Spatial analysis
Because it is a multi-diverse forest with low species dominance, the small differences in vegetation structure and composition observed between hillslope positions were not sufficient to configure significant variations for the TF DOC concentrations, without effects arising from allometric variations between tree-species as observed between conifers and hardwoods (Liu & Sheu, 2003; Inagaki et al., 1995). The spatial heterogeneity of canopy leaching is high, but it occurs on a small spatial scale (punctual) and the means obtained from 300 m2 area were very similar when comparing to the hillslope positions. The variance of the TF DOC data within each hillslope position reflects that the local minutiae in the dry deposition, canopy washing, and rainwater interception control the DOC enrichment process in this compartment.
The slightly larger canopy opening in the SMT did not materialize lower DOC values as expected. Probably, the canopy opening variation between the SMT and the other hillside positions is not sufficient to generate differences as observed by Costa et al (2016) between native forests and cacao agroforestry, since the largest native forest biomass would manifest in less canopy opening (unmeasured).
An apparent duality regarding the relationship between litter decomposition rates and DOC leaching seems to manifest itself prominently on the studied hillslope. Lower DOC concentrations from birch forest floor compared to conifers were reported by Froberg et al (2011) and argued for the higher biological activity on birch litter (lower C / N ratio), resulting in thin soil litter stocks above ground. Similarly, lower DOC concentrations have also been reported for Mull humus form compared to the adjacent area with Moder humus (the difference between humus forms is due to the greater accumulation of the sub-horizon H for Moder type) by McDowell and Wood (1984). The key issue is the maintenance or not of a MO source stock above ground and the greater leaching probability when percolating thicker litter stocks. However, it is noteworthy that the non-observation of the litterfall rates can be a large error source factor when relating the litter accumulation above ground and DOC leaching in the comparison between areas.
On the other hand, some studies showed the positive effect of decomposition for the DOC leaching in a comparison between broadleaved forest areas (Camino-Serrano et al., 2014; Guggenberger & Zech, 1993; Zhou et al., 2015), as well as by experiments with controlled leaching and/or decomposition. (Don & Kalbitz, 2005; Lee et al., 2018). According to Zhou et al. (2015), at the first moment, the DOC leaching is high due to the great solubles availability in fresh leaves, with subsequent rate decreases. Continuity of DOC leaching depends on the availability of new solubles, which is greater in situations where biochemical activity in material degradation is more intense. The key issue here is the availability of new solubles by chemical rework on the material, mobilizing C for leaching. However, laboratory-controlled experiments may not consider the effect of constant accumulation of semi-degraded material in topsoil that occurs under natural field conditions. Additionally, the positive relation between decomposition and leaching of DOC may be affected by considering losing massive C by microbiota respiration (mineralization) in environments where decomposition is more intense (e.g., tropical forests).
Camino-Serrano et al. (2014) comment about this duality from the idea of concurrent pathways to C: being breathed by the microbiota or being leached by water. Efforts to modeling the DOC leaching behavior from the coníferous forest floor also face this duality, represented by the challenge to combine the leaching of semi-decomposed material accumulated in topsoil to the large contribute by fresh leaves (Ǻgren, Kleja, & Bosatta, 2018).
In this way, by analyzing our data, it not surprising to us that the lower DOC concentrations in the VBT litter leachate, where decomposition is faster and there is little accumulation of OM material as DOC source, with a high loss of C by microbiota respiration. It is also not surprising that the largest litter accumulation in the SMT did not materialize in higher DOC concentrations, as the availability of new solubles is slower. Both mid-hillslope positions, with similar litter stock between themselves and intermediates compared to the summit and bottom valley, showed very close DOC concentrations between themselves and higher compared to geomorphological extremes.
Our inference, considering the fact that hillslope positions have similar litterfall, is that the mid-hillslope conditions, with intermediate decomposition rate, presents a balance between source material accumulation and availability of new solubles that favors DOC leaching. The leaching is lower at the valley bottom due to the little material accumulation (less C available) but is also limited at the summit by the lower biochemical work over the material (fewer solubles available).
The second-degree polynomial tendency presented in figure 9 does not represent a proposed mathematical model of this relation, being only a tool to emphasize the nonlinear character of the behavior for the studied area. Given the studied area representativeness to the Atlantic forest domains of the mountainous environment (Miranda & Avelar, 2019), we believe such nonlinear behavior may be a standard to be considered for the large extension of the forested domains of the Serra do Mar and Mantiqueira, main actual Atlantic Forest strongholds.
Higher soil solution DOC concentrations at 50 cm depth in summit compared to the VBT result from the joint action of higher inputs (UL DOC) and lower adsorption rates once the higher MO content incorporated into the soil results in higher occupation of sorption sites (kaiser et al., 1996; Singh et al., 2016). At higher depth (100 cm) the difference between hillslope positions reduces its significance due to the soil acting as a sorption environment, a similar observation as reported by Park and Matzner (2003). In both cases (50 and 100 cm) the reported differences (SMT vs VBT) are further influenced by occasional increases in-depth concentration that has been treated as an interaction of spatio-temporal factors.