RELATIONSHIP BETWEEN ALLOCHTHONOUS DOC CONCENTRATION AND A SPECIFIC UV254 ABSORBANCE (SUVA) AT A MESO-STRATIFIED RESERVOIR

This study aimed to identify the origin of DOC loading, by which the extent of SUVA representing of humic content or aromaticity of natural organic matter, NOM concentration in lake, could be determined and the effect of hydraulics conditions on the dynamic behavior of SUVA in the Daecheong reservoir, Korea. The type of Daecheong reservoir was meso-stratified as the ratio of inflow rate to reservoir capacity, a, was under 10 but it was temporally classified into intermediate type during storm season by heavy rainfall. Over 90% of total DOC loading attributed to allochthonous sources in upstream and it concentrated in storm season. While, autochthonous DOC by production of algae in Daecheong reservoir appeared only in short period of algal bloom. And the proportion of autochthonous DOC to total organic carbon loading varied significantly depending on spatial locations but was not significant. The extent of SUVA in lake water depth was severely deteriorated by inflow rate and the location of intrusion of storm flows into lake layer during storm event because density of inflow storm water is usually heavier than that of the surface water. It has been found that SUVA in Daecheong reservoir strongly attributed to allochthonous DOC from upstream watershed

RELATIONSHIP BETWEEN ALLOCHTHONOUS DOC CONCENTRATION AND A SPECIFIC UV 254 ABSORBANCE (SUVA) AT A MESO-STRATIFIED RESERVOIR S.J.Yu*, S.R.Ha**, S.U.Cheon*, J.Y.Hwang*. *Kum River Water Quality Research Lab. National Institute of Evironmental Research 395-1 Annemyen, Okchungun, Chungbuk 373-810, Korea (E-mail: ysu1221@.me.go.kr) **Department of Urban Engineering, Chungbuk National University San 48 Keshindong, Heungdukgu, Cheongju, Chungbuk 361-763, Korea E-mail: simplet@cbucc.chungbuk.ac.kr ABSTRACT This study aimed to identify the origin of DOC loading, by which the extent of SUVA representing of humic content or aromaticity of natural organic matter, NOM concentration in lake, could be determined and the effect of hydraulics conditions on the dynamic behavior of SUVA in the Daecheong reservoir, Korea. The type of Daecheong reservoir was meso-stratified as the ratio of inflow rate to reservoir capacity, α, was under 10 but it was temporally classified into intermediate type during storm season by heavy rainfall. Over 90% of total DOC loading attributed to allochthonous sources in upstream and it concentrated in storm season. While, autochthonous DOC by production of algae in Daecheong reservoir appeared only in short period of algal bloom. And the proportion of autochthonous DOC to total organic carbon loading varied significantly depending on spatial locations but was not significant. The extent of SUVA in lake water depth was severely deteriorated by inflow rate and the location of intrusion of storm flows into lake layer during storm event because density of inflow storm water is usually heavier than that of the surface water. It has been found that SUVA in Daecheong reservoir strongly attributed to allochthonous DOC from upstream watershed. KEYWORDS Allochthonous; organic matter; COD Mn ;DOC; POC; Reservoir; SUVA; UV 254 INTRODUCTION Organic matter in natural water body used to be classified according to origin, such as allochthonous and autochthonous, physical state, such as particulate and dissolved, and nutrient content, such as carbon/phosphorous ratio. There are two main types of organic matter in lakes with forest zone: organic matter from primary or secondary production in water body, and humic matter from forest soil, paddy land, and marginal rooted vegetation (Brenig, 1987). As origin and composition of organic matter in freshwaters have not been largely explored, it is not possible to identify the sources of all organic compounds in lakes. However the description of gross properties may help to discriminate chemical composition of organic matter from different source (Cole et al., 1984). Various methods have been applied to determine the concentration of organic matter in water body. Despite of lower yielding of organic content by COD Mn than by dichromate method, COD Cr , the method of COD Mn has been used to measure the content of organic compounds in lakes’ water as a water quality criteria index in Korea. However it is not still clear if COD Mn could represent the content of organic carbon in water since of insufficient number of data to ensure the thorough relationship. - 111 - While, dissolved organic carbon, DOC in waters has been recognized principally to represent the content of aquatic humic substances with organic compounds of different molecular weights which could be of autochthonous or allochthonous in origin. Autochthonous DOC is mainly composed of photosynthetic inputs of the littoral and pelagic flora through secretion and autolysis of cellular contents. The humic compounds as DOC attributed to the excreted matter by zooplankton and higher animals, the products by bacterial chemosynthesis of organic matter with subsequent release of DOC, and so on (Wetzel. 1983). On the other hand, allochthonous DOC, which is of terrestrial origin consists mainly of humic substances synthesized biologically and chemically from the degradation products of terrestrial plant, animals and microorganism, nonhumic substances discharged through human activities. While humic content or aromaticity of natural organic matter, NOM concentration used to be monitored based on a specific ultraviolet absorbance at 254nm, SUVA that can be determined by dividing UV 254 with DOC concentration. Namely SUVA value depends on the amount of DOC and UV 254 in water body. As UV 254 value of allochthonous dissolved organic matter containing aromatic compounds is higher that value of autochthonous organic matter by aliphatic compounds, the extent of SUVA of allochthonous can be higher than of autochthonous organic matter when the portion of allochthonous organic matter become greater. Most organic matter in water is being dissolved and consists of protein, carbohydrate, fat and humic substances and so on. And furthermore, about 70 to 80% of total dissolved organic matter is occupied by humic substances. In addition, since aquatic humic substances are polar, straw-colored, organic acids that are derived from soil humus and terrestrial and aquatic plants, humic substances generally comprise one third to one-half of DOC in water. Public concerns on the role of humic substances in water chemistry is increasing because those are a source of methyl groups for the production of chlorinated methane in water treatment process, and implicated in the complexation or solubilization of pesticides and hydrocarbons in the aqueous environment (Thurman et al., 1981). The former view as a source of methyl groups is interested in drinking water treatment with respect to produce disinfection-by-products (DBPs) such as halogenated carcinogens which are then directly introduced into the public drinking water with obvious health consequence. The view of latter as being implication in solubilization of pesticides is regarding to humic substances among organic matter contain long-lived of free radicals, which are capable of reducing inorganic species such as Hg, Cr, and Pu as well as interacting with anthropogenic organic compounds (Senesi et al., 1994). Objectives of this study are to identify the origin of DOC loading could be determined and to explore the effect of hydraulics conditions on the dynamic behavior of SUVA in the Daecheong reservoir. METHODS Field observation and preservation of samples: As shown in Fig. 1, there were 6 sampling points in Daecheong reservoir and samples were collected at upper, middle, and lower layer in depth. Sampling frequency was once per month from May to October 2000. And inflow rate was measured for 4 times (March, June, September and October) to estimate allochthonous DOC of each stream during this period. Samples were immediately cooled in an icebox after sampling and brought back to laboratory. Hydrolgical parameter: The ratio of inflow to reservoir capacity, α (Aki and Shirasuna, 1978), as hydrology parameter was used to define reservoir type. Q in α = ------- (1) Vo In which, Q in is an inflow rate, m 3 /sec and Vo is a volumetric capacity of reservoir, m 3 . And DOC loading in Kuemkang river and its tributaries to reservoir was calculated by multiplied flow rate with DOC concentration measured. - 112 - Chemical analysis: For pretreatment of DOC and UV 254 , samples were filtered with Whatman GF/F filters prepared which has effective pore size 0.7µm and was pre-combusted at 450°C for 3 hours before use. The filtrates and raw water samples were kept at 4°C in glass bottles until analysis. Both DOC by filtrate and TOC were measured by TOC combustion analyzer. Analytical precision was ±1% based on three or more measurements of each sample. Particulate organic carbon, POC contents were determined with the difference of between TOC and DOC. UV 254 absorbance was measured with UV/Vis Spectrophotometer by filtrates. SUVA was calculated by dividing UV 254 absorbance with DOC concentration. Water sample to measure the concentration of Chlorophyll-a was filtered through Whatman GF/C filters prepared. Filters were ground by using tissue homogenizer with 6ml of 90% acetone and then centrifuged to remove turbidity. The concentration of Chlorophyll-a was analyzed with UV/Vis Spectrophotometer. Table 1. General information on Daecheong reservoir Fig. 1 Location of sampling stations in study watershed. RESULTS AND DISCUSSION Hydraulic characterics of Daecheong reservoir Daechoung reservoir is an artificial dam reservoir constructed for multi-purpose such as flood control, securing municipal water supplies, power generation and so on. It has a dendritric shaped stream network and locates in the central part of South Korea. Watershed consists of about 74.5% of forest and mountain area, 16.3% of paddy area, 9.2% of urban area. It is suspected that nutrient loading in lake is mainly originated from diffuse pollution sources such as forest and agricultural field. 0 20 40 60 80 100 120 140 160 180 JunJulyAugSepOct date mm 0 1000 2000 3000 CMS rainfall inflow Fig. 2 Seasonal variation of rainfall and inflow rate in Keumkang river Annual precipitation in this region is about 1400mm/year but more than half of precipitation is concentrated on storm season from July to September. Approximately, delivery time of storm peak in main Trophic state Meso-eutrophic state Reservoir area 64.3 km 2 Watershed area 4,166.8 km 2 Reservoir length 86 km Average depth 23.6 km Water capacity 1,490×10 6 m 3 Effective capacity 790×10 6 m 3 Yearly average inflow, outflow 2.76×10 9 , 2.75×10 9 m 3 /s Hydraulic residence time 0.68 year Urban area in drainage basin 55 km 2 ( 9.2%) Paddy area in drainage basin 678 km 2 (16.3%) Forest and mountain area 3,101 km 2 (74.5%) Total population in drainage area 332,580 persons - 113 - stream, Kuemkang is comparatively short to response against short-term rainfall events and it takes less than 2 days (Fig. 2). Through application of the ratio of inflow to reservoir capacity, α defined by Eq. 1, it was found that Daecheong dam reservoir was classified to be slightly stratified or mixed because of α>20 but to be intermediate type by heavy rainfall on August 1998 (Fig. 3). Hence it can be grouped to meso- stratified type. Statistical correlation of different indices representing of organic matter COD vs. TOC: The correlation analysis in terms of different organic indices was carried out statistically for the data obtained in dry season and wet season and results were summarized in Table 2. In upper layer of lake, there was significantly big difference of the correlation of COD Mn and TOC in dry and wet season. Even though, in wet season, correlation coefficient, r was 0.93 in upper layer between TOC and COD Mn . But, in dry season, the slope decreased to 0.79 slightly from 3.16. Table 2. Correlation and statistically significant regressions IndepandentDependantKinds of dataSlopeInterceptr*Size of sample All data0.7290 Upper layerWet season3.16-3.720.9318COD Mn TOC Dry season0.790.920.7912 All data0.6090 Upper layerWet season0.882.300.6018 Dry season1.85-1.240.9112 Middle layerWet season6.60 -11.020.8518 Dry season1.67-0.240.8112 Bottom layerWet season6.49 -10.010.9518 DOCUVA 254 Dry season1.63-0.380.9212 All data at Jannge site0.8015 Wet season -0.173.20 -0.25**9DOC Dry season0.541.670.756 All data at Jannge site -0.8015 Wet season0.26-0.470.699 Log Q POC Dry season-0.631.35 -0.77**6 *P<0.001, **P>0.05 DOC vs. UVA 254 : While, the correlation between DOC and UV 254 absorbance (UVA 254 ) was good except that of upper layer in wet season. Although summer season in Korea has high water temperature to give better condition for blooming of algae but it is just time of rainfall too. Hence it is common that algae blooming used to be occurred at the temporally intermediate period between the end of wet season and early of dry season. It could be suspected of that organic matter in water attributed mainly to autochthonous organic matter by algae at the temporally intermediate period. Statistics between DOC and UVA 254 supported this suspicion that slope was quite steep as 1.85 (with r=0.91) at dry season whereas it was gentle as 0.88(with r=0.6) at wet season. Whereas slopes of regression equation between two organic indices were high for whole seasons as ranged in 6.49-6.6 (r=0.85-0.95) for wet ad 1.63-1.67 (r=0.81-0.92) for dry in the middle and lower layer. Namely, the tendency of statistics between DOC and UVA 254 in middle layer and lower layer was similar. As a result, condition of both DOC and UVA 254 at middle and bottom layer was almost homogenous and the extent of NOM in wet season was about 4 times greater than that in dry season. It meant that allochthonous DOC in either middle or bottom layer dominated the extent of DOC. DOC loading: Regarding to loading, it had been believed that most of allochthonous DOC loading came from upper region of Keumkang river and about 85 to 95% of DOC in the reservoir was occupied by allochthonous DOC. In particular, about 90% of the allochthonous carbon loading to Daecheong Reservoir from watershed were concentrated at the summer season in July and September. From this reason, it was clear that carbon loading in Daecheong reservoir was largely depending on heavy storm runoff as shown in Fig. 4 and Table 3. While, despite of good relationship between inflow and POC, POC loading was lower than DOC loading comparatively. The ratio of DOC vs. POC varied in the range of 8.6 to 95.4 (as average - 114 - 0 10000 20000 30000 40000 MAMJJASO Month DOC Load(kg/day) substream Upstream 35.5) in wet season and 1.8-6.93 (as average 3.5) in dry season. Table 3. Allochthonous DOC loading in Daecheong reservoir MonthMarchJuneSeptemberOctober Flow rate(CMS) & Loading (kg/day) Flow rate DOC Loading Flow rate DOC Loading Flow rate DOC Loading Flow rate DOC Loading Total (main + tributary)9.793224.55,854153.934,07122.63,644 From Main stream9.187723.65,547148.332,64319.43,096 (%)94.394.196.194.896.495.885.985.0 Sum from tributaries0.6550.73075.61,4283.2548 (%)5.75.93.95.23.64.214.115.0 Fig. 3 Monthly retention time, α of water Fig. 4 Allochthonous DOC loading Fig. 5 SUVA variation of each layer. Fig. 6 Monthly SUVA variation in main stream of Daecheong Reservoir. Fig. 7 Variation of chlorophyll-a concentration in upper layer Fig. 8 Plot of chlorophyll-a concentration vs. POC in upper layer upper middle lower wet season dry season 0 1 2 3 SUVA layer Jangge C h ud o ng C h uso H o inam D a m M u nyi 2 2 - M a y 3 - J u l 3 1 - J u l 4 - S e p 2 3 - O c t 0 1 2 3 4 SUVA 0 20 40 60 80 100 5.227.037.319.0410.23 date chlorophyll-a(ug/L) Jannge Chuso Hoinam Chudong Dam Munyi 0 5 10 15 20 JuneJulyAugSepOct Month The ratio of monthly inflow to reservoir capacity, α 2000 1998 y = -0.0014x + 7.967 R = 0.12 y = 27.951x + 8.5395 R = 0.42 0 20 40 60 80 100 0500100015002000 POC(ug/L) Chl-a(ug/L) wet season dry season - 115 - Seasonal variation of SUVA Range of SUVA in reservoir: The broad diversity of the chromophores in NOM molecules causes the different absorbance bands to overlap and merge so that the overall UV absorbance spectrum of natural water resembles a single, smooth hump. The absorbance of distinct peaks in the spectrum has dissuaded researchers from analyzing it in detail. The use of UV spectroscopy to study NOM has consisted almost entirely of reports of the UV absorbance at a single, characteristic wavelength (λ), typically 254nm, or SUVA at the same wavelength (Li et al. 1998). SUVA in Daecheong reservoir was in the range of 1.465- 2.696 through the whole observation period. Particularly, SUVA of middle and lower layer in wet season ranged in 2.696 to 2.686. In Daecheong reservoir, storm water intruded into middle layer of the lake during storm season because the density of inflow water was usually greater than that of the surface water (Fig. 5). Impact of SUVA: Those allochthonous organic pollutant has been considered to influence on algae production in the surface layer of lake or to react as refractory portion to provide source of DBPs such as THM in water treatment process. In natural lakes, inflows usually are restricted to overland flow and small streams, and their impact is confined to be the littoral zone and surface waters (Ford, 1990). In contrast, in artificial dam reservoirs such as Daecheong, normally located at the upstream part of a river in mountainous terrain, inflows usually enter at upper end of the reservoir. Galapate et al. (1997) reported that among the diffuse pollution sources, forest litter exhibited the highest UV 260 and DOC concentration in Kurose River. In addition organic carbon budgets from diffuse pollution sources during heavy rainfall may help in the development of general mass-balance model with respect to carbon dynamics because the storm runoff has a dramatic impact on Daecheong reservoir. Spatial variation of organic carbon in Daecheong reservoir Overall trend of spatial changes: Spatial variation of SUVA along to main stream was plotted in Fig. 6. Although the pattern of SUVA variation corresponded with inflow pass caused by rainfall, there was clear different in the level of SUVA depending on the local condition such as water depth, temperature, and pollution sources. For instance, the first rainfall in dry season caused to produce high organic matter concentration in upstream. While around the location of Chuso monitoring station, the algal bloom used to be grown luxuriantly every year as water depth was shallow and nutrients were rich because of discharging from sewage treatment facility and releasing from sediment. From Fig. 7, it was distinct that distribution of chlorophyll-a concentration in upper layer varied depending on site to site as well as time to time. Namely, chlorophyll-a concentration might be high level when discharge of organic matter by continuous inflow just after first flush out by rainfall was coincided with other conditions. Significance of autochthonous DOC: The correlation coefficient between chlorophyll-a and POC was 0.42 in upper layer in wet season whereas that at dry season was also very low (in Fig. 6) in comparing with 0.90(r 2 ) in euphotic zone of Soyang reservoir (Kim et al. 2000). Consequently, it was quite hard to estimate exactly the extent of autochthonous organic matter within total organic matter existed in reservoir water. So that it was convinced that autochthonous organic matter by algae production in Daecheong reservoir would be effected in short-term only during peak of algal blooming. And autochthonous organic carbon loading would occupy only little part of total organic carbon loading. Even if primary productivity of autochthonous organic matter by algae, which was cited from reference, was applied, it would be difficult to estimate the extent of autochthonous organic carbon loading exactly as having great spatial variance depending on local conditions such as watershed characteristics and algal species and population. Autochthonous organic matter by algae production would occupy only limited portion of TOC loading because of great changes by location as well as time for instance being negligible during winter. - 116 - CONCLUSION This study aimed to identify the origin of DOC loading, by which the extent of SUVA representing of humic content or aromaticity of natural organic matter, NOM concentration in lake, could be determined and the effect of hydraulics conditions on the dynamic behavior of SUVA in the Daecheong reservoir, Korea. Major results summarized as follows: The type of Daecheong reservoir was meso-stratified as the ratio of inflow rate to reservoir capacity, α, was ranged in 10-20. Through statistical analysis in terms of indices of organic compounds, such as COD, TOC, DOC, and UVA 254 , it had been found that the correlation between DOC and UVA 254 in middle and bottom layer was good for whole season. Whereas, in upper layer in storm season, correlation was not significant since intrusion of storm water flowed into beneath middle water layer in the reservoir. Over 90% of total DOC loading attributed to allochthonous sources in upstream and it concentrated in storm season. While, autochthonous DOC by production of algae in Daecheong reservoir appeared only in short period of algal bloom. And the proportion of autochthonous DOC to total organic carbon loading varied significantly depending on spatial locations but was not significant. The extent of SUVA in lake water depth was severely deteriorated by inflow rate and the location of intrusion of storm flows into lake layer during storm event because density of inflow storm water is usually heavier than that of the surface water. It has been found that SUVA in Daecheong reservoir strongly attributed to allochthonous DOC from upstream watershed. SUVA was in the range of 1.465-2.696 through the whole period and that of middle and lower layers in wet season were 2.696 and 2.686 respectively. REFERENCES Aki, S. and T. Shirasuna, (1973). Investigation and analysis of turbid flows in reservoir, Part 1, Technical Report 73518, Civil Engineering Research Institute, Central Research Institute of Electric Power Industry, Japan. Kim, B. C., Choi, K.S., Kim, C. G., Lee, U. H. and Kim, Y. H. (2000). Effects of the summer monsoon on the distribution and loading of organic carbon in a deep reservoir, Lake Soyang, Korea. Wat. Res. 34(14), 3495-3504. Bruenig, E. F. (1987). The forest ecosystem : Tropical and boreal. Ambio, 16, 68-79. Cole, J. J., W. H. McDowell & G. E. Likens. (1984). Sources and molecular weight of ‘dissolved’ organic carbon in an oligotrophic lake. Likos, 42, 1-9. Danielsson, L. G. (1982). On the use of filters in distinguishing between dissolved and particulate fractions in natural waters. Water Res., 16, 179-182. Ford, D. E. (1990). Reservoir transport processes. In Reservoir Limnology : Ecological Perspectives, K. W. Thornton (ed.), John Wiley & Sons, Inc., pp 15-41. Galapate, R. P., Kitanaka, A., Itao K., Mukai, T., Shoto, E. & Okada, M. (1997). Origin of trihalomethane (THM) precusors in Kurose river, Hiroshima, Japan. Wat. Sci. Tech. 35(8), 15-20. Li, C.W., Gregory V. Korshin and Mark M. Benjamin, (1998). Monitoring DBP formation with differential UV spectroscopy. Jour. AWWA, 90(8), 88-100. Senesi, N. and Miano, T. M. (1994). Humic Substances in the Global Environment : Implications for Human Health. Elsevier, Amsterdam. Thurman E. M. and Malcolm R. L. (1981). Preparative Isolation of Aquatic Humic Substances. Environ. Sci. Technol., 15(4), 463-466. Wetzel, R. G. (1983). Limnolgy. W. B. Saunders College Publ., Philadelphia. - 117 - . season6.60 -11 .020.8 518 Dry season1.67-0.240. 811 2 Bottom layerWet season6.49 -10 . 010 .9 518 DOCUVA 254 Dry season1.63-0.380.9 212 All data at Jannge site0.8 015 Wet. 2000 19 98 y = -0.0 014 x + 7.967 R = 0 .12 y = 27.951x + 8.5395 R = 0.42 0 20 40 60 80 10 0 050 010 0 015 002000 POC(ug/L) Chl-a(ug/L) wet season dry season - 11 5