Application of multiple biogeochemical techniques for analyzing stream and riverine carbon from forest watersheds in Korea = 다중 생지화학적 기법을 활용한 우리나라 산림 하천탄소 분석

다국어 초록 (Multilingual Abstract)

Forests act as natural carbon sinks by capturing atmospheric CO2 through vegetation and soil processes, after which sequestered carbon in terrestrial ecosystems is returned to the atmosphere via respiration, stored in soils, or exported laterally through river networks. Although riverine carbon export is smaller in magnitude than vertical carbon fluxes and soil carbon inputs, it is substantial enough to influence inland carbon balances, being comparable to net ecosystem exchange (NEE). Moreover, rivers provide a critical linkage between terrestrial and marine carbon cycles. I focused on forest streams and rivers to investigate how plant-derived carbon is mobilized, as reflected in stream water chemistry and dissolved organic matter (DOM) composition, using advanced techniques such as dual carbon isotope analysis, lignin phenol analysis, and molecular-level indices.
First, I investigated stream carbon concentrations and DOM characteristics at Mt. Jeombong, a Korean long-term ecological research (KLTER) site. From May 2019 to October 2024, dissolved inorganic carbon (DIC), dissolved organic carbon (DOC), and particulate organic carbon (POC) concentrations ranged from 1.3 to 9.5 mg L-1, 0.5 to 7.5 mg L-1, and 0.1 to 4.1 mg L-1, respectively. Based on discharge measurements in 2024, carbon export from April to December 2024 totaled 69.2 kg DIC, 20.9 kg DOC, and 11.5 kg POC, equivalent to 31.5, 9.5, and 5.2 kg ha-1 for the 2.2 ha catchment. The NEE from April 20 to December 31, 2024, was -1,423.0 kg-C ha-1. The sink strength shifted to -1,378.2 kg C ha-1 after subtracting DOC, POC, and DIC loads, which corresponds to a 3.1% decrease in the forest carbon sink. Dual-carbon isotope ratios and lignin phenol ratios showed C3 angiosperm plant-dominated signals across baseflow and stormflow, while Fourier transform ion cyclotron resonance mass spectrometry (FT-ICR MS) reflected shifts toward more N-rich compound classes under stormflow conditions.
Secondly, I examined how seven decades of contrasting political and forest management regimes have influenced DOM characteristics between North and South Korea. The study included two transboundary watersheds dominated by North Korean land cover (N1: 3,131 km2; N2: 8 km2) and two watersheds located primarily within South Korea (S1: 2,080 km2) or entirely within South Korean territory (S2: 262 km2). Although DOC concentrations were similar across sites, river Δ14C-DOC varied significantly depending on the contributing watershed: 683 years before present (yr BP) in river N1, 85 yr BP in river S1, and modern carbon signatures in stream N2 and S2. Combined with δ13C-DOC, SUVA254, and molecular characteristics of DOM, the results suggest that forest watersheds across the border differ in DOM sources as a consequence of long-term policy divergence.
Thirdly, I analyzed sixteen forest streams to compare DOM characteristics between coniferous- and deciduous-dominated watersheds. This study applied lignin phenol analysis to Korean forest watersheds for the first time. The syringyl-to-vanillyl (S/V) ratio revealed that watershed-scale dominance of coniferous and deciduous forests is reflected in stream DOM. In contrast, neither SUVA254 nor bulk DOC concentration differed consistently between forest types, and FT-ICR MS analyses indicated only minor effects of dominant tree species on overall DOM composition.
These three chapters collectively illustrate how spatial and ecological variability, tree species, and seasonality affect the composition and export of carbon in forest streams.

국문 초록 (Abstract)

산림은 이산화탄소를 흡수하는 자연 탄소 흡수원으로 기능한다. 그러나, 육상생태계에 흡수된 탄소는 식물이나 토양의 호흡과정에서 다시 대기로 방출될 수 있다. 또한, 하천을 통해 빠져나가는 탄소 유출량은 육상생태계 탄소 균형에 영향을 미칠 만큼 충분히 클 수 있다. 이러한 하천탄소 유출은 육상과 해양을 연결하는 탄소 순환의 핵심 경로로 작용한다. 이 연구는 산림하천과 강을 통해 식물이 대기 중 이산화탄소를 흡수하여 만들어낸 유기물이 어떻게 유출되는지를 이중탄소동위원소비, 리그닌 페놀 지표, FT-ICR MS 분석을 이용해 조사했다.
첫째, 점봉산 장기생태연구지에서 산림하천으로 빠져나가는 탄소농도와 용존유기물의 특성을 조사했다. 2019년 5월부터 2024년 12월까지 점봉산 산림하천에서는 1.3~9.5 mg L-1, 0.5~7.5 mg L-1, 0.1~4.1 mg L-1의 용존무기탄소(DIC), 용존유기탄소(DOC), 입자유기탄소(POC)가 빠져나갔다. 유량을 고려하면, 탄소 유출량은 DIC 69.2 kg, DOC 20.9 kg, POC 11.5 kg로 2024년 4월 20일부터 12월 31일까지 점봉산 순생태계교환량(NEE; -1,423.0 kg C ha-1)의 3.1%가 산림계류수로 빠져나간 것이다. 이중탄소동위원소비(평균 Δ14C-DOC: +13.7‰, δ13C-DOC: -27.0‰)와 리그닌 페놀 지표는 기저유출과 강우유출 모두에서 C3 식물이 점봉산 하천수 성분을 이루고 있다는 것을 보여줬다. FT-ICR MS 분석에서는 통해 강우 시에 하천수 내 질소 함유 화합물의 비율이 증가함을 확인했다.
둘째, 약 70년에 걸쳐 서로 다른 정치 체계와 산림관리 체계에서 관리되어온 남-북한 산림하천수의 DOM 특성을 검토했다. 북한 산림이 유역 면적의 대부분을 차지하는 두 개의 접경 유역(N1, 3,131 km²; N2, 8 km²)과 거의 대부분이 남한인 유역(S1, 2,080 km²), 온전히 남한에 속한 유역(S2, 262 km²)을 대상으로 조사했다. DOC 농도는 모두 비슷했으나, 하천 Δ14C- DOC 값은 크게 달랐다. N1 하천에서는 방사성탄소동위원소 연령이 683 yr BP, S1에서는 85 yr BP였던 반면, N2와 S2에서는 최근 광합성된 탄소에 해당하는 Δ14C 값이 측정되었다. δ13C-DOC, SUVA254, FT-ICR MS, 방사성탄소동위원소비 분석 결과를 종합하면, 70여년에 걸친 남-북한의 산림유역 관리 방법 차이와 그에 따른 유기물 기원 차이가 하천수 DOM 성분에도 반영됐을 가능성을 시사한다.
셋째, 침엽수림과 활엽수림이 우점한 유역 간의 하천 DOM 특성을 우리나라 16곳 산림하천수로 비교, 분석했다. 그 결과, 우리나라 산림하천수 DOM에서 얻은 시린질/바닐린(S/V) 리그닌 페놀 비를 이용하여 침엽수림과 활엽수림 유역을 처음으로 구분할 수 있었다. 반면, SUVA₂₅₄와 DOC 농도는 침엽수림과 활엽수림 간에 차이가 없었고, FT-ICR MS 분석에서도 수종이 하천수 성분에 미치는 영향은 제한적이었다.
이 연구를 통해, 산림유역의 규모 차이, 생태계의 변화와 수종 차이, 그리고 계절이, 산림하천 탄소 성분과 유출에 어떤 영향을 미치는지 종합적으로 보여주고자 했다.

목차 (Table of Contents)

• Table of Contents
• Abstract ii
• Table of Contents iv
• List of figures vii
• List of Tables xii
• 1. Introduction １
• 2. Biogeochemical properties of dissolved carbon in a forest stream at Mt. Jeombong, a Korean long-term ecological research (KLTER) site 5
• 2.1. Introduction 5
• 2.2. Materials and Methods 8
• 2.2.1. Study site 8
• 2.2.2. Sample collection, hydrologic classification, and analytical design 9
• 2.2.3. Stream carbon concentration and flux calculations 11
• 2.2.4. DOM compositional characterization 12
• 2.2.5. Statistical analysis of multi-technique results 18
• 2.3. Results 18
• 2.3.1. Stream carbon concentrations (2019 – 2025) and loads (April 2024 – December 2024) 18
• 2.3.2. Stream [DOC] and SUVA254 under baseflow conditions (2019 – 2024) 22
• 2.3.3. Dual-carbon isotope ratios in stream DOC under baseflow conditions 24
• 2.3.4. Lignin phenol indices in stream DOC under baseflow conditions 26
• 2.3.5. Molecular-level characterization using FT-ICR MS under baseflow conditions 30
• 2.3.6. Stream carbon properties under stormflow conditions 34
• 2.3.7. Lignin-phenol indices in stream DOC under stormflow conditions 38
• 2.3.8. Molecular-level characterization using FT-ICR MS indices in stream DOC under stormflow conditions 41
• 2.4. Discussion 46
• 3. Multiple techniques for analyzing riverine dissolved organic carbon to trace integrated characteristics of forest watersheds in South and North Korea across seven decades after the Korean War 52
• 3.1. Introduction 52
• 3.2. Materials and Methods 55
• 3.2.1. Study sites 55
• 3.2.2. Sample collection and pretreatment prior to specific analyses 57
• 3.2.3. Dual-carbon isotope analysis 58
• 3.2.4. Lignin phenol analysis 59
• 3.2.5. FT-ICR MS analysis 60
• 3.2.6. Statistical analysis 62
• 3.3. Results and Discussion 63
• 3.3.1. DOM characteristics of North and South Korean rivers 63
• 3.3.2. Molecular-level drivers of DOM aromaticity across contrasting forested watersheds 67
• 3.3.3. Principal component analysis (PCA) 72
• 3.4. Conclusion 75
• 4. Comparative biogeochemical characterization of riverine DOM across deciduous and coniferous forest watersheds 76
• 4.1. Introduction 76
• 4.2. Materials and Methods 78
• 4.2.1. Study sites 78
• 4.2.2. Sample collection and pretreatment prior to specific analyses 82
• 4.2.3. Dual-carbon isotope analysis 83
• 4.2.4. Lignin phenol analysis 83
• 4.2.5. FT-ICR MS analysis 84
• 4.2.6. Statistical analysis 85
• 4.3. Results 86
• 4.3.1. Forest stream [DOC] and SUVA254 87
• 4.3.2. Total dissolved nitrogen 91
• 4.3.3. Carbon-normalized lignin (Λ8) and Δ14C-DOC 93
• 4.3.4. Molecular-level indices of stream DOM by plant species 95
• 4.4. Discussion 105
• 4.5. Conclusions 108
• 5. Conclusions 110
• References 112
• 국문초록 136
• List of figures
• Figure 1.1. Conceptual diagram of lateral carbon export from forest watersheds, highlighting the pathways of dissolved organic carbon (DOC), particulate organic carbon (POC), and dissolved inorganic carbon (DIC). 2
• Figure 2.1. Forest watershed at the V-notch weir at Mr. Jeombong, Gangwon-do, South Korea. 8
• Figure 2.2. Temporal variations in [DOC], [POC], and [DIC] in stream water collected at the V-notch weir, Mt. Jeombong from 2019 to 2025. 19
• Figure 2.3. Temporal variations in stream carbon load, discharge, and precipitation at Mt. Jeombong in 2024. 20
• Figure 2.4. Seasonal distribution of dissolved organic carbon concentration ([DOC]) under baseflow conditions in the Mt. Jeombong stream. 23
• Figure 2.5. Seasonal variation in SUVA254 under baseflow conditions in the Mt. Jeombong stream. 24
• Figure 2.6. Dual carbon isotope ratios of the forest stream at the V-notch weir at Mt. Jeombong and other rivers. 26
• Figure 2.7. Lignin phenols in forest stream DOM at the V-notch weir of Mt. Jeombong and in world rivers. 29
• Figure 2.8. Van Krevelen diagrams of stream DOM collected at the V-notch weir at Mt. Jeombong, categorized by major natural compound classes (Table 2.5). 32
• Figure 2.9. Relationships between discharge and carbon concentrations from April 20 to December 31, 2024. (a) Dissolved inorganic carbon concentration ([DIC]), (b) dissolved organic carbon concentration ([DOC]), and (c) particulate organic carbon concentration ([POC]) as a function of discharge (m3 hr-1). 36
• Figure 2.10. Relationship between dissolved organic carbon ([DOC]) and particulate organic carbon ([POC]) from April 20 to December 31, 2024. The solid red line indicates the linear regression, showing significant co-mobilization of [DOC] and [POC]. 37
• Figure 2.11. Lignin phenols in forest stream DOM at the V-notch weir at Mt. Jeombong under baseflow and stormflow conditions, and in world rivers. 39
• Figure 2.12. Seasonal variation in the acid-to-aldehyde ratio of vanillyl phenols (Ad/Al(v)) in stream DOM at Mt. Jeombong under baseflow (gray) and stormflow (black) conditions. 40
• Figure 2.13. Van Krevelen diagram of stream DOM collected at the V-notch weir of Mt. Jeombong, categorized by major natural compound classes. 42
• Figure 2.14. Van Krevelen diagrams of DOM molecular formulae under stormflow (top) and baseflow (bottom) conditions in the Mt. Jeombong stream. 45
• Figure 3.1. Locations of the four study sites in Korea. 56
• Figure 3.2. δ¹³C-DOC versus Δ¹⁴C-DOC of Korean rivers compared with major global rivers. 64
• Figure 3.3. DOM molecular composition and elemental ratios across the studied rivers and streams. 70
• Figure 3.4. (a) Correlations between Δ¹⁴C-DOC and the H/C ratio, and (b) correlations between Δ¹⁴C-DOC and carbon-normalized lignin. 71
• Figure 3.5. Principal component analysis (PCA) biplots of dissolved organic matter (DOM) characteristics from four forested watersheds (N1, N2, S1, and S2). 73
• Figure 4.1. Relationship between the syringyl-to-vanillyl (S:V) ratio and angiosperm cover (%) across the studied watersheds. Forest composition and watershed information are provided in Section 4.2.1 and Table 4.1. 87
• Figure 4.2. Relationships between angiosperm tree cover (%) in watersheds and [DOC] (a) and SUVA254 (b). Colors denote seasons: spring (green), summer (yellow), autumn (purple), and winter (orange). 90
• Figure 4.3. (a) [DOC] and (b) SUVA254 in watersheds dominated by coniferous, mixed, and deciduous tree species. Green triangles denote mean values, and orange lines denote median values. 91
• Figure 4.4. Relationships between angiosperm tree cover (%) in watersheds and total dissolved nitrogen (TDN) concentrations. 92
• Figure 4.5. Relationships between angiosperm tree cover (%) in watersheds and Λ8 (a) and Δ14C-DOC (b). 94
• Figure 4.6. (a) Λ8 and (b) Δ14C-DOC in watersheds dominated by coniferous, mixed, and deciduous tree species. 95
• Figure 4.7. Relationships between angiosperm tree cover (%) in watersheds and the relative abundance of CHO (a), CHON (b), CHOS (c), and CHOP (d). Colors denote seasons: spring (green), summer (yellow), autumn (purple), and winter (orange). 97
• Figure 4.8. (a) CHO, (b) CHON, (c) CHOS, and (d) CHOP in watersheds dominated by coniferous, mixed, and deciduous tree species. Green triangles denote mean values, and orange lines denote median values. 98
• Figure 4.9. Seasonal variation in CHO and CHON by angiosperm cover group. 99
• Figure 4.10. Relationships between angiosperm tree cover (%) in watersheds and the relative abundance of lignin-like (a) and tannin-like (b) fractions. Colors denote seasons: spring (green), summer (yellow), autumn (purple), and winter (orange). 101
• Figure 4.11. Relative abundances of CHO (a), CHON (b), CHOS (c), and CHOP (d) in watersheds dominated by coniferous, mixed, and deciduous tree species. Green triangles denote mean values, and orange lines denote median values. 102
• Figure 4.12. Seasonal variation in lignin-like and tannin-like compounds by angiosperm cover group. Panels (a) and (d) show results for the 0–30% group, with measurements taken in spring (n=3) and summer (n=3). 103
• Figure 4.13. Relationships between angiosperm tree cover (%) in watersheds and the relative abundance of lipid-like fractions. 105
•  
• List of Tables
• Table 2.1. Abbreviations of lignin phenol parameters (Burns et al., 2023). 15
• Table 2.2. Classification of DOM compounds by key natural compound classes (Xu et al., 2023). 17
• Table 2.3. Total stream carbon loads and hydrological balance for a 2.2 ha forest watershed at Mt. Jeombong (April – Dec 2024). 20
• Table 2.4. Monthly stream carbon loads of DIC, DOC, and POC in a 2.2 ha forest watershed at Mt. Jeombong (April – Dec 2024). 21
• Table 2.5. The number of classified forest stream DOM compounds by key natural compounds (see Table 2.2). 33
• Table 2.6. Molecular-level characteristics of dissolved organic matter (DOM) under baseflow and stormflow conditions in the Mt. Jeombong stream. 44
• Table 3.1. River and stream sampling sites and their watershed characteristics. 57
• Table 4.1. Summary of catchment characteristics and forest composition across the studied watersheds. 80
• Table 4.2. Seasonal [DOC], SUVA254, and sample numbers by dominant tree species 89