• Home
• Program
• Search
• Author Index
• Sponsors
• Committees
• Contact Us
• Tech Support

Session:

Posters

Abstract No.:

P-018

Title:

Evaluation and Calibration of Dynamically Downscaled Precipitation over Norwegian Mainland

Author(s):

O. Haug, Norwegian Computing Center (NO)
E. Orskaug, Norwegian Computing Center (NO)
I. Scheel, University of Oslo (NO)
A. Frigessi, University of Oslo (NO)
D. Maraun, Kiel University (DE)
P. Guttorp, University of Washington (US)

Abstract:

The intensification of climate research over the past decade produces a steadily increasing number of data sets combining different global circulation models, CO2 emissions scenarios and downscaling techniques. Turning future projections into robust and reliable information available at a local scale is imperative for the successful modeling of impacts of climate change. The comprehensive financial and safeguarding challenges of mitigation and adaptation call for thorough validation, improvement and extensions of current downscaling techniques. In the evaluation part of our work we report on a systematic approach to identify discrepancies in downscaled climate data. Using reanalysis data as forcing, we have investigated by statistical techniques how well the Norwegian regional model HIRHAM compares to triangulated and aggregated station measurement data on a 2525 km2 grid over Norwegian mainland. Methods considered are among others the Kolmogorov-Smirnov two-sample test, a Fisher exact test for equality of quantiles, an Extreme Value Theory test, where equality of the one-year return levels are tested, and equality of wet day frequency. The regional model is skillful in describing the lower quartile of the precipitation distribution, but underestimates higher levels of precipitation. Our results indicate that the regional model has too many but too small rain events for all seasons. The results of the evaluation underlines the need for enhanced climate projections at a local scale. Generally, discrepancies between the two distributions exist for the whole range of data, leaving demand for a full quantile calibration function. We address this issue through developing a spatially smoothed set of transformations that will make the model distribution closer to that of the observed precipitation data. The transfer functions between the two distributions are characterised using Doksum's shift function. Model parameters are estimated by means of Integrated Nested Laplace Approximations (Rue H., Martino S. and Chopin N., 2009).