Throughfall and stemflow chemical dynamics of Satoyama,a traditional secondary forest system under threat in Japan

Satoshi Asaoka·Fuyuhiko Sumikawa·Yoshifumi Watanabe·Waqar Azeem Jadoon·Masaki Ohno·Nobumichi Shutoh·Yuki Wakamatsu·Lawrence M.Liao·Akane Kanazawa·Yuka Sato·Natsumi Fujiwara

Abstract The term ’Satoyama’ refers to traditional and unique secondary forests in Japan that occupy intermediate zones between villages (’sato’) and hills or mountains(’yama’).Satoyama landscapes help sustain ecosystem services and the diversity of secondary natural environments.As Japan relies more heavily on foreign timber imports,the traditional role of Satoyama in providing forest products has diminished,and this has led to their abandonment and poor management.The chemical behavior of cations,anions,and dissolved organic matter in throughfall and stemflow from one such threatened Satoyama system in central Japan was investigated.From autumn to winter,the atmospheric deposition of sulfates and nitrates was 2.5–6.0 times higher compared to the amounts in summer due to the intrusion of air masses from the Asian continent.The dissolved organic matter in the throughfall and stemfolw was composed mainly of humic substances and protein derivatives.The deposition fluxes of dissolved organic carbon from throughfall(7.31–10.1 g m?2 a?1) and stemflow (1.79–3.84 g m?2 a?1)in this study were within ranges seen in temperate forests in previous studies.The deposition flux of sulfates was low compared to that in other forest types because canopy interaction was lower,suggesting higher canopy openness than in primary forests.If a shift from a mixed species Satoyama forest to a conifer-dominated forest occurs after the mass mortality of oak,the deposition flux of dissolved organic carbon and K+ might decrease by 33% and 62%,respectively,while NO3? might increase by 20%.In the near future,the degradation of Satoyama landscapes might change the levels of dissolved organic carbon and nitrogen loads,resulting in imbalances in river-ocean linkages affecting forested catchments and aquatic ecosystems in Japan.

Keywords Atmospheric deposition·Dissolved organic matter·Humic substance·Mass mortality·Oak trees ·Sustainable development goals

Introduction

Japan has one of the highest forest to open land ratios in the world with forests comprising 68.2% of the land area(FAO 2005).’Satoyama’ refers to forests at the border of or in intermediate zones between one or more villages (’sato’)and a hill or mountain (’yama’),representing a common,traditional and rural landscape.A Satoyama landscape is composed of several habitat types such as paddy fields,secondary forests and grasslands,and ponds and streams maintained by human activities.Satoyama upkeep is closely linked with supporting livelihoods and agricultural production consistent with maintaining sustainable ecosystem services and the natural diversity of secondary environments(Takeuchi 2010;Kamiyama et al.2016).In essence,high biodiversity is sustained by anthropogenic activities related to agriculture (Katoh et al.2009).Cultivated trees in the Satoyama are cut every 15–30 years to obtain timber for building materials,bed logs for cultivating mushrooms,charcoal production,and other wood-derived products(Katoh et al.2009;Gan and Tsing 2018).After clearcutting,trees regenerate from the stumps so that the stand is maintained (Washitani 2001).However,Japan currently depends on imports for most of its timber needs,and thus,Satoyama forests have been largely abandoned without management.In addition,increasing urbanization has seen a declining need for fuelwood while a rapidly aging society has contributed to the decline in interest in the maintenance of Satoyama forests (Jiao et al.2019).Consequently,their deterioration has triggered a domino effect of biodiversity decline due to the simplification of forest structure,mortality of oak and pine trees,rapid expansion of bamboo forests and increasing landslides (Katoh et al.2009;Indrawan et al.2014;Ito 2016;Koganezawa 2016;Kamada et al.2017;Nakajima 2019).Therefore,the traditional Japanese Satoyama system is facing a serious crisis.Globally,the deterioration of many forests,including the mortality of oak trees,has become a serious environmental issue in the USA,Italy and Switzerland (Bendixsen et al.2015;Colangelo et al.2018;Dey and Schweitzer et al.2018;Etzold et al.2019).As forested catchments are primary sources of nutrients and dissolved organic matter for aquatic systems,the intimate link between upland and coastal areas is further emphasized.The riverocean linkages involving nutrients and dissolved organic matter are particularly critical,particularly those involving massive terrigenous inputs from large river systems such as the Amazon and Changjiang rivers (Gomes et al.2018)and the Congo River (Holtvoeth et al.2003),often affecting cross-border bodies of water.Hence,river-ocean linkages have shown that coastal ecosystems are significantly affected by terrigenous or marine deposited nutrients and other materials as well as by carbon cycling (Cao et al.2018;Muller 2018;H?der and Barnes 2019).Therefore,the deterioration of secondary forests such as those in the Satoyama will directly affect the fate of nutrients and dissolved organic matter and further change the dynamics of river-ocean linkages.In Japan,the impact will be strongly felt as Satoyama landscapes comprise approximately 40% of the country’s land area.However,nutrient and dissolved organic matter dynamics in Satoyama ecosystems are not fully understood.Therefore,the purposes of this study were:(1) to characterize the chemical composition of anions,cations,and dissolved organic carbon of throughfall and stemflow in a typical Satoyama forest facing deterioration;and (2) to clarify the effect of any large-scale changes of tree species composition in Satoyama forests on the changes in deposition of nutrients and dissolved organic matter and subsequent changes in river-ocean linkages.

Materials and methods

Study site

Data was collected on total rainfall,throughfall,and stemflow at a Satoyama site,a mixed forest of 70,000 m2located on the campus of the Naragakuen Junior and Senior High School,Nara Prefecture (34.63400° N,135.73515° W;Fig.1).The soil is brown forest soil.The forest was mainly Japanese red pine (Pinus densifloraSiebold &Zucc.) and Konara oak (Quercus serrata) used for friewood and bed logs for cultivating mushrooms before the second deforestation in the 1960s.After the 1960s,the species consisted of Japanese red pine,Konara oak and Japanese cypress (Chamaecyparis obtusa).Currently,the deteriorated forest is maintained by the local high school for education and promotion of the Satoyama concept.However,mortality of Japanese oak trees has increased due to infection by the fungusRaffaelea quercivora(Kobayashi and Ueda 2005).At present,the area ratio of broad-leaved forest to coniferous forest is approximately ten to one.The mixed temperate forest has a mean annual precipitation of 1385 mm and an average yearly temperature of 14.9 °C (Nara Local Meteorological Office 2021),with elevations 120–210 m a.s.l.To avoid possible errors arising from edge effects (Wuyts et al.2008),all samples were taken 60–120 m from the forest edge.

Fig.1 Study sites of Satoyama forest.Site R:rainfall sampling site;Site B:throughfall sampling site in the broad-leaved forest;C:throughfall sampling site in the coniferous forest;SC:throughfall sub sampling site in the coniferous forest

Throughfall samples were collected at two sites in the mixed forest:a coniferous species-dominant forest (coniferous forest:site C) of mostly Japanese red pine and Japanese cypress,and a broad-leaved species-dominant forest (broadleaved forest:site B) of mainly Konara oak and sawtooth oak (Quercus acutissimaCarruth.).To check the spatial heterogeneity of the sampling sites,sub samples were collected 250 m from site C,which is composed of a coniferous species-dominant forest (Site SC) from September to November in 2016.The sampling sites for collection ofthroughfall were located at the center of both forest types.Average leaf area index (LAI),canopy openness and canopy closure values were 1.8 ± 0.2 m2m?2,(27.6±4.8)% and(78.9±6.7)% for the coniferous species-dominant forest and 2.8±0.7 m2m?2,(13.4±3.8)% and (84.5±2.2)% for the broad-leaved species-dominant forest.Rainfall samples without any canopy obstructions (site R) were taken 100 m from the mixed forest.

For the collection of stemflow,four major tree species with diameter at breast height (DBH) 22.5–29.1 cm were selected to represent trees growing within the Satoyama:Japanese red pine (DBH:27 cm),Japanese cypress (DBH:20 cm),Konara oak (DBH:22 cm),and Japanese hill cherry(Prunus serrulata,DBH:33 cm).

Sampling procedures and analyses

Rain samples were collected to measure pH,electric conductivity (EC),and dissolved ion concentrations in total rainfall,throughfall,and stemflow in 19 of 43 total rainfall events that occurred from June 13,2016 to January 23,2017.However,some samples were lost due to a typhoon and some quantities were insufficient for chemical analysis.Samples for DOC (dissolved organic carbon) and EEM (Excitation-Emission Matrix) determination were collected four times during the sampling period.

Throughfall was collected using a polyethylene bottle(120 mm tall,80 cm2) placed under both coniferous and broad-leaved species.Rainfall at site R was collected at ground level using the same bottle type.Stemflow was collected with a sheet of pre-cleaned medical gauze (FC gauze,Hakujuji Co.,Tokyo) wrapped around the trunks at 1.0 m height and connected to a 5-L polyethylene bottle.The gauze had been pre-cleaned by boiling in pure water for 20 min and drying prior to use.Both medical gauze and sampling bottles were installed before rain events.At the same time,sampling bottles to collect throughfall and total rainfall were placed inside and outside sites B and C.After a rain event had ceased,rain samples were transferred to the laboratory.Hence,throughfall,total rainfall and stemflow samples were collected at the same time from one rainfall event.The samples were filtered through a PTFE 0.45-μm-pore size membrane filter (Omnipore,Merck,Darmstadt,Germany) and the filtrates stored in the dark at 4 °C before analyses.

The concentrations of anions(Cl?,NO3?,and SO42?)and cations (Na+,K+,Mg2+,and Ca2+)were measured by an ion chromatograph (LaChrom Elite,Hitachi High-Tech,Tokyo,Japan).The limits of detection calculated by 3σ were Na+(0.03 mg L?1),K+(0.05 mg L?1),Mg2+(0.03 mg L?1),Ca2+(0.06 mg L?1),Cl?(0.02 mg L?1),NO3?(0.01 mg L?1),and SO42?(0.01 mg L?1).The EC and pH of non-filtered samples were determined using portable water quality meters (EC11,Kenis and LAQUA D-50,Horiba,Kyoto,Japan).The concentrations of DOC were measured with a total organic carbon analyzer (TOCV,Shimadzu,Kyoto,Japan).EEM spectra were measured with a fluorescence spectrophotometer (FR-6200,Jasco,Tokyo,Japan) and obtained by collecting emission scans (λEm 300–550 nm,1-nm intervals) at 5-nm excitation wavelength intervals between λEx 240 and 450 nm.UV–Vis absorbance spectra were collected prior to the measurement of EEM fluorescence spectra using a double-beam spectrophotometer (UV-2600,Shimadzu,Kyoto,Japan) at 240–550 nm.Primary and secondary inner filter effects,Raman scattering,and Rayleigh-Tyndall scattering of the EEM fluorescence spectra were corrected (Larsson et al.2007).The EEM was characterized by parallel factor analysis,PARAFAC (Stedmon et al.2007;Stedmon and Bro 2008) to identify fluorescent components.The DOM Fluor v.1.7 Toolbox (University of Copenhagen) was used for MATLAB (R 2013a,MathWorks,Natick,MA,USA)to fit the PARAFAC model over a dataset comprising all rain water samples.The components were verified by a random initialization and split-half analysis (Stedmon and Bro 2008).

Deposition flux

The deposition flux of the Satoyama forest was estimated approximately.The yearly total rainfall (G:mm a?1) was frist calculated using a 5-year average (2015–2019) of precipitation provided by the government offices of Nara Prefecture.The rain partitioning of throughfall,rainfall interception,and stemfolw were calculated according to Eqs.1,2,3 and 4.The throughfall ratio was estimated by obtaining the correlation between the canopy openness and the ratio of throughfall to total rainfall (Noguchi et al.2007;Eqs.1 and 2).

where,Rb,Rc,CbandCcrepresent the ratio of throughfall in the broad-leaved forest,the ratio of throughfall in the coniferous forest,canopy openness of the broad-leaved forest (%) and canopy openness of the coniferous forest (%),respectively.

The throughfallin thebroad-leavedforest(Tb:mma?1)and in the coniferous forest(Tc: mm a?1)werecalculated byEq.3.

where,GandRborcshow the yearly total rainfall(G:mm a?1)of the Satoyama and the ratio of throughfall in the broad-leaved forest or in the coniferous forest,respectively.

The rainfall interception ratio was calculated as the average ratio of the interception loss to the total rainfall (Toba and Ohta 2005).The interception ratios to the total rainfall were 0.147 for site C and 0.210 for site B.Rainfall interception in site C(Ic: mm a?1) and in site B (Ib: mm a?1) were calculated by multiplying the rainfall interception with the yearly total (G:mm a?1).

Stemflow quantities were calculated by subtracting both throughfall and rainfall interception from the total (Eq.4).

where,S borc,G,T borcandIborcindicate stemflow in site B(mm a?1) or siteC(mma?1),G istotalrainfall(mma?1),throughfall in site B (mm a?1) or site C (mm a?1) and rainfall interception in site B (mm a?1) or site C (mm a?1),respectively.

The average concentrations of the cations,anions,and DOC obtained were used for the calculation of each flux(g m?2a?1).

Dissolution test of anions and cations from leaves and bark samples

The amounts of anions and cations from leaves and bark samples were determined by adding 50 g L?1of leaves or bark to pure water,and letting stand for one day at ambient temperature (approx.25 °C).The leachate was then filtered through a PTFE 0.45-μm-pore size membrane filter (Omnipore,Merck).The filtrates were stored in the dark at 4 °C before analyses.The concentrations of anions and cations were measured by ion chromatography (LaChrom Elite Hitachi High-Tech).

Data processing

A multiple comparison test was carried out among rain fractions with normal distribution.If the fraction showed a non-normal distribution,Kruskal-Wallis test was carried out to detect statistical differences among fractions.Probability (P)-values <0.01 or <0.05 were considered significant.Hemispherical canopy photographs of the forest were used to calculate LAI and canopy openness,and imaging software,i.e.,Gap Light Analyzer ver.2.0 (Frazer et al.1999) was applied to extract canopy structure and gap light transmission.

The back trajectory calculation of each air mass was performed using the Centre for Global Environmental Research(CGER) Meteorological Data Explorer (METEX) provided by Japan’s National Institute for Environmental Studies(CGER-METEX 2016).

Results and discussion

pH and concentrations of cations and anions

Average pH values of rainfall,throughfall and the stemflow were 5.90,5.55–5.56 and 3.74–5.27,respectively(Table 1).Concentrations of cations in rainfall and throughfall in site C and site B are shown in Fig.2 a.Average concentrations of Na+,K+,Mg2+and Ca2+in the rainfall and the throughfall were 1.36?1.58 mg L?1,0.70–7.13 mg L?1,0.17–0.33 mg L?1and 0.92–1.20 mg L?1,respectively.Average concentrationsof Cl?,NO3?and SO42?were1.40–2.42mg L?1,1.05–2.82mgL?1,and1.42–2.61mg L?1,respectively(Fig.2 b).To determine the spatial heterogeneity of the sampling sites in this study,anion and cation concentrations of sub samples collected from site SC were compared to those from site C.The average concentrations of Cl?,NO3?,SO42?,Na+and K+were not statistically different between site C and site SC.The average concentrations of Mg2+and Ca2+in site SC were 0.59 mg L?1and 1.86 mg L?1,which were higher than those of site C.However,values obtained from site SC were almost within the concentration range of Mg2+and Ca2+in site C.

Fig.2 Cation a and anion b concentrations of rainfall and throughfall.B:throughfall in the broad-leaved forest,C:throughfall in the coniferous forest,R:rainfall,?:outliers

Table 1 pH and electro conductivity of each fraction

The average concentrations of cations in the stemflow are illustrated in Fig.3 a.The ranges of the concentrations from the four trees were 1.02–2.38 mg L?1for Na+,1.84–10.9 mg L?1for K+,0.21–0.60 mg L?1for Mg2+and 0.70–2.60 mg L?1for Ca2+.The average concentrations of anions in the stemflow were 1.72–5.70 mg L?1for Cl?,0.19–2.31 mg L?1for NO3?and 0.68–2.87 mg L?1for SO42?,respectively (Fig.3 b).

Fig.3 Cation a and anion b concentrations of stemflow.Jc:stemflow from Japanese cypress,Jh:stemflow from Japanese hill cherry,Ko:stemflow from Konara oak,Rp:stemflow from Japanese red pine,?:outliers

The sources of NO3?and SO42?were estimated by calcu-lating the non-sea salt sulfate (nss-SO42?)in the rainfall.The range of the concentration was 0.46–3.36 mg L?1,accounting for 91.2–98.6% of the total SO42?(0.47–3.69 mg L?1).Thus,the forest was affected mainly by atmospheric deposition due to anthropogenic activities.The effect of sea spray on the forest is negligible because the study site is located approximately 30 km from the sea and is surrounded by mountains.

A 5-day backward trajectory analysis was performed to identify the source(s) of NO3?and SO42?.In summer,air masses flowing to the study site originate from the Pacific Ocean,the Sea of Japan,and from regional emissions (Fig.4 a),whereas during the autumn to winter months,air masses come mainly from the Asian continent(Fig.4 b).The different sources of air masses were reflected in the chemical compositions of the rainfall,throughfall and stemflow.The average concentrations of Na+,Cl?,NO3?,and SO42?in the rainfall and throughfall collected in the autumn and winter were 2.5–6.0 times higher than those collected during the summer (P<0.01 orP<0.05,Table 2).Notably,the NO3?and SO42?concentrations of the throughfall in the broad-leaved forest collected in the autumn to winter months was 5–6 times higher than those in the summer (P<0.01).From autumn to winter,the average concentration of SO42?tended to increase in the fol-lowing order:rainfall(1.84 mg L?1)

Table 2 Cation and anion concentrations (mg L?1) of rainfall and throughfall

Fig.4 The 5-day backward trajectory of the air mass at the study site calculated by CGER-METEX

The dry deposition of various elements increased with increasing leaf area index (LAI) (Devlaeminck et al.2005).In this study,the LAIs of the broad-leaved forest and the coniferous forest were 2.8 m2m?2and 1.8 m2m?2,respectively;thus,the broad-leaved forest was affected by significant dry deposition compared to the coniferous forest.This is consistent with increases in the monthly concentrations of SO42?in the air of Nara city (near the study site) during autumn and winter.In the spring to summer months,the concentrations of SO42?are approximately 45–70 nmol m?3(Matsumoto and Murano 1998).During autumn andthroughout the winter,the concentrations increase to around 60–140 nmol m?3(Matsumoto and Murano 1998).It is suspected that the origin of air pollutants are mainly domestic emissions during the spring and summer.

In contrast,from autumn to the winter months,the origin of air pollutants can be traced to both domestic and continental Asian sources (e.g.,coal combustion) (Wang et al.2018).This enrichment from atmospheric depositions was also found in the stemflow data obtained in the present study,especially those for the Japanese hill cherry and Japanese red pine (Table 3).However,the enrichment of the atmospheric deposition in the stemflow compared to the throughfall was not clear.This is because tree stems might be less affected by atmospheric deposition compared to the canopies.

Table 3 Cation and anion concentrations (mg L?1) of stemflow

Dissolved organic matter

The average DOC concentrations were 1.9 mg L?1for rainfall,10.6 mg L?1for throughfall in site B,and 6.6 mg L?1for site C (Fig.5).Regarding stemflow,the average concentrations of the four trees were 11.2–26.8 mg L?1(Fig.5).The increasing DOC concentration in stemflow and throughfall might have resulted from the accumulation of organic matter from bark,litter and fauna supplemented by the presence of epiphytes on the trees (Stubbins et al.2017).The average DOC concentrations in the throughfall(6.6–10.6 mg L?1) and stemflow (11.2–26.8 mg L?1) were both within the ranges reported for throughfall (3–19 mg L?1,Le Mellec et al.2010;Van Stan and Stubbins 2018;Chen et al.2019) and stemflows (6–332 mg L?1,Stubbins et al.2017;Van Stan and Stubbins 2018).

Fig.5 Dissolved organic carbon (DOC) concentrations of rainfall,throughfall and stemflow.B:throughfall in the broad-leaved forest,C:throughfall in the coniferous forest,Jc:stemflow from Japanese cypress,Jh:stemflow from Japanese hill cherry,Ko:stemflow from Konara oak,R:rainfall,Rp:stemflow from Japanese red pine

In this study,the average SUVA254(Specific UV Absorb-ance at 254 nm) ranges of throughfall and stemflow were 1.5–2.7 and 1.7–3.1 L mg?1m?1,respectively (Fig.6).Although there were no significant differences among all fractions,the values tended to be high compared to that of the total rainfall (0.3 L mg?1m?1),indicating tree-DOM(tree-dissolved organic matter) was rich in aromatic organic matter compared to rainwater.The SUVA254values were close to those reported for throughfall(2.0–3.3 L mg?1m?1)and stemflow (1.6–5.1 L mg?1m?1,Stubbins et al.2017;Van Stan et al.2017;Wagner et al.2019).However,the S UVA254of the throughfall in site B varied widely from 0.2 to7.9L mg?1m?1,indicating that the aromatic carbon content in the dissolved organic carbon had fractured.

Fig.6 SUVA254 of rainfall,throughfall and stemflow.B:throughfall in the broad-leaved forest,C:throughfall in the coniferous forest,Jc:stemflow from Japanese cypress,Jh:stemflow from Japanese hill cherry tree,Ko:stemflow from Konara oak,R:rainfall,Rp:stemflow from Japanese red pine,SUVA254: Specific UV Absorbance at 254 nm

PARAFAC analyses were carried out to identify fluorescent components obtained by the EEM (Table 4 ;Figs.7 and 8).Component 1 was derived from tyrosine-like and protein-like compounds (Coble et al.1998).Components 2,5,and 7 represent UV humic compounds (Coble et al.1998;Stedmon et al.2003).Component 3 is humic-like compounds derived from microbial and photochemical degradation (Stedmon et al.2007).Component 4 is derived from protein contained in tryptophan (Coble et al.1998),and component 6 is protein-like (Sanchez et al.2013).The results of the PARAFAC analyses demonstrate that,compared to total rainfall,the throughfall in the broad-leaved forest was enriched with humic substances (components 2,3,and 5) and protein derivatives (component 6) (P<0.05)(Fig.8).The stemflow were enriched with humic substances(components 2,3,and 5 for the Japanese cypress and the Konara oak) and protein derivatives (components 1,4 and 6 for the Japanese cypress;component 6 for the Japanese red pine) compared to the rainfall (P<0.05).The correlation efficient for the seven components obtained by the PARAFAC analyses or UV absorbance at 254 nm (a254),and the DOC concentration of each fraction are provided in Table 5.High correlations were revealed between DOC concentration and component 7 (r=0.964,P<0.05) for total rainfall,components 3,5 and a254 (r=0.969–0.989,P<0.05) for the throughfall in site C,component 6 (r=0.958,P<0.05)for the stemflow from the Japanese hill cherry,and components 5 and 6 (r=0.977–0.996,P<0.05) for the stemflows from the Japanese red pine.Additionally,the stemflows from the Konara oak showed a weak positive correlation between the DOC concentration and components 2,3,5 and a254(r=0.913–0.942,P<0.1).Similarly,the DOC concentrations of the throughfall in site C and the stemflow from the Konara oak were loosely correlated with component 2(r=0.908–0.919,P<0.1).

Table 4 Characteristics of the components identified by PARAFAC analyses

Fig.7 The spectral properties of the seven fluorescent components identified by the PARAFAC analysis.Ex.:excitation wavelength,Em.:emission wavelength

Fig.8 The Raman unit fluorescence intensity of rainfall and throughfall,stemflow of each EEM components obtained by PARAFAC analyses.B:throughfall in the broad-leaved forest,C:throughfall in the coniferous forest,Jc:stemflow from Japanese cypress,Jh:stemflow from Japanese hill cherry tree,Ko:stemflow from Konara oak,R:rainfall,Rp:stemflow from Japanese red pine

Therefore,it was concluded that the DOC concentrations of total rainfall were affected by humic substances.Regarding the stemflow from the Japanese hill cherry,the DOC concentration was influenced by protein derivatives.With throughfall in the coniferous forest,DOC concentrations were affected by humic substances and a254,representing colored dissolved organic matter (CDOM),which has UV absorbance.For stemflow from the Konara oaks,DOC concentrations were affected by humic substances and CDOM with UV absorbance.For stemflow from Japanese red pines,DOC concentrations were affected by humic substances,protein derivatives,and CDOM with UV absorbance.The source of DOC was considered to be from the trees and supplemented by inputs from epiphytes and atmospheric deposition (Van Stan and Stubbins 2018).As previously noted,this site is affected by air masses from the Asian continent and by local emissions,therefore,atmospheric deposition is likely to be one of the contributors of humic materials as humic-like substances in aerosol accounts for 9–72% of water soluble organic matter (Zheng et al.2013).

Deposition flux

The deposition flux was approximately estimated(Table 6).The deposition flux of the DOC from throughfall (7.31–10.1 g m?2a?1) and stemflow (1.79–3.84 g m?2a?1) were within the range of values from various temperate forest studies (4.1–34.0 g m?2a?1for throughfall;0.1–5.6 g m?2a?1for stemflow;Van Stan and Stubbins 2018).However,the deposition flux of the DOC from the stemflow in this study was relatively higher than those of previous studies.

The deposition fluxes of cations and anions (except for K+,Ca2+and SO42?)were similar to those of conif-erous and broad-leaved forests in Japan(Na+:1.7–2.4,K+:2.2–3.1,Mg2+:1.2–1.4,Ca2+:3.1–3.2,Cl?:3.4–5.3,N O3?:2.6–6.0,SO42?:8.5–11.8gm?2a?1;Oura2010).In contrast,the deposition flux of K+from the broad-leavedforest (site B) in this study (8.35 g m?2a?1) was high compared to those of coniferous and broad-leaved forests in Japan.The high deposition flux of K+in the broad-leaved forest has been attributed to leaching from the canopy(Devlaeminck et al.2005).The amounts of ions dissolved from the bark and leaves were investigated and it was found that the high amounts of K+were from the Konara oaks and Japanese hill cherry (Table 7).The deposition flux of Ca2+was low compared to those of other coniferous and broad-leaved forests in Japan.The deposition flux of SO42?at the study site (1.78–2.58 gm?1a?1) was low compared to most Japanese forests(8.5–26.7g m?1a?1;Kobayashi et al.1999;Oura 2010),a forest in the San Bernardino Mountains,USA (4.8–29 g m?1a?1,Fenn et al.2000) and relatively low compared to forests in Flanders,Belgium (2.6–10.6 g m?1a?1,Wuyts et al.2008).The canopy openness of the Satoyama forest in this study was 13.4–27.6%,higher than that of many natural forests worldwide.Therefore,the canopy interaction was considered to be low compared to that of natural forests,resulting in a low deposition flux of SO42?in Satoyama forests.However,as previously mentioned,the traditional Japanese Satoyama is facing a serious crisis.In this study site,mass mortality of oak trees has been widely observed.If the forest composition changes to coniferous due to the mortality of oak trees,the deposition flux of DOC and K+might decrease by 33% and 62%,respectively (Table 6 as described in Coniferous forest/Present).One reason for the decrease in DOC deposition flux may be the decreased amount of pollen produced from the oaks.It has been reported that large amounts of oak pollen significantly contributed to the highest DOC levels in throughfall in forests of Georgia and South Carolina in the southeastern USA (Van Stan et al.2017;Chen et al.2019) and in New England,USA (Decina et al.2018).In contrast,the deposition flux of NO3?might increase by 20% due to mortality of oak (Table 6).As shown in Table 2,the main source of NO3?in our study site was atmospheric deposition.Coniferous canopies show a higher filtering capacity to atmospheric deposition than broadleaved canopies because the structure of twigs with needle-shaped leaves favors deposition through impaction and diffusion processes (Le Mellec et al.2 010),which is consistent with the higher throughfall depositions observed in coniferous forests in Germany (Rothe et al.2002;Le Mellec et al.2010).Hence,the deposition flux of NO3?would favor coniferous species due to oak mortality.Similar to this study,the mass mortality of oak species has been observed worldwide,such as in North America (Dey and Schweitzer et al.2018),central USA (Bendixsen et al.2015;Wood et al.2018) and southern Europe (Colangelo et al.2018;Etzold et al.2019).Ultimately,the decrease of DOC deposition from these sites will result in the decrease of riverine DOC levels onto the coasts,accompanied by changes in the terrestrial and marine carbon cycles as DOC is an important source of humic substances with chelating functions in river-ocean systems.Organic Fe-binding ligands are especially important in iron uptake for marine phytoplankton(Gledhill and Buck 2012).Furthermore,the increase in deposition of NO3?from these forests will lead to nitrogen saturation,which in turn enhances the acidification of soils and leaching of NO3?into surface and ground water.In the near future,the further deterioration of these forests might change river-ocean linkages affecting forested catchments and aquatic ecosystems.

Table 5 The correlation efficient between seven components obtained by PARAFAC analyses or UV absorbance at 254 nm (a 254) and DOC concentration of each fraction

Table 6 Deposition flux(g m?2 a?1) of DOC,cations and anions in Satoyama forest

Table 7 Amounts of cations and anions dissolved from bark or leaves (μg g?1)

Conclusions

Chemical behavior of major ions and dissolved organic matter were identified in throughfall and stemflow in a typical Satoyama forest facing imminent deterioration.Average dissolved organic carbon concentrations in throughfall and stemflow site were within range of values reported elsewhere.The Satoyama forest was affected by both atmospheric deposition due to anthropogenic activities and by seasonal changes in air masses.The deposition of sulfates was low compared to that in other forest types because the canopy interaction of the Satoyama forest was low.If the forest composition changes into a pure coniferous forest from the present mixed Satoyama forest due to oak mortality,the deposition of dissolved organic carbon and K+might decrease by 33% and 62%,respectively,and NO3?might increase by 20%.In the near future,the degradation of Satoyama landscapes might change dissolved organic carbon and nitrogen loads,resulting in imbalances in river-ocean linkages affecting forested catchments and aquatic ecosystems in Japan.

AcknowledgementsWe thank Michiko Kato,Hiroki Yamamoto,Mayu Inoue,Ayane Kawaguchi,Kentaro Onishi,and Moe Hirano for their help with rain sampling and LAI measurements.Meteorological Data Explorer (METEX) was provided by Dr.Zeng Jiye,(Centre for Global Environmental Research National Institute for Environmental Studies).

Author contributionSA:Conceptualization,methodology,formal analysis,writing original draft,supervision;FS:Conceptualization,supervision;YW:Conceptualization,supervision,data quality checking;WAJ:Formal analysis;MO:Data processing;NS:Data processing;YW:Data processing;LML:Data processing,English review;AK:Investigation,formal analysis;YS:Investigation,Formal analysis;NF:Investigation,formal analysis.

Declarations

Conflict of interestsThe authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.