Estimation of surface air temperature in the Canadian Arctic Archipelago using a semi-empirical, knowledge-based "topoclimatic" model

The existing network of Atmospheric Environment Service Canada (AES) weather observing stations in the Canadian Arctic Archipelago (CAA) is of low density and possesses coastal bias. These factors limit the applicability of surface data recorded at these sites for use in short-term analyses because 1) a data set of low spatial density limits the resolution of subsequent analyses, and 2) coastal bias further reduces station representativeness. A topoclimatic model has been developed to estimate surface air temperatures at high spatial resolution for this region. Model operation centers around solving an equation that represents the mean environmental lapse rates using a high-resolution digital elevation model (USGS GTOPO30). Mean environmental lapses rate for time periods being modeled are based on functions fitted to upper-air temperature ascents taken at nine stations throughout the CAA and northern Alaska. These functions are interpolated over the entire region, such that there exists a lapse rate function that is unique to each pixel in the elevation model. The effects of mean wind direction and speed, site proximity to water and ice fields, and 'thermal mass' of smaller islands are factored in to the final estimate. Model results agree well with comparisons conducted against surface observations from AES stations and against a larger network of non-standard, short-term observing stations operated by the Polar Continental Shelf Project, a Canadian federal organization that provides logistial support to arctic research.