Journal Search Engine
Search Advanced Search Adode Reader(link)
Download PDF Export Citaion korean bibliography PMC previewer
ISSN : 1229-3431(Print)
ISSN : 2287-3341(Online)
Journal of the Korean Society of Marine Environment and Safety Vol.21 No.4 pp.327-338

Spatial Variability of in situ and GOCI and MODIS Chlorophyll and CDOM in Summer at the East Sea

Mi-Ok Park*,Woo-Chul Shin**,Young-Baek Son***†,Tae-Geun Noh****
*Department of oceanography, Pukyong National University, Busan 48513, Korea
**BLUGEN KOREA 107, Busan, Korea
***The Jeju International Marine Science Center for Research & Education, KIOST, Jeju 63349, Korea
****Ocean Observation & Information Section, KIOST, Ansan 15627, Korea
Corresponding Author :>, 064-798-6071
July 14, 2015 July 30, 2015 August 27, 2015


Because of impact on the underwater light field, CDOM can influence the accuracy of global satellite-based measurement of ocean chlorophyll and primary productivity. So we investigated the distribution and seasonal variation of CDOM in the East Sea during summer 2009 and 2011. Among them we report two distinctively different summer cases between 2009 and 2011 year, in which showed the different main sources for CDOM. Regulating factors and sources of CDOM in the East Sea were examined. Comparison between in situ and satellite derived Chl a and CDOM were made to find an influence of CDOM on measurement of satellite derived Chl a. Similar pattern and matching of MODIS Chl a with in situ Chl a 2009 was comparable, but significant discrepancy between MODIS Chl a and in situ Chl a was found, when CDOM was high in summer of 2011. GOCI data showed better matching with in situ data for both Chl a and CDOM, compared to MODIS data in summer of 2011. The presence of high CDOM at the surface layer supplied by vertical mixing seems to affect on the overestimation of Chl a by satellite data.

여름철 동해의 현장측정치와 GOCI와 MODIS 위성 자료로 측정한 엽록소와 유색용존유기물의 공간 변동성

*국립 부경대학교 해양학과
**블루젠 코리아
***한국해양과학기술원 제주국제해양과학연구 지원센터
****한국해양과학기술원 해양관측정보과


해수 중 광학적 특성 때문에 유색 용존유기물은 위성자료에 기반한 해양의 엽록소와 일차생산력의 정확한 측정에 영향을 미칠 수 있다. 따라서 본 연구에서는 2009과 2011년 여름철을 대상으로 서로 특이하게 다른 결과를 보고하고자 하였다. 이 두 시기의 여름철 차이는 것은 용존유기물에 영향을 주는 주요 공급원이 다른 것으로 나타났다. 그리고 현장 측정치와 위성자료로 부터 얻어진 엽록소 농도를 비교하여 위성자료로부터 구한 엽록소 농도 측정에 대한 용존유기물의 영향을 보았다. 그 결과, 2009년 MODIS를 이용한 엽록소 농도와 현장 측정된 엽록소 a 농도는 서로 유사하였으나, 2011년과 같이 유색용존유기물의 농도가 높았던 시기에는 이 두 농도 간에 유의한 차이가 나타났다. 2011년 여름 MODIS 자료와 비교하였을 때, GOCI 자료는 엽록소와 유색용존유기물 모두 현장 측정치 자 료와 잘 일치하였다. 수직 혼합에 의해 공급된 표층 해수 중 높은 유색용존유기물의 존재는 위성자료에 의한 엽록소 농도의 과대평가 에 영향을 주는 것으로 보인다.

    Pukyong National University


    The terrestrial CDOM (Chromophoric Dissolved Organic Matter) input impacts the DOM (Dissolved Organic Matter) dynamics in the sea. Kowalczuk et al.(2001; 2005) reported that annual CDOM cycling was determined by the annual maximum riverine discharge into the southern Baltic. Although the riverine source of CDOM is the most important, it is not the only factor to affect CDOM pools. in situ production of CDOM is also significant in remote area which is not directly influenced by riverine plumes (Kowalczuk, 1999; 2001). These constituents cause an increase in the absorption of blue bands and influence the spectral ratio of reflectance, which in turn, made it difficult to provide precise evaluations of Chlorophyll-a (Chl a). Therefore, it is necessary to discriminate different sources of CDOM in the surface layer, in order to understand the CDOM dynamics in the coastal region and open ocean. Recenlty the photobleaching, photohhumification and microbial degradation of CDOM was known to either reduce the CDOM or transform in a short time scale (~ days) and the dynamics and source of CDOM become a very active research field, especially in high latitude region related with climate sensitive change (Ortega-Retuerta et al., 2010).

    The East Sea has unique optical properties because of a very low input of freshwater from the land area and limited water exchange with the North Pacific through the Tsugaru Straits and Korea Strait. However, in summer season local upwelling phenomena are frequently observed (Yoo and Park, 2009; Yoo et al., 2000), which is caused by persistent southwesterlies on the surface the East Sea. As a result the enhanced Chl a patches are observed from satellite images near Gampo area in every summer (Yoo and Park, 2009). The river run-off and upwelling phenomena in summer affects on the abundance of phytoplankton and primary production in the East Sea.

    In this study, we examined the temporal and spatial variation of CDOM and Chl a in the East Sea and compared the distribution pattern of in situ measured and satellite derived CDOM and Chl a during summer of 2009 and 2011. The seasonal variation of CDOM and distribution pattern was also confirmed by in situ measurement and compared with multi-year satellite data (moderate resolution imaging spectroradiometer (MODIS) and geostationary ocean color imager (GOCI)) in the study area. To discriminate the difference source of CDOM and find the controlling factors to affect the CDOM, we used environmental data (sea surface temperature (SST) and salinity) to find their effects on the distribution of the CDOM, i.e., dilution process of terrestrial CDOM and mixing event by local upwelling event. The study area has known that the riverine discharge is very low and insignificant as source of DOM to the coastal water because the rivers along the east coast of Korea are relatively small in size, except Nakdong river. However we find that in summer the riverine discharge can be a significant source to the coastal water in the study area. The correlation of CDOM with salinity and Chl a was examined to evaluate a contribution from terrestrial sources and in situ production of CDOM from biological activity

    2.Material and Methods


    Location of stations, where seawater samples were collected and hydrographic conditions were measured in summer of 2009 (August 8-18) and 2011 (July 25-26) is shown in Fig. 1. Seawater samples were collected at 46 stations of 8 transect by Haeyang 2000 (Korea Hydrographic and Oceanographic Administration, KHOA) 2009 (August 8-18) and 21 stations of 4 transect on the southwestern East Sea survey cruises by R/V Tam-Yang (PKNU) (July 25-26) in the East Sea. The seawater samples were collected from Niskin bottles mounted on a conductivity-temperature-depth (CTD) rosette sampler at seven depths (0, 10, 20, 30, 50, 70, 100 m) and analyzed for CDOM, FDOM (Fluorescence Dissolved Organic Matter) and Chl a. Temperature and salinity were measured with Sea Bird 911 CTD.


    Seawater samples for analysis of CDOM were filtered under a gentle vacuum (<5 in Hg) through pre-combusted (6 h at 450°C) Whatman GF/F glass fiber filters (diameter: 25 mm) and collected directly into pre-cleaned and pre-combusted sample glass bottles (125 mL Clear narrow neck bottle). Until analysis, sample bottles were stored in a freezer (4°C) after covered with foil. Filtering set was conducted following the method of Mitchell et al.(2002). Absorbance spectra of CDOM were measured using a double-beam Cary 100 spectrophotometer (Varian) with cylindrical 100 mm path length cells with Milli-Q water as blank (Mitchell et al., 2002). Instrument scan settings were as follows: 200-800 nm wavelength scan range, 1 nm data interval, 100 nm min-1 scan rate. The data were corrected for scattering and baseline fluctuations by subtracting the absorption 750 nm (Green and Blough, 1994). The absorption coefficients were calculated from the following expression (Mitchell et al., 2002):

    aCDOM(λ) = 2.303A(λ) × b-1

    where A(λ) is the absorbance of filtered seawater at a specific wavelength measured across path length b (m). Instrument performance tests (wavelength accuracy and reproducibility, photometric noise, and baseline flatness) were conducted each day prior to analysis.

    2.3.Chl a

    Chl a samples were filtered under a gentle vacuum (< 5 in Hg) through Whatman GF/F glass fiber filters (diameter: 47 mm), the filters were put into vials after covered with foil. Until analysis, sample vials were stored in a deep-freezer (< -80°C) or in a liquid nitrogen. After removal from liquid nitrogen or freezer, the pigments were extracted by placing the filter in 10 mL of 90 % Acetone. The samples are covered with cab to reduce evaporation, allowed to extract for 24 hr in the freezer. After 24 hrs the samples were centrifuged for 10 minutes at 3,000 rpm. Remove tubes carefully, so as not to re-suspend particulates, and place tube in rack check to be sure particulates are settled out. Chl a was measured by using 10-AU Fluorometer (Turner Designs, Sunnyvale California) and 5 mL cell with 90 % acetone as blank (Jeffery and Humphrey, 1975). After the first measurement (Rb) of a sample, 2~3 drop 1 N HCl was added in the tube and mixed. Wait 2 minutes and measure again (Ra). Chl a was calculated from the following equation:

    Chlorophyll-a (gL-1) = Fd × r·(r-1)-1× (Rb-Ra) × Va·Vs-1

    Fd : Calibration coefficient

    (g Chl a per mL 90 % acetone per instrument fluorescence units

    r : maximum acid ratio (Rb/Ra) of pure chlorophyll a standard.

    Rb : sample fluorescence before acidification.

    Ra : sample fluorescence after acidification.

    Va : extraction volume in mL.

    Vs : filtered volume in L.

    2.4.Satellite data

    To estimate absorption of CDOM and compare between in situ CDOM and satellite derived CDOM, daily MODIS and GOCI satellite images were obtained from Ocean Color Web ( and GOCI ( Level 1 data recorded at the top of the atmosphere (TOA) were atmospherically corrected and processed to Level 2 data by use of SeaWiFS Data Analysis System (SeaDAS) and GOCI Data Processing System (GDPS). Remote-sensing reflectance (Rrs) was derived from MODIS and GOCI level 2 data with a spatial resolution of ~1 km/pixel (MODIS) and ~500 m/pixel (GOCI) at nadir. A match-up dataset was generated between in situ CDOM and MODIS and GOCI Rrs(λ) synchronously obtained. To obtain reasonable match-up data, Son et al.(2009) suggested that the matching of spectral data with our in situ sampling data was limited with satellite data collected within ±3 hr window. For additional condition of satellite data, 3 × 3 pixel grid was used to satisfying the requirement for reasonable balance between the geophysical homogeneity of the sampling matrix and number of clear pixels in the spatial window.

    However, 355 nm band (ultra violet) is not incorporated in the MODIS and GOCI sensor, which had only blue, green, red and near-infrared bands. To calculate the satellite-derived CDOM absorption 355 nm (aCDOM(355)), Quasi-Analytical Algorithm (QAA) (Lee et al., 2009) were applied using MODIS and GOCI Rrs(λ). aCDOM(443) is determined by QAA method and then converted to aCDOM(355) using the below equation:

    aCDOM(355) = aCDOM(443)e-s(355 - 443)

    where S is the exponential slope parameter (Lee et al., 2005; 2009). The composite images are made from daily data for 8 ~ 11 days (August 8 ~ 18, 2009 and July 22 ~ 29, 2011).


    3.1.Oceanographic conditions in summer of 2009 and 2011

    Surface distributions of oceanographic conditions in the study area are shown in Fig. 2. SST was measured as 22.82 ~ 26.15°C in August, 2009 and 20.05 ~ 25.12°C in July, 2011. The salinity was 30.73 ~ 34.61 psu in August 2009 and 32.57 ~ 33.63 psu in July, 2011 (Table 1). The horizontal distribution pattern of salinity in 2009 (Fig. 2) showed a low saline water (< 32 psu), which spread to eastward offshore and high saline water are at the offshore to northeastern direction.

    In summer of 2011, the high saline water was near 36.5 °N and 130 °E, but in general the horizontal distribution of salinity was relatively homogeneous and the average value was about 33 psu. Differences in SST (ΔT) were higher in 2011 compared to the summer of 2009, which seems related with upwelling phenomena near Gampo (Table 1). However difference in salinity (ΔS) of surface water was higher in summer of 2009, i.e., 3.88 psu compared to 1.06 psu in summer of 2011. The salinity at the surface layer was in most stations less than 34 psu, which implies that the surface layer in summer of 2009 and 2011 was under the influence of fresh water input.

    3.2.Comparison of aCDOM in summer 2009 and 2011

    The horizontal distribution of in situ aCDOM(355), MODIS and GOCI image are shown in Fig. 3-6. Two high aCDOM(355) patch were observed from MODIS and GOCI image at the coastal region centered Gampo and eastward to offshore at 37.5 °N (Fig. 3-6). The horizontal distribution of Chl a and aCDOM(355) by MODIS data was very similar with distribution of in situ Chl a and aCDOM(355). Two high-concentration Chl a patches near Ulsan was observed at 35.5°N with (> 3 μgL-1). However the distribution of in situ Chl a distribution was different with distribution of in situ aCDOM(355). At offshore, in situ Chl a was very low (< 0.3 μgL-1) which are well agree with MODIS data. The high aCDOM(355) patch at 38 °N, 130 °E was not detected on MODIS data.

    In summer of 2009 (Fig. 3 and 4), distribution pattern of in situ aCDOM(355) was similar with MODIS data except the high CDOM patch located at offshore. According to the MODIS data, the high CDOM was confined on the coastal region of the shelf area, but the in situ measurement of aCDOM showed two additional high CDOM patches at offshore at 35.5 °N - 132 °E and 38 °N - 130.5 °E. Relatively low aCDOM(355) was observed at the center or the study area (< 0.14 m-1). The discrepancy of aCDOM(355) between in situ and MODIS data was great, especially where CDOM values were low. in situ aCDOM(355) in summer of 2009 was 0.1971, and MODIS aCDOM(355) was 0.1202, which is about 60 % of the in situ aCDOM(355). Distribution patterns of in situ aCDOM(355) was compared with aCDOM(412) and aCDOM(443) and the differences were minimal (Fig. 7).

    In summer of 2011 (Fig. 5 and 6), the horizontal distribution of in situ aCDOM(355) was compared with MODIS and GOCI data with their satellite images. First of all, the MODIS data were very scarce because of cloud and fog in the study area. On the other hand, GOCI images provided more information on CDOM compared to MODIS data, though there is a clear discontinuity between coverage. The high CDOM patch from near-shore of Ulsan is shown in in situ measurement and both satellite data. Although MODIS data in general very few because of screening by cloud, the high aCDOM(355) from Ulsan was clearly shown with very clear boundary. The GOCI satellite images and data provided whole high aCDOM(355) patch from Ulsan to northeastward direction.

    The ratio aCDOM(355) between MODIS/GOCI was 0.77, which means underestimation of aCDOM(355) by MODIS by 30 %. The comparison with in situ data gives 0.45 for aCDOM(355) (MODIS/in situ) and 0.50 for aCDOM(355) (GOCI/in situ). Direct matching with in situ data with GOCI and MODIS data revealed that satellite derived aCDOM(355) are significantly underestimated compared in situ CDOM up to 55 %. This underestimation was clear at offshore (Table 2).

    3.3.Seasonal variation of aCDOM

    In order to know the seasonal variation of CDOM in the southwest East/Japan Sea, four cruises were made on spring (May), summer (July), fall (September), 2011 and winter (January), 2012 (Table 4). The maximum aCDOM(355) was measured on spring and minimum aCDOM(355) was measured on winter. The satellite data for multi year showed that aCDOM(355) was minimum in summer and maximum in spring (personal comm.). The cruise in May (spring) 2011 was limited only in the northern part (D, E, F transect) because of bad weather situation. So we compared only with data from D transect (Fig. 8), which are common on the four seasons. With average value, the aCDOM(355), aCDOM(412), aCDOM(443) are the same order of spring > fall > summer > winter.

    As shown in Table 3, maximum aCDOM(m-1) was measured in spring and minimum in winter. It is also confirmed by GOCI multi-year data, with maximum aCDOM(412) in spring bloom (April) and second peak in fall (November), which coincides the phytoplankton bloom period and minimum aCDOM(412) was found in summer. However our measurement of aCDOM(355) reveals that aCDOM(355) was lowest in winter instead of summer season. In many studies reported (Hansell and Carson, 2002; Ortega-Retuerta et al., 2010) that seasonal variations of aCDOM(355) showed the minimum of aCDOM(355) in the summer because of the photobleaching at the sea surface layer.

    In Table 3, the maximum in situ aCDOM(355) (0.4205) in 2011 was well match with MODIS data (0.4306) than GOCI (0.2336). However, the average values showed highly underestimated aCDOM(355) by satellite-derived data for both GOCI and MODIS data. The deviation of satellite derived aCDOM(355) from in situ aCDOM(355) was great in lower CDOM concentration. The in situ aCDOM(355) in summer of 2009 was 0.1971, and MODIS aCDOM(355) was 0.1202, which is about 60 % of the in situ aCDOM(355).

    Correlation between aCDOM(355) and Chl a showed positive correlation for both 2009 (R2 = 0.6401) and 2011 (R2 = 0.6566), but at offshore no clear correlation was found (Fig. 9). FDOM distribution also showed very similar spatial pattern with Chl a (Fig. 10).

    In summer of 2009, CDOM showed conservative mixing along the salinity decrease from near-shore to offshore (Fig. 11a). However, the correlation in summer 2011 was different. No close correlation was observed at coastal water and weak inverse relationship with salinity at offshore water, which suggests more saline water has lower CDOM. The higher aCDOM(355) in 2011 was measured than in 2009. Salinity was lower in 2009 which means more influence of freshwater input (Table 1).

    Good correlation between FDOM and Chl a (Fig. 12) was observed at the coastal water for both 2009 and 2011 (R2 > 0.78) and weak positive correlation at offshore (R2 = 0.25 ~ 0.51). For FDOM, Excitation Emission Matrices(EEMs) of the C-peak is known to be originated from humic-like chromophores and it was in good correlation with Chl a. It suggests that fresh riverine water from the coastal region, with high FDOM, CDOM and nutrients supply promoted the biological activity in coastal water and resulted in the high Chl a.

    Finally aCDOM(355) from GOCI and MODIS were compared each other and also with in situ data (Fig. 13). There was weak correlation between in situ aCDOM(355) and satellite-derived aCDOM(355) at the coastal water in summer of 2009, but there was no clear correlation in offshore in 2009 and 2011 was observed(Fig. 13).

    Thus, the accurate estimation of Chl a during summer in the study area, will be highly interfered by light absorption by CDOM especially when the vertically mixing provides the aCDOM from the subsurface, because of relatively high CDOM and low Chl a in summer. Chl a values were highly overestimated by MODIS satellite data up to 340 %, on a while GOCI showed a better correlation than MODIS(Fig. 14). Still Chl a values were slightly overestimated by GOCI satellite data.


    In summer of 2009, relatively warm and less saline water patch (< 32 psu) was observed at the tip off Gampo area and area with high Chl a coincided with the low saline surface water (Fig. 2-4). These facts suggest that high concentrations of Chl a was either originated from terrestrial source with high abundances of phytoplankton, or the nutrients supplied by the river runoff promoted the growth of phytoplankton. CDOM (R2 = 0.6481) as well as Chl a showed an inverse correlation with salinity (Fig. 15). The increase in SST in summer (August) would warm the surface water and stratification causes a depletion of nutrients at the surface water. In this conditions supply of the nutrients from terrestrial source could support the growth of phytoplankton and increase in Chl a might resulted in.

    The higher temperature (24.29℃) and lower salinity (32.12 psu) as average values in summer of 2009 showed that the surface water is warmer and lighter than in summer of 2011, which implies the stratified water at the surface layer. Not only the higher SST and lower salinity in average, but also the lower in CDOM and higher concentrations of Chl a were measured in summer of 2009 compared to those in summer of 2011. The range of aCDOM(355) in 2009 and 2011 was different in that background aCDOM(355) in 2009 was much lower than in 2011 (Table 3). The average aCDOM(355) in August of 2009 was 0.1971 and is also lower than 0.2876 in 2011 This low aCDOM(355) in summer of 2009 might be resulted from the photobleaching of CDOM at the stratified surface water.

    In general coastal area, where receives a high river input exhibit high levels of CDOM absorption. For coastal waters strongly influenced by river input, CDOM absorption usually dominates the total light absorption, not only in the UV-B and UV-A but also portion of the visible spectrum, where phytoplankton also absorbs. As De Grandepre et al.(1996) found that the aCDOM(442) was 2 to 3 fold greater than the particulate absorption coefficient ap(442) (which includes phytoplankton absorption), during August on the shelf of the US middle Atlantic Bight. According to their results, during a fall bloom in November did ap(442) exceed aCDOM(442) and still the aCDOM(442) represented a substantial portion of the total absorption. As De Grandepre et al.(1996) pointed out the absorption by aCDOM(442) is significant especially during summer season, because of low ap and high riverine input from land.

    The correlation between Chl a and CDOM in summer of 2011 was quite similar with the case with 2009 (Fig. 11). At coastal region, CDOM and Chl a showed a good correlation (0.6566) for aCDOM(355), but no significant correlation was found at offshore region. However the correlation between CDOM and salinity in summer of 2011 was not clear, in contrast with the case which showed linear inverse relationship in summer of 2009 (Fig. 11). In particular, at transect of C line, vertical profile of aCDOM(355) showed high CDOM extended up to 100 m at station C4 and station C5 (Fig. 15). As Son et al.(2011) pointed out that increased vertical mixing can optically increase CDOM and detrial absorption at shorter wavelengths, producing false satellite high Chl a water as noise. The vertical mixing phenomena of CDOM at station C4 might be the source of the high evaluation of Chl a.

    The water column in the center of the study area was relatively cool (< 22°C) and has higher salinity(> 33 psu) than the surrounding water in summer of 2011 (Fig. 2). The salinity at the surface layer is almost homogeneous, but the temperature difference was rather high (5°C). And also, the vertical mixing was active in this cold water region (Fig. 10). The CDOM was high at the cold water and at the coastal region (Fig. 5). In average, aCDOM(355) was 1.5 times higher in summer of 2011 compared to those in 2009. The maximum aCDOM was similar but the lowest aCDOM(355) was much lower in August 2009 than July 2011 (Table 3).

    Possible reasons for the high CDOM in July 2011 and high level of background aCDOM are (1) decrease in CDOM in August, 2009 due to photobleaching by increased stratification under stronger insolation (2) additional input of CDOM from subsurface by vertical mixing in summer 2011. In contrast high CDOM in summer of 2011, concentrations of Chl a was very low compared to those of 2009 (Table 4). This can be attributed to the water column stability and optical properties, such as turbulent water masses and less PAR in summer of 2011.

    This area is known to have frequent upwelling by persistent winds from southeast. The vertical mixing in the center of the study area might be restricted from upwelling phenomenon. Surface salinity was low within shelf and nearshore (< 33 psu) increasing slightly to offshore direction. The near-shore shelf waters had low salinity due to inputs of freshwater from Hyeongsan River at 36 °N and Wangpi streamlet at 37 °N in summer 2009. However the salinity in summer 2011 was rather uniform except relatively cool (< 23°C) and saline water at the center of the study area. The CDOM showed a conservation mixing in summer of 2011. Salinity and DOC values also fall closely a mixing line between coastal water and offshore water.

    The cruise in July 2011, corresponded to the period of frequent local upwelling which induces vertical mixing and cruise in August, 2009 corresponded to a period of stratification. Thus the two of cruises represents different situation for the physical environments in summer season. The two different cases of the spatial and temporal distribution of CDOM in summer season gave the opportunity to understand the controlling factors of distribution and sources of CDOM in the study area.

    To understand and predict CDOM distributions, the major sources of CDOM and their relative importance for contribuiton must first be determined. It is unclear to what extent coastal and oceanic CDOM in remnant of diluted freshwater input from the terrestrial biosphere, is the product of in situ biological process (Høierslev 1982). The accuracy of these algorithm in accounting for the effects of CDOM in upper ocean optics is limited by our knowledge of CDOM dynamics in coastal region (Twardowski and Donaghay, 2001; Twardowski et al., 1999).

    The linear inverse correlation (R2 = 0.4641) between CDOM and salinity signature of the freshwater and oceanic members implies the CDOM is behaving conservatively at the coastal region in summer of 2009 (Fig. 12). Little correlation between in situ CDOM and salinity in 2011 (Fig. 12) was observed at the offshore (R2 = 0.2351). Export of dissolved and particulate OM from continental shelves is thought to be an important contribution to the biogeochemical cycling of carbon and nitrogen in the open ocean (Walsh, 1991). Mixing of organic rich shelf waters with offshore water is a mechanism for transferring carbon to open ocean.

    Seasonal variation was clearly shown by in situ measurement of aCDOM(355) from 4 different cruise data set (Table 4). Although the seasonal variation from satellite data during 2003 ~ 2010 in the southwestern East Sea showed the minimum level in summer (August) and high in spring (personal comm.), our data showed the minimum aCDOM(355) in winter. This might be related with the region of our study is more coastal area and the satellite data was collected the whole southwestern East Sea, which is more oligotrophic offshore area. That means the supply from the terrigeneous OM is very scarce. The range was 0.2266 ~ 0.3095 for average aCDOM(355) and 0.1096 ~ 0.5511 for individual measurement of aCDOM(355). The level was relatively low for coastal region. Although the seasonal variation was evident, and the discrepancy in aCDOM(355) was not great.

    During winter, the CDOM decreased overall. It seems that not only the lower biological activity and decease in terrestrial input from river run-off , but also increase in the mixed depth by the vertical mixing by the loss of density stratification affect on the low aCDOM(355) in winter time (Table 4). By the average, the highest in situ CDOM was in the spring and the lowest in the winter. In summer, CDOM in the surface layer was lower than spring and fall. The highest CDOM was measured in spring for all wavelengths (355, 412, 443 nm). The multi-annual satellite data showed (personal commun.) maximum CDOM in spring (April) and lowest CDOM in summer (August) for the southwestern East Sea. The discrepancy might be resulted from the input from land and upwelling input of CDOM by vertical mixing in our study which showed minimum CDOM in winter. Since our sampling in summer was July 2011, the extended exposure of CDOM in the surface water under solar radiation during mid summer of August could reduce the CDOM significantly by photobleaching.

    In summer of 2011, low SST and high saline water at the center of the study area showed high Chl a, CDOM and low DOC, which implies high biomass of phytoplankton with high CDOM. The vertical structure of nutrient (Oh et al. in preparation) and CDOM showed the mixing up to 100 m depth in this low SST patch at the center of the study area. CDOM, in general is very low in summer because of photobleaching of the surface layer. However CDOM is comparable to fall season (> 0.2), especially at this low SST patch. Fig. 10 shows the vertical distribution of aCDOM(355) for transect A and C.

    It is regarded that high CDOM at the surface large is provided by vertical mixing from subsurface with high CDOM . Nutrient provided from the vertical mixing enables the growth or phytoplankton. The difference of SST is not large enough (~ 5°C) to be considered as an upwelling, but nutrients and CDOM showed surface maxima at station C4 (Fig. 15). Comparison between station C4 and other stations of C transect showed different distribution patterns. At offshore aCDOM was high at 20m (> 0.5 m-1) and 70 m respectively, which is different from C4, with the surface maximum. However the surface maximum values is lower than (~ 0.4 m-1) the subsurface maximum CDOM (~ 0.6 m-1). This seems dilution by mixing (CDOM is high at 20 m and 75 m at the most transects).

    The subsurface maximum layer in A transect are formed at depth 40~50 m and relatively well-stratified structure with deepening toward offshore direction. But transect C showed dramatically different distribution from A transect. The vertical mixing at station C4 destroyed the stratification of the water column. Especially at station C4 and station C5 the high aCDOM was extended to 100 m depth.

    This vertical mixing of water column provided a high background level of CDOM at the surface layer in July 2011, and as a result, the minimum level of CDOM in summer of 2011 was higher than in summer of 2009. The weak correlation with salinity might be related with this high saline water from the subsurface layer. The upwelling phenomena in this region provided an additional input of nutrients and in turn, enhance the primary production near Gampo area. The fogs are used to be formed because of high temperature difference between the air and the cold upwelled surface water (> 10°C ~ 15°C difference). Although the observed difference SST in 2011 is not big enough to point out as local upwelling water, vertical structure of CDOM and nutrients provides the evidence for the vertical mixing up to 100 m and upwelling.

    In 2009 summer, the surface water was stratified and high temperature and low salinity water was observed at near-shore. High difference in salinity (3.88 psu) was observed by increase in discharge from the Hyeongsan River and Wangpi streamlet, and inverse relationship between (Chl a, CDOM) and salinity was observed. According to the inverse relationship between salinity and CDOM, the conservative mixing process is confirmed. The close correlation can be resulted from nutrient supply by fresh water, which can enhance the phytoplankton growth. However the contribution from phytoplankton to CDOM cannot be excluded, because DOC vs salinity didn’t show a significant correlation (R2 = 0.0108) (personal commun.). The autochthonous CDOM production is dependent of three primary processes occurring consecutively. The first is the fixation of carbon in photosynthesis, the second is the production of extracellular DOM, and the third is CDOM formation from DOM precursors. Yentsch and Reichert(1961) observed the rapid formation of colored humic–like substances in the laboratory from algal cell constituents and Sieburth and Jensen (1970) from incubating algal exudate.

    In this study, two possible different sources of CDOM for the spatial variation during 2009 and 2011 seem to present : in summer of 2009, it is the input of the fresh water loaded with high in DOM. This extra source of CDOM affected the level of CDOM in surface layer. On the other had in summer of 2011, local upwelling increase the background concentrations of CDOM through the injection of CDOM from subsurface layer.



    Map of sampling locations in the East Sea. (a) August, 2009; (b) July, 2011.


    Horizontal distributions of temperature (°C) and salinity (psu). (a) and (b) August, 2009; (c) and (d) July, 2011.


    Horizontal distributions of aCDOM(355) (m-1) in situ (a), satellite-derived data derived from MODIS (b), and MODIS satellite image (c) in August, 2009.


    Horizontal distributions of Chl a (μgL-1) in situ (a), MODIS (b) and GOCI satellite images (c) in July, 2011.


    Horizontal distributions of Chl a (μgL-1) in situ (a), satellite-derived data derived from MODIS (b), and MODIS satellite image (c) in August, 2009.


    Horizontal distributions of CDOM (m-1) in the study area. (a) ~ (c) August, 2009; (d) ~ (f) July, 2011.


    Horizontal distributions of aCDOM(355) (m-1) in situ (a), MODIS (b) and GOCI satellite images (c) in July, 2011.


    Horizontal distributions of Chl a (μgL-1) and FDOM (C peak: Ex/Em=345/440 nm) in the study area. (a), and (b) August, 2009; (c) and (d) July, 2011.


    Correlations with CDOM and Chl a at the coastal and offshore. (a) and (b) August, 2009; (c) and (d) July, 2011.


    Horizontal distributions of Chl a (μgL-1) and FDOM (C peak: Ex/Em = 345/440 nm) in the study area. (a), and (b) August, 2009; (c) and (d) July, 2011.


    Correlations with CDOM and Salinity at the coastal and offshore. (a) and (b) August, 2009; (c) and (d) July, 2011.


    Correlations with FDOM and Chl a at the coastal and offshore. (a) and (b) August, 2009; (c) and (d) July, 2011.


    Correlations with aCDOM(355)in situ and aCDOM(355)sattllite at the coastal and offshore. (a) and (b) August, 2009; (c) and (d) July, 2011.


    Correlations with Chl ain situ and Chl aMODIS, Chl aGOCI at the coastal and offshore. (a) and (b) August, 2009; (c) and (d) July, 2011.


    Vertical distributions of aCDOM(355) in the A and C transect of the study area(July, 2011).


    Oceanographic conditions in the study area

    Comparisons of aCDOM(355) in situ and satellite-derived data in the study area. MODIS data is composite data 8th-18th August, 2009 by QAA(Quasi-Analytical Algorithm) algorithm. GOCI data is composite data 22nd-29th July, 2011 by QAA algorithm

    *aCDOM(355) : average of 355 15nm
    **aCDOM(355)MODIS and aCDOM(355)GOCI were estimated

    Seasonal variation of CDOM in the study area

    Comparisons of Chl ain situ and satellite-derived data in the study area. MODIS data is composite data 8th-18th August, 2009 by MODIS standard chlorophyll algorithm. GOCI data is composite data 22nd-29th July, 2011 by GOCI chlorophyll algorithm


    1. De Grandpre MD , Vodacek A , Nelson RK , Blough N (1996) Seasonal seawater optical properties of the US middle Atlantic Bight , Journal of Geophysical Research, Vol.101 (C10) ; pp.22727-22736
    2. Green SA , Blough NV (1994) Optical absorption and fluorescence properties of chromophoric dissolved organic matter in natural water , Limnology and Oceanography, Vol, Vol.39 ; pp.1903-1916
    3. Hansell DA , Carlson CA (2002) Biogeochemistry of Marine Dissolved Organic Matter, Academic Press, ; pp.509-601
    4. Kowalczuk P , Ston-Egiert J , Cooper WJ , Whitehead RF , Durako MJ (2005) Characterization of chromophoric dissolved organic matter (CDOM) in the Baltic Sea by excitation-emission matrix fluorescence spectroscopy , Marine Chemistry, Vol.96 ; pp.273-292
    5. Kowalczuk P , Donaghay PL (2001) Separating in situ and terrigenous sources of absorption by dissolved materials in coastal waters , Journal of Geophysical Research, Vol, Vol.106 ; pp.2545-2560
    6. Lee ZP , Du KP , Arnone R (2005) A model for the diffuse attenuation coefficient of downwelling irradiance , Journal of Geophysical Research, Vol, Vol.110 ; pp.C02016
    7. Lee ZP , Lubac B , Werdell J , Arnone R (2009) An update of the quasi-analytical algorithm (QAA_v5) ,,
    8. Mitchell BG , Kahru M , Wieland J , Stramska M (2002) Determination of spectral absorption coefficients of particles, dissolved material and phytoplankton for discrete water samples , NASA Technical Memorandum 2002-21004 Rev3 Vol, Vol.2 ; pp.231-257
    9. Ortega-Retuerta E , Reche I , P Villena E , Agusti S , Duarte CM (2010) Distribution and photoreactivity of chromophoric dissolved organic matter in the Antarctic Peninsula , Marine Chemistry, Vol.118 ; pp.129-139
    10. Sieburth JM , Jensen A (1970) Production and transformation of extracellular organic matter from littoral marine algae: a resume, in Organic Matter in Natural Water , College Alaska, Alaska, ; pp.203-223
    11. Son YB , Ishizaka J , Jeong JC , Kim HC , Lee TH (2011) Cochlodinium polykrikoides Red Tide Detection in the South Sea of Korea using Spectral Classification of MODIS Data , Ocean Science Journal, Vol.46 (4) ; pp.239-263
    12. Son YB , Gardner WD , Mishonov AV , Richardson MJ (2009) Multispectral remote-sensing algorithms for particulate organic carbon (POC): The Gulf of Mexico , Remote Sensing of Environment, Vol.113 ; pp.50-61
    13. Twardowski MS , Donaghay PL (2001) Separating in situ and terrigenous sources of absorption by dissolved materials in coastal waters , Journal of Geophysical Research, Vol.106 (2) ; pp.2545-2560
    14. Twardowski MS , Sullivan JM , Donaghay PL , Zaneveld JRV (1999) Microscale quantification of the absorption by dissolved and particulate material in coastal waters with an ac-9 , Journal of Atmospheric and Oceanic Technology, Vol, Vol.16 ; pp.691-707
    15. Walsh JJ (1991) Importance of continental margins in the marine biogeochemical cycling of nitrogen , Nature, Vol.350 ; pp.53-55
    16. Yentsch CS , Reichert AC (1961) The interrelationship between water-soluble yellow substances and chloroplastic pigment in marine algae , Botanica Marina, Vol.3 ; pp.65-74
    17. Yoo SJ , Park JS , Kim HC (2000) Validation of Ocean Color Algorithms in the Ulleung Basin, East/Japan Sea , Journal of the Korean Society of Remote Sensing, Vol.16 (4) ; pp.315-325
    18. Yoo SJ , Park JS (2009) Why is the southwest the most productive region of the southwestern East Sea/Sea of Japan , Journal of Marine Systems, Vol.78 ; pp.301-315