Litterfall at different successional stages in a tropical rain forest in southern Brazil

Produção de serapilheira em diferentes estádios sucessionais da Floresta Ombrófila Densa no sul do Brasil

The litterfall constitutes the central reservoir of organic and mineral elements in tropical forest ecosystems where soils are chemically poor and, their deposition enables elements released from plant biomass to be re-incorporated into the system. The importance of evaluating the litterfall lies in the understanding of the reservoirs and nutrient flow in the ecosystems, which are the main route of nutrient supply through the mineralization of the vegetal remains (NUNES; PINTO, 2007).

Several factors may affect the litterfall, such as vegetation type, latitude, altitude, temperature, precipitation, light availability during the growing season, photoperiod, evapotranspiration, relief, deciduousness, successional stage, water availability, soil nutrient stock and herbivory. In this way, the litterfall is considered as the result of the interaction of these factors and, according to the peculiarities of each system, one factor may prevail over the others (VITOUSEK JUNIOR; SANFORD, 1986; PORTES et al., 1996).

In soils in environmental recovery, litterfall is the largest source of organic matter, and its quantity and nature play an important role in the maintenance and fertility of the soil and, consequently, in the supply of organic matter to the flora and fauna (SOUZA; DAVIDE, 2001; MACHADO et al.; 2008).

Litterfall is one of the important indicators of evaluation and monitoring of post-revegetation phases, aiming at the restoration of degraded areas, as it allows assessing the control of the superficial erosion, as well as the whole process of forest dynamics (RODRIGUES, 1998). In restored areas with a plant community already formed (four years or more post-planting), the effectiveness of the restoration can be evaluated in relation to its physiognomic aspects, such as plant stratification and nutrient cycling (BRANCALION et al., 2012).

The litterfall is considered an excellent indicator of productivity and nutrient cycling in native forest fragments or in recovered areas, making it possible to evaluate the state of environmental conservation and/or the rehabilitation of structural and functional characteristics related to soil stability in reforested areas. Thus, besides representing a fundamental link in the energy flow and matter cycle, the litterfall produced by the vegetation gives the ecosystem greater stability, suffering a variation according to its successional stage (NUNES; PINTO, 2007).

Numerous studies address the litterfall in fragments or forest remains in Brazil (DICKOW et al., 2012; VENDRAMI et al., 2012, VIDAL et al., 2007). In Southern Brazil, in the states of Rio Grande do Sul and Santa Catarina, such studies are more scarce (CITADINI-ZANETTE, 1995; MELLO, 1995; SANTOS, 1997; MARTINELLO et al., 1999; VIBRANS; SEVEGNANI, 2000), and studies involving the litterfall in degraded and rehabilitated areas after open-pit coal mining with built soils cannot be found. However, one of the causes that lead to the low success of recovery projects in degraded areas is the lack of knowledge about the factors that support a high production of biomass and, consequently, soil fertility.

The specific objectives of this study are: (1) to estimate and compare the litterfall of three areas in different successional stages in a tropical rain forest; (2) to verify the seasonal variation of litterfall in these areas and; (3) to investigate the existence of correlations between litterfall and climate variables in the municipality of Siderópolis, southern Santa Catarina state

The annual litterfall was estimated at 8.13 Mg ha^-1 for the AR area, followed by A2 (4.26 Mg ha^-1) and A1 (2.74 Mg ha^-1), being significantly higher in AR when compared to the other areas (F[_2.33] = 15.83 p <0.05), as shown in Table 2. The peak litterfall in A1 occurred in September (0.42 Mg ha-1) and October (0.27 Mg ha^-1), in A2 in September (0.47 Mg ha-¹) and in November (0.45 Mg ha^-1), while in AR in the months of October (1.11 Mg ha^-1), August (1.05 Mg ha^-1) and January (1.02 Mg ha^-1). The lowest litterfall in A1 was recorded in June (0.12 Mg ha^-1) and July (0.14 Mg ha^-1), in A2 in July (0.17 Mg ha-1) and June (0.19 Mg ha^-1), whereas in AR in the months of March (0.24 Mg ha^-1) and July (0.35 Mg ha-1), as shown in Figure 1 a. The Pearson’s correlation analysis did not show significant correlations between litterfall and climatic variables in the three areas studied, except in AR where litterfall was significant correlated to wind speed (Table 3).

The leaf fraction was dominant in the three areas (Figure 1b), and in A1 the contribution was 60%, in A2 58% and in AR 65%. The second most representative fraction in the three study areas was branches that contributed 30% to A1, followed by A2 (29%) and AR (16%).

The reproductive material fraction contributed to A1 with 1%, A2 with 4% and AR with 7%. According to the variance analysis, there was a significant difference between AR and other areas (F[_2.33] = 11.91 p < 0.05). With a lower participation in total litterfall and a very irregular deposition pattern during the year, the reproductive material fraction presented peaks in A1 and A2 in April and in AR in January (Figure 1c).

The residue fraction did not present a well-defined pattern of seasonality and was represented in area A1 by 9% of the total litterfall annually, followed by A2 (10%) and AR (11%). The waste deposition peaks at A1 and A2 were reached in May and in AR in December (Figure 1d).

According to the ANOVA analysis of variance, with the exception of the branches fraction, significant differences of the other fractions were found in A1 and A2 in relation to AR (p = 0.000), as shown in Table 2.

Table 2
Mean annual litterfall in Mg ha-1 (± SD) and its fract ions in the three sampled areas in the municipality of Siderópolis, SC

* Values followed by the same letter do not differ statistically from each other (Tukey: p > 0.05)

Table 3
Pearson’s correlation analysis between monthly litterfall and climatic variables of the three studied areas

* Significant correlation. Ppt (precipitation), Tmax (average maximum temperatures), Tmed (average monthly temperature), Tmin (average minimum temperatures), UR (relative humidity) and Vent (wind speed)

Figure 1
Monthly total litterfall (a), leaves fraction (b), reproductive material fractions (c) and residues (d) in the three study areas, in the municipality of Siderópolis, SC

Studies assembled by Haag (1985) show that the litterfall in tropical forests ranges from 7.0 to 12.0 Mg ha^-1 yr^-1, and the leaf fraction represents the largest proportion. The references found in the specific literature with respect to annual amount of litterfall in Ombrophilous Dense Forest vary according to forest formation, ranging from 5.4 to 12.2 Mg ha^-1 yr^-1 in Submontane Ombrophilous Dense Forest (ARAÚJO et al., 2006; DICKOW et al., 2012; BIANCHIN et al., 2016; SLOBODA et al., 2017), from 4.9 to 5.6 Mg ha^-1 yr^-1 in Montane (GOMES et al., 2010; FREIRE et al., 2014), and from 6.1 to 7.1 Mg ha^-1 yr^-1 in Mixed Ombrophilous Forest (FIGUEIREDO FILHO et al., 2005; ANTONELI; THOMAZ, 2012; SANQUETTA et al., 2016). In studies carried out in old-grown forests in the north coast of Parana state, litterfall was estimated at 8.4 Mg ha-1 yr-1 (SLOBODA et al., 2017), and in advanced secondary forests in the southern region of Santa Catarina at 7.5 to 10.55 Mg ha^-1 yr^-1 (CITADINI-ZANETTE, 1995; SANTOS, 1997; ROSSO, 2011), being close to that sampled in AR in the present study, with 8.13 Mg ha^-1 yr^-1.

The results of the study of litterfall in the south of Santa Catarina showed that there is a marked deposition in the months of October and November for the primary forest fragment (CITADINI-ZANETTE, 1995; SANTOS, 1997). In these studies, the higher litterfall in the spring shows that many species undergo leaf exchange in this period. On the other hand, in a fragment in an advanced stage of natural regeneration (ROSSO, 2011) and in this study, although the highest deposition also occurs in the month of October, peaks of deposition were observed throughout the year (April and August), which is a general trend in rainforests (MEGURO et al. 1978) and may be related to the absence of extreme climatic conditions such as prolonged cold periods and drought, combined with the great floristic diversity of these forests (SELLE, 2007).

According to Nascimento (2005), areas of the Atlantic Forest in initial regeneration stage show lower litterfall than those in more advanced stages, corroborating this study, where the initial stages of succession (A1 and A2) show values below that found in the forest in an advanced stage of natural regeneration. The fact that the vegetation of A1 and A2 has a higher density of thin trees, a more open canopy and lower richness and diversity of plant species may have contributed to a lower biomass production. The litterfall in A1 (2.74 Mg ha^-1 yr^-1) and A2 (4.26 Mg ha^-1 yr^-1) were similar to those observed by Barbosa and Faria (2006) in a revegetated area of three years and in a forest area of 20 years, where the litterfall was 3.0 and 5.5 Mg ha^-1 yr^-1, respectively.

The highest levels of litterfall in the most advanced stages of tropical rainforest were shown in studies in Amazonia (MARTIUS et al., 2004), in Mexico (SANCHES et al., 1995) and in Jamaica (McDONALD; HEALY, 2000). Boeger et al. (2000) also found significantly higher results in advanced stages of a tropical rainforest in Itapoá, SC. However, in São Paulo, Leitão-Filho (1993) found greater litterfall in initial successional stretches in the rainforest in Cubatão. The authors relate this tendency of decrease in litterfall as the community approaches its climax to the great abundance of pioneer species in the initial stages of succession, since pioneer species grow fast and need a high photosynthetic rate, which is achieved through the high biomass of leaves that are rapidly replaced.

The leaf fraction generated percentages close to 70%, the suggested value for forest ecosystems (MEENTMEYER et al., 1982) and those observed in Brazilian forests (FERNANDES et al., 2006). Regardless of the biome studied, the leaf fraction is usually the largest proportion of litterfall fractions produced (WERNECK et al., 2001; TIENNE et al., 2002), which shows the importance of this fraction to nutrient cycling in these ecosystems.

The results of this study do not corroborate other studies, which observed that the production of the leaf fraction decreased with the advance of the successional phase in tropical forests (PEZZATO; WISNIEWSKI, 2006; ROCHA, 2006; KÖHLER et al., 2008). Investigating the litterfall in different successional stages of a Subtropical Forest in Paraná, Dickow et al. (2012) found a higher leaf deposition in initial stages of succession in relation to the medium and advanced stages of natural regeneration. The higher deposition of leaves found in early stages of succession was justified by the authors as a consequence of the accelerated growth in the vegetation, mainly due to the high abundance of pioneer species, which need a high photosynthetic rate achieved through the high biomass in leaves that are rapidly replaced. However, in this study, the occurrence of strong winds from August, 2012 to January, 2013 may have been a mechanical factor, causing a higher deposition of leaves in AR (advanced regeneration area).

Regarding the branches fraction, the results found in this study were expected, since the vegetation of areas A1 and A2 was predominantly composed of herbaceous and shrub species. According to König et al. (2002), the deposition pattern of the branches is marked by temporal heterogeneity with low correlation with climatic events. Green (1998) and Moraes (2002) showed that the first strong rains after the dry season cause the fall of many dry branches, which still remained united to the plant. Therefore, the intense rains and the strong winds, which occurred from August, 2012 to January, 2013, were probably responsible for the greater production of the branch fraction. This effect was most pronounced in areas A1 and A2, due to their greater exposure to external factors and due to the condition of greater isolation in the landscape without a closed canopy.

According to Diniz and Pagano (1997), the temporal variation of the production of reproductive material may be related to the variety of species in the study areas and to the phenology of these species rather than to the successional stage. Thus, the production values of this material in the same forest type may vary according to the area studied, as they depend on local abiotic conditions, floristic composition and demography of each species.

Tropical forests have a very complex structure and floristic composition, which directly affects the litterfall, decomposition and release of nutrients into the environment (ALVAREZ-SÁNCHEZ; GUEVARA, 1999; VASCONCELOS; LUIZÃO, 2004). The litterfall of an area depends primarily on the productivity of the plant community, and the main abiotic factor determining this production is climate, with precipitation and temperature as the main components (FACCELLI; PICKETT, 1991). The temporal variation in litterfall is correlated with the amount of rainfall and has been reported for almost all tropical forests in the world.

Although the pioneer species in A1 and A2 areas grow fast and have a short life cycle, investing heavily in biomass production in a short period, the highest annual litterfall was found in the forest fragment (AR) that significantly produced more leaves, residues and reproductive material. This can be explained by its community structure, with large trees, more closed canopy and greater number of epiphytes. On the other hand, areas A1 and A2 produced significant amounts of thin stems. These areas have a higher proportion of species in early successional stages, a higher density of thin trees and a more open canopy, being subject to winds, which could result in greater production and fall of branches.

The trend of increased litterfall in the spring season indicates the renewal of leaves and a possible response to water stress regarding the end of winter with low temperatures and low rainfall, which could explain the lack of direct correlation between litterfall and climatic variables in the three areas studied. Considering that the peak litterfall was in September and the driest months were from April to June, there was a delay of two to three months in the vegetation response to water stress in the study area.

https://periodicos.ufsm.br/cienciaenatura/article/view/26019/pdf (pdf)