Utilizing North American Regional Reanalysis for modeling soil moisture and evapotranspiration in reconstructed watersheds

Abstract

Oil sands development in Canada presents overwhelming challenges for boreal forest conservation and reclamation due to the disturbance of watershed function. Modeling hydrological processes for reclaimed landscapes is essential for assessing the performance of different reclamation strategies and their evolution over time, and requires a reliable continuous source of input data. This study evaluated the utility of precipitation and temperature data from the North American Regional Reanalysis (NARR) for the hydrological modeling of two reconstructed sites located north of Fort McMurray, Alberta, Canada. NARR precipitation and air temperature data were in good agreement with the observed onsite datasets. A generic system dynamics watershed (GSDW) model was initially calibrated using the onsite meteorological data, resulting in simulated soil moisture values that slightly overestimated observed values. The model was re-calibrated using the NARR temperature and precipitation data as inputs while the remaining data (e.g., net radiation, wind speed, and relative humidity) were obtained from the weather station. The re-calibrated model was a good alternative to onsite weather station data for simulating soil moisture patterns and evapotranspiration fluxes, particularly in cases of data scarcity. The re-calibrated model was used along with the NARR long term dataset (1979–2006) to evaluate the current/future hydrological performance of existing reconstructed watersheds with respect to soil moisture deficit using a probabilistic framework, and the results compared with long term performance estimates using the onsite-calibrated GSDW model. The study showed overall average correlation coefficients of 0.96 and 0.64 between the NARR and the onsite observed datasets for temperature and precipitation, respectively, and average root mean squared error (RMSE) values of 4.2 °C and 2.3 mm for the same datasets for the period of 2000–2006. The current study demonstrates and validates the application of NARR data to hydrological modeling for future reclaimed sites in the same region. Adopting NARR data for the assessment of the long term hydrological performance of reconstructed watersheds can inform decision making with respect to the design of reconstructed soil covers.