3), so the mechanisms for climatic effects remain uncertain. We were limited in our analysis

to using climate variables based on monthly data and, therefore, could not assess storminess which may better relate to allochthonous sediment transfer. Although it is widely known that short-term rainfall events can be a more dominant control on sedimentation, the data constrained us to only explore the potential influence of long term precipitation change which XL184 supplier would largely control cumulative runoff at coarse temporal scales. Process-based studies of lake catchments are needed to understand the mechanisms of how climate-driven changes may affect sedimentation and to differentiate between autochthonous production and allochthonous inputs. The lack sediment source discrimination is a major limitation of our study. The Spicer (1999) analyses for Vancouver Island and central to eastern Interior Plateau lakes included systematic, LOI-based estimates of organic content. Regression models by Spicer (1999) yielded better fits between land use and inorganic sedimentation,

suggesting that forestry activities may have elevated mineralogenic sediment delivery. It is important to note, however, that changing organic fractions could also influence composition trends and that organic sediment sources can be aquatic or terrestrial based. Significantly more sediment analyses would be needed for any possible attempt of such discrimination. Inconsistent LOI measurements from our other regional records showed that organic matter tended to increase up core. Such a trend could be associated with increased Y-27632 chemical structure Pregnenolone autochthonous production or allochthonous inputs over time, both of which could be related to land use by nutrient or debris transfer. Alternatively, diagenesis could be influencing some of the sediment composition trends (e.g. decomposition of organics over time). To account for the potential effect of diagenesis or some other unknown linear control over time on the sediment records (Fig. 4) (e.g. a bias associated with the sampling or dating methods), we tried adding a

standardized time variable (interval year) as a fixed and random effect to our best models. For both the complete inventory and the Foothills-Alberta Plateau subset models, estimates of land use and temperature fixed effects were greatly reduced, although most remained as positive coefficients. Even with this addition of a linear trend in time, the continued inclusion of all fixed effect variables continued to yield better overall models (based on AIC), than with any combination removed. This could further support the land use and climate relations with sedimentation; however, those environmental changes are correlated with time and multicollinearity inhibited model interpretation. We noted that model fits were significantly improved with time included, suggesting that a highly time correlated process or methodological artifact remains undefined.