Long-Term Patterns of Dissolved Organic Carbon in Boreal Lakes
I analyzed the 21 year dynamics of dissolved organic carbon (DOC) in 55 lakes in five sites across Eastern Canada in relation to regional and global variables. Regional variables included total solar radiation (TSR), precipitation (PPTN), air temperature (T) and sulfate deposition (SO4). Global variables included the Southern Oscillation Index (SOI), North Atlantic Oscillation (NAO) and Pacific Decadal Oscillation (PDO). A synchronous pattern in DOC was found among lakes within each region; however, a synchronous pattern in DOC was not found between sites, except for Kejimkujik and Yarmouth which were only 80 km apart from each other. This suggested that the variation of the long-term DOC pattern was in response to the temporal pattern of regional variables, and it supports the recent understanding that regional factors have a strong influence on many lake properties. Significant long-term trends in DOC were not evident except at the Experimental Lakes Area (ELA), where an increase in DOC was observed together with a decrease in summer TSR and an increase in summer precipitation. Annual mean air temperature has increased at the Nova Scotia and Turkey lakes sites over the study period. The relationship between the long-term pattern in DOC with the regional and global variables was analyzed for each study site to determine the key variables that could best explain the variation in the long-term pattern in DOC. TSR and PPTN were important independent variables across all sites, except for the Turkey Lakes Watershed site (TLW). Summer TSR (annual TSR for Kejimkujik and Yarmouth) had a negative relationship, while summer precipitation had a positive relationship with the long-term DOC pattern for all sites except TLW. TSR and PPTN explained 78%, 49% and 84% of the variation in the long-term DOC pattern at Dorset, ELA, and Nova Scotia (NS) sites, respectively. In contrast, the long-term pattern in DOC at TLW only had a weak relationship with the regional and global variables considered. A General model was developed to compare the strength of the response of DOC to the regional variables among sites. Therefore, only the variables which had a significant linear correlation with DOC across sites were selected. If a site had no variables in common with other sites, it was excluded from the general model. This resulted in TLW being excluded from the general model because the long-term DOC pattern at TLW was not significantly correlated with any regional variables. The best general model included TSR from Dorset, ELA and NS sites and precipitation from only the NS site. The strengths of the response of DOC to precipitation were weak at Dorset and ELA compared to NS, therefore, they were excluded. The general model explained 91% of the site-to-site variation of DOC among sites. Among them, TSR was an important negative variable which contributed 56% of the explanation to the general model. Precipitation at NS was an important positive variable for the general model. It contributed 34% of the explanation to the model. As the response of the long-term DOC pattern to the changes of environmental variables (TSR and PPTN) was very strong at NS, the NS site dominated the general model, and its temporal (year-to-year) variation in the long-term DOC pattern explained 60% of the site-to-site variation of DOC in Eastern Canada. The other two sites, Dorset and ELA, had weak contributions (20% and 11%, respectively) to the general model.
DegreeMaster of Science (M.Sc.)
CommitteeDavis, John-Mark; Neal, Dick; Sheard, John; Bonham-Smith, Peta
North Eastern North America
dissolved organic carbon