Regional, national and international climate modeling


We have been applying the thin plate spline smoothing algorithms (ANUSPLIN) developed by Dr. Michael Hutchinson of The Australian National University to Canadian and U.S. climate data. This work is in partnership with Dr. Hutchinson, the Monitoring Strategies and Data Management Division and Climate Research Division of Environment Canada and Custom Climate Services Inc. We thank them for access to data, specialized preparation of station observations, and support with the assessment of climatological and meteorological validity of our spatial climate models. We also would like to thank staff at the US National Climatic Data Center for their assistance with US data for our North American models.

Thin plate splines should not be confused with simple univariate cubic splines. They can be thought of as a non-parametric, multi-dimensional curve fitting technique for application to noisy multi-variate data. ANUSPLIN offers an operationally efficient means to develop spatially continuous climate models ("surfaces"). This is especially important to many forest applications because weather station data are rarely available in remote forest locations.

The approach is also useful for mapping climate. Gridded climate data is often used for running spatial process models and other ecological applications. Many articles are available which describe the methods and applications in detail.

Our aims are to support climate impact and climate change analyses in forestry including species modeling, exotic species risk assessments and productivity studies.

Many of our climate grids are available for downloading, others require special permission because they have not been completely tested and/or peer reviewed. Point estimates at specific locations (eg. research plots, field locations) may also be available upon request from Dan McKenney