Using neural networks to assess the influence of changing seasonal climates in modifying discharge, dissolved organic carbon, and nitrogen export in eastern Canadian rivers

Abstract

Changes in both air temperature and precipitation are expected as atmospheric CO2 increases because of anthropogenic activities. These changes will not be evenly distributed across seasons because of the complexity of climate patterns. Using climate, hydrology, and water chemistry data, we investigated how changes in the seasonal climates of the Atlantic Provinces of Canada would change both hydrology and geochemical cycling of 14 river basins influenced by wetlands. We calculated monthly discharge (Q) and dissolved organic carbon (DOC) and dissolved organic nitrogen (DON) exports from 1983 to 1992 and developed three neural network models relating climate variability to hydrologic and geochemical fluxes from the basins. We then identified a series of potential temperature/precipitation scenarios which we applied to the models. We then studied how changes in climate regimes would affect water discharge and aquatic dissolved carbon and nitrogen released from these basins. Our results predict a complex series of possible outcomes for water and elemental fluxes. Warming or cooling in each season combined with potential changes in precipitation will cause completely different outcomes as snowpack melting dominates winter and early spring conditions and evapotranspiration controls summer and fall. A warming winter would cause less snow storage and more runoff, with little evapotranspiration. Warmer summers, on the other hand, will be influenced by evapotranspiration, and water flows will be lower. Carbon and nitrogen exports from basins generally mirror those of Q. Our results suggest that major hydrological changes in basins may lead to significant ecological and water resource impacts, especially in the spring and early summer.