Forest litter decomposition as an effect... FOREST LITTER DECOMPOSITION AS AFFECTED BY EUCALYPTUS STAND AGE AND TOPOGRAPHY IN SOUTH-EASTERN BRAZIL1

Forest litter decomposition is a major process in returning nutrients to soils and thus promoting wood productivity in the humid tropic. This study aimed to assess decomposition of eucalypt litter in the Rio Doce region, Brazil. Leaf litter was sampled under clonal eucalypt stands aged 2, 4 and 6 years on hillslopes and footslopes. Soil and soil+litter samples were incubated at two levels of soil moisture, temperature and fertilization. C-CO2 emissions from soil measured during 106 days were higher at 32 °C than at 23°C, mainly for the 2-yr-old stand on footslope. When leaf litter was added on soils, C-CO2 emissions were eight times higher, mainly on footslopes, with no effect of stand age. Leaf decomposition in situ, assessed with a litterbag experiment showed a mean weight loss of at least 50% during 365 days, reaching 74% for 2 yr-old stands on footslopes. In comparison with data from the native forest and the literature, no apparent restrictions were found in eucalypt litter decomposition. Differences between in vitro and in situ results, and between eucalypt and native forest, were most likely related to the response of diverse decomposer communities and to substrate quality.


INTRODUCTION
Forest litter decomposition plays an important role in returning nutrients to soil, and is critical to the productivity of fast-growing tree plantations in the humid tropics.However, there have been reports of intense litter accumulations (>25 Mg ha -1 ) under eucalypt stands in Southeastern Brazil (Gama-Rodrigues et al., 2008;Palha Leite, F., unpublished data), raising concern about proper nutrient cycling.Identifying the causes of such unbalances can be difficult since factors such as phenology and litter quality interact with the biotic and abiotic environment, and also because of seasonal and yearly climate variation.Measuring litter decomposition processes is important in assessing litter dynamics, but difficult because decay rates and decomposer organisms vary over time and space (PRESCOTT, 2005).Additionally, stand age and position in the landscape can affect the decomposition process because of their effect on litter production and site properties, but this has seldom been studied (BARGALI, 1996;ALVAREZ-SÁNCHES;ENRÍQUEZ, 1996).
This study aimed to investigate potential restrictions to eucalypt litter decomposition and nutrient cycling in the Doce River region, which may be related to stand age, topographic position, temperature, soil moisture and fertility.To test this hypothesis, leaf litter from different combinations of ages and topographic positions were incubated at different temperatures, soil moistures and fertilizations.In addition, a litterbag experiment was established to assess decomposition and nutrient release in situ.

MATERIAL AND METHODS
The study area is located in the Doce River near Belo Oriente, Minas Gerais, Brazil (19°18'S and 42°22'W), with a mean altitude of 336 m above sea level.The natural vegetation is semi-deciduous tropical forest.According to the Koeppen classification, the climate is a tropical humid, with mean annual temperature and precipitation of 25.2°C and 1,000-1,280 mm, respectively.In general, the rainy season is from October to March, comprising about 80% of mean annual precipitation.Potential annual evapotranspiration varies from 950 to 1,200 mm, with an annual water deficit between 30 and 90 mm.The dominant soils in the area are Oxisols and Ultisols of clayey texture developed from Proterozoic granite and gneiss.Soil used in this incubation has pH values from 4.4 to 5.9, and soil organic matter between 13 and 18 g kg -1 under eucalypt stands.Under native forest, soil organic matter is higher (28 g kg -1 ), and soil pH is 5.9.
Soil and leaf litter were sampled in December 7 th 2000, under Eucalyptus grandis x E. urophylla hybrid clonal stands (3 x 2.5 m spacing) aged 2, 4 and 6 yrs on footslope and hillslope positions, and under a native forest on the footslope.For each combination of stand age and topographic position, recently abscised leaves, varying from green to yellow-green colors, were collected in triplicate on the forest floor and stored at 4-7 o C for the CO 2 -C evolution assessment.Leaf lignin was determined by the acid detergent fiber method of Van Soest and Wine (1968).In each sampling plot described above, soil from the 0-20 cm depth was sampled and stored at 4-7 °C for two weeks before incubation.Soils were incubated with and without leaf litter fragments.The sampled leaf material was air-dried and cut to approximately 0.5 x 0.5 cm fragments.Subsequently, 15 g of leaf fragments were brought to 80% water retention capacity and placed onto 50 g of air-dried soil sieved <2 mm contained in 0.6 dm³ glass jars.Incubations were conducted at two levels of temperature, moisture and fertilization.The average minimum and maximum monthly temperatures of the region, respectively 23 and 32°C, were used.Soil moistures were 40 and 80% of the water holding capacity.Fertilization levels were 0 and 5 mL of a solution with 50 mg L -1 P(KH 2 PO 4 ), 80 mg L -1 N(NH 4 Cl), and 80 mg L -1 K(KCl).
The CO 2 evolved was trapped with NaOH and titrated with HCl using phenolphthalein as indicator, after addition of BaCl 2 to precipitate sodium carbonate (ANDERSON, 1982).The first titration was done after 2 days of incubation, followed by a new incubation for 3 days and titration, and so on, for 4, 6, 8, 10, 12, 14, 20, and 27 days, in a total of 106 days.Results were expressed in CO 2 -C mg g -1 soil evolved, according to the formula: Forest litter decomposition as an effect...The experimental design was a completely randomized factorial design, in which the main factors were stand ages in three levels (2, 4 and 6 yrs) and topographic positions in two levels (footslope and hillslope), summing up to six main plots.Additionally, a split-plot was used, as each main plot was subdivided into a combination of three incubation factors in two levels: soil moisture, temperature and fertilization, resulting in eight subplots.As a reference, a nearby native forest on the footslope was sampled and compared with the eucalypt stands on the footslope.Data on accumulated C-CO 2 after 106 days were submitted to analysis of variance followed by the least significant difference (LS Means STDERR PDIFF) or the Ryan-Einot-Gabriel-Welch Q (REGWQ) test of means (SAS, 1990).
A litterbag experiment was set to determine in situ decomposition rates of leaf litter.In each of the eucalypt plots mentioned above, 24 nylon bags of 23 x 23 cm (3-mm mesh) filled with 40 g of recently fallen leaves were placed on the soil between tree rows.Four litterbags were retrieved at intervals of 52 (December.12 th ), 64 (February.20 th ) days, 46 (April 7 th ) and 203 days (October.24 th 2001).The total number of litterbags was 432 (3 ages x 2 slopes x 3 replicates x 6 samplings x 4 subsamples).Due to technical difficulties, the 5 th and 6 th samplings were done together in Oct. 24 th , 2001.In addition, 54 litterbags (3 replicates x 6 samplings x 3 subsamples) containing a mixture of recently fallen leaves were distributed randomly in the native forest, and retrieved as above.The leaf material in litterbags was oven-dried at 75°C, weighted and analysed chemically.About 0.5 g of fresh ground litter material was digested with concentrated HNO 3 and HClO 4 (JONES; CASE, 1990).From the digest, P (Olsen;Sommers, 1982) and B were determined by spectrometry; K and Na by flame photometry, Ca, Mg, Cu, Mn and Zn by atomic absorption.Total nitrogen was determined by the Kjeldahl method (BREMNER; MULVANEY, 1982).Total carbon was estimated by weight loss on ignition at 550 o C (JONES; CASE, 1990).

Leaf litter quality
Table 1 shows the initial nutrient, carbon and lignin concentrations in the leaf litter according to stand age and topographic position.Total N was higher in leaf litter of 2-yr-old stands compared to 4-and 6-yr-old stands.There was no effect of topographic position on nutrient concentrations.No significant differences were found for carbon, lignin concentration and C:N, C:P and lignin:N ratios in eucalypt leaf litter due to stand age and topographic position.The mean C:N ratio of eucalypt leaf litter was 40:1, similar to values reported by Sodré (1999) in Southeast Bahia, Brazil.Leaf litter from the native forest showed high concentrations of Ca, Mg, N, K, Cu and B compared to eucalypt litter.

C-CO 2 evolved
The accumulated soil respiration after 106 days was affected mainly by temperature and stand age (P<0.01), and secondarily by slope (P=0.052).In general, soil respiration was significantly higher at age 2 yrs on the footslope (Table 2), and was not affected by moisture and fertilization.When leaf fragments were added on soils, the total C-CO 2 released was affected significantly by temperature and slope (P<0.01),but not stand age and moisture (Table 2).Presence of leaf litter fragments on soils increased the C-CO 2 evolved up to 8-fold, when compared to soil respiration (Table 2).
The C-CO 2 evolved was compared among the three eucalypt stand ages and the native forest on footslope position.Despite the lower soil respiration for stand age 4 yrs, there were no major differences under the two vegetation types (not shown).The total C-CO 2 evolved during the experiment was better described by linear functions (Figure 1), in which intercepts and slopes were higher at 32 o C compared to 23 o C, in both topographic positions and vegetation types.

Litterbag experiment
Litterbag weight loss was fast during the first 60 days, and then slowed down, with little change after 154 days.Weight loss after 120 days was generally lower for stands aged 4 and 6 yrs, compared to the 2-yr stands, reaching ca.48 % by 365 days.The weight loss was higher under 2-yr-old stands, reaching 74% on footslopes, and 58% on hillslopes, in the 116-365 days period.As a comparison, litterbags in the native forest on footslope showed a weight loss of 68%.However, decomposition rates in the native forest followed a contrasting pattern, being initially lower, increasing in the 52-116 day period, then becoming comparable to eucalypt decomposition rates in the 116-365 day period (Figure 2).SKORUPA, A.L.A. et al.

C-CO 2 evolved from soil and leaf litter
The non-significant effect of substrate moisture and fertilization indicates that 40% water holding capacity and current soil fertility were sufficient for an optimal soil respiration in this study.Correlations between CO 2 emission rates and water content vary according to soil moisture conditions.Kiese and Butterbach-Bahl (2002) reported that CO 2 emission rates were positively correlated with moisture in dry soils, but negatively correlated when water-filled pores were>50-60%, because of limited O 2 diffusion in the soil matrix.Other authors proposed that C-mineralization and CO 2 emission rates are incremented by drying and wetting cycles (VANGESTEL et al., 1993;FIERER;SCHIMEL, 2002).  1 -Mean concentrations of nutrients, carbon, lignin and C:N, C:P and lignin:N ratios in leaf litter from eucalypt stands aged 2, 4 and 6 yrs, and native forest.Tabela 1 -Média de concentrações de nutrientes, carbono, lignina e taxas de C:N, C:P and lignin:N em frações foliares de povoamentos de eucalipto nas idades de 2, 4 e 6 anos e em uma floresta nativa.
Means (n = 3) at the same column followed by the same letter are not significantly by REGWQ test (P< 0.05).
Forest litter decomposition as an effect...
In the present study, C-CO 2 evolved from soils under ages 2 and 6 yrs were respectively 60 and 100% higher at Despite the initially higher N, P and Ca concentrations in leaf litter under 2-yr-old stands (Table 1), there were no differences due to age in total C-CO 2 evolved when leaf fragments were added on soils (Tables 2).This indicates that substrate quality was not a limiting factor to microbial activity, which is corroborated by the nonsignificant effect of fertilization.Fertilization of eucalypt stands has been shown to increase mineralization of added nutrients, but not litter decomposition rates (RIBEIRO et al., 2002;O'CONNELL;MENDHAM, 2004).Soil respiration on footslopes was higher than on hillslopes, despite the similar substrate chemistry (Table 1).This difference may be caused by a diverse microbial composition, since more mycelia were visible on footslope samples.The constant rates of C-CO 2 release from soils and soil + leaf litter (Figure 1) indicated a homogeneous lability of the organic substrates respired during the 106 days incubation.
The similar soil respiration rates between eucalypt and native forest on footslopes contrasts with the reportedly higher values for soils under native forest, usually higher in soil organic carbon, fertility and pH (DELLA BRUNA et al., 1991;FIALHO et al., 1991;CARVALHO et al., 1997;SODRÉ, 1999;ASSIS JÚNIOR et al., 2003).This trend indicates that labile C pools are equivalent, and the higher soil organic matter under native forest is mostly in stable forms that did not undergo decomposition during the incubation period.In Oxisols, the recalcitrant pool of soil organic matter associated to Fe/Al hydroxides and clays controls mineralization of N (SIERRA; MARBÁN, 2000) and probably also C. The similarity between the two vegetation types also occurred when leaf litter was added on soil, which suggests similar leaf litter quality, despite the lower nutrient contents and higher C:N, C:P and lignin:N ratio in eucalypt litter (Table 1).Accordingly, earlystage CO 2 release from eucalypt and other tree species litter was not affected by initial N and lignin contents (BERNHARD-REVERSAT, 1998).Fisher and Binkley (2000) pointed out that litter quality indicators such as C:N and lignin:N do not always relate well to field and laboratory decomposition, which restricts their wide use.Luo and Zhou (2006) agree that multiple interacting factors render soil respiration understanding very difficult.Literature data from in vitro studies showed that mean daily soil respiration for eucalypt stands and native/secondary forests in Brazil is 0.015 and 0.03 mg C-CO 2 g -1 soil d -1 , respectively (Table 3).These values are proportional to the mean soil organic carbon concentrations of 20.5 and 26.5 g kg -1 , for eucalypts and native vegetation, respectively.In this study, mean soil respiration from eucalypt and native forest were respectively 0.007 and 0.008 mg C-CO 2 g -1 soil d -1 .These values are low compared with the means and similar only to those of low-carbon soils in Table 3, and may be caused by lower soil organic carbon contents.

Litterbags
The significant effect of topography on C-CO 2 evolved in vitro by both soil and soil+leaf fragments was reflected into higher decomposition in situ only for 2-yr old stands, which lost 74% of litterbag weight on footslopes, compared to 58% on hillslopes.
Leaf litter decomposition was faster in the youngest eucalypt stands, as reported for branch litter in India (BARGALI, 1996).The effect of stand age on litter decomposition in situ, but not on C-CO 2 emission in vitro, suggests differential responses by decomposer Forest litter decomposition as an effect... communities.The remaining litter mass after 365 days was higher for eucalypt (41%, mean of three ages) on footslopes than for the native forest (32%).These values represent a much faster decomposition than reported by Costa et al., (2005) for eucalypt plantations in Rio de Janeiro.Differences in nutritional quality between vegetation types, and also among eucalypt ages, may have influenced the activity of larger decomposers (meso and macrofauna) rather than soil microbiota.Litter decomposition is influenced by substrate quality and climate (AERTS, 1997), but also by the physical-chemical environment and decomposer organisms (GONZALES; SEASTEDT, 2001).Lack of decomposer fauna was suggested by Mesquita, Workman and Neely (1998) as responsible for slow decomposition in a Cecropia-dominated secondary rainforest in Amazon.
The main decomposing fauna may vary according to the landscape, vegetation and period.On footslopes in this study, ants were common in both native forest and eucalypts, whereas earthworms were noted only under eucalypt stands.In an Amazon upland rainforest, Cornu et al. (1997) observed the dominance of earthworms and termites, the latter responsible for 40% of the litter disappearance in the wetter period (LUIZÃO; SCHUBART, 1987).Decomposition of eucalypt leaf litter by termites was reported by Bernhard-Reversat and Schwartz (1997) in Congo.Paul and Polglase (2004) suggested that rainfall and temperature explain better decomposition in situ than litter chemistry alone.Investigating the initial three months of Eucalypt litter decomposition in Congo, Ngao et al. (2009) reported that C mineralization were first driven by soluble organic compounds and then by soluble phenolic compounds, but relations between litter quality and decomposition varied strongly in the dry and rainy seasons.Sierra and Marbán (2000) reported higher N mineralization in clayey Oxisols with increasing water content, especially for acid soils and temperatures >30 o C. Sierra (2002) reported that N mineralization was higher under fluctuating than under constant temperature, and higher temperatures had a stronger effect on re cal cit rant than on labile N. Eu cal ypt leaf decomposition in this study was faster than reported for native species in Caribbean islands (LORANGER; PONGE, 2002) and Amazon rainforest (MESQUITA; WORKMAN; NEELY, 1998;SMITH et al., 1998), comparable to legume tree plantations in Amazon (SMITH et al., 1998), and lower than in Mexican rainforest (ALVAREZ-SÁNCHES; ENRÍQUEZ, 1996).

CONCLUSIONS
Data presented here suggest that no inherent limitations to decomposition of eucalypt litter in the studied sites occur.Thus, the large litter layer stocks reported are most likely due to intense litter production in the area.The differences in decomposition patterns of eucalypt and native forest litter are most likely related to the respective soil decomposer communities and their response to rainfall.
where [HCl] = HCl concentration; f = correction factor for HCl concentration; 6 = equivalent mg of C-CO 2 Vb = volume (mL) of HCl spent with blank; Va = volume (mL) of HCl spent to titrate NaOH from samples.
32 o C compared to 23 o C. Vasconcellos (1994) and Grisi et al. (1998) incubated different Brazilian soils at 15 o C and 35 o C and found differences between treatments only at 35 o C.This suggests that microbial activity for these tropical soils is triggered above 15 o C, most likely between 15 and 23 o C.
Fernando P. Leite for providing all facilities during the sampling and part of laboratory work.The Incubation was conducted in the Departamento de Microbiologia at Universidade Federal de Viçosa (UFV), MG, Brazil.Our thanks to Drs.Maria C. M. Kasuya and Marcos R. Tótola for the laboratory support and to Dr. Haroldo N. de Paiva (Dept.Engenharia Florestal, UFV) for the suggestions.

Table 3 -
Overall means for soil organic carbon, C-CO 2 evolved from soil and leaf litter on soil, under eucalypt plantation, and native/secondary vegetation in Brazil.Tabela 3 -Médias gerais de carbono orgânico do solo, C-CO 2 emitido do solo e de folhas fragmentadas sobre o solo, sob plantação de eucalipto e vegetação nativa ou secundária, no Brasil. .