No description available.

Rivers are a significant source of particulate and dissolved organic carbon (POC, DOC) into inland waters and coastal systems and provide a fundamental linkage between the terrestrial, oceanic, and atmospheric carbon reservoirs. Recent studies have examined the relationship between the quantity and form (POC vs. DOC) of carbon delivered to the aquatic system; however, little is known about the age of POC and DOC exported and how the radiocarbon age may vary with latitude, topographic gradient, vegetation, and land use. I provide the first global synthesis of published radiocarbon values of POC and DOC (∆14C). Inclusion of DOC and POC parameters (µM, δ13C, ∆14C) reveal significant driving forces of DOC (µM), latitude, and elevation (m) as capable of explaining 25% of the variability in DO14C in rivers and POC (µM) and latitude accounting for 15% of the variability in PO14C. When δ13C of DOC and POC and latitude were incorporated with ∆14C of DOC observations, 61% of the variability in DOC age was explained revealing the necessity to include dissolved and particulate fractions of organic carbon to yield the most robust predictive models. This study found a global trend of increasing age of DOC and increasing δ13C of DOC and POC with increasing latitude. My study suggests future research should incorporate both particulate and dissolved OC parameters along with elevation, vegetation, land cover, and climate zones to increase understanding of what drives the age of carbon exported in riverine systems.

Allochthonous carbon can be a large proportion of the carbon budget in northern temperate and boreal lakes. This thesis uses stable carbon and oxygen isotopes to examine the fate of allochthonous dissolved organic matter (DOM) in northern temperate lakes, and to determine the importance of dissolved organic carbon (DOC) in lake carbon mass balances and in the δ¹³C of lake sediments. To use stable isotopes as a tool for studying DOC loss and sedimentation within lakes requires an understanding of processes that affect the δ¹³C and δ¹⁸O in aquatic systems. Photolysis is one mechanism that can account for the large allochthonous DOC loss within northern temperate lakes. There is, however, little research examining the effects of photolysis on stable isotopes (e.g. δ¹³C and δ¹⁸O) in aquatic systems, or how photodegradation of DOM affects the δ¹³C of lake sediments. To study the effects of DOM photodegradation on carbon and oxygen isotopes, stream waters from catchments with varying peatland coverage were incubated in Tedlar bags placed in water baths under natural sunlight. Results from three streams flowing into two oligotrophic headwater lakes (Harp and Dickie Lakes) indicate that O₂ consumption rates and dissolved inorganic carbon (DIC) production rates were an order of magnitude greater in light exposed treatments than in dark treatments, suggesting that light mediated processes control O₂ consumption and DIC production in incubations. The similarity between filtered, inoculated, and sterile treatments, indicate that photolysis was the dominant O₂ consuming and DIC producing process in the incubations, while the contribution of respiration to these processes was not detectable. Differences in both O₂ consumption rates and DIC production rates (normalized to DOC loss) among streams suggest that DOM photolability was an important factor in both O₂ loss and DIC production on a volumetric basis. A concomitant increase in δ¹⁸O-O₂ was observed with O₂ loss indicating that during the photo-oxidation of DOM, the lighter ¹⁶O isotopomer was preferentially consumed in the oxidation of DOC to CO₂. Fractionation factors for respiration, photolysis and other abiotic reactions were not a function of O₂ consumption rates and ranged between 0.988 and 0.995, which lies outside the range published for respiration (0.975-0.982). These are the first published photolytic fractionation factors. The δ¹³C-DIC produced collectively by photolysis, respiration, and other abiotic reactions in incubations exposed to natural sunlight ranged between –23‰ and –31‰, and were similar in the light incubations for each treatment, but different among streams. Together, the light and dark incubation data suggest that photolysis and other abiotic reactions were largely responsible for the DIC concentration and δ¹³C-DIC changes observed, while respiration is a relatively minor contributor. During the incubations, as DOC photodegraded to CO₂, the lighter ¹²C isotope was preferentially mineralized (or the moieties cleaved were depleted in ¹³C) leaving the residual δ¹³C-DOC 1‰ to 4‰ enriched, creating enrichment (ε) values up to ~–3‰. The change in final δ¹³C-DOC after DOM photodegradation was different for each inflow, ranging from ~1 ‰ to 8.0 ‰, and provides an average enrichment of –2.1‰ (Harp Inflows ε: –1.2‰; Dickie Inflows ε: –3.4‰). These ε values are in agreement with the average ε from previous incubations on 3 of the inflows and 3 published studies based on UV exposed bog water (Osburn et al., 2001), riverine waters (Opsahl and Zepp, 2001), and lyophilized Juncus leachate dissolved in lake water (Vähätalo and Wetzel, 2008) (average ε = –2.9‰). The structure of DOM changed during photolysis. Absorbance data indicated that the aromaticity, colour, UV absorption and the average molecular size of the DOC decreased. Additionally, after exposure to sunlight, C/N ratios of the DOC changed from high values (24-55), indicative of terrestrial inputs, to lower values (4-13) traditionally thought to be representative of algal or microbial inputs. This contradicts the conventional view that terrestrial DOC has C/N ratios >20, and shows that abiotic processes can alter allochthonous carbon structure and the residual allochthonous carbon can have C/N values similar to, or overlapping with, C/N ratios expected from algal or microbial carbon. With the loss of 61-90% of the DOC, the particulate organic carbon (POC) created accounted for 20-90% of the DOC lost. Values of δ¹³C-POC ranged from –25.7‰ to –27.7‰, with 80% of the samples within 1‰ of the initial δ¹³C-DOC indicating that the particulate carbon created from the photodegradation of DOM that settles to the lake sediments could be isotopically similar to the source DOC. Overall, these incubations indicate that the photodegradation of DOM can affect both concentrations and isotopes of O₂, DIC, DOC, and POC of the stream waters flowing into Harp and Dickie Lakes and are important to consider in lake dynamics of high DOC retention lakes. Two independent methods were used to examine the importance of allochthonous DOC to lake sediments. The first method used a two end-member mixing model to estimate the proportion of allochthonous and autochthonous carbon within the lake sediments. Inflow δ¹³C-POC data, δ¹³C-leaf litter measurements, and DOC photodegradation experiments were used to calculate average annual δ¹³C-POC values for the allochthonous end member. The average annual δ¹³C-POC values for the autochthonous end member were calculated using estimates of productivity, surface δ¹³C-CO₂ values and estimated average annual fractionation factors. Average annual δ¹³C-POC values from allochthonous and autochthonous sources for these lakes were distinct. Using the end members to calculate the relative contributions of allochthonous and autochthonous carbon to lake sediments revealed that the δ¹³C of the lake sediment can be significantly affected by the ratio of autochthonous and allochthonous contributions. Furthermore, peaks in the allochthonous contributions of carbon accompany the δ¹³C peaks in the sediment records to the lake sediments. This suggests that climate change and/or anthropogenic changes to the landscape, and the concomitant changes in DOC inputs to lakes, can be recorded in the sediment record indicating that sediment records are not just productivity signals, but also mass balance signals in high DOC retention lakes. In the second method carbon isotope budgets were completed to accompany the carbon mass budgets for Harp and Dickie Lakes. Mass-weighted average annual δ¹³C-DOC values from the inflows and outflows and δ¹³C-DIC values from the inflows varied by 0.2‰ to 1.3‰, suggesting the values are well constrained. Conversely, the range of weighted δ¹³C-DIC values from the outflows were larger (2.2‰) than those of the inflows. Calculated δ¹³C values of the lake sediment were not equal to the measured δ13C values of the lake sediments for either Harp or Dickie Lakes suggesting a problem lies within the mass balances, or the weighted average annual δ¹³C values used in the isotope budgets. To examine the sensitivity of the average annual weighted δ¹³C values for the carbon entering and exiting the lakes, and the mass of carbon entering the lakes δ¹³C of the lake sediments, a mass and isotope budget model was created. The model indicated that the δ¹³C of the lake sediments is sensitive to a number of parameters including the amount of DOC entering the lake, the δ13C-CO2 evaded from the lake, the areal water discharge rate (qs), the gas exchange coefficient (k), and pH. Many of these parameters required adjustments for the masses of carbon to match those presented in the mass balances suggesting that the mass balances averaged over 8 years have errors associated with them. However, changing the DOC load to the lakes in the model by the variability observed over all the years of the mass balances) indicates that the isotopic signature of the lake sediment could change by up to 2.5‰. This isotope change is large enough to account for the historical δ¹³C changes observed in the δ¹³C sediment record, suggesting that allochthonous DOC can drive the sediment record.

Allochthonous carbon can be a large proportion of the carbon budget in northern temperate and boreal lakes. This thesis uses stable carbon and oxygen isotopes to examine the fate of allochthonous dissolved organic matter (DOM) in northern temperate lakes, and to determine the importance of dissolved organic carbon (DOC) in lake carbon mass balances and in the δ¹³C of lake sediments. To use stable isotopes as a tool for studying DOC loss and sedimentation within lakes requires an understanding of processes that affect the δ¹³C and δ¹⁸O in aquatic systems. Photolysis is one mechanism that can account for the large allochthonous DOC loss within northern temperate lakes. There is, however, little research examining the effects of photolysis on stable isotopes (e.g. δ¹³C and δ¹⁸O) in aquatic systems, or how photodegradation of DOM affects the δ¹³C of lake sediments. To study the effects of DOM photodegradation on carbon and oxygen isotopes, stream waters from catchments with varying peatland coverage were incubated in Tedlar bags placed in water baths under natural sunlight. Results from three streams flowing into two oligotrophic headwater lakes (Harp and Dickie Lakes) indicate that O₂ consumption rates and dissolved inorganic carbon (DIC) production rates were an order of magnitude greater in light exposed treatments than in dark treatments, suggesting that light mediated processes control O₂ consumption and DIC production in incubations. The similarity between filtered, inoculated, and sterile treatments, indicate that photolysis was the dominant O₂ consuming and DIC producing process in the incubations, while the contribution of respiration to these processes was not detectable. Differences in both O₂ consumption rates and DIC production rates (normalized to DOC loss) among streams suggest that DOM photolability was an important factor in both O₂ loss and DIC production on a volumetric basis. A concomitant increase in δ¹⁸O-O₂ was observed with O₂ loss indicating that during the photo-oxidation of DOM, the lighter ¹⁶O isotopomer was preferentially consumed in the oxidation of DOC to CO₂. Fractionation factors for respiration, photolysis and other abiotic reactions were not a function of O₂ consumption rates and ranged between 0.988 and 0.995, which lies outside the range published for respiration (0.975-0.982). These are the first published photolytic fractionation factors. The δ¹³C-DIC produced collectively by photolysis, respiration, and other abiotic reactions in incubations exposed to natural sunlight ranged between –23‰ and –31‰, and were similar in the light incubations for each treatment, but different among streams. Together, the light and dark incubation data suggest that photolysis and other abiotic reactions were largely responsible for the DIC concentration and δ¹³C-DIC changes observed, while respiration is a relatively minor contributor. During the incubations, as DOC photodegraded to CO₂, the lighter ¹²C isotope was preferentially mineralized (or the moieties cleaved were depleted in ¹³C) leaving the residual δ¹³C-DOC 1‰ to 4‰ enriched, creating enrichment (ε) values up to ~–3‰. The change in final δ¹³C-DOC after DOM photodegradation was different for each inflow, ranging from ~1 ‰ to 8.0 ‰, and provides an average enrichment of –2.1‰ (Harp Inflows ε: –1.2‰; Dickie Inflows ε: –3.4‰). These ε values are in agreement with the average ε from previous incubations on 3 of the inflows and 3 published studies based on UV exposed bog water (Osburn et al., 2001), riverine waters (Opsahl and Zepp, 2001), and lyophilized Juncus leachate dissolved in lake water (Vähätalo and Wetzel, 2008) (average ε = –2.9‰). The structure of DOM changed during photolysis. Absorbance data indicated that the aromaticity, colour, UV absorption and the average molecular size of the DOC decreased. Additionally, after exposure to sunlight, C/N ratios of the DOC changed from high values (24-55), indicative of terrestrial inputs, to lower values (4-13) traditionally thought to be representative of algal or microbial inputs. This contradicts the conventional view that terrestrial DOC has C/N ratios >20, and shows that abiotic processes can alter allochthonous carbon structure and the residual allochthonous carbon can have C/N values similar to, or overlapping with, C/N ratios expected from algal or microbial carbon. With the loss of 61-90% of the DOC, the particulate organic carbon (POC) created accounted for 20-90% of the DOC lost. Values of δ¹³C-POC ranged from –25.7‰ to –27.7‰, with 80% of the samples within 1‰ of the initial δ¹³C-DOC indicating that the particulate carbon created from the photodegradation of DOM that settles to the lake sediments could be isotopically similar to the source DOC. Overall, these incubations indicate that the photodegradation of DOM can affect both concentrations and isotopes of O₂, DIC, DOC, and POC of the stream waters flowing into Harp and Dickie Lakes and are important to consider in lake dynamics of high DOC retention lakes. Two independent methods were used to examine the importance of allochthonous DOC to lake sediments. The first method used a two end-member mixing model to estimate the proportion of allochthonous and autochthonous carbon within the lake sediments. Inflow δ¹³C-POC data, δ¹³C-leaf litter measurements, and DOC photodegradation experiments were used to calculate average annual δ¹³C-POC values for the allochthonous end member. The average annual δ¹³C-POC values for the autochthonous end member were calculated using estimates of productivity, surface δ¹³C-CO₂ values and estimated average annual fractionation factors. Average annual δ¹³C-POC values from allochthonous and autochthonous sources for these lakes were distinct. Using the end members to calculate the relative contributions of allochthonous and autochthonous carbon to lake sediments revealed that the δ¹³C of the lake sediment can be significantly affected by the ratio of autochthonous and allochthonous contributions. Furthermore, peaks in the allochthonous contributions of carbon accompany the δ¹³C peaks in the sediment records to the lake sediments. This suggests that climate change and/or anthropogenic changes to the landscape, and the concomitant changes in DOC inputs to lakes, can be recorded in the sediment record indicating that sediment records are not just productivity signals, but also mass balance signals in high DOC retention lakes. In the second method carbon isotope budgets were completed to accompany the carbon mass budgets for Harp and Dickie Lakes. Mass-weighted average annual δ¹³C-DOC values from the inflows and outflows and δ¹³C-DIC values from the inflows varied by 0.2‰ to 1.3‰, suggesting the values are well constrained. Conversely, the range of weighted δ¹³C-DIC values from the outflows were larger (2.2‰) than those of the inflows. Calculated δ¹³C values of the lake sediment were not equal to the measured δ13C values of the lake sediments for either Harp or Dickie Lakes suggesting a problem lies within the mass balances, or the weighted average annual δ¹³C values used in the isotope budgets. To examine the sensitivity of the average annual weighted δ¹³C values for the carbon entering and exiting the lakes, and the mass of carbon entering the lakes δ¹³C of the lake sediments, a mass and isotope budget model was created. The model indicated that the δ¹³C of the lake sediments is sensitive to a number of parameters including the amount of DOC entering the lake, the δ13C-CO2 evaded from the lake, the areal water discharge rate (qs), the gas exchange coefficient (k), and pH. Many of these parameters required adjustments for the masses of carbon to match those presented in the mass balances suggesting that the mass balances averaged over 8 years have errors associated with them. However, changing the DOC load to the lakes in the model by the variability observed over all the years of the mass balances) indicates that the isotopic signature of the lake sediment could change by up to 2.5‰. This isotope change is large enough to account for the historical δ¹³C changes observed in the δ¹³C sediment record, suggesting that allochthonous DOC can drive the sediment record.

Physical erosion can mobilise particulate organic carbon (POC) from vegetation and soil, representing an export of primary productivity from ecosystems, and a lateral transfer of carbon recently-derived from the atmosphere. These carbon transfers are thought to be enhanced in mountain forests where erosion rates are high. However, the rates and controls on POC transfer remain poorly constrained, as does the impact of POC export on carbon cycling at regional and global scales. This thesis takes an interdisciplinary approach to address this issue, using remote sensing, river geochemistry, river hydrology, and geomorphic mapping in the Kosñipata Valley, in the Central Andes of Peru. Its main aims are to: 1) estimate stream discharge throughout the year and to evaluate the water balance and sources; 2) quantify the source of riverine POC, accounting for POC derived from sedimentary rocks (POC_fossil) to examine the POC eroded from soils and vegetation (POC_non-fossil); 3) quantify river POC yields; 4) assess the hillslope processes that erode POC; and 5) assess how POC export impacts the carbon balance of mountain forest, and how fluvial transfer impacts the wider carbon cycle. Stream flow was monitored from January 2010 to February 2011 at two newly installed river gauging stations in the Kosñipata Valley at 2250 m (Wayqecha, 48.5 km²) and 1360 m (San Pedro, 164.4 km²). Then annual water balance for the San Pedro catchment was quantified. Rainfall inputs of 3028 mm and cloud water inputs of 308 ± 97 mm were balanced by outputs via stream runoff (2721 mm) and actual evapotranspiration (907 mm), leaving a residual of -294 ± 97 mm (< ~10 % of water inputs). The source of POC in river suspended sediment samples was quantified using radiocarbon (Δ¹⁴C, ‰), stable carbon isotopes, and the nitrogen to carbon ratio. This revealed that river POC_non-fossil was sourced from very young organic carbon in the valley (Δ¹⁴C ~50 ‰) and that POC_fossil comprised 43 % of total POC. Combining the hydrometric measurements with river samples, annual particulate load fluxes were quantified. The vast majority (73 % to 77 %) of the annual suspended sediment transfer and POC (both POC_fossil and POC_non-fossil) occurred in the wet season over a period of 4 months. The suspended sediment yield for the valley (960 – 1200 t km^-2 yr^-1) was consistent with those for the Andean portion of the Madre de Dios River into which the Kosñipata River drains. The river POC_non-fossil yield was 5.2 – 6.9 tC km^-2 yr^-1. Landslides are likely to have played an important role in the mobilisation of POC_non-fossil. A detailed landslide mapping using 25 years of remote sensing data revealed that on average 0.09 % of the valley per year is impacted by this mass-wasting process. These landslides mobilise ~28 tC km^-2 yr^-1 of soil and vegetation valley-wide. The discrepancy between the landslide erosional flux and fluvial POCnon-fossil export suggests an important fraction of the POCnon-fossil harvested by landslides is either exported as coarse debris (not quantified in the fluvial POC_non-fossil flux), remains buried onsite, or is degraded and respired onsite. Landslides also played an important ecosystem function, turning over some sections of the mountain forest within ~625 years, with a 1200 year valley-wide mean. On the basin scale, the Madre de Dios River drains ~ 6 % of the Amazonian Andes. This study enables estimation of the delivery of POC to the lowland Amazon Basin. Using the observation that POC_non-fossil and POC_fossil fluxes were closely linked with suspended sediment transfer, total yields of ~0.22 MtC yr^-1 and ~0.17 MtC yr^-1, respectively, were estimated from this section of the Andes. The export of POC_non-fossil from mountain forests by rivers represents 0.4 – 1.0 % yr^-1 of the net primary productivity of Andean forest and so even if only a small portion of this is buried in sedimentary deposits, it may promote the Andes as a carbon sink. These results demonstrate the long-term influence of erosional processes in the cycling of carbon in the Amazon Basin.

The spread of surface coal mining has resulted in loss of forests in the Appalachian region. The Forestry Reclamation Approach (FRA) was developed to provide guidance for restoring forests on reclaimed mined land. This study hypothesizes that the FRA will result in larger magnitude of sediment accumulation rates in reclaimed mine sites compared to those reclaimed using grassland reclamation. Three sediment cores and six trenches were sampled within four reclaimed mined and three previously logged sites in eastern Kentucky. Samples were processed for radionuclides, grain-size, stable isotopes (δ13C), and POC. LIDAR data were used to identify valley fills, while historical aerial photography was used to identify changes in vegetative cover from 1994 to 2016. Radionuclide dating was used to determine sediment accumulation rates over the previous 100 years. Results from logged sites are inconclusive. δ13C data for all sites fall within the range expected for forested landscapes (C3), and do not show any transitions from grassland to forests. POC data indicates that inventories and fluxes were the same for mined and logged sites. Sediment accumulation rates for reclaimed mined lands show elevated values after the implementation of the FRA, compared to grassland reclamation, thus supporting the hypothesis for previously mined sites.

Arctic soils store half of the global soil organic carbon (OC) pool and twice as much C as is currently present in the atmosphere. A considerable part of these carbon pools are stored in permafrost. Amplified climate warming in the Arctic will thaw permafrost and remobilize some of these substantial carbon stocks into the active carbon cycle, potentially causing positive feedback to global warming. Despite the global importance of this mechanism, our understanding of the fate of these thawing organic carbon (OC) pools is still poor, particularly regarding its degradation potential. This makes good estimates on greenhouse gas emissions versus coastal reburial impossible. This doctoral thesis aims to improve our understanding on the fate of high-latitude terrestrial OC during fluvial and coastal transport. In two study regions, the Bothnian Bay and the East Siberian Sea, we apply a wide range of bulk, molecular and isotopic geochemical analyses to reveal information on sources, age, degradation and transport routes. Our results show that both study regions receive and store large amounts of terrestrial OC, largely derived from peatlands (paper I, II and IV). This terrestrial matter undergoes extensive degradation in both the water column and surface sediments (paper I, III and IV). Surface sediments in the East Siberian Sea show a offshore-decreasing input of riverine OC and a considerable and constant input of OC from coastal erosion. The strong imprint of rapidly settling coastal OC far out on the shelf may be explained by a strong benthic boundary layer transport in combination with offshore ice-transport and selective preservation of erosion OC compared to riverine OC (paper IV). Molecular radiocarbon data allowed us to distinguish between two (sub-)Arctic soil OC pools that show a remarkably different susceptibility to degradation upon arrival in the coastal system; a young and easily degradable pool originating in surface peatlands, and an old and recalcitrant pool originating in deep mineral soils and coastal mineral Pleistocene deposits (paper III and IV). Our first estimates suggest that, in the Bothnian Bay coastal system, mineral soil OC is at least 20 times less susceptible to degradation than peatland OC (paper III). Hence, a considerable part of the thaw-released mineral OC pool may simply be relocated to coastal sediments instead of being emitted to the atmosphere. / At the time of the doctoral defense, the following paper was unpublished and had a status as follows: Paper 3: Accepted. Paper 4: Manuscript.

Abstract The South China Sea (SCS) is the largest marginal sea in the world and connects with the East China Sea (ECS) through the Taiwan Strait (TS). This study investigates the distribution and biogeochemical behavior of both particulate and dissolved organic matter in the SCS and the TS based on samples collected on several cruises of the R/V Ocean Researchers I and III. Dissolved inorganic nitrogen and phosphorus (DIN and DIP), particulate organic carbon and nitrogen (POC and PON) as well as dissolved organic nitrogen and phosphorus (DON and DOP) concentrations were determined. Concentrations of DON and DOP in the SCS are in the range of 1.2-9.9 mMN and 0.04-0.21 mMP, respectively. The surface DON concentration is the highest in the northern SCS, whereas it is the lowest in the southern part. The DOP does not show a similar trend. DON and DOP concentrations all decrease with depth but increase slightly near the bottom, perhaps on account of sediment resuspension. Because of the preferential degradation of DOP over DON, the maximum concentration of DOP appears at a shallower depth than that of DON. Approximately 11 % and 2 % of DIN and DIP respectively are attributed to the degradation of DON and DOP above 500 m in the SCS. Concentrations of POC and PON in the SCS are in the range of 1.06-2.84 mMC and 0.07-0.36 mMN, respectively. The distributions of POC and PON show similar patterns with a correlation coefficient of 0.97. The concentrations of these are the highest at the surface layer, decrease with depth, but then increase slightly near the bottom, perhaps again because of resuspension of the bottom sediments. The ratio of PON/POC is 0.138 in the euphotic zone, a value close to the Redfield ratio of 0.15. In the TS and the adjacent coastal zones, the effect of terrestrial input is obvious and results in higher POC, PON, DON and DOP nearshore. Ranges of these concentrations are 0.06-59.6 mMN, 0.01-1.29 mMP, 3.80-57.1 mMC and 0.19-3.4 mMN, respectively. There was an attempt to use the one-dimensional diffusion-advection model to estimate the DIN and DIP production rates and the DON and DOP consumption rates over the depth range of 900-2500 m. These values are, respectively, 0.036, 0.006, 0.021 and 0.002 mmol/kg/yr.

High Arctic permafrosted soils represent a massive sink in the global carbon cycle, accounting for twice as much carbon as what is currently stored as carbon dioxide in the atmosphere. However, with current warming trends this sink is in danger of thawing and potentially releasing large amounts of carbon as both carbon dioxide and methane into the atmosphere. It is difficult to make predictions about the future of this sink without knowing how it has reacted to past temperature and climate changes. This dissertation summarizes the results of the first study to look at long term, fine scale organic carbon delivery by the high-Arctic Colville River into Simpson’s Lagoon in the near-shore Beaufort Sea. Modern delivery of organic carbon to the Lagoon was determined to come from a variety of sources through the use of a three end-member mixing model and sediment biomarker concentrations. These sources include the Colville River in the western area of the Lagoon near the river mouth, marine sources in areas of the Lagoon without protective barrier islands, and coastal erosional sources and the Mackenzie River in the eastern area of the Lagoon. Downcore organic carbon delivery was measured on two cores in the Lagoon, one taken near the mouth of the Colville River (spans about 1800 years of history) and one taken on the eastern end of the Lagoon (spans about 600 years of history). Bulk organic parameters and biomarkers were measured in both cores and analyzed with Principle Component Analysis to determine long-term trends in organic carbon delivery. It was shown that at various times in the past, highly degraded organic carbon inputs of what is likely soil and peat carbon were delivered to the Lagoon. At other times, inputs of fresher, non-degraded, terrestrially-derived organic carbon inputs of what are likely higher amounts of plant and vegetative material was delivered to the Lagoon. Inputs of degraded soil carbon were also shown to correspond to higher temperatures on the North Slope of Alaska, likely indicating that warmer temperatures lead to a thawing of permafrost and in turn organic carbon mobilization to the coastal Beaufort Sea.

Conselho Nacional de Desenvolvimento Científico e Tecnológico / Soil carbon (C) sequestration in agriculture soil is a low cost option to mitigate global climatic change. No-till (NT) associated with good husbandry practices could compensate up to 15% of the anthropogenic CO2 emissions by storing the C from atmosphere as soil organic carbon (SOC). To be fully accepted as a mitigation alternative, research must be conducted to improve the accuracy of soil C sequestration estimates on field experiments as well as those made by mathematical models at regional and local scales. Complementarily, is necessary to improve the knowledge about the SOC stabilization mechanisms, delimiting the real soil´s capacity into accumulate C, quantifying how much of the stored C could be re-emitted to the atmosphere by changes in soil management. The present work is divided in four chapters with the objective to answer these questions. The first chapter has the objective to discuss the importance of sampling depth (0-0.30, 0-0.60, and 0-0.90 m) and the definition of a reliable and adequate baseline for the calculation of the C sequestration rates. Two long-term field experiments from a temperate (Mollisol) and a sub-tropical (Oxisol) climate soil were selected for this research. The experiments tested soil tillage systems (conventional tillage (CT) and NT) (Mollisol and Oxisol) and sources and rates of nitrogen amendment on corn in the Mollisol (control without N, 168 kg N ha-1 as ammonium sulfate, and 168 kg N ha-1 as organic fertilizer) and different crop rotation systems in the Oxisol (R0:soybean-wheat, R1:soybean-wheat-soybean-oat, and R2:soybean-oat-soybean-oat+vetch-corn-radish-wheat). The increase of sampling depth provided limited contribution to the estimates of C sequestration rates due to the increase of the error on SOC stocks estimates at deeper soil depths. To improve the C sequestration rate estimates, SOC temporal dynamic analysis should be preferred rather than the comparison of the SOC stocks of paired plots at a unique time point. The second chapter had the objective to apply simple mathematical equations to describe the SOC dynamics and improve the estimates of C sequestration rates and also to understand the role of the macroaggregate formation on SOC accumulation and saturation. The use of linear and kinetic (exponential growth) equation was adequate to describe the SOC dynamics increasing the accuracy of the C sequestration rate estimates by reducing errors promoted by soil spatial variability. The SOC accumulation was a function of the amount of C input to the soil and the macroaggregate formation to protect SOC. The SOC saturation process occurred from the smaller to the larger aggregate size fraction, limiting the capacity of a given superficial soil layer to accumulate SOC. However, the SOC saturation at superficial soil layers did not indicate the end of C sequestration in the soil, since the SOC accumulation occurred at sub-superficial soil layers. In the third chapter, the mathematical approach to determine changes on SOC stocks and the SOC saturation-induced limitation for C sequestration were applied to improve the accuracy of the Hénin e Dupuís (1945) one- 8 compartmental mathematical model into predict future soil C sequestration rates. The long-term field experiment from the Mollisol was selected for this research because of the better data availability (sampling years) and also by the presence of two treatments under CT and NT with SOC saturated soil layers. The mathematical adjustment (by linear equations) of the SOC dynamic coefficients improved the adjustment of the model‟s predictions. The SOC saturation-induced restriction for SOC accumulation on the mathematical model avoided the overestimation of the soil‟s potential for C sequestration. The predictions of the mathematical models indicate that the Mollisol‟s superficial layer (0-0.05 m) under NT could maintain significant C sequestration rates for up to 50 years as a function of the amount of C input to the soil. For the fourth chapter, a detailed study of the SOC pools in water-stable aggregate size fractions by granulometric and densimetric fractionation was carried out. The objective was to identify the pools where SOC accumulation was occurring and what SOC stabilization mechanisms were present. This will provide estimates of the NT potential to promote long-term C sequestration. SOC accumulation occurred preferentially in the more stable and recalcitrant SOC fractions (Mollisol and Oxisol) or in microaggregate and macroaggregate physically protected fractions (Mollisol). In the Oxisol, the SOC enrichment occurred mostly in the mineral associated-SOC fractions extra-microaggregates occluded within meso- and macroaggregates, while in the Mollisol, the SOC accumulation occurred in both intra- and extra-microaggregate mineral associated-SOC fractions. More than 78 and 92% of the C sequestration verified in the Mollisol and Oxisol, respectively, were considered as long-term by occurring in stable SOC fractions. / O seqüestro de carbono (C) em solos agrícolas é uma opção de baixo custo para mitigação das mudanças climáticas globais. O plantio direto (PD), associado a boas práticas agronômicas, pode compensar até 15% das emissões antrópicas de CO2 ao armazenar o C drenado da atmosfera na forma de carbono orgânico (CO) do solo. Para ser amplamente aceito como alternativa de mitigação, pesquisas devem ser conduzidas a fim de melhorar a precisão das estimativas de taxas de seqüestro de C em experimentos de campo, assim como as previsões feitas por modelos matemáticos em escalas regionais e locais. Complementarmente, é necessário aprimorar o conhecimento sobre os mecanismos de estabilização do CO, delimitando a capacidade real do solo em acumular C e quantificando quanto do C acumulado no solo pode ser re-emitido para atmosfera por mudança no manejo do solo. Desta maneira, o presente trabalho se divide em quatro capítulos com o objetivo de abordar estas questões. O primeiro capítulo tem por objetivo discutir a importância da profundidade de amostragem (0-0,30; 0-0,60; 0-0,90 m) e da definição de situações de linhas-base confiáveis a adequadas para o cálculo das taxas de seqüestro de C. Para isto, foram utilizados dois experimentos de longa duração sobre um solo de clima temperado (Mollisol) e outro de clima sub-tropical (Oxisol). Os experimentos testaram efeitos de sistema de preparo do solo (preparo convencional (PC) e PD) (Mollisol e Oxisol) e fontes e doses de nitrogênio para o milho no Mollisol (testemunha (T), 168 kg N ha-1 na forma de sulfato de amônia (AM) e 168 kg N ha-1 na forma de adubo orgânico (AO)) e diferentes sistemas de rotação de culturas no Oxisol (R0:soja-trigo, R1:soja-trigo-soja-aveia e R2:soja-aveia-soja-aveia+ervilhaca-milho-nabo-trigo). O aumento da profundidade de amostragem não contribuiu com a melhoria das estimativas de taxas de seqüestro de C devido ao aumento do erro nas estimativas dos estoques de CO nas camadas mais profundas de solo. Para melhoria das estimativas das taxas de seqüestro de C devem-se preferir análises temporais da dinâmica do CO no solo ao invés da comparação de estoques de CO em um único momento. O segundo capítulo tem por objetivo aplicar equações matemáticas simples para descrever a dinâmica do CO e melhorar as estimativas taxas de seqüestro de C e também entender o papel da formação de macroagregados no acúmulo e saturação de CO no solo. O uso de equações lineares e cinéticas (crescimento exponencial) foi adequado para descrever a dinâmica do CO, aumentando a precisão das estimativas de taxas de seqüestro de C ao reduzir os erros de estimativa por variabilidade espacial do solo. O acúmulo de CO no solo mostrou-se uma função da quantidade de C aportada ao solo por resíduos vegetais e a formação de macroagregados no solo para proteção do CO. O processo de saturação do solo ocorreu das menores para as maiores frações de agregados do solo, limitando a capacidade de uma 6 determinada camada de solo em acumular CO. No entanto, verificou-se que a saturação de camadas superficiais de solo não indica o fim do seqüestro de C neste solo, visto que o acúmulo de CO passa a ocorrer em camadas sub-superficiais. No terceiro capítulo, a aproximação matemática para determinar mudanças nos estoques de CO e a limitação na capacidade do solo em acumular CO promovida pelo processo de saturação dos agregados do solo foram aplicados para melhorar a precisão do modelo matemático uni-compartimental de Hénin e Dupuís (1945) em prever futuras taxas de seqüestro de C. O Mollisol foi escolhido para este estudo em função da maior disponibilidade de dados (anos de amostragem) e também pela presença de dois tratamentos em PC e PD com camada de solo saturada por CO. O ajuste matemático (por equações lineares) dos coeficientes da dinâmica do CO melhorou o ajuste das previsões do modelo com os dados observados. A restrição do modelo matemático quanto à capacidade do solo em acumular CO (saturação de CO) evitou a superestimação do potencial de seqüestro de C deste solo. As previsões do modelo matemático indicam que a camada superficial (0-0,05 m) do solo sob PD pode apresentar taxas significativas de seqüestro de C por até 50 anos, em função da quantidade de C adicionada ao solo. No quarto capítulo, foi realizado um estudo detalhado dos compartimentos do CO em função da sua distribuição em classes de tamanhos de agregados estáveis em água e o fracionamento granulométrico e densimétrico do CO. O objetivo foi identificar em quais compartimentos está ocorrendo o acúmulo de CO no solo, os mecanismos de estabilização do CO, estimando o potencial do PD em promover sequestro de C de longa duração. Verificou-se que o acúmulo de C ocorre preferencialmente em frações mais estáveis e recalcitrantes do CO (Mollisol e Oxisol) ou em frações protegidas fisicamente por micro e macroagregados (Mollisol). No Oxisol, o enriquecimento de CO ocorre principalmente nas frações de CO associadas aos minerais extra-microaggregados oclusas em meso e macroagregados de solo, enquanto que no Mollisol, o acumulo de CO ocorre tanto na fração intra como extra microagregados. Mais de 78 e 92% do seqüestro de C verificado no Mollisol e Oxisol, repectivamente, foi considerado de longa duração por ocorrer em frações estáveis do CO.