First, we assess wood biomass stocks by diameter, tree height and wood density applying two allometric models.Second, we use tree-ring data to estimate stand age and to model changes in biomass stocks over time.We than relate productivity and stand age to soil and hydrology conditions of the studied sites.Finally, we compare our results with other data of wood biomass productivity from different forest types of the Amazon basin and discuss their application in the context of conservation and forest management.

Study region
The study was carried out in the interfluvial area of the Purus-Madeira Rivers in the Amazonas State, Brazil, that is intersected by the BR-319highway connecting the cities Introduction

Conclusions References
Tables Figures

Back Close
Full of Manaus and Porto Velho, the capitals of the Amazonas and Rondonia states, respectively (Fig. 1).The dominant vegetation type in the northern part of this region is dense lowland rainforests, while in the south a transition towards open lowland rainforests occurs, probably caused by increasing rainfall seasonality (IBGE, 1997).The predominant soil type for the whole interfluvial region is Plinthosol/Gleysol (Martins et al., 2012).These soils have a firm plinthite layer that can change to hardpan if exposed to repeated wet and drying cycles (Quesada et al., 2011).On a local scale, frequent variations in the topography of a few meters create temporary pools on the lower and poorly drained areas during the rainy season (Rosetti et al., 2005).The duration and intensity of the rainy season varies strongly from north to south along the interfluvial area declining from 2800 to 2100 mm annual precipitation with increasing rainfall seasonality (ANA, 2011; Table 1).
As indicated by future scenarios the interfluvial region between the Madeira and Purus rivers will suffer huge impacts as a consequence of paving the BR-319 highway and the human occupation following in the next decades (Laurance et al., 2001;Soares-Filho et al., 2006;Fearnside et al., 2009;Davidson et al., 2012).To avoid deforestation, several conservation units (Sustainable Development Reserves, Extractive Reserves and State Forests) have been established along the highway.However, longterm scientific data on floristic composition, forest structure and dynamic are essential as a decision tool for the sustainable development of the region, as criteria for discrimination between areas of strict conservation or management priorities.In this context, a set of 11 research modules of the PRONEX project "Rapid Assessment for Long Duration Ecological Projects" (RAPELD) was established along the Purus-Madeira interfluvial region as part of the research modules network of the Research Program in Biodiversity (PPBio, http://ppbio.inpa.gov.br/sitios/br319).

Sampling design
In each of the 11 modules established along the interfluvial region of the Purus-Madeira rivers, ten 1 ha plots (250 × 40 m) were installed at every kilometer on two 5 km-long Introduction

Conclusions References
Tables Figures

Back Close
Full trails (Magnusson et al., 2005), totalizing 110 plots in the interfluvial region.Each plot followed an isocline to minimize variation in the topography and soil conditions within a plot (Magnusson et al., 2005).For this study, we selected four of the 11 modules (Table 1).The average distance between the modules was 140-200 km, spanning approximately 600 km distance (Fig. 1).At each area, we selected two of the 10 permanent plots, one on the highest altitude (well drained) and the other plot on the lowest topography (poorly drained),using SRTM-DEM data (Shuttle Radar Thematic Mapper Digital Elevation Model) (USGS, 2000;Farr et al., 2007).

Soil data
Soil samples were obtained every 50 m along the central 250 m long transect in all plots, totalizing six samples per plot.Soil samples of 30 cm depth were extracted with a borer and labeled and kept in sealed plastic bags for 2-5 days.On arrival at the laboratory, the samples were air dried at ambient temperature.After drying, composite samples were prepared from all points, resulting in one sample per plot.Soil texture was then analyzed following standard protocol of total dispersion using sodium pyrophosphate to obtain clay, sand and silt percentages (EMBRAPA, 2011).The soil phosphorus concentration was also analyzed, following protocol by EMBRAPA (2011).To analyze soil water saturation pits of 2 m depth were dug in modules M01, M05 and M08.To increase spatial coverage of soil water saturation scores corer samplings in each plot were used.Soil water saturation in each plot was scored using the classification index of Quesada et al. (2010) (Table 2), using soil samples up to 7 m depth collected by the HIDROVEG project and 2 m depth pit descriptions dug in modules M01, M05 and M08.Soil water saturation conditions classified by an index are based on the effective depth of the soil, hydrological properties and the presence of a plinthite layer, pointing to soil hydromorphic features.This index may be an important edaphic parameter as it appears to be related to vegetation parameters (Martins et al., 2012).All soil samples were collected in cooperation with the HIDROVEG project and analyzed at the Thematic Soils Laboratory of the National Institute for Amazon Research (INPA).Introduction

Conclusions References
Tables Figures

Back Close
Full

Topographical and hydrological conditions
Terrain hydrological conditions were assessed by using remote sensing and field data.
The plots were preselected based on SRTM data to capture a wide range of topographies within and between the selected modules.Afterwards we applied the HAND model (height above the nearest drainage) that indicated the vertical distance of the plot from the nearest water-table as an indicator for the hydrological conditions (Renn ó et al., 2008).The HAND value, based on SRTM data, was calculated for all plots (Moulatlet, 2010).We also used the soil water saturation index (Table 2) as a reliable indicator for the terrain's hydrological condition, to relate wood biomass productivity to soil hydrology.The difference between this index and the HAND value is that the HAND data describe hydrology based on topography, and therefore will be most effective in terrains with pronounced topographical variations, since the level of the groundwater table may vary with the elevation depending on the soil conditions.The water saturation index, on the other hand, is based on soil features that were developed by long-term underground water fluctuations, such as the depth of the plinthite layer (see Quesada et al., 2010), and is therefore a more reliable indicator for variations in soil water saturation of the study sites.HAND data can be obtained with less effort than the soil water saturation index, however, they are less reliable in terrains with a smooth topography as it is the case in our study region.

Field measurements
The RAPELD program records all trees with diameter at breast height (DBH) above 30 cm in the installed 1 ha plots.Trees with a DBH of 10-30 cm are considered on two 10 m large sections on both sides of the 250 m long transect in the middle of the plot (0.5 ha) (Magnusson et al., 2005).In this study, we considered all trees with DBH > 30 cm of the forest inventory (Schietti et al., unpublished data) in order to obtain data on wood densities as well as to estimate ages and diameter increment rates for each tree by growth-ring analysis.A total of 554 trees were sampled in the eight Introduction

Conclusions References
Tables Figures

Back Close
Full plots, from which 22 individuals were not considered due to very low distinction of the growth rings.Table 3 shows the number of sampled trees with DBH > 30 cm in each plot.The sampling effort per plot was 47 to 87 trees with DBH > 30 cm per hectare, corresponding to 53-96 % of all individuals of this diameter class.Further we sampled 30 randomly selected trees within the DBH classes 10-30 cm in each plot.Palm trees (Arecaceae) were not considered in this study.
The DBH of all trees was measured by a diameter tape.In case of buttresses, diameter was measured above them to avoid overestimates of basal area and wood biomass.Tree height was estimated using a height measurement device (Blume Leiss BL6).Two wood samples were extracted from the trunk of each tree using an increment borer of 5.15 mm internal diameter.The sampling of wood samples was made 10 cm below the DBH (120 cm above the forest floor) to avoid errors for future repeated diameter measurements with the aim to monitor forest dynamics.One sample was extracted for wood density determination.To avoid dehydration these samples were labeled with the plot and tree number and stored in closed plastic bags.The second sample was extracted to estimate radial increment rates.These wood samples were glued on wooden supports with identity numbers for plot and tree.All wood samples were transported to the Dendroecological Laboratory of the scientific cooperation between INPA and MPIC (Max Planck Institute for Chemistry) in Manaus for further analyses.

Botanical data
All sampled individuals are currently being identified by Priscila Souza and Carolina Levis.The botanical material was pre-identified in field with the help of a parataxonomist.After a preliminary identification, the botanical identification was confirmed with the aid of specialists, identification guides and by comparing the vouchers collected to specimens at the INPA Herbarium (Manaus, Brazil) and virtual herbariums (http://sciweb.nybg.org/science2/vii2.asp).Fertile specimens will be deposited at INPA and sterile material will be stored in an adjacent working collection.Plants were identified in accordance with the APGIII (Angiosperm Philogenetic Group III) Introduction

Conclusions References
Tables Figures

Back Close
Full classifications.For name correction of the taxa the Brazilian Flora Species List was consulted (http://floradobrasil.jbrj.gov.br/2012/).

Biomass and carbon stocks estimates
Information on the total basal area per hectare for the diameter classes 10-30 cm and above 30 cm was obtained from forest inventory data (J.Schietti et al., unpublished data).Since trees with DBH of 10-30 cm are sampled only on 0.5 ha in the existing inventory, information on basal area for this cohort was multiplied by factor two to obtain estimates for the entire hectare.
To calculate wood density (ρ) the fresh volume (V fresh ) of each sample was determined by the water displacement method.The wood sample was mounted on a needle and immersed into a recipient filled with water on an analytic balance calibrated to zero.The sample volume is equal to the indicated weight of the displaced water after submerging the sample entirely into the water without touching the side or the bottom of the recipient.After this step of analysis the samples were dried at a temperature of 105 • C to obtain their dry weight (W dry ) (Chave et al., 2005;Sch öngart et al., 2005).
Wood specific density was than calculated as: For each plot we calculated the mean wood density and standard deviation.
Since there are large distances between the four studied modules and the vegetation type shifts from the north to the south within the interfluvial region, possible differences in tree height and wood density between the areas can be expected (Chave et al., 2005;Wittmann et al., 2006;Nogueira et al., 2008a, b;Feldpausch et al., 2011).As there are no specific allometric models available for the studied forest types, two models from other biogeographic regions were used for estimates of aboveground wood biomass (AGWB) (Cannell, 1984;Chave et al., 2005).These models use diameter, height and wood density as independent parameters to enhance the quality of the Introduction

Conclusions References
Tables Figures

Back Close
Full AGWB estimates (Chave et al., 2004(Chave et al., , 2005;;Sch öngart et al., 2010Sch öngart et al., , 2011) ) since the use of allometric equations with less than three independent parameters would lead to strong biases in the data (Feldpausch et al., 2012).For all of the equations described below, the parameters are referred as: aboveground coarse wood biomass (AGWB in kg), diameter at breast height (DBH in cm), tree height (H in m), and wood density (ρ in g cm −3 ).Cannell (1984) used a constant form factor (F = 0.06) to estimate woody biomass for tree species from the pantropics as the following product: Chave et al. ( 2005) developed different allometric models for forests types submitted to different climate conditions.For lowland rainfall forests with a marked dry season of 1-4 months and 1500-3500 mm yr −1 rainfall, the following equation was developed: As not all trees of the plot were sampled we estimate the AGWB in relation to the basal area of the sampled trees.This was performed separately for the two diameter classes DBH > 30 cm (> 30) and trees with DBH 10-30 cm (< 30).Table 3 indicates the number and percentage of all sampled trees, trees which were sampled for additional wood density determinations and not sampled trees.For this study we assume that the carbon content (C (%) ) of the AGWB depends on wood density (Elias and Potvin, 2003) estimated by: C (%) = (ρ + 16.21)/0.3732(4) To account for possible measurement errors in the field we performed an error propagation for our AGWB estimates.Errors in measurements of DBH were assumed as 1 %, 10 % for tree height and the standard deviation of the wood density of the plot for Introduction

Conclusions References
Tables Figures

Tree ring analysis and growth modeling
The prepared wood samples for tree-ring analysis were sanded and polished to produce a plain surface that enabled the visualization of the annual rings based on the wood anatomical analysis.Wood anatomy of tree rings was characterized following Worbes (2002): (1) intraannual variations of wood density within a tree ring where wood density increases from earlywood to latewood, typical for the families Annonaceae, Myrtaceae and Lauraceae; (2) tree rings delimited by marginal parenchyma bands commonly observed for species from the families Fabaceae, Meliaceae and Bignoniaceae; (3) alternated bands of parenchyma and fiber tissue, usually with higher fiber content at the beginning of the tree ring and higher parenchyma contents at the growth boundary commonly observed for the families Sapotaceae, Lecythidaceae, Combretaceae and Moraceae, and (4) ring-porous tree rings with larger vessels in the earlywood and smaller vessels in the latewood, rarely observed in the humid tropics (Fig. 2).
Ring widths were measured by a digital measuring device (LINTAB) with 0.01 mm precision attached to a computer with the software Time Series Analysis and Presentation (TSAP-WIN) to determine mean radial increments (Sch öngart et al., 2004).On Introduction

Conclusions References
Tables Figures

Back Close
Full samples which contained the pith tree age was estimated by direct ring count.For samples with missing pith we estimated tree age by dividing the obtained average diameter increment rates by the measured DBH in the field.The mean tree age per plot was calculated as the average of the ages from all sampled trees with DBH > 30 cm (Age >30 ) and with DBH of 10-30 cm (Age <30 ), weighted by their basal areas (BA >30 and BA <30 ) in the plot: Non-linear regressions were carried out between DBH and tree height using potential equations to produce DBH-height relationships for each plot and module.The relationship between tree age and DBH was fitted to non-linear regression models (Sch öngart, Estimation of forest productivity and carbon sequestration To estimate the AGWB productivity the cumulative diameter growth curve of a tree was combined with the standspecific DBH-height regression model.Together with the information of wood density it is then possible to predict for every age along the entire life span the AGWB by the allometric models of Eqs. ( 2) and (3) (Sch öngart et al., 2011).With these models we estimated the age-related aboveground wood biomass production of each tree (AGWBP tree ) calculating the average of the difference between the AGWB of consecutive years (t) (Eq.9).
Stand productivity was then calculated in two different ways, one to indicate the current total productivity of the stand and the other for data analysis to account for the long-term influence of environmental conditions on tree growth and wood biomass increments.
To indicate the current total productivity of the plot (AGWBP c ), AGWBP tree was calculated considering the last five growth rings of each tree (Eq.9, t = 5).As not all trees of the plot were sampled we estimated the productivity in relation to the basal area for all sampled trees.This was performed separately for the two diameter classes DBH > 30 cm (> 30) and trees with DBH 10-30 cm (< 30) (Eq.10).Relating the AGWB productivity per m 2 basal area multiplied by the total basal area in each diameter class gives an estimate of the stand's AGWB production.To account for the long-term influence of environmental conditions on tree growth and wood biomass increment (AGWBP m ), AGWBP tree was calculated considering the whole life span of each tree (Eq.9).For data analysis, AGWBP m was calculated by Eq. 11 considering the mean value of both allometric equations.Then, in Eq. ( 11) AGWBP m was also divided by the stand's basal area (SCF -structural conversion factor; Malhi et al., 2006) to account for high structural differences between the plots: BA Total (11) Data analysis were carried out using R-statistics.Graphs were built using R-statistics.

Results
Table 4 shows basal area, mean canopy height, mean wood density, mean tree age and the non-linear regression models between DBH and tree height for each site.Tree Introduction

Conclusions References
Tables Figures

Back Close
Full height was measured for a total of 630 trees.In all four modules DBH explains 56-62 % of the variability in tree height, with exception of module M08 where the multiple R-squared is only 0.30 (Table 4).
Of all sampled trees one has 214 rings including the pith.Another has 239 rings without reaching the pith, resulting in an estimation of more than 500 yr by extrapolating the number of counted rings in relation to the mean radius of the tree trunk.By this calculation, we estimate that from all the sampled trees, 33 would be more than 200 yr old.The non-linear relationship between DBH and tree age from all plots is significant (n = 534, df = 532, F = 108.65,R 2 = 0.29, p < 0.01) (Fig. 3a).Only about 30 % of the variability of tree age among different species and varying growth condition can be explained by DBH, due to differences in growth rates between species and individuals of the same species growing in different environmental conditions.However, the strength of correlation is not sufficient to consider the model to estimate tree ages in the field only by measuring DBH, independent of the species.Table 5 indicates mean wood densities and radial increment rates for the most common species.Among the most abundant tree species wood density varies between 0.44± 0.11 g cm −3 (Apeiba echinata) and 0.90 ± 0.05 g cm −3 (Licania oblongifolia).The lowest mean annual radial increments are observed for Pseudolmedia laevis (1.1 ± 0.2 mm) and Eschweilera coriacea (1.1± 0.4 mm), while A. echinata presents radial increment rates of 3.4± 2.4 mm.On the stand level mean wood density is significantly related to the mean radial increments (n = 8, df = 7, F = 11.12,R 2 = 0.65, p = 0.015) (Fig. 3b).

Conclusions References
Tables Figures

Back Close
Full 2.4-2.9Mg ha −1 yr −1 (Eq.2) and 2.7-3.2Mg ha −1 yr −1 (Eq. 3) in the other plots (Table 6).Our estimations of total AGWBP C result in much lower values for the M01 seasonally floodplain plots compared to the plots of other modules.The estimates of AGWB and AGWBP produced by Eqs. ( 2) and ( 3) indicate only small differences.
AGWBP is here hypothesized to vary with soil structure and hydrology.A strong relation is presented between AGWBP m (calculated from the average of all rings of a tree) and the soil water saturation index (Fig. 4a, n = 8, df = 6, R 2 = 0.83, F = 30.44,p < 0.01) representing the soil structure (effective depth) and hydrology.The soil water saturation index significantly affects the mean biomass production of the trees, once that increasing values of the index enhance AGWBP m (Fig. 4a) and decreases plot's mean age (n = 8, df = 6, R 2 = 0.7, F = 13.98,p < 0.01) (Fig. 4b).This pattern is not observed when we related the biomass production of the trees to the HAND data (n = 8, df = 6, R 2 = 0.04, F = 0.25, p > 0.05).However, when the two seasonally flooded plots of M01 are removed from the analysis, we observed a relatively strong relation between biomass production of trees and the HAND data, close to significance (n = 6, df = 4, R 2 = 0.61, F = 6.3, p = 0.06), following the same pattern as observed with the water saturation index.Table 7 summarizes the soil data obtained for each plot.The soil texture of all plots is mainly composed by silt, with varying small percentages of sand and clay.The concentration of iron molecules, which is the main element associated with the formation of plinth soils, ranges considerably among the plots between 66.7 and 388.8 mg g −1 soil .We find no correlation between productivity or stand age and any soil chemical property.However, since flooding on the M01 module is influenced by large river waters, soil chemical properties at M01 do not follow the same patterns as in other modules.Therefore, excluding the M01 plots, we noticed a very strong correlation between total productivity and available phosphorus in the soil (n = 6, df = 4, R 2 = 0.76, F = 13.21,p < 0.01) (Fig. 5a).We find no relation between the water saturation index and the phosphorus concentration in our plots, but on a regional scale, there is a positive correlation between the concentration of available phosphorus and the water saturation index (n = 42, df = 40, R 2 = 0.16, F = 7.76, p < 0.01) (Fig. 5b).Introduction

Conclusions References
Tables Figures

Back Close
Full

Discussion
The presented data on tree ages, diameter growth, estimates on AGWB and forest productivity indicate old-growth forests with large stocks of AGWB and relatively high AG-WBP, at the range of the nutrient-rich v árzea floodplains (Nebel et al., 2001;Sch öngart et al., 2010) and paleov árzeas (Stadtler, 2007) (Table 6), despite the low soil fertility of the area (RADAMBRASIL, 1978).However, the AGWBP of the studied forests varies considerably between 3.4-6.8Mg ha −1 yr −1 (Table 5).We find increasing wood biomass productivity with increasing water-logging indicated by the soil water saturation index.This is surprising, since flooding or watersaturated soils cause anoxic conditions (Lambers et al., 2007;Kursar et al., 2008;Haase and R ätsch, 2010;Piedade et al., 2010), leading to a decrease in diameter growth (Sch öngart et al., 2002) and a decline in forest productivity, as Stadtler ( 2007) indicated for the nutrient-poor black-water floodplain forests along a hydrological gradient in the Aman ã Sustainable Development Reserve, Amazonas state, Brazil.However, the surveillance of tree growth in floodplain forests reveal that trees start growing within the dry season, when forests are still flooded (Sch öngart et al., 2002).That could indicate that the conditions of saturation and flooding may generate different responses of tree growth, with water saturation being favorable for tree growth while flooding being unfavorable for tree growth.
The increase in the wood biomass productivity is not observed when using HAND data.However, not considering the two seasonally flooded plots in the analysis, the HAND data reveals a negative correlation with the total ABWBP in accordance with the results obtained using the soil water saturation index.Those two referred plots, despite being seasonally flooded, are scored with a lower soil saturation index than another plot that is saturated but not flooded.Therefore, the structure of the soil related to hydrology (i.e.soil depth, drainage and duration of the saturation period) seems to be more important for the productivity of trees than just the water level described by the HAND data and height of the water column.Indeed, during field work it became

BGD Introduction Conclusions References
Tables Figures

Back Close
Full apparent that the HAND data does not describe well the local hydrology related to the smooth topography of the terrain and the inundation of the seasonally flooded plots, and thus is not a reliable data base to relate hydrology with estimated forest productivity alone.
Soil water saturation may vary with soil type and topography of the area.As described by Quesada et al. (2010Quesada et al. ( , 2011) ) plinthosols, which are quite common in the study region, develop under non-optimal soil conditions by deposition of nutrients like iron.The phosphorus in these soils remains mainly associated with iron and is liberated when the oxidized iron is reduced in the flooded plinthite layer (Chacon et al., 2006).Under these conditions it is thinkable that hydromorphic soil characteristics influence the phosphorus availability in the soil, which in addition with the poor soil structure presented on each area, affects wood productivity of the different stands.This hypothesis is corroborated by the observed correlations between total wood biomass productivity and available phosphorus as well as between the regional variation of available phosphorus and the water saturation index.On the most saturated areas with poor soil structure, more plant-available phosphorus can be found.In the generally nutrientpoor soils of the study region (RADAMBRASIL, 1978) the release or reduction of one growth-limiting nutrient could possibly be sufficient to have strong impacts on forest growth.On well-drained soils, there is less reduction of the oxidized iron and as a consequence phosphorus would become more and more unavailable for the roots due to the strong fixation by the present oxidized iron, leading to reduced tree growth.A similar pattern was indicated by Clawson et al. (2001) at a semi-deciduous forest in the SW-USA indicating an increase in forest productivity with decreasing soil drainage, where phosphorus releasing was associated to reduction of oxidized iron in bad drained soils.The variations in wood biomass productivity cannot, however, be associated with the chemical reactions of the soil alone.We show that there is indeed a variation of available phosphorus with the water saturation index, but that variation is still weak.We think the physical constrains related to the bad hydromorphic structure of the soil is also a very strong factor.It is possible that the water-logging of the soils at different Introduction

Conclusions References
Tables Figures

Back Close
Full depths of the plinthite layer acts as an environmental filter, creating a trade-off where plants would grow faster and present shorter life-cycles.This hypothesis is supported by the strong trend we found between the soil water saturation index and stand age.
A high productivity means that trees grow faster and achieve larger diameters at lower ages.Water-logging can lead to shifts in species compositions as it is well known form Amazonian floodplain forests (Wittmann et al., 2006(Wittmann et al., , 2010) ) towards tree species with higher growth rates and shorter life cycles.In this case we would expect a lower wood density for those tree species compared to slow-growing tree species (Gourlet-Fleury et al., 2011) that attain high ages, which is not the case for our studied forests.Still, despite the low variation of mean wood density, we find a significant negative relationship between mean radial increments and wood density between the studied sites which suggests that there is possibly a shift in tree species composition.Such variation could be a consequence of water-logging in the soils, leading to species assemblies with faster growth rates and shorter life spans due to special adaptations as it is described for trees in the central Amazonian floodplains (Parolin et al., 2004).
A third explanation could be that during the dry season water-logged sites provide a better water supply for tree growth than well-drained sites where tree growth is reduced during the dry season (Brienen and Zuidema, 2005).In this case we would expect varying climate-growth relationships between well and poorly drained soils.This is a hypothesis to be tested in future studies applying dendroclimatology for characteristic tree species from different ecotypes between well-drained and water-logged soils in the interfluvial region using traditional tree ring analysis and stable isotopes analyses.We predict that the vegetation period of tree species varies temporarily between different forest types depending on the soil type and hydrological regime such as it was observed between floodplain forests and adjacent terra firme forests in Central Amazonia (Sch öngart et al., 2004(Sch öngart et al., , 2010)).
In Table 8 we compare the results for AGWB and AGWBP in this study with other studies in different regions of the Amazon basin.The AGWB stocks in the studied forests of the interfluvial landscape of the Madeira-Purus region are lower than in Introduction

References
Tables Figures

Back Close
Full terra firme forests, but higher than in floodplain forests and old fluvial terraces (paleov árzeas).Comparing the woody biomass productivity our studied forests indicate a similar AGWBP as paleov árzeas and v árzeas.In comparison to the seasonally flooded and nutrient-poor igap ó and the Central Amazonian terra firme forests the studied forest present a higher AGWBP.Only the Southwestern Amazonian terra firme forests seem to be more productive, however, as Malhi et al. (2004) applied allometric models which do not consider tree height for the estimates of AGWBP, the values could present biases in the estimates (Sch öngart et al., 2010;Feldpausch et al., 2012).
We conclude that AGWBP in the study region is as high as in v árzea and paleov árzea regions that are among the most productive forest types in the Amazon basin.Productivity varies with the hydrological conditions, but it seems that variation in the soil water saturation has even a bigger impact on forest productivity then simply the topographic variation of the terrain.However, it remains unclear to us whether hydrological, climatic or edaphic properties or a combination of all control productivity of these forests, since the structure of soils in the area is strongly determined by underground water fluctuations.
A high natural productivity of a forest ecosystem is one of the criteria for the development of sustainable forest management plans (Sch öngart and Queiroz, 2010).However, the extraction of timber resources in tropical forests as practiced in general is not sustainable (Brienen andZuidema, 2006, 2007;Sch öngart, 2008;Shearman et al., 2012).In our study region, the most productive sites are encountered on poorly structured soils.The poor structure will probably lead to slow recover of the forest after extracting, and therefore make extraction of timber resources not sustainable.Species and site specific forest managements have to be developed as it was formulated by the GOL-concept (Growth-Oriented Logging) for the high-productive v árzea floodplain forests of Central Amazonia (Rosa, 2008;Sch öngart, 2008Sch öngart, , 2010)), that resulted in the Normative Instruction (IN)no.009.In this forest legislation timber resource management was differentiated for fast-growing tree species with low wood densities (madeira branca) and those with high wood densities above 0.60 g cm −3 (madeira pesada), Introduction

Conclusions References
Tables Figures

Back Close
Full applying felling-cycles of 12 and 24 yr, respectively, and species-specific minimum logging diameters.Species-specific and site-specific growth models based on tree rings together with studies on the regeneration dynamics and population structure should be performed to evaluate the potential of forest management plans for the region where huge complexes of conservation units have been established allowing the sustainable use of timber resources.General global climate models predict a huge impact for the Amazon basin and its carbon stocks and uptakes, mainly due to shifts in the precipitation and hydrological regimes (Dufresne et al., 2002;Betts et al., 2004;Cook et al., 2012).These changes are mainly caused by the increase of sea surface temperature (SST) anomalies in the Equatorial Pacific (Sombroek, 2001;Foley et al., 2002;Marengo, 2004;Sch öngart and Junk, 2007) and the tropical Atlantic Oceans (Tomasella et al., 2010;Yoon and Zeng, 2010).During the severe drought events in 2005 and 2010, which affected between 2.5 and 3.2 million square kilometers in the Amazon basin, respectively, the interfluvial region between Purus and Madeira Rivers suffered negative precipitation anomalies in 2005 and 2010 (Phillips et al., 2009;Lewis et al., 2011).However, it is unknown how such severe droughts affect the patchwork of floodplain forests, forests on waterlogged soils and well-drained site.Tree species in Amazonia within the same stand and between ecosystems present varying climate-growth relationships (Worbes, 1999;Sch öngart et al., 2002Sch öngart et al., , 2004Sch öngart et al., , 2010;;D ünisch et al., 2003;Brienen and Zuidema, 2005) and may present varying responses to soil water saturation (Rodríguez-Gonz ález et al., 2010).More field data are necessary to relate recruitment and mortality rates to interannual climate and hydrological variation.Such data are essential for the sustainable development of this particular region in terms of timber resource management and conservation as well as future scenarios for carbon stocks, emissions and uptakes due to changes in land-use and climate change.

Conclusions References
Tables Figures

Back Close
Full

Conclusions References
Tables Figures

Back Close
Full  Full  Plot ID M01-TN1500 M01-TN2500 M05-TN(-)500 M05-TN1500 M08-TS2500 M08-TS4500 M11-TN1500 M11-TN2500 Full Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | providing laboratories and logistics; the PPBio/CENBAM for financing and providing all field infrastructure and the HIDROVEG project for logistics support for all fields expeditions in the BR-319.The service charges for this open access publication have been covered by the Max Planck Society.
Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper |

Fig. 1 .
Fig. 1.Map of the Purus-Madeira interfluvial area that is crossed by the BR-319 Highway.The study sites are indicated by white asterisks.All study sites are sampling modules (M) of the PRONEX Project Rapid Assessment for Long Duration Ecological Projects (RAPELD).

Table 1 .
(Renn ó et al., 2008;, hydrology and rainfall patterns of the selected study sites.HAND (Height Above Nearest Drainage) model data indicate the elevation above the nearest drainage(Renn ó et al., 2008; Nobre et al., 2011); number in parenthesis indicates the flood height of seasonally inundated forests measured in the field.SRTM (Shuttle Radar Topographic Mission) digital elevation model estimates the elevation above sea level of each plot.The Soil Water Saturation (SWS) index is explained in more detail in Table2(no data of the SWS is available for the plot M05.TN500).Rainfall data were obtained from the Brazilian Waters Agency (Ag ência Nacional de Águas -ANA).

Table 4 .
Basal area per hectare for the diameter classes DBH > 30 cm and < 30 cm, mean canopy height (H, trees with DBH > 30 cm), mean wood density (ρ) with standard deviation of each plot, mean tree age(Age) and linear regression models between DBH and tree height for each module.