The Arctic Ocean plays an important role on the global hydrological and carbon cycles. It contributes 5–14% to the global balance of CO2 sinks and sources. Carbon is also cycled in the Arctic Ocean through the primary producers, with high primary production observed in the marginal ice zones, ice-free zones and melt ponds, with increased biogenic carbon export to the deep layers. Although being the smallest ocean basin, the Arctic Ocean receives ~11% of the global riverine runoff. Along with the freshwater, high loads of organic carbon are introduced in the Arctic Ocean. Most of it is observed in the fraction of dissolved organic matter (DOM). With the ongoing global warming, glacier melt and permafrost thaw are observed and pointed as the main drivers for increasing the freshwater discharge into the Arctic basin. Along side, permafrost thaw coupled with increased coastal erosion lead to an increase in mobilization of carbon from permafrost, which could have critical implications for microbial processes, primary production, terrestrial carbon fluxes to the shelf seas and, thus, carbon cycling in the Arctic. This thesis is focusing on tracing the mixing of DOM along the Siberian shelves and developing potential applications of DOM as an environmental tracer. Four main objectives have been pursed: (1) to quantify, characterize and assess the distribution and transformation of DOM across the river-shelf transition and provide insights into the fate of Arctic riverine DOM; (2) to assess the potential of DOM, especially its fluorescent fraction (FDOM), as a tracer of freshwater in the surface layers in the Arctic Ocean; (3) to characterize the non-water absorption in the surface central and eastern Arctic Ocean and further test whether bio-optical properties (such as absorption and reflectance) can reproduce hydrographical variability; (4) to evaluate the performance of ocean color algorithms frequently applied for studies in the Arctic Ocean using novel data from a central-eastern Arctic expedition. In the first study the fluorescent components of DOM isolated with PARAFAC model were characterized along the river to sea transition in the Laptev Sea and Lena River delta region. Results showed a strong dominance of visible wavelength DOM fluorescence (VIS-FDOM), which is associated to terrestrial signal (or humic-like compounds). The results corroborate previous reports showing strong removal of DOM at low salinity. However, our results showed that the removal occurs preferentially for VIS-FDOM, whereas ultraviolet wavelength FDOM (UV-FDOM, associated to autochthonous marine production) differed in behavior, with an increase during estuarine mixing. DOM removal occurred primarily in the surface layer, under direct influence of the Lena River runoff (salinity <10), which indicates that it was mainly driven by photodegradation and flocculation. The second study explored the potential of VIS-FDOM components isolated with PARAFAC analysis as an environmental tracer in the Fram and Davis Straits. VIS-FDOM was strongly correlated to the fractions of meteoric water (fmw) in polar waters. Furthermore, a pattern allowed the distinction between the sources of polar waters exiting the Fram Strait as being from the Eurasian or Canadian basins. In the bottom waters of the Davis Strait, VISFDOM was correlated to apparent oxygen utilization (AOU), tracing deepwater turnover of DOM and production of VIS-FDOM fluorescence. The findings presented in this study show which wavelengths carry information on sources and mixing of DOM, which therefore can be applied to monitor freshwater and carbon export to the North Atlantic. The third study shows that colored DOM (CDOM) dominates the nonwater absorption in the surface waters of the central and eastern Arctic. Spatial variability observed in the non-water absorbers (phytoplankton, CDOM and non-algal particles–NAP) clustered the sampling sites in agreement with hydrographic variability. Such variability was also detected by the analysis of hyperspectral remote sensing reflectance (Rrs). The empirical and semianalytical ocean color algorithms frequently applied in studies in the Arctic Ocean were applied to in situ measured Rrs to evaluate their performance. The retrievals (chlorophyll-a, and the absorption coefficients of CDOM and phytoplankton) were then validated to the correspondent in situ measurements. The results showed that empirical algorithms have poor performance, whereas the semi-analytical algorithms appeared to be robust for application in the Arctic Ocean; however still with considerable errors embedded to the retrievals. The main findings of this thesis are that bio-optical measurements have strong potential to trace environmental variability in the Arctic Ocean, and those can therefore provide insights on the Arctic hydrological and biogeochemical cycles. These parameters can be monitored by bio-optical sensors (e.g., radiometers, transmissometers, fluorometers, etc.). Such sensors can be further coupled to autonomous platforms such as satellites, gliders, automated underwater vehicles (AUVs) and ice-tethered profilers (ITPs), and significantly increase the amount of biogeochemical data in the Arctic Ocean, filling the gap left by classical sampling methods (i.e., oceanographic expeditions) and ocean color remote sensing, restricted to spring and summer seasons.