Home        >    Projects

Projects

Development of Weather Generators & Statistical Techniques for Downscaling Climate Extremes for Canada (2005)

This project was a mix of applied and basic work. One task undertook the improvement of existing weather generators. Another component of this project investigated the downscaling of extremes.

This project was a mix of applied and basic work. One task undertook the improvement of existing weather generators. Specifically, the authors tuned the WGEN weather generator for Canadian stations for the following variables: precipitation, temperature, wind speed, dewpoint temperature, and solar radiation. The results for most variables were quite satisfactory, but this weather generator does not accurately capture the variability of precipitation. The project addressed this issue by exploring the fundamental underpinnings of weather generators. Specifically, the researchers compared various mathematical/statistical approaches for the modelling of daily precipitation. They recommended specific approaches that appear to be suited to Canadian climate conditions. A manuscript describing the results has been accepted in Atmosphere-Ocean; a review by the evaluation team suggests that this work is methodologically sound and an original contribution to the state-of-the-art in weather generators. The authors state that FORTRAN software is available for simulating precipitation based on the results of their work. The evaluation team did not review the software, but in order for the results of the weather generator work to yield dividends, it is important that the software be user-friendly for the I&A community, many of whom will not be familiar with the mathematical basis of these tools.

Another component of this project investigated the downscaling of extremes. The researchers used sophisticated techniques to examine potential changes in extreme precipitation. Specifically, generalized extreme value functions are used to model the probability distributions of precipitation extremes. A unique feature is that atmospheric circulation measures are incorporated as co-variates in the parameters of the distribution. Changes in atmospheric circulation projected by GCMs were then input into the extremes model to estimate changes.

In another aspect of the extremes work, the researchers performed a study of methods for detecting trends. This study was published in the Journal of Climate. In the future, the team plans to study the uncertainty of scenarios and to tackle the special interpretation of weather generator parameters over data sparse regions.

The results of this project provide a sound mathematical/statistical basis for advancing the science of scenarios development in Canada. However, it is important that interactions between this group and other scenarios development groups continue so that these advances can be rapidly incorporated into appropriate efforts by other groups.

This project contributed to the training of graduate students and post-doctoral scientists.

This information is dated January 31st, 2005 and was found here

Home | Contact Us