CONICET | Buscador de Institutos y Recursos Humanos

INVESTIGADORES

PENALBA Olga Clorinda

datos académicos

artículos

capítulos de libros

congresos y reuniones científicas

convenios, asesorías y/o servicios tecnológicos

informe técnico

congresos y reuniones científicas

Título:

How well do GCMs represent present climate characteristics in Southern South America?

Autor/es:

BETTOLLI ML, PENALBA O.C, KRIEGER PA

Reunión:

Conferencia; WCRP-Conference for Latin America and the Caribbean, developing, linking, and applying climate knowledge.; 2014

Resumen:

The outputs of the new generation of global circulation models (GCMs) are available for different climate-related studies through the WCRP CMIP5 (World Climate Research Programme Coupled Model Intercomparison Project Phase 5). Dynamical and statistical downscaling are examples of these studies. Downscaling applications assume that GCMs realistically simulate climate conditions. In this context, it is important to evaluate the GCMs skills to simulate regional patterns, temporal variations and distribution of climate variables in order to assess their representation of the observed climatology in a specific region. The main objective of this study is to provide an evaluation of the GCMs in representing air temperature fields at 850 hPa in southern South America. Daily fields of air temperature at 850 hPa were considered in this analysis. They were found to be one of the main predictors of local maximum and minimum temperature over southern La Plata Basin using the analogue method for statistical downscaling. NCEP reanalysis 2 were used to represent observed temperatures for the period 1979-1999. The corresponding model outputs from the historical experiment were obtained from the WCRP CMIP5. The analysis was focused on austral winter (June, July, August) and summer (December, January, February). Reanalysis and GCMs fields were interpolated to the MIROC5 grid (1.4008 x 1.4063) with an inverse distance weighting method in order to facilitate comparisons. In order to analyze the ability of the GCMs to reproduce the basic characteristics of air temperature at low levels, the comparison between reanalyses and GCMs was performed by means of the principal component analysis of each dataset. Mean fields for summer and winter were also compared by means of Taylor diagrams, which convey information with clarity and quantify the degree of statistical similarity between two fields considering the correlation coefficient, the standard deviation and the root mean squared error. Results show that the main characteristics of air temperature at 850 hPa are reproduced by the GCMs. However, their performance depends on the model analyzed and the season of the year. The climatic characteristics of the mean temperature patterns, represented by the first PC of NCEP datasets, reflect the notorious seasonality that characterized temperature fields. The variance explained by each PC can be interpreted as a measure of the strength of each spatial pattern. The percentages of variance explained by the first PC for the NCEP dataset reach 92.4% and 88.9% for summer and winter respectively, indicating that the cold season is more perturbed than the warm season. In both seasons, the percentage of variance explained by the 15 GCMs is below than observed. Thus, GCMs tend to represent a lesser incidence of the mean pattern and, therefore, a higher presence of perturbations represented by high order PCs. Taylor diagram for winter shows the smallest dispersion of the cloud of points. Summer higher dispersion is mainly due to the range in standard deviation values of the GCMs fields. Correlation values and root mean square errors are similar in both seasons.

enviar mensaje