Temporal analysis
Since our study area is located in the inter-tropical zone, with mild winter, the bulk precipitations not show seasonal variation for DOC concentrations caused by pollution emissions from domestic heating, as reported in the studies in higher latitudes as by Inagaki, Sakai, and Ohnuki (1995) in Japan, Pan et al. (2003) in China, and Siudek et al. (2015) in Poland. However, one of two rainfall events with higher DOC concentrations occurred in winter (EV3). The tendency of less precipitation - drier vegetation and litter - coupled with the practice of balloon release typical of the São João festivities that occur at this time in Brazil (June to August), makes winter the highest occurrence time of forest fires in the Atlantic Forest, in particular in the Tijuca Forest. It seems that DOC concentrations in BP tend to be close to 3 mg.L-1 in our study area, however, they may present higher values linked to stochastic causes.
Storms formed over the ocean usually have lower DOC concentrations than those originated over the continent (Kieber et al., 2002). However, according to reports from the Rio de Janeiro State Meteorological Monitoring System, only EV5 was not caused by cold fronts. This indicates similar origins between events in relation to rainfall formation including the DOC-peak events (EV3 and EV7), considering that these cold fronts that usually reach Rio de Janeiro by sea winds from the southwest (Dereczynski et al., 2009), this is not an explanatory factor for our DOC peaks
There was a dilution effect, but not associated with rainfall volume as reported by Li et al. (2016), being such an effect linked to the rainfall intensity. Iavorivska et al. (2017) reported the trend of decreasing DOC concentrations during rainfall events which seems to be an important factor for the occurrence of a dilution effect linked to rainfall volume. These authors also mention an interactive set of factors that can contain the downward trend or even generate increases in DOC concentrations over the course of rainfall events that may help explain our weak correlation between DOC concentrations and rainfall volumes. In particular, the low-temperature reduction from the rainfall beginning (maintenance of the dilution capacity) and the maintenance of atmospheric OM sources (urban spatial context) are factors, among those mentioned by the authors, which may be key elements in this matter.
The TF DOC showed a bigger dilution effect relative to the rainfall intensity and the maintenance of the absence of correlation with the total rainfall volumes, in comparison to occurs in BP. The rainfall intensity protagonism for the dilution effect in detriment of the total rainfall volume infers the idea of some temporal continuity about the OM sources to be considered for both wet and dry deposition.
The two events with the highest wet deposition (BP data) lost prominence in comparison to the dry deposition (CL data), demonstrating that the factors behind the wet deposition “peaks” did not conduce the same increase in dry deposition. Similarly to that observed by Kieber et al. (2002), this difference can be explained by the constant action of sea breezes pushing the air laden with organic components into inland areas, which adds to the fact that rainfall is often caused by cold fronts that reach the sea (Dereczynski, 2009).
The arise of the pre-wash effect in TF, not observed in BP, is evidenced by the statistical correlation of concentrations and some antecedent humidity parameters. That is, the pre-wash effect occurs similarly to the reported by Turgeon and Courchesne (2008), which found greater DOC flows in a dry climatic context. However, here, this effect was manifested statistically related to the previous rainfall volume and not to the number of previous drought days, indicating that small rainfall volumes may not be effective for washing. In this context, the previous time range that stood out as the most relevant for this rainfall accumulated volume was 15-days.
About litter leaching, authors such as Dawson et al. (2008) and Jiang et al. (2014) found higher DOC concentrations from the temperate forest floor in the summer, when litter decomposition rates and DOC solubility are highest. The low thermal amplitude typical of the area reduces the temperature effect (freezing and solubility) relevance on litter leaching described by Zhou et al. (2015).
The evergreen character of the forest in question results in a strong temporal correlation between litterfall and rainfall since stormy winds mechanically act for the fall of senile leaves (constantly produced). Therefore, the occurrence of rainfall tends to be accompanied by fresh leaves with high leaching potential, which helps to explain the pre-wash absence in this compartment. That is, the small variance in temperature, with the absence of litter freezing at the winter, and the poor seasonality of litterfall with constant new leaves replacement throughout the year result in no seasonal-tendency for litter DOC leaching in our study area.
The absence of a dilution effect on litter leachate indicates that the DOC available on the forest floor does not deplete easily (Schmidt, Wang, Chang, & Matzner, 2010). Michalzik, Kalbitz, Park, Solinger, and Matzner (2001) also found no dilution effect in litter leachate in a temperate forest. In this situation, higher amounts of rainfall result in more DOC infiltration into the soil or export through the drainage channel (dependent on the soil DOC retention). The continuity and efficiency of biochemical action on litter can be an important factor for the maintenance of leaching rates. No dilution or prewash effects were observed for soil solution.
The stream water DOC concentration is largely given by the balance between atmospheric deposition, biomass leaching, soil sorption, and biodegradation processes on the OM in the terrestrial environment (Nosrati, Govers, & Smolders, 2012; Jiang et al., 2014; Oni et al., 2014). In the post-event baseflow condition, the SW DOC concentrations were similar between the analyzed events, which highlights the soil capacity to dampen the oscillation in DOC flows from the other compartments.
DOC modulations correlated to variations in canal discharge, as observed by Turgeon and Courchesne (2008) and Roig-Planasdemunt et al. (2016), were not contemplated in the present study since water in stormflow condition was not collected. However, hortonians runoffs, the main sources of DOC to the canal during rainfall events according to these authors, are uncommon in our area due to the high water infiltration capacity in the soil (Coelho Netto, 1987).
A study of variations of inorganic compounds chemistry of stream water in association with stream water discharge modulation was carried out in the same (our) area by Ovalle (1985). This author reported the tendency of stream water, in the first moment of increased discharge, to be chemically close to the soil solution on the influence of more intense rainfalls (“piston” effect pushing edaphic water), and that the water started to return to the chemical condition of the baseflow even before the water discharge reduction. When the rainfall was intense enough to generate overland flows, either Hortonian or saturation flows, a large amount of OM was transported to the stream in addition to typical inorganic solubilization of elements of rocky outcrops.