Downscaling of the global climate change projections: Some approaches

Some approaches, used in the Republic of Moldova for downscaling of the GCM-scale climate changes to a finer resolution, are demonstrated. The proposed methods include development of climate change projections for a whole country, for any site within its territory, and for the satisfaction of the user requirements. The climate change value for a country is estimated as a weighted sum of the values in the nearby grid boxes, herewith "weight" is a function of territory being included in a corresponding box. To estimate a local value the statistical dependencies of climate change on latitude and longitude were studied. These dependencies, approximated by the second order polynomial, served as a basis to automate compilation of the climate change fields. As an example of more complete use of the models' output for the impact assessment, the statistical transformation of surface air temperature and precipitation changes into changes of some air humidity parameters (potential evaporation, coefficient of humidification, and climatic diagrams) is shown.