Heavy storm events may
increase the amount of organic matter in runoff from forested watersheds as
well as the relation of dissolved to particulate organic matter. This study
evaluated the effects of monsoon storm events on the runoff fluxes and on the
composition of dissolved (< 0.45
As much of the dissolved organic matter (DOM) in aquatic systems originates
from soil-derived organic matter, the export of terrestrial carbon (C) and
nitrogen (N) into aquatic environments is a primary link between these
systems (Bauer and Bianchi, 2011; Bianchi, 2011; Camino-Serrano et al., 2014; Canham et al.,
2012). The export of terrestrial C and N occurs in the form of dissolved and
particulate organic carbon and nitrogen (DOC, DON, POC, PON). Particulate
organic matter can be operationally classified into fine (0.1 to 63
In regions with seasonally large differences in precipitation, most of the annual organic C export from forested watershed to steams is driven by heavy storm events with cyclones and hurricane (Dhillon and Inamdar, 2013; Lloret et al., 2013). Such conditions are pronounced in the Korean Peninsula, where the monsoon season (Jeong et al., 2012; Kim et al., 2010) represented 52 and 83 % of the annual DOC and POC runoff fluxes. During storm events, a change in hydrological flow paths in watersheds has often been observed from deeper to upper soil layers (Bass et al., 2011; Sanderman et al., 2009; Singh et al., 2014). Surface flow-inducing storm events can alter the fluxes and concentrations of DOC and POC in runoff by shifting preferential flows through macropores, surface runoff, and lateral flow (Katsuyama and Ohte, 2002; Kim et al., 2010; McGlynn and McDonnell, 2003).
In the case of organic N export, DON was the major form of N in runoff from pristine forested watersheds (Alvarez-Cobelas et al., 2008; Frank et al., 2000; Kaushal and Lewis, 2003; Pellerin et al., 2006; Yates and Johnes, 2013). Only few data are available on the partitioning of DON and PON fluxes in runoff from forested watersheds, like Inamdar et al. (2015). They reported that particulate N composed 39–87 % of the storm event N export. The question remains open of whether organic N in runoff – either dissolved or particulate – from forested watersheds behaves similar to organic C or not. Some studies reported that concentrations of DON and DOC correlated strongly (von Schiller et al., 2015), but weak relationships were also found (Singh et al., 2015).
Considering an effect of watershed characteristic, tree species might influence the export of DOM from forested watersheds. DOM from coniferous litter generally comprises more refractory (e.g., hydrophobic acid, lignin) and aromatic compounds and a relatively larger proportion of high molecular weight compounds than DOM from deciduous litter. It is also more acidic than DOM from deciduous litter (Don and Kalbitz, 2005; Hansson et al., 2011; Kiikkilä et al., 2013). Moreover, higher DOC and DON concentrations were found in oak, beech, and silver birch forest floors compared to Norway spruce, Douglas fir, and Scots pine (Smolander and Kitunen, 2011; Trum et al., 2011). Amiotte-Suchet et al. (2007) found higher annual DOC concentrations and fluxes in runoff at a deciduous forested watershed than at a watershed dominated by coniferous species.
As a result of global warming, heavy storm events have occurred more frequently and become stronger in recent decades (IPCC, 2013). Furthermore, forest management, namely the selection of tree species, might influence the export of organic matter from forested watersheds. Understanding the influence of both drivers is needed for a better prediction of the link between terrestrial and aquatic ecosystems and to support an efficient downstream water quality management. The goal of this study was thus to investigate the influence of tree species and heavy storm events on the fluxes of dissolved and particulate forms of C and N from a mixed coniferous/deciduous and a deciduous forested watershed in South Korea during the 2013 monsoon season.
The Lake Soyang basin area (Fig. 1) is located in the upstream region of
the Han River, which is the main source of drinking water for about 23
million citizens of South Korea
(Lee et al., 2013; Park et al., 2010). The average annual temperature of the Lake
Soyang watershed in western Gangwon province is 11
Location and tree species composition of the two studied forested watersheds. Lake Soyang map was modified from Jung et al. (2015).
The mixed coniferous/deciduous forested watershed (mixed watershed; Fig. 1)
is located in Seohwa, the Gangwon province (38
Tree species composition and geomorphological characteristics of the studied forested watersheds.
The slope direction of the coniferous part at the mixed watershed is towards the MD plot. Lateral flow from the coniferous part to the MD plot can only influence deeper soil solution characteristics as near-surface flow was never observed. Our data (see results) indicate significant quality differences of soil solutions between the MD and MC plots which suggest only a minor influence on soil solution chemistry at the MD plot from lateral flows. Furthermore, the quality parameters of soil solutions at the MD plot were similar to those of the DD plot, the latter being not influenced by lateral flows from coniferous sites. Thus, it is unlikely that the MC plot did affect the MD plot.
The deciduous forested watershed (deciduous watershed; Fig. 1) is located
in Haean, the Gangwon province (38
Bulk precipitation samplers (
Before storm events in June 2013, throughfall, forest floor leachate, and
soil solution were collected at about weekly intervals, and runoff samples
were collected two to three times per week. During storm events in July 2013,
throughfall, forest floor leachate, and soil solution were collected after
each storm event so that these samples represent cumulative water samples
during the entire storm event. In case of runoff, samples were taken in July
2013 at the weir using automatic collectors (6712 Portable Sampler, Teledyne
Isco Inc., Lincoln, NE, USA) before, during, and after each rain event at
intervals of 1 or 2 h. Discharge at the outlet of the watersheds was
measured by a v-notch weir. During routine runoff sampling, water
temperature, pH, and electrical conductivity were measured in situ. Water
samples were cooled at 4
Precipitation data (total and hour unit; Table 2) at the study area were used from the automatic weather station of the Korean meteorological administration at the point “Seohwa 594” and “Haean 518” for the mixed watershed and for the deciduous watershed, respectively. Those data were also comparable to ours from bulk precipitation measurements at the field sites.
The total stock of organic horizons (Oi: slightly decomposed recognizable
litter; Oe: moderately decomposed fragmented litter; Oa: highly decomposed
humic material) was collected at each plot in a 20
In June 2013, before the monsoon storm events, the fluxes of DOC were
calculated on a weekly basis by multiplying the DOC weekly mean
concentration in runoff by the weekly mean discharge. The concentrations of
DON, NO
The normality of data was tested with the Shapiro–Wilk test. When the normality was assured, the Holm–Šidák test was used for both pairwise comparisons and comparisons to a control group. When the normality test failed, the Dunn's test was used for all pairwise comparisons and comparisons against a control group with rank-based ANOVA.
After filtration through a pre-rinsed cellulose acetate membrane filter
(0.45
In this study, the POC and PON fraction is defined as the size class 0.7
The absorption spectra of DOM were obtained at wavelengths from 200 to 600 nm
using a UV–visible spectrophotometer (DR5000, HACH). Specific ultraviolet
absorbance (SUVA
For fluorescence excitation–emission matrices, fluorescence intensities were recorded with a luminescence spectrometer (LS-55, Perkin-Elmer, USA) following the method of Baker (2001), Chen et al. (2007), and Hur and Cho (2012). Excitation and emission slits were both adjusted to 10 nm. DOM samples were diluted under the ultraviolet absorbance of 0.1 at 280 nm to avoid inner-filter correction and then were adjusted to pH 3.0 for the fluorescence measurements. The fluorescence intensities of all samples were normalized to units of quinine sulfate equivalents. The humification index (HIXem) was calculated by dividing the emission intensity from 435 to 480 nm region by intensity from 300 to 345 nm (Zsolnay et al., 1999). Fluorescence characteristics of water samples were interpreted as fulvic-like fluorescence (FLF), humic-like fluorescence (HLF), and protein-like fluorescence (PLF) (Fellman et al., 2010; Singh et al., 2014).
After filtration (0.45
The morphologies of the organic layers at the MC, MD, and DD plots were
similar with distinct Oi layers
and less distinct Oe and Oa layers. However, the depth of O layer in the MC
plot (ca. 3 cm) was thinner than in the MD and DD plot (ca. 4–5 cm). The
typical soil type at both watersheds is Dystric Cambisol (FAO, 2014). Soil
texture at all plots ranged from 40–44, 30–38, and 18–22 % for
sand, silt, and clay, respectively. The C content of the organic layers at
all plots ranged from 45 to 48 % in the Oi and from 34 to 38 % in the
Oe
Soil profiles of
The average discharge in June 2013 before storm events was 0.03 mm h
Hydrological characteristics of sampled storm events and maximum concentration of dissolved organic carbon (DOC) and nitrogen (DON), particulate organic carbon (POC), and nitrogen (PON) in runoff.
The increase of the DOC concentrations in runoff with discharge was steeper
at the deciduous watershed (e.g., 1.9 to 6.9 mg C L
Concentrations of
At discharges from
The runoff DOC concentrations in response to discharge had a clockwise hysteretic loop with higher concentrations on the rising than on the falling limb (Fig. 3a). No hysteretic loops were observed for DON, POC, and PON (Fig. 3b, d, e).
The DOC
The fluxes of DOC, DON, and NO
Fluxes of carbon (dissolved organic carbon (DOC) and particulate
organic carbon (POC)) and nitrogen (dissolved organic nitrogen (DON),
particulate organic nitrogen (PON), and nitrate (NO
The integrated C and N fluxes over the study period from both watersheds
were in the order of DOC > POC and NO
Total precipitation, total runoff, and integrated fluxes of
dissolved organic carbon (DOC) and nitrogen (DON), nitrate (NO
The chemical properties of DOM changed with increased discharge at the
deciduous watershed, while no significant changes were observed at the mixed
watershed (Fig. 5). At the deciduous watershed, SUVA
At the mixed watershed, the ranges of the DOC
Specific ultraviolet absorbance (SUVA
Range of dissolved organic carbon and nitrogen ratio (DOC
The patterns of DOC
The range of the POC
Range of particulate organic carbon and nitrogen ratio (POC
We intensively sampled four heavy rainfall events during the monsoon season, the events representing a substantial proportion of the annual precipitation in the region. While the number of events was rather small, consistent patters emerged documenting the response of N and C fluxes to precipitation and discharge changes. The increase of DOC concentrations and fluxes in runoff induced by heavy storm events with increased discharge is consistent with the findings of previous studies (Dhillon and Inamdar, 2013; Jeong et al., 2012; Johnson et al., 2006; Lloret et al., 2013). In our study, the response to discharge and the integrated fluxes of DOC in runoff were much larger at the deciduous than at the mixed watershed. Similar to our results, larger annual DOC fluxes at a deciduous forested catchment than at a mixed coniferous catchment were reported by Amiotte-Suchet et al. (2007).
The different response of DOM in runoff to discharge between the two
watersheds, such as the large response of runoff DOC concentration to
discharge at the deciduous watershed (Fig. 3a) and the significant change
in runoff DOC quality parameters (Fig. 5), is likely caused by a shift of
hydrological flow paths to more surficial layers at the deciduous watershed.
Also, the comparison of DOC quality parameters in runoff with those in
forest floor leachates and soil solution at the deciduous watershed (Fig. 6)
indicated that a larger proportion of the DOC in runoff from forest floor
leachates at the deciduous. Previous studies have also reported a positive
relationships between discharge and DOM concentrations in runoff as a
consequence of changing hydrologic flow paths from deeper soil to upper soil
layers and forest floors at high discharge
(Aitkenhead-Peterson
et al., 2005; Bass et al., 2011; Sanderman et al., 2009). As several
watershed characteristics (slope and soil textures) and the precipitation
regime at both watersheds were similar, the differences between the
watersheds are likely due to the tree species effects on the infiltration of
precipitation water into the soil and on the mobilization of DOM. The tree
species effect became obvious although the proportion of coniferous tree
species was only 39 % of the watershed area. Several processes might be
involved to explain the tree species effect. (i) In the deciduous litter
layer the leaves are overlapping and partly impermeable, which may cause
more surface near flow than in coniferous litter layers with relatively
large pore spaces in between needles. (ii) The relatively higher level of
hydrophobicity of coniferous forest floors compared to deciduous forest
floors (Butzen et al., 2014) can
result in less DOC release from coniferous forest floors. (iii) The
mobilization of DOC in soils depends on throughfall chemistry
(Kalbitz et al., 2000). Throughfall at the MC plot was
more acidic (pH 4.7
At both watersheds, NO
The integrated fluxes of POC and PON during the study period were much less
than those of the dissolved elements and did not differ significantly
between the watersheds. POC and PON fluxes exceeded their dissolved
fractions only for a short time during heavy storm events with more than 100 mm
precipitation except one storm event at the deciduous watershed on
14 July 2013 (Table 2). Previous studies in the nearby region considered 100 mm
precipitation as the threshold that would induce large POC fluxes
(Jeong et al., 2012; Jung et al., 2012). Our finding indicates that POM fluxes from forested
watershed are unlikely regulated solely by precipitation amount, but slope
and river bench characteristics will interfere. The small proportion of
particulate fluxes in our study seems to be mainly caused by the relatively
moderate precipitation events during the study period. The POC
Our study emphasized the role of heavy precipitation events and vegetation
cover for the export fluxes of particulate and dissolved organic C and N
with runoff from forested watersheds. Our results suggest that changes of
the precipitation regime, with more severe monsoon storms in the future as
predicted, will increase the export of dissolved and particulate organic
matter from these watersheds. The proportion of coniferous tree species at
the mixed watershed was sufficient to induce less DOC fluxes and larger
NO
Average annual and monthly temperatures of the western Gangwon province from
1981 to 2010 were obtained from the Korean Meteorological Administration
(
Mi-Hee Lee carried out the experimental work and data evaluation and prepared the manuscript with contribution from all co-authors. Egbert Matzner and Ji-Hyung Park contributed to the design of this study, to data evaluation, interpretation of results, and writing of the manuscript. Jean-Lionel Payeur-Poirier supported the field work and provided the discharge data.
This study was accomplished within the framework of the International Research Training Group TERRECO (GRK 1565/1) and funded by the German Research Foundation (Deutsche Forschungsgemeinschaft; DFG) at the University of Bayreuth and the Korean Research Foundation (KRF) at Kangwon National University. We acknowledge the BayCEER Laboratory of Isotope Biogeochemistry for the isotope abundance analysis and the Central Analytical Department of BayCEER for mineral-N measurements at the University of Bayreuth. We are grateful to other TERRECO colleagues for the comprehensive support and to Uwe Hell for the sampler installation of soil solution. We also appreciate the international collaboration with Bomchul Kim, Youngsoon Choi, and Jaesung Eum from Kangwon National University (Chuncheon) and with Jin Hur and Bomi Lee from Sejong University (Seoul). The publication was supported by the University of Bayreuth in the funding program Open Access Publishing. Edited by: M. Sarin Reviewed by: D. Kumar and one anonymous referee