Evaluating the suitability of three gridded‐datasets and their impacts on hydrological simulation at Scotty Creek in the southern Northwest Territories, Canada

• Published in: Hydrological Processes
• Main Authors: Persaud, Bhaleka D., Whitfield, Paul H., Quinton, William L., Stone, Lindsay E.
• Other Authors: Natural Sciences and Engineering Research Council of Canada
• Format: Article in Journal/Newspaper
• Language: English
• Published: Wiley 2020
• Subjects: Northwest Territories
• Online Access: http://dx.doi.org/10.1002/hyp.13663; https://api.wiley.com/onlinelibrary/tdm/v1/articles/10.1002%2Fhyp.13663; https://onlinelibrary.wiley.com/doi/pdf/10.1002/hyp.13663; https://onlinelibrary.wiley.com/doi/full-xml/10.1002/hyp.13663

Abstract In the southern Northwest Territories (NWT), long time series of historical observations of climate and hydrology are scarce. Gridded datasets have been used as an alternative to instrumental observations for climate analysis in this area, but not for driving models to understand hydrological processes in the southern NWT. The suitability of temperature and precipitation from three‐gridded datasets (Australian National University Spline [ANUSPLIN], ERA‐Interim, and Modern‐Era Retrospective Analysis for Research and Application, Version 2 [MERRA‐2]) as forcings for hydrological modelling in a small subcatchment in the southern NWT are assessed. Multiple statistical techniques are used to ensure that structural and temporal attributes of the observational datasets are adequately compared. Daily minimum and maximum air temperatures in gridded datasets are more similar to observations than precipitation. The ANUSPLIN temperature time series are more statistically similar to observations, based on population statistics and temporal structure, than either of ERA‐Interim or MERRA‐2. The gridded datasets capture the seasonal and annual seasonal variability of precipitation but with large biases. ANUSPLIN precipitation compares better with observations than either ERA‐Interim or MERRA‐2 precipitation. The biases in these gridded datasets affect run‐off simulations. The biases in hydrological simulations are predictable from the statistical differences between gridded datasets and observations and can be used to make informed choices about their use.