On the need of bias adjustment for more plausible climate change projections of extreme heat

ABSTRACT: The assessment of climate change impacts in regions with complex orography and land-sea interfaces poses a challenge related to shortcomings of global climate models. Furthermore, climate indices based on absolute thresholds are especially sensitive to systematic model biases. Here we asse...

Descripción completa

Detalles Bibliográficos
Autores: Iturbide Martínez de Albéniz, Maialen|||0000-0002-5048-0941, Casanueva Vicente, Ana|||0000-0002-7568-0229, Bedia Jiménez, Joaquín|||0000-0001-6219-4312, Herrera García, Sixto|||0000-0002-5384-179X, Milovac, Josipa|||0000-0001-5648-5279, Gutiérrez Llorente, José Manuel
Tipo de recurso: artículo
Fecha de publicación:2022
País:España
Institución:Universidad de Cantabria (UC)
Repositorio:UCrea Repositorio Abierto de la Universidad de Cantabria
Idioma:inglés
OAI Identifier:oai:repositorio.unican.es:10902/23915
Acceso en línea:http://hdl.handle.net/10902/23915
Access Level:acceso abierto
Palabra clave:Bias correction
Climate impacts
Climate indices
CMIP5
Descripción
Sumario:ABSTRACT: The assessment of climate change impacts in regions with complex orography and land-sea interfaces poses a challenge related to shortcomings of global climate models. Furthermore, climate indices based on absolute thresholds are especially sensitive to systematic model biases. Here we assess the effect of bias adjustment (BA) on the projected changes in temperature extremes focusing on the number of annual days with maximum temperature above 35°C. To this aim, we use three BA methods of increasing complexity (from simple scaling to empirical quantile mapping) and present a global analysis of raw and BA CMIP5 projections under different global warming levels. The main conclusions are (1) BA amplifies the magnitude of the climate change signal (in some regions by a factor 2 or more) achieving a more plausible representation of future heat threshold-based indices; (2) simple BA methods provide similar results to more complex ones, thus supporting the use of simple and parsimonious BA methods in these studies.