Statistical climate modeling

STAtistical Regional climate model (STAR) was developed to project regional climates for the near future by resampling observed records with the biased bootstrapping method. It generates a date-to-date mapping through which a date from the observational period is assigned a date of the future projection period to fit a prescribed linear temperature trend (or trend of any other variable) that serves as the only constraint. Not only projections of station data but also of meteorological/hydrological/climatological variables can be obtained.

Owing to short CPU time required, STAR can produce a large number of realizations for assessing projection-related uncertainties. It has been successfully applied for different regions in the world. For details please visit the project website of STAR - STAtistical Resampling scheme. We are currently applying STAR for the Tibetan Plateau, the Amazon basin, and the Tarim basin. Details follow soon.