Extended summary of "Diplom" thesis

Accumulation of nutrients within the aboveground biomass of the land use system’s commercial species is one main path of the nutrient fluxes investigated by ENV-45. Its determination at an age of the plants of 4 years was the final aim of this study. The accumulated biomass - pre-requisite to determine accumulated nutrients - was also of interest for comparing the performance of the different species and of the fertilizer treatments. Since the systems are also to be investigated in future, a minimal-invasive method had to be used. Thus, precise allometric biomass models for the non-destructive determination of the aboveground biomass were developed. These models were applied to biometric data measured in field; nutrients accumulated within the aboveground biomass of the trees were determined from plant tissue samples taken from the living trees. Since the differentiation between long-living and short-living plant organs was of need, separate models for leaves/leaflets, branches/leafaxis, and stem of the species were developed. The biomass of the single trees was to be determined as precisely as possible since the data was also to be used at the level of the single trees and (by other members of ENV-45) at the level of single branches. Another aim of this study was to provide more information and data about autecological characteristics of the species; rather few is known so far, compared to their actual and prospected importance as crop.

Allometric biomass models were developed for castanha, cupuaçu, pupunha (managed for fruit production), and urucum. 7 (castanha, cupuaçu), 5 (pupunha (fruit)), and 4 (urucum) borderline trees - representative for the population - were cut in the experimental plots and biometric variables and biomass data were determined. As raw models, the "general allometric model" (b^.x^c) or the simple linear model was found to be appropriate in most cases. To account for heteroscedasticity, a new weight (1/Predicted²) was proposed. This weight was derived to be appropriate for all allometric data sets that yield homoscedastic residuals under log-log-transformation - which is the case for most allometric biomass relationships. By applying weighed non-linear regression the disadvantages confronted within the log/log-transformation are avoided. To allow for a quantitative comparison among different fits and different data-sets, a new statistic was suggested; r², COV and others were found to be inappropriate. This statistic, the Adjusted Coefficient of Variation (ACOV), is a modified COV accounting for the weighing process. In addition to models using data of the cut trees only, "improved models" using additionally data of the living trees were developed resulting in less biased estimates of the weaker models.

At the level of the single branch/frond, the branch/rachis diameter was the best single variable for estimating its leaf(let) and wood biomass. For pupunha (fruit), also the absolute number of the respective leaf was found a good and simple variable. Multi-variable models including the length of the branch/rachis did not substantially improve the models. At the level of the whole tree, single or several stem diameters at relative stem height resulted in good models for the estimation of leaf-, branch-, and stem biomass of castanha. For cupuaçu, the stem diameter at 20 cm aboveground resulted in quite precise models for the plant organs leaves and wood. For pupunha (fruit) the tree height variable DiffHeight and also the total number of stem-inserted leaves formed by the tree so far were found to be good variables for estimating stem biomass. The precision of the best models for the respective plant organ was found to lie between 2.6 % relative error for castanha stems estimated from the approximated stem volume, to (approximately) 20 % for the biomass of a single pupunha (fruit) leaf estimated from the rachis diameter. Only the prediction of the biomass of the urucum prunings - using the sum of three stem diameters - had an (expected) higher relative variation of around 35 %.

The biomass of the different aboveground plant organs of the single trees of castanha, cupuaçu and pupunha (fruit) growing in the investigated experimental plots was estimated allometrically and corrected to a common date. The within-plot variation in biomass of trees of the same species was very high. Between the two fertilizer treatments of the AF-System some - but no significant - differences of the mean aboveground biomass of the three species were found. Mean aboveground biomass of 4-year-old castanha trees of the AF-System (both treatments) was 30.4 kg (32.5 kg in the 100%, and 28.3 kg in the 30% treatment). The mean pupunha (fruit) tree weighed 33.6 kg (with 40.8 kg in the 100%, 26.5 kg in the 30% treatment), the mean cupuaçu tree 4.14 kg (with 4.51 kg in the 100%, and 3.77 kg in the 30% treatment). Mean biomass of pupunha (fruit) was about the same in the monoculture (38.8 kg), whereas the cupuaçu trees of the respective monoculture were surprisingly significantly smaller (2.02 kg) than the trees of the AF-System 100% (ANOVA, tested pairwise as planned comparisons, a = 0.05). This was the only significant difference in mean tree biomass found among the treatments. For urucum a common value of 13.7 kg before and 6.5 kg after pruning was estimated from biometric data of 36 borderline trees. For pupunha (palm-heart) harvest data of C. R. de A. Moraes (SHIFT) allowed to approximate an aboveground biomass of 7.0 kg as mean value per plant position just before harvest.

Summing up the estimated aboveground biomass of all commercial species, the mean biomass of the respective treatment per ha (at a plant age of around 4 years) was obtained. Highest aboveground biomass was found for the right half of the pupunha monoculture (37.4 t/ha, sub-system "8a") with the pupunha (fruit) trees in high density, interplanted with pupunha (palm-heart) plants. Also the biomass of the left half of that system with all 2500 plants/ha managed for palm-heart (sub-system "8b") was approximated to have been high (17.5 t/ha). The two treatments of the AF-System had an aboveground biomass of 10.4 t/ha (100% fertilizer) and 8.84 t/ha (30% fertilizer); this difference was not significant. The monoculture of cupuaçu had the lowest aboveground biomass of only 0.447 t/ha; the low planting density (223.2 trees/ha) and the low biomass of the single trees of the slower growing species cupuaçu were the reason. From root/shoot ratios calculated by Haag (1997) [unpubl.], the belowground biomass of the trees of the commercial species was approximated. Aboveground biomass of the secondary growth within the plots was approximated to be between close to zero (in the pupunha monoculture) and up to about 2.3 t/ha (in the cupuaçu monoculture).

In addition to biomass data, the leaf area, the crown-diameter and -volume, the LAI of the single trees and of the system, and the leaf volume index (LVI) was determined. Looking at the distribution of biomass and crown areas in the space of the systems, the layout of the AF-System and of the pupunha monoculture were seen to be generally well chosen, although in future some problems of shading of cupuaçu and urucum probably reducing fruit production are to be expected for the AF-System. In the cupuaçu monoculture, in contrast, much space between the trees remained unused (coverage of the cupuaçu crowns in May 1996: 2.8% of the system’s area); secondary growth - although regularly cut down - had an approximately up to 5 times higher aboveground biomass within the system than the cupuaçu trees. The cupuaçu monoculture was therefore suggested to be extended to an agroforestry system, cropping annual species between the cupuaçu trees (and/or lowering the planting density of these). For the AF-System a prolonged cropping of annual species was suggested, too. A quite important finding for all systems was the high small-scale heterogeneity of biomass and crown diameters of the single trees. Differences among the species were found to be smaller than those among the single trees of the same species. This spatial heterogeneity disturbs the layout of the system, which is planned for the average growth of the respective species. Some suggestions were made to manage this heterogeneity in future experiments.

Preliminary absolute and relative growth rates as well as the Unit Leaf Rate of the single trees and of the species were calculated around the first half of 1996. The mean relative growth rate was highest for castanha trees of the AF-System-30% (0.00388 d^-1) and lowest for the cupuaçu trees of the monoculture (0.00220 d^-1). Differences among the treatments were small for castanha and high for cupuaçu and pupunha (fruit). Among the individuals a considerable variation was found. The lower relative growth rate in the cupuaçu monoculture suggested, that the relative differences in biomass will even become higher in future. The highly variable and low average relative growth rate of pupunha (fruit) was attributed to the drought-stress during the dry season, in which period growth of this species was determined. The average Unit Leaf Rate was (nevertheless) highest for the population of pupunha (fruit) in the AF-System-30% (2.59 g^.d^-1.m^-2), and lowest for cupuaçu of the monoculture and of the AF-System 100% (0.96 g^.d^-1.m^-2 and 0.94 g^.d^-1.m^-2, respectively).

From the presence of leaves at lower branches close to the stem, for castanha a leaf-life time of surprisingly more than two years was derived.

Accumulated biomass and relative growth rates of the single trees were used to investigate possible competition within the AF-System. It was found that a relevant competition among the species castanha, cupuaçu and pupunha (fruit) was unlikely to have occurred until May 1996. An approximation of the aboveground biomass of the secondary growth suggested some competition with the pupunha (fruit) trees - considering the extension of pupunha’s root-system found by Haag (1997) [unpubl.]. For the pupunha (fruit) trees of the monoculture the expected intraspecies belowground competition was found to obviously have no (formal) effect on the mean biomass of the trees by May 1996; this was about the same as in the more open AF-System. The fruit trees of the pupunha (monoculture were assumed to have profited from fertilizer applied to their small and shaded (and thus less demanding) palm-heart neighbors.

Drought-stress during the dry season of 1996 was found for pupunha (fruit) resulting in a reduction of the living leaves for most trees. The reduction was up to more than a third - especially in bigger individuals. Observations made on the foliage pattern of castanha and data on LVI of cupuaçu in the monoculture, suggested a probable relevance of drought stress for these species as well.

About 10% of the pupunha (fruit) trees of the AF-System and approximately 17% of the individuals of the monoculture were felled by wind until September of 1996. No trees of the other species were lost.

Composed samples of leaves/leaflets and branches/leaf-axes collected from all inner trees of castanha, cupuaçu, and pupunha (fruit) in the investigated plots were used to calculate accumulated nutrients in the aboveground tissue. The nutrient concentrations of the stems were approximated from nutrient relationships between branches and stem established from the trees cut. For urucum, nutritional data of the trees cut was used - the nutrient concentrations for pupunha (palm-heart) were approximated by (modified) nutritional data of the pupunha (fruit) trees. The amounts of nutrients accumulated within the species and in the whole plantation systems were found to be about proportional to the respective biomass; for the pairs pupunha/cupuaçu and castanha/cupuaçu, differences in nutrient concentrations between the species were much smaller than differences in their mean biomass. Since variation of the single trees’ biomass was high, also the variation of accumulated nutrients was high. Hence, for determining the spatial distribution of nutrient sinks in the system, the importance of a precise determination of biomass of the single trees is stressed; the determination of the average tree biomass and nutrient-content of the respective species is sufficient at the system-level only.

When pruning urucum by May 1996, about 50% of the aboveground biomass and about two thirds of the nutrients accumulated aboveground were cut.

Differences in nutrient concentrations among the species were found to be significant for the majority of the nutrients and tissues; the species thus have to be sampled separately. Pupunha had rather low mean concentrations in total aboveground biomass but high concentrations in the (short-living) leaves, whereas cupuaçu had rather higher overall concentrations for most nutrients, but the gradient between (long-living) wood and (short-living) leaves was smaller. Also a comparison of the concentrations among the treatments showed several significant differences; the variation among the plots was high, too. It appears to be inevitable to analyze tissue samples of the plots separately. A relevant variation was found also among individual trees; depending on the required precision, single trees might have to be analyzed separately when accumulated nutrients of the single trees are to be determined. Correlation of tissue concentrations with biomass showed a likely (long-term) deficiency of Mn and Ca in cupuaçu and of Mn in castanha. K in branches of castanha was found to be significantly negatively correlated with biomass.

A comparison of the time course of fertilizer application with the approximated amount accumulated within the growing trees of castanha and cupuaçu was carried out using additional biometric data of C. D. M. Nunes and R. M. B. de Lima (SHIFT). This comparison showed, that in the first years of growth a high (formal) surplus of fertilizer was applied compared to the amount accumulated. A formal surplus by a factor of 40 to 50 for N and a factor of 100 for K was found for an average castanha or cupuaçu tree during the first 21 months of growth. In the last year before May 1996 the fertilizer-supply of these two nutrients met about the actual demand of an average tree of these species, or was much lower (for N in castanha). It was suggested to modify the distribution of fertilizer application in time to obtain a more economic use of the fertilizer.

Facing the high heterogeneity of the demand of the single trees it was also suggested to differentiate the amounts fertilized depending on the size of the single trees. A simple procedure thought to be practically applicable was suggested to this end. Further considerations including the low CEC_eff of the soil led to the suggestion, of applying slow-release fertilizers both to guarantee a constant supply of the trees and to avoid a high, unproductive loss of nutrients leaching from the system.

WOLF, M.-A. (1997). Accumulation of biomass and nutrients in the aboveground organs of four local tree species in monoculture and polyculture systems in central Amazonia. German "Diplom"-thesis [unpubl.]. Technische Universität Braunschweig.

Contact me: