Naslov
Bias correction of monthly temperature and precipitation from regional climate model – calibration and validation

Autori
Sokol Jurković, Renata ; Güttler, Ivan

Vrsta, podvrsta i kategorija rada
Sažeci sa skupova, sažetak, ostalo

Izvornik
Kratki sažetci Meteorološki izazovi 6: Napredne tehnologije u rješavanju meteoroloških izazova / - , 2018, 49-49

Skup
Znanstveno-stručni skup s međunarodnim sudjelovanjem: Meteorološki izazovi 6: Napredne tehnologije u rješavanju meteoroloških izazova

Mjesto i datum
Zagreb, Hrvatska, 15.11.2018. - 16.11.2018

Vrsta sudjelovanja
Poster

Vrsta recenzije
Domaća recenzija

Ključne riječi
bias correction, regional climate model, calibration, validation

Sažetak
The use of climate projections in different impact models requires input with very low or without any bias that usually global and regional climate models contain. Therefore, the bias in climate model results should be corrected. Data corrected by univariate bias correction methods, i.e. methods which correct each variable individually (e.g. temperature and precipitation), in combinations that are found in impact models can cause unrealistic final results. The aim was that the physical relationship between the two variables confirmed in the observations retains in the model results, that is, the variables should be corrected together to preserve the physical relationship between them. Using bivariate bias correction method (variables are corrected simultaneously) with Gauss copula function and gamma and normal distribution as marginal distribution for precipitation and temperature, respectively, the regional model was corrected. The RegCM4 regional model with boundary conditions from the MOHC- HadGEM2-ES global model was corrected. Monthly precipitation and mean air temperature correction were performed for each season. Historical simulations were used and 1971-1990 as calibration and 1991-2004 as validation period. The correction was performed according to E-OBS with data at 0.25×0.25°. Spearman's coefficient of correlation was observed as a validation measure. In comparison to univariate bias correction, results show significant improvement in Spearman’ s correlation coefficient after applying bivariate method.

Izvorni jezik
Engleski

Znanstvena područja
Geofizika

POVEZANOST RADA