North American Precipitation Isotope (δ18O) Zones Revealed in Time Series Modelling across Canada and the Northern United States

Wednesday, 17 December 2014
Carly J Delavau1, Kwok Pan Chun2, Tricia A Stadnyk1, S Jean Birks3 and Jeffrey M Welker4, (1)University of Manitoba, Winnipeg, MB, Canada, (2)University of Saskatchewan, Saskatoon, SK, Canada, (3)Alberta Innovates-Technology Futures, Calgary, AB, Canada, (4)University of Alaska Anchorage, Department of Biological Sciences, Anchorage, AK, United States
Delineating spatial patterns of precipitation isotopes is increasingly important to understanding the processes governing the modern hydrologic cycle. However, the extent to which patterns can be empirically predicted across Canada and the northern United States (US) has not been fully articulated. This study focuses on creation of empirically derived δ18O precipitation (δ18Oppt) isoscapes at a monthly frequency across the study domain utilizing CNIP (Canadian Network for Isotopes in Precipitation) and USNIP (United States Network for Isotopes in Precipitation) measurements. Five regionalization approaches are used to separate the study domain into isotope zones to explore the effect of spatial grouping on empirical model parameterization and performance.

Time series simulations are evaluated at CNIP and USNIP stations, and model structure uncertainty and prediction bounds are assessed. Results show models are able to capture the annual cycle in observed δ18Oppt at Arctic and eastern stations, despite the large variability in observations. Arctic models have relatively simplistic parameterizations, with precipitable water and latitude consistently explaining the largest portions of δ18Oppt variability, suggesting fewer complicating factors influencing precipitation isotopes in northern latitudes. The eastern region requires more complex model parameterizations but still performs well, likely due to interannual consistency in this area. The continental west and north-west show less agreement between models, overall poorer model performance, and higher prediction uncertainty, suggesting more complex isotope-climate relations. This is the initial step towards empirically derived time-series δ18Oppt for iso-hydrological modelling; however results can provide insight into processes driving spatial and temporal δ18Oppt variability in Canada and the northern US, contributing to improved use of paleoprecipitation archives as indicators of hydro-climate variability.