Sources and Deposition Fluxes of PCDD / Fs in the Largest Reservoir System in Taiwan before and after Typhoon Morakot

The Tsengwen Reservoir (TWR) is positioned at 23°16′N, 120°34′E and is the largest reservoir in Taiwan. During 2009 to 2010, deposition flux of polychlorinated dibenzo-p-dioxin and dibenzofuran (PCDD/F) in atmosphere, water bodies and sediments were measured in the TWR area for this study. The monthly ambient air PCDD/F deposition in the surrounding area of the TWR was found to range from 0.206 to 2.43 pg I-TEQ/m/day. In addition, the PCDD/F levels (11 to 12 fg I-TEQ/L) in the shell water (water depth: 0.5 m) were found to be uniform and well mixed within the upstream, central and lower reaches of the TWR. The PCDD/F input flux into the reservoir was investigated in two sediment cores sampled at the TWR. The year dating within the sediment cores at different depths was projected from the sedimentation rate (> 17 cm/y) and calculated by Pb analysis. The PCDD/F content within the sediment cores at different depth were 0.136 ng I-TEQ/kg d. w. to 0.413 ng I-TEQ/kg d. w. Before the Morakot typhoon event (June, 2009), the PCDD/F concentration of the surface sediments were 0.091 to 0.187 ng I-TEQ/kg d. w. as detected at the upstream, central and lower reaches of the TWR. Interestingly, the PCDD/F content (0.132 to 0.222 ng I-TEQ/kg d. w.) of surface sediments increased somewhat after typhoon event (January, 2010). During the severe typhoon periods, the PCDD/F input contributed by the enhanced catchment landslides dramatically increased to 99%.


INTRODUCTION
Polychlorinated dibenzo-p-dioxins (PCDDs) and polychlorinated dibenzofurans (PCDFs) are characterized as the persistent organic pollutants (POPs) that have been aimed for international source reduction by the Stockholm Convention.PCDD/Fs are generated as unwanted impurities during different industrial and thermal processes and are released into the atmosphere through natural or anthropogenic combustion processes (Kulkarni et al., 2008).As sediment is the final sink for heavy metals and POPs, the vertical distributions of residues in dated sediment cores have been investigated as historical records of pollution (Okumura et al., 2004;Suarez et al., 2006).Significantly higher persistence of PCDD/Fs results in a very slow decomposes in the concentrations of these pollutants in soils and sediments (Masunaga et al., 2001).
Recently, the United Nations Environment Programme (UNEP/AMAP, 2011) proposed the theoretical demonstration of key factors influencing the environmental fate and transport of PCDD/Fs under a climate change situation.Taiwan is located in the subtropics, off the southeast coast of mainland China.During summer season, Taiwan is seriously exposed to natural hazards including flood, typhoon and landslides.The relevant extreme events including the 2001 Toraji Typhoon, the 2004 Mindulle Typhoon, and the 2009 Morakot Typhoon, all caused terrible damage (Juang, 2011).The occurrence and the severity of damage of those events over the last decade are of significant concern to the government agencies, engineering society and the public issues.Recent study demonstrated that over 70% of atmospheric PCDD/Fs are fundamentally particle-bound (Chi et al., 2011).These aerosols are transported into the atmosphere by the wind and then eventually deposit into water bodies or other regions in the environment via either dry or wet depositions.In addition, it has been shown that the PCDD/F concentrations shortly after storms and typhoons are elevated in surface soil, sediment and runoff-water by several folds (Presley et al., 2006).In regions with no precipitation or during the periods of low rainfall, atmospheric PCDD/Fs are transported more effectively.Additionally, the PCDD/F depositions from the atmosphere into surface media are decreased with decreasing precipitation.Precipitation forecast with respect to climate change differ significantly around the globe and appear to point out both decreased (20%) and increased (40%) trends as reported by the Intergovernmental Panel on Climate Change (IPCC, 2007).Hence, a major impact of these variations in precipitation on PCDD/F concentrations will be by reason of changes in wet deposition.
In addition, the effect of changing precipitation on soil moisture will be another factor; this will affect the microbial communities in the sediment that decompose PCDD/Fs (UNEP/AMAP, 2011).Hence, the influence of typhoons on the long-term remobilization of PCDD/Fs supports the hypothesis that relevant events have the possibility to reactivate pollutants that have been accumulated in the past.Oram et al. (2008) has indicated that extensive measurement of POPs in water of shore-side, freshwater, sediment, ambient air, wastewater outflow and sludge can be integrated via a model of mass budget providing a synthetic observation of these emerging pollutants.The Tsengwen Reservoir (TWR) is the most important reservoir in southern Taiwan; it is a multifunction reservoir the primary purpose of which is the supply of water to southern Taiwan.As a result of the TWR's importance, the potential risks due to PCDD/F contamination in this watershed cannot be underestimated in terms of environmental health.Hence the measurement of the deposition fluxes of PCDD/Fs within this reservoir system will be useful when evaluating the effectiveness of legislative action in terms of the inputs of contaminants.In this study, the comprehensive deposition of atmospheric PCDD/Fs and systematic monitoring of PCDD/F sedimentary was measured to quantify the comparative PCDD/F deposition fluxes and identify probable sources in the TWR.

Description of Tsengwen Reservoir (TWR)
Tsengwen Reservoir (TWR) is located upstream of Tsengwen River at 23°16′N, 120°34′E.The TWR scenic area is enclosed by precipitous green mountains.Endangered black-faced spoonbills overwinter at the TWR every year as migrants and inhabit the downstream area near the estuary, where many other types of waterfowls are also found.A conservation area has been set up to protect the spoonbills.Reservoir structure began to construct in mid-1967 and started to storage water in 1973.The TWR belong to the Tsengwen River in the tropical mountain zone of southwestern Taiwan.The river with 140 km length originates in the Wansui Mountain of Central Mountain Range (altitude: 2,440 m).It drains the watershed including of rural, agricultural, and suburban zones with 1,177 km 2 catchment area, that is over 60 times larger than the surface area of TWR (17.1 km 2 ).In addition, there are two municipal wastes incinerators and two electric arc furnaces in the southwest and northwest of the reservoir (about 30-50 km).The possible emission sources of PCDD/Fs in the vicinity of TWR ate these four facilities.

Sample Collection and Analysis
To determine the atmospheric PCDD/F concentration in the surrounding area of the TWR in southern Taiwan, sampling site A was selected at the operations building of the TWR (Fig. 1).Atmospheric PCDD/Fs of both the vapor and solid phases were sampled via Shibata HV-700F for semivolatile pollutants.Quartz fiber filters and polyurethane foam (PUF) plugs were utilized for collection of all particle-bound and the vapor-phase PCDD/F compounds, respectively.During the monthly sampling duration (7 days), the total sampling volume of each sample was greater than 1,000 m 3 (gas flow rate: 100 L/min).In 2010, atmospheric PCDD/F samples (n = 5) were collected during February, April, June, August and October.The ambient air samples were collected during the second or third week of the month at the operations building of TWR.To measure the atmospheric depositions of PCDD/Fs, Site A and Site B were set up in the vicinity of the TWR (Fig. 1).For atmospheric PCDD/F deposition measurements, two atmospheric sampling stations were established at the operations building (Site A) and a primary school (Site B) in the vicinity of downstream and upstream reservoir, respectively.The monthly (30 days) atmospheric PCDD/F deposition flux was measured via the cylindrical vessel (D: 500 mm, H: 600 mm) made by mirror-polished stainless steel.In addition, around 10 L of de-ionized water was added to each vessel before sample collection to cover the bottom of sampler.After 30 days sampling, we used a portable sampling system via two different sizes of quartz fiber filters, polyurethane foam (PUF) and gear pump to recover the remaining water and dust in the cylindrical.During 2010, around eighteen samples of atmospheric PCDD/F deposition flux were collected each month at Site A and Site B. The detail information (Chi et al., 2009b, Chi et al., 2011) of atmospheric PCDD/F depositions measured in the vicinity of TWR can be calculated as follows: ( ) Flux day : Atmospheric PCDD/F deposition flux (pg I-TEQ/m 2day); M: PCDD/Fs (pg I-TEQ) collected by stainless steel cylindrical vessel; A: surface area of stainless steel cylindrical vessels (m 2 ); t: sampling duration (days).
On June 2009 and on Jan 2010, PCDD/F levels in the surface water (n = 6) and sediments (n = 12) were also considered (Fig. 1) in the upstream, central and lower reaches of the TWR, respectively.In this study, we used a portable sampling system with a quartz fiber filter, PUF and gear pump to collect the surface water with water depth around 0.5 m.According to the water circulated through the filters and PUF, the particles were captured by filter, and the dissolved PCDD/Fs were sampled by PUF.Two sediment (97 cm and 77 cm) cores were collected from positions in lower reaches and south bay of the TWL during June 2009; sample collection was directly collected by cautiously inserting plastic tubes into the bottom sediment by manpower.After the sampling, the sediment core was delivered to the laboratory within 120 minutes and sectioned instantaneously into slices with 1-2 cm thickness.After the Typhoon Morakot event, we tried to collect the deep sediment core samples from positions in lower reaches and south bay of the TWL.However, the sampling condition at the lower reaches of TWR is not appropriate especially for deep sediment core collection after the Morakot typhoon event.Hence, the information and PCDD/F concentration of surface sediment samples collected after Typhoon Morakot is provided in this study.The sediment samples were then freeze-dried and ground to 100-200 mesh-sized powder using an agate mortar and pestle.

Tsengwen Reservoir
Based on the variation of the sediment mass concentration before and after the freeze drying process, the water content of the sediment was calculated.In addition, total organic carbon (TOC) was analyzed using a EMIA CS500 equipped with a resistance furnace and an ND-IR detector.About 100 mg of sediment was fumed with HCl to remove carbonates and then the de-carbonated sample was combusted at 1,350°C in the analyzer.Based on the PCDD/F concentrations obtained from different depths and mean sedimentation rate, the annual input fluxes of PCDD/Fs (Chi et al., 2009a(Chi et al., , 2013b) ) for the TWR can be calculated as follows: PCDD/F input flux (ng I-TEQ/m 2 /y) = PCDD/F concentration (ng I-TEQ/kg d.w.) of each sediment core × (Dry weight of each part in sediment core (kg d.w.)/Height of each part in sediment core (cm)) × (Sedimentation rate/Area of sediment core (6.36 The characteristics of all of the environmental samples collected in the vicinity of the TWR are listed in Table 1.Regarding to the analysis of PCDD/Fs, the atmospheric (ambient air and deposition) and environmental (water flow and sediment) samples were spiked with internal quantification standards according to USEPA methods 23 and 1613, respectively.The relevant information related to the extracted and clean-up process of the PCDD/F samples is provided elsewhere (Chi et al., 2007(Chi et al., , 2009a)).After the clean-up process completed, all the samples were analyzed using a high-resolution gas chromatography system (HRGC)/ high-resolution mass spectrometer (HRMS) (Thermo DFS) equipped with a fused silica capillary column (60 m × 0.25 mm × 0.25 µm, J&W).

Quality Control and Data Analysis
In this study, the concentrations of PCDD/Fs in part of the sediment core samples (the depth ranging from 0 to 10 cm) were analyzed in duplicate to diminish the uncertainty of the measurements.The variability levels in the duplicate analysis of PCDD/F concentrations were not higher than 10% (3.85-9.3%).In addition, a laboratory blank and a matrix spiked sample (2.0-20 pg/µL PCDD/Fs) were used in the analytical procedure for every eight to ten samples for quality control purpose.Method detection limits (MDL) were around 0.04-1.30pg/g and calculated based on the measurements of the blanks and quantified as three times the standard deviation of the concentration in the blanks.In this study, the concentrations of all laboratory blank samples were less than 1.0 pg (PCDD/Fs).The mean recoveries of the standards for all 13 C 12 -2,3,7,8-substituted PCDD/Fs were 53% to 107%.Therefore, the analyzed results were all within the acceptable 40%-130% range recognized by the US EPA in Methods 23 and 1613.For data analysis, Toxic Equivalency Factors (TEFs) are adopted to evaluate the potential toxicity of each PCDD/F and dioxin-like PCB congener in a concoction related to the well-studied and known toxicity of TCDD, the most toxic congener of the group.Researchers have reviewed toxicological databases with particular reference to chemical structure and have agreed to assign to TEFs a specific order of magnitude of toxicity for each dioxin-like congener relative to that of 2,3,7,8-TCDD, which is assigned a TEF of unity (1.0).The other congeners have TEF values ranged from 0.00001 to 1.0.The TEF of each PCDD/F congener in attendance in concoction is multiplied by the individual mass concentration, and the products are summed to yield the 2,3,7,8-TCDD Toxic Equivalence (TEQ) of the mixture.

Year Determination of Sediment Core
In this study, 210 Pb was measured by alpha spectrometry via its daughter isotope 210 Po (half life: 138 days).Because the half life of 210 Po is greatly shorter than that of its parent element, 210 Po and 210 Pb are typically found in radioactive balance within the sediments.In general, the relationship between the concentration of excess 210 Pb and sedimentation rate can be defined by: C = C 0 × exp[(-λ/S)/z] based on the assumption of a constant 210 Pb flux and the stable sedimentation rate.Where z (cm) is the depth in the core; C(dpm/g) is the activity of excess 210 Pb at z; C 0 is C at depth z = 0; λ is the decay constant of 210 Pb (0.0311 y -1 ) and S (cm/year) is the sedimentation rate.Excess 210 Pb (dpm/g) = 210 Pb (α) -226 Ra (γ) was measured by gamma counting of the 214 Pb (daughter isotope).The profiles of 210 Pb in selected sediment core samples sampled at the TWR south bay are presented in Fig. 2(a).Similar depth profiles of excess 210 Pb were also obtained by gamma counting of 210 Pb (Fig. 2(b)).As shown in Fig. 2, the sedimentation rates in average of the sediment core collected at the south bay of the TWR was higher than 17.0 cm/y.Overall, the sedimentation rate calculated by 210 Pb is according to the stable deposition.However, 137 Cs gives more accurate dates, and is particularly useful for recent sedimentation studies.Fig. 2 shows the down core distribution of 137 Cs activity and this indicates that there are no significant peaks across all depths of the sediment core obtained from the TWR.In general, the peaks in 137 Cs activity corresponded to a historical fallout maximum in 1963.From the 137 Cs profile (Fig. 2(c)), we cannot estimate the sedimentation rate at the south bay of TWR in average, but the result is similar with those derived from 210 Pb dating.Therefore, sedimentation rate at the south bay of TWR was calculated to be 17.0 cm/y, according to the depth profile conducted by gamma counting of 210 Pb.The results (Fig. 3) regarding to the deposition flux of atmospheric PCDD/Fs indicated that around 0.435 to 1.86 pg I-TEQ/m 2 /day and 0.206 to 2.42 pg I-TEQ/m 2 /day, respectively, at Sites A and B measured in the vicinity of the TWR.In addition, the annual atmospheric PCDD/F input flux was 0.421 ng I-TEQ/m 2 /y in the vicinity of TWR.Relevant measurements regarding to the deposition fluxes of atmospheric PCDD/Fs conducted in other Asian countries (Seike et al., 1998;Ogura et al., 2001;Sakai et al., 2001;Moon et al., 2005;Ren et al., 2007;Fang et al., 2011) demonstrated that the deposition flux of atmospheric PCDD/Fs measured in the vicinity of the TWR in southern Taiwan was considerably lower.According to other previous studies measured in Taiwan, PCDD/F depositions (0.206-2.42 pg I-TEQ/m 2 /d), as measured in this study, were also considerably lower than the results measured in an industrial  The deposition flux of atmospheric PCDD/Fs measured in summer (August, 2010) was the lowest.Numerous mechanisms could be answerable for the seasonal prototype in air pollutant variation observed in this study.These include lower oxidation reaction rate and atmospheric mixing layer height in winter, a higher atmospheric oxidant concentration in summer and seasonal variations in meteorological surroundings (Chang et al., 2003).In addition, Wang et al. (2010) demonstrated that the PCDD/F concentrations in ambient air measured in the surrounding area of industrial park and municipal waste incinerators in northern Taiwan were around 82.9 to 112 fg I-TEQ/m 3 .Previous studies (Wu et al., 2010;Huang et al., 2011) have demonstrated that the PCDD/F concentrations in ambient air sampled in industrial parks in central and southern Taiwan were 33.9 to 72.7 fg I-TEQ/m 3 and 34.8 to 106 fg I-TEQ/m 3 , respectively.The results of our previous measurements (Chi et al., 2008) also demonstrated that the PCDD/F concentrations in ambient air measured in Taiwan urban were 20 to 110 fg I-TEQ/m 3 .Another study (Chi et al., 2013a) also demonstrated that the considerably lower PCDD/F contents (3.45-49.1 pg I-TEQ/g-TSP) in total suspended particles (TSPs) and ambient air concentrations (1.24-7.75fg I-TEQ/m 3 ) were measured at background areas in Taiwan and remote islands around Taiwan.In Korea and Japan, the ambient air PCDD/F concentrations in urban areas were ranged from 28 to 120 fg I-TEQ/m 3 (Makiya, 1999;Lee et al., 2007).The ambient air PCDD/F concentrations measured in the surrounding area of the TWR in southern Taiwan are thus overall significantly lower than the measurements observed in in Taiwan and other countries.Because of the lack of emissions and combustion sources within 30 km of the TWR, the significantly lower atmospheric deposition and ambient air concentration of PCDD/Fs were observed in this study.

Atmospheric Deposition Flux and Ambient Air Concentration of PCDD/Fs over TWR
In our measurement, the solid-phase PCDD/Fs were characterized as the ambient air PCDD/Fs sampled by the quartz fiber filter, whereas the vapor-phase PCDD/Fs were characterized as the ambient air PCDD/Fs that were not collected by the filter but by the PUF.Fig. 4 show the partitioning of atmospheric PCDD/Fs between the vapor phase and the solid phase with ambient air temperature and TSP variation.In Fig. 4(a), the temperature describes the ambient temperature in average over the sampling duration (around seven days) in the surrounding area of the TWR.From February to August, the ambient air temperature increases by 10°C and in parallel, the solid-phase PCDD/Fs decreases by almost 25%.In most cases, the partitioning of semi-volatile compounds (like PCDD/Fs and PCBs) distributed in particles is generally affected by the vapor pressure of the compound involved (Pankow, 2001).The vapor pressures of organic compounds decrease as the temperature decreases, which results in a higher partitioning of PCDD/F congeners distributed in vapor phase.Additionally, the partitioning of solid-phase PCDD/Fs is also affected by the suspended particle concentration like TSP in the ambient air.As shown in Fig. 4(b), as the TSP concentration decreases, the partitioning of PCDD/Fs distributed to particles also decreases.

PCDD/F Circulation in Reservoir Water and Sediment over TWR
In the surface water, the PCDD/F levels in average sampled at upstream, central and lower reaches of the TWR, were 12 fg I-TEQ/L (n = 2), 11 fg I-TEQ/L (n = 2) and 11 fg I-TEQ/L (n = 2), respectively.The measurements demonstrated that the PCDD/F levels in the surface water were sufficiently mixed and homogeneous across the various diverse locations in the TWR.In addition, our water measurements in this study also indicate that the distribution of PCDD/Fs between dissolved (collected by PUF) and solid (collected by filter) phases within the water flow are dissimilar within the upstream, central and lower reaches of the TWR.The solid-phase distribution of PCDD/Fs (91%) in the surface water, as sampled within the lower reaches of the TWR is lower than that sampled within the central (92%) and upstream reaches (98%) of the TWR.The measurements also show that the sediment PCDD/F concentrations varied at different locations within the TWR.On June 2009, the highest PCDD/F concentration (0.199 ng I-TEQ/kg dry weight, d. w., n = 2) in the surface sediment was observed within the lower reaches in TWR.The average PCDD/F concentration (0.091 ng I-TEQ/kg d.w., n = 2) of the surface sediment observed in upstream of the TWR was lower than that observed within the central reaches of the TWR (0.115 ng I-TEQ/kg d.w., n = 2).According to the sediment PCDD/F concentration observed in United States (1.47-1.99ng I-TEQ/kg d.w.), Italy (1.3-13 ng I-TEQ/kg d.w.), Japan (1.9-5.3 ng I-TEQ/kg d.w.) and China (1.9-89 ng I-TEQ/kg d.w.) (Fattore et al., 2002;Kiguchi et al., 2007;Liu et al., 2007;Loganathan et al., 2008), the PCDD/F concentrations in surface sediments measured in TWR in southern Taiwan were considerably lower.Relevant studies (Chi et al., 2009a(Chi et al., , 2011) ) demonstrated that the PCDD/F concentrations in sediments observed in reservoirs were 0.53 ng I-TEQ/kg d.w. and 0.393 ng I-TEQ/kg d.w. in northern and central Taiwan, respectively.In this study, the considerably lower PCDD/F concentrations in sediments observed in TWR may be due to the lack of any other PCDD/F emission sources within around 30 km of the reservoir.Compared to the sediment measurements in the reservoir, the PCDD/F levels in the surface water were sufficiently mixed and homogeneous at various dissimilar locations in the TWR area.Upstream of the TWR, the PCDD/Fs in solid phase were found to be distributed as fine and small particles within shallow water and similar particles were not observed in the deeper water, nor were they deposited on the bottom of reservoir.Hence, the significantly higher distribution of solid-phase PCDD/Fs in the water flow and a lower PCDD/F concentration in the sediments has been discovered in the upstream of the TWR.
Typhoon Morakot struck Taiwan from August 6-10, 2009, and it had a slowly moving route that was accompanied by a strong southwesterly monsoon.With a radius of over 1,000 km., Typhoon Morakot provided over 2,000 mm of rainfall, over 70% of the area's average annual rainfall, on southwestern and southeastern Taiwan (Juang, 2011).This extreme rainfall not only caused significant flooding, but also triggered slip of earth and debris flows throughout southern Taiwan especially in the area of the TWR.The extreme floods and debris flows caused massive destruction to roadways, levees, bridges, and buildings in southern and central regions of Taiwan (Juang, 2011).The heavy rainfall brought by Typhoon Morakot is characterized as high intensity and long duration, with a return period well over 200 years.The landslide distribution correlates well with the heavy rainfall distribution.The significant rainfall and floods triggered 12,697 landslides, including four enormous landslides.The cumulative area of the landslides was 183.1 km 2 (Wu et al., 2011).On January 2010, our measurements indicated that the PCDD/F concentration in average of surface sediment sampled at the different location within the TWR area increased before and after typhoon Morakot event from 0.091-0.199to 0.132-0.222ng I-TEQ/kg d.w..In addition, the PCDD/Fs increase was also observed in the upstream of the TWR.These measurements speculate that a remarkable increase of the PCDD/F input flux is likely to have occurred due to a deeper turbid layer forming upstream after landslides and the presence of catchment erosion during the Typhoon Morakot event.The discharge, disseminating and degradation of POPs are significantly dependent on environmental conditions and therefore climate change and variability have the possible to affect POPs contamination via changes to transport processes, pathways, and routes of degradation (Bogdal et al., 2010).In addition, successive and intensive typhoons that result in serious rainfall do seem to cause increased landslide and this in turn resulted in a drastically higher input flux of PCDD/Fs into the sediment of the reservoir under study.Chi et al. (2013b) has demonstrated that significant increases in PCDD/F input fluxes and elements levels of mineral-derived do occur; these seem to be attributed to the landslides and the presence of catchment erosion during the typhoon season.
The distribution of total PCDDs and PCDFs (only seventeen 2,3,7,8-chlorosubstituted congeners) in the ambient air, atmospheric deposition, surface water and sediment measurements were presented in Fig. 5(a).The measurements demonstrated that the lowest distribution of PCDDs (26%) was observed in the solid-phase sample in ambient air.An decreased distribution of PCDFs was also observable in the aerosols, atmospheric deposition, surface water and sediments samples.The origins of this trend towards an decrease in PCDF distributions are twofold.Firstly, PCDFs are primarily distributed in the vapor phase in ambient air (Lohmann et al., 1999) and therefore deposited particles collected in the ambient air and in water lead to an enhancement in PCDD distribution.The second aspect reflects the consequence of landslide around the reservoir, especially after the Typhoon Morakot event.iIn general, the atmospheric deposition and soil erosion of the reservoir's catchment area were defined as the PCDD/F sources in the reservoir.In addition, the significantly higher distribution of PCDDs was measured in soil samples (Jou et al., 2007;Kiguchi et al., 2007).The results of our measurements indicated that increases in PCDD/F concentration were observed in the surface sediment of the TWR after the Typhoon Morakot event.Hence, we considered that the catchment erosion of the reservoir is likely to have resulted in an increase in the distribution of PCDDs in the sediment.A further investigation regarding the distributions of PCDD/F congener across dissimilar environmental receptors in this study was carried out and this also demonstrated that the distribution of OCDD was higher than 75% measured in sediments (Fig. 5(b)).On the other hand, the distribution of OCDD measured in ambient air deposition and water flow were not higher than 60%.A high distribution of OCDD in homologue profiles of archived soil samples in Taiwan are have been described (Cheng et al., 2003) and based on these it would seem that catchment erosion of the TWR increases the of PCDD distribution in deposited particles in water flow and in sediment.

PCDD/F Input Flux Derived from Sediment Cores
Fig. 6(a) indicates that the TOC level increases slightly as the depth of the sediment core increases.On the other hand, the TOC level (0.3%-0.9%) observed at dissimilar depths in cores that were sampled within the lower reaches and south bay of the TWR show the same pattern.Significantly higher TOC levels (0.8-0.9%) were measured in the deeper part of the core (80-85 cm) that was sampled within the lower reaches of the TWR.Additionally, the water content (Fig. 6(b)) of the sediment cores varied considerably at the dissimilar locations near the TWR.The water content (34%-62%) at different depths in the core sampled at the lower reaches of the TWR was higher than that measured for the south bay (27%-58%) of the TWR.We considered that is because of the different grain size compositions with the sediment core collected in the TWR.Therefore, the coarse particles in the water flow of the south bay of the TWR seem to have mostly sunk to the bottom.Fig. 5(c) shows that the concentrations of PCDD/Fs (0.162-0.413 ng I-TEQ/kg d.w.) in the sediment cores measured within the lower reaches were slightly higher than those measured at the south bay (0.136-0.367 ng I-TEQ/kg d.w.) of the TWR.The results obtained from the measurements of surface sediments are similar to those observed in the sediment core of the TWR.Our previous measurements (Chi et al., 2009) demonstrated that PCDD/F concentrations of a sediment core were 0.948 to 14.4 ng I-TEQ/kg d.w.measured in the surrounding area of the municipal waste incinerator in northern Taiwan.In this study, there are no major PCDD/F emissions sources exist or have existed within approximately 30 km of the vicinity of the TWR.Hence, PCDD/F concentrations in the sediment cores measured in southern Taiwan were considerably lower than those measurements conducted in other area in Taiwan.According to the mean sedimentation rate (17.0 cm/y) and the PCDD/F concentrations at different depths, the annual PCDD/F input for the TWR can be estimated.The flux of PCDD/F input obtained from the surface sediment (depth: 0-2 cm) at the south bay of the TWR was around 165 ± 19 ng I-TEQ/m 2 /y.In Europe, the PCDD/F deposition flux in atmosphere measured in urban areas were range from 1.2 to 13.9 ng I-TEQ/m 2 /y in Germany and 0.73 to 4.38 ng I-TEQ/m 2 /y in Denmark (Fiedler et al., 2000;Vikelsoe et al., 2003).Based on the measurements from a dated sediment core sampled at a reservoir in northern Taiwan, Chi et al. (2013b)  This finding confirms the effect of typhoon events on the long-term remobilization of PCDD/Fs as well as supporting the hypothesis that such events such have the possible to reactivate pollutants that have been deposited in the past.
In addition, the water depths observed at two sediment cores sampled at lower reaches and south bay of the TWL were over 30 m.At that deep water height, we considered the PCDD/F release of surface sediment due to vigorous perturbation during intensive typhoon periods was not significant.
The total input flux of PCDD/Fs measured at the TWR is dramatically higher than that observed in these other countries.In TWR, no major PCDD/F emission sources exist in the vicinity.However, the input flux (165 ± 19 ng I-TEQ/m 2 /y) of PCDD/Fs obtained from the surface sediments at south bay of the TWR dramatically exceeds the level of PCDD/F deposition (0.421 ng I-TEQ/m 2 /y) in atmosphere.In general, the steep landscape of Taiwan is responsible for significant erosion by the river upstream of the reservoir, especially during a typhoon event.In this study, the sedimentation rate in the TWR exceeds 17 cm/yr and dramatically higher than that observed in a lake in the United Kingdom (Rose et al., 1997).Additionally, the catchment area of the TWR is larger than 1,100 km 2 and is thus 60 times the reservoir surface area (17.1 km 2 ).Hence, we considered that the input flow of PCDD/Fs in this investigation involves atmospheric deposition and landslide of the surface soil upstream of the TWR.

CONCLUSIONS
Based on the findings obtained for the investigated reservoir, there is a dramatically lower deposition flux and concentration of PCDD/Fs in atmosphere and surface sediment in the surrounding area of the TWR.However, the input flux (165 ± 19 ng I-TEQ/m 2 /y) of PCDD/Fs obtained from the surface sediment was dramatically higher than the PCDD/F deposition (0.41 ng I-TEQ/m 2 /y) in atmosphere.The PCDD/F input flux observed in the TWR may to a large extent represent the sum of the PCDD/F input flux to the whole catchment area.Besides the depositions in atmosphere, impure compounds in specific pesticide products may be another source of PCDD/F contamination in the soil.Therefore, we considered that over 99% of the PCDD/Fs input into the TWR to have been contributed by enhanced catchment erosion during severe typhoon periods.These findings imply that pollutants such as PCDD/Fs deposited in the watershed were transported downstream by typhoon runoff and by associated physical erosion and both play would seem to have most important roles in the accumulation of PCDD/Fs in the downstream reservoir.

Fig. 2 .
Fig. 2. The depth profile of 210 Pb ex activity in the sediment core from the lower reaches of the Tsengwen Reservoir (TWR).

Fig. 4 .
Fig. 4. Comparison of PCDD/F partitioning between the vapor and solid phases with (a) ambient air temperature and (b) total suspended particle variation.

Fig. 6 .
Fig. 6.(a) Water content, (b) total organic carbon content and (c) PCDD/F concentration measured at different levels of the sediment core sampled from the lower reaches and south bay of the Tsengwen Reservoir (TWR).
also indicated that the significant increases in input fluxes of mineral-derived elements levels and PCDD/Fs in 1990 (20 ng I-TEQ/m 2 /year), 2001 (17 ng I-TEQ/m 2 /year), 2004 (16 ng I-TEQ/m 2 /year) and 2005 (15 ng I-TEQ/m 2 /year).Those phenomenons seem to be caused by a deep turbid layer formed upstream due to landslides and catchment erosion during the typhoon season.

Table 1 .
The characteristic of all the environmental samples collected in the vicinity of the Tsengwen Reservoir (TWR).