Delineation of precipitation regions using location and atmospheric variables in two Canadian climate regions: the role of attribute selection |
| |
| Authors: | Sarah Irwin Roshan K Srivastav Slobodan P Simonovic Donald H Burn |
| |
| Institution: | Department of Civil and Environmental Engineering, Western University, London, ON, Canada |
| |
| Abstract: | The identification of homogeneous precipitation regions is essential in the planning, design and management of water resources systems. Regions are identified using a technique that partitions climate sites into groups based on the similarity of their attributes; the procedure is known as regionalization. In this paper the ability of four attribute sets to form large, coherent precipitation zones is assessed in terms of the regional homogeneity of precipitation statistics and computational efficiency. The outcomes provide guidance for effective attribute selection for future studies in Canada. The attributes under consideration include location parameters (latitude, longitude), distance to major water bodies, site elevation and atmospheric variables modelled at different pressure levels. The analysis is conducted in two diverse climate regions within Canada including the Prairie and the Great Lakes–St Lawrence lowlands regions. The method consists of four main steps: (i) formation of the attribute sets; (ii) determination of the preferred number of regions (selection of the c-value) into which the sites are partitioned; (iii) regionalization of climate sites using the fuzzy c-means clustering algorithm; and (iv) validation of regional homogeneity using L-moment statistics. The results of the attribute formation, c-value selection, regionalization and validation processes are presented and discussed in a comparative analysis. Based on the results it is recommended for both regions to use location parameters including latitude, longitude and distance to water bodies (in the Great Lakes region) to form precipitation regions and to consider atmospheric variables for future (climate change) applications of the regionalization procedure. |
| |
| Keywords: | Spatial temporal regional frequency climate clustering L-moments atmospheric variables precipitation |
|
|
