Calcium and Magnesium Dynamics in Litter in a Successional Forest Ecosystem, Under Hydroperiod

This study was part of the Manipulation of Moisture and Nutrient Availability in Young Regrowth Forests in Eastern Amazonia Project (MANFLORA). The experiment was designed in completely randomized blocks containing control and irrigated treatments during the dry period (5 mm of water/day), with four repetitions each. The monthly mean litter values ranged from 316.10 to 997.90 kg ha month. The magnitude of this phenomenon can be explained by the functional role of the floristic structure, represented by the species Myrcia sylvatica (G. mey) DC., Myrcia bracteata (Rich) DC., Miconia ciliata (Rich) DC., Lacistema pubescens Mart., Lacistema aggregatum (Berg.) Rusby, Vismia guianensis (Aubl.) Choisy, Cupania scrobiculata Rich. and Ocotea guianensis Aubl., which constituted the determinant factors, associated with the hydroperiodic effect and ecosystem manipulation. The monthly mean of the analytical results of mass treatments were significant (P < 0.05), however, when compared annually there was no significance, which indicates seasonal influence, since the period of greatest deposition is the dry one, regardless of the water manipulation along the period studied. Only in time the mass values of Ca and Mg were not significant for treatment (P < 0.05). The amount of Ca was significantly (P < 0.05) higher than that of Mg.

amount of nutrients in the soil. Litter deposition results from the interaction of these factors and, according to the peculiarities of each system, one factor may prevail over the others (Vitousek & Sanford, 1986;Odum, 1988;Golley et al., 1978;Mateus et al., 2013;Zhang et al., 2014;Silva et al., 2015).
Low water availability is considered one of the main conditions of environmental stress (Santos, 1996) and it is of extreme importance for metabolical activity and leaf survival, being the survival ability during hydric stress determinant for the distribution and productivity of plants (Pimentel et al., 1990). In regions prone to water deficit, due to low rainfall or irregular rainfall distribution, the growth of planted forests in height and diameter is limited . For this reason, water scarcity is a concern in tropical forests and may lead to an environmental collapse, in which several species may disappear if these ecosystems undergo consecutive droughts (Santos, 1996;Nepstad et al., 1998). Therefore, the necessary amount of water and nutrients are major factors, as their lack and/or excess may compromise the production and/or productivity of forest species (Sobrinho et al., 2020).
Studies on nutrient cycling are of fundamental importance for the knowledge of the structure and functioning of any ecosystem. Phosphorus (P), calcium (Ca), magnesium (Mg), potassium (K) and nitrogen (N) are essential elements which are determinants in plant growth, because they have considerable importance in plant metabolism and function (Barroco Neta & Nishiwaki, 2018). Considering the importance of nutrient cycling for the soil, this is directly related to seasonality, as the amount and decomposition of litter varies according to rainfall.
The aim of this study was to quantify the monthly accumulated litter production and the concentration of Ca and Mg in successional forest, under water manipulation, with the following question: can water availability during the dry period alter the biogeochemical cycles of Ca and Mg in a successional ecosystem? If biogeochemical cycles vary according to the hydroperiodic phenomenon, then they have seasonal dynamics.

Study Area
The study was developed in the Manipulation of Moisture and Nutrient Availability in Young Regrowth Forests in Eastern Amazonia Project (MANFLORA), which began in 1999, when forest regeneration was 12 years old. The experiment was carried out at the Fresh Water Fish Farming Station (EPAD), which belongs to the Federal Rural University of the Amazon (UFRA), in the region of the middle Apeú River, Castanhal, in the Praquiquara River basin (1º19′S, 47º57′ W), 80 km away from Belém.
The surrounding landscape is marked by secondary forests, agroecosystems and pastures. According to Falesi et al. (1980), in the Bragantina Zone there existed the humid tropical forest that, with colonization and agriculture practice, through successive cuts and burnings, was modified, which caused the formation of a mosaic of different successional stages, with predominance of several agroecosystems, mainly pasture. The settlement under study was modified due to shifting agriculture, which includes the cultivation cycle of corn, cassava and bean, for 1 to 2 years, followed by fallow, which was abandoned in 1987. The region's relief is slightly wavy to wavy, under forest vegetation, mostly constituted by a flattened surface, dissected in flat top hills, with small altimetric variation. It presents a dystrophic yellow latosol of clayey texture and concretionary laterites (Tenório et al., 1999).
The climate, according to the Köppen classification, is type Am3, with annual mean rainfall of 2000-2500 mm; 70-90% of the annual rainfall occurs between January and July, while the dry period occurs from August to December, with dry months being considered when rainfall was lower than 100 mm (Table 1).  The monthly variation of total litter production during the periods from January to December 2000 and 2006 in the control and irrigated plots is shown in Figure 3b. The monthly litter production in 2000, in the control treatment, varied from 421.20 kg ha -1 (March) to 997.90 kg ha -1 (September), with a mean value of 644.40±225.90 kg ha -1 , representing a total of 7.73 Mg ha -1 year -1 (Figure 3b). In this case, it is observed that the values are in the same period, but in this year the rainfall indices have been low since June (Figure 3a), indicating that the hydroperiodic phenomenon influences the litter dynamics in this ecosystem, evidencing the seasonality in the different periods studied (2000 and 2006).    Table 3. Variance analyses with associated significance levels for the effects of treatments (control and irrigated), sampling time and their interaction on non-woody mass and nutrients in a secondary rain forest in eastern Amazon, Brazil. The significance level is indicated (*: P < 0.05, **: P < 0.01, ns: not significant) In the control treatment, Ca production in 2000 was 5.74 kg ha -1 year -1 , ranging from 0.034 g m -2 to 0.071 g m -2 , with a mean value of 0.048±0.017 g m -2 . In 2006, the mean production was 0.047±0.017 g m -2 , ranging from 0.019 g m -2 to 0.074 g m -2 with a total amount of 5.58 kg ha -1 year -1 (Table 4, Figure 3c). In the control treatment, the annual production of Mg in the litter flow was 1.88 kg ha -1 year -1 , with a monthly mean of 0.016±0.007 g m -2 in 2000, and 1.53 kg ha -1 year -1 in 2006, with a monthly mean of 0.013±0.005 g m -2 . On the other hand, in the irrigated treatment the Mg production was 1.79 kg ha -1 in 2000, with a mean value of 0.015±0.005 and in 2006 the total production was 1.44 kg ha -1 year -1 , with a mean value of 0.012±0.005 g m -2 ( Figure 4b).

Discuss
As for the secondary horizontal (Correia & community The most a  Vol. 13, No. 11; Leaf fall is caused by senescence, resulting from a series of metabolic processes linked to the physiology of each species, as well as by environmental stimuli such as photoperiod, temperature and water stress (Kramer & Kozlowski, 1960). However, in 2006, the annual litter quantity in the control treatment reduced to 6.91 Mg ha -1 year -1 when compared to the irrigated treatment, ranging from 316.10 kg ha -1 (January) to 937.90 kg ha -1 (July), with monthly mean of 575.96±202.60 kg ha -1 . In addition, the litter flow varies according to the ecosystem considered and its successional stage (Delitti, 1989).
The annual production in successional ecosystems in plateau areas in the Amazon is in the range of 7-10 Mg ha -1 year -1 (Klinge & Rodrigues, 1968;Stark, 1971;Herrera et al., 1978;Vitousek, 1984;Vasconcelos et al., 2008), but it can vary considerably from one year to the next, depending on the phenology of tree species and, mainly, on rainfall patterns, because there is a strong seasonal control of thin litter production: larger productions are measured in the driest periods of the year (Luizão, 1989). Therefore, the litter layer on the soil also presents a strong seasonal pattern, decreasing its thickness at the end of the rainy season and increasing it in the dry period (Luizão & Schubart, 1987).
The irrigated treatment had a greater litter production than the control plot, that is, the artificial hydroperiodic pulses probably contributed to a higher phytomass deposition in this ecosystem, which may have caused an ecophysiological disturbance on the plants, in relation to water availability outside the rainy season. However, there was no significant interaction on treatment × time (Table 3, Figure 3b). For the annual mass, only for time it was highly significant (P < 0.01).
In forestry cultivation, irrigation has been used mainly in the first year of stomatal formation, presenting stomatal closure under conditions of high evaporative demand, which may have variable or gradual effects depending on the stage of development (Amaral, 2019). The variation of water in the soil directly influences transpiration and conductance.
For Delitti (1984), the increase in litter quantity with low rainfall is common in tropical regions and reflects a strategy of minimizing the effects of water scarcity. In tropical regions such as the Amazon, the influence of rainfall is of fundamental importance for the accumulation, decomposition rate and release of nutrients from the litter (Luizão, 1982). The monthly distribution of Ca and Mg masses are shown in Figures 3c and 3d, respectively, and were highly significant (P < 0.01) only in time (Table 3).
Ca dynamics were significant both in the dry season and in the rainy season. However, considering that it is proportionally associated to the pulses of monthly litter production, which are higher in the dry period, the dynamics of Ca become more pronounced in the latter. The monthly water pulses in the irrigated treatment, the Ca mass, in 2000, ranged from 0.031 g m -2 to 0.076 g m -2, with a mean value of 0.050±0.015 g m -2 , totaling 5.99 kg ha -1 year -1 . In 2006, the average monthly production was 0.048±0.018 g m -2 , corresponding to a total of 5.71 kg ha -1 year -1 .
Comparing the monthly distribution in the years studied (2000 and 2006) in the dry period (August to December), when rainfall is lower than 100 mm, it is observed that the deposited amount of Ca in both treatments was higher than the one of Mg. Calcium dynamics is high in most of the tropical forests studied (Vitousek, 1984), because it is a fixed element in plant tissues (leaf, bark, wood and branches), however, there was not a marked difference between the treatments.
The concentration and nutrient content of the litter varies according to soil type, vegetation, population density, species ability to absorb, utilize and redistribute nutrients, natural habitat and tree age. The monthly concentration of Ca ( Figure 3) and Mg ( Figure 3) varied significantly (P < 0.05) in treatment and in time, but only Mg concentration was highly significant (P < 0.01) and only by treatment (Table 3, Figures 4a and 4b).
According to Yavitt et al. (2004) irrigation did not have a significant effect on the concentration or fall of leaves or amount of nutrient return annually by leaf fall. The temporal patterns in nutrient concentrations tended to accompany those in the leaf fall. The biogeochemical cycle encompasses the processes of nutrient transferring within the soil-plant system. The nutrients cycling depends on several factors, including their mobility within the plant.
According to Mengel and Kirkby (1982), the biochemical cycle, which represents the movement of translocation of nutrients from old to new tissues of the plant, is of fundamental importance for high mobility nutrients such as jas.ccsenet.org Journal of Agricultural Science Vol. 13, No. 11; Mg. Nonetheless, it is of less significance to those of limited redistribution such as Ca. In addition, much of the nutrients is allocated in the trees. Among the components of the aerial part of the tree, the highest nutrient content is found in the leaves, then in the branches and the rest in the wood and trunk (Vieira, 1998). Therefore, Ca values are higher in relation to Mg due to the higher deposition of senescent tissues, such as leaves, and their low mobility.

Conclusion
The seasonality of total litter production was significative, being higher in dry season, regardless of irrigation performed in dry period, in the irrigated treatment. The annual litter production did not show significant differences between control and irrigated treatments. However, a longer observation period would be necessary to confirm the information more conclusively or to obtain more significant results for irrigated treatment.