Estudio de la incidencia de episodios de calima en Canarias mediante modelos climáticos globales

Abstract

At present, the study of atmospheric aerosols has aroused great interest, especially in places where, due to their geographical location, there are many invasions. One example is the Canary Islands, which suffers episodes of desert dust from the African continent. These episodes, known as calima, affect the radiative balance and cloud formation, as well as influence human health and ecosystems. The study of desert dust intrusions has evolved over the years thanks to advances in observational methods and numerical models. In the present study, the potential of the GCMs (Global Climate Models) of the new phase of CMIP (Coupled Model Intercomparison Project), CMIP6, will be evaluated. For this purpose, the results of the simulations of these models will be compared with the observations of recent past years. In particular, the only three models that have made daily data on dust concentrations available, i.e. IPSL, GFDL and MIROC6, will be used. These three models allow us the analysis of aerosol transport and generation through simulations. For the observations, the data studied are from MERRA version 2 (Modern-Era Retrospective analysis for Research and Applications). They are obtained from the reanalysis of space-based aerosol observations. For the simulated models and the observations we worked with column dust concentrations (kg m -2 ) and in order to study the incidence of calima episodes the data associated with a focused grid point in the Canary Islands was chosen. This study begins by studying the percentile associated with the concentration corresponding to an atmospheric aerosol episode, i.e. the 60th percentile. Once the percentile was determined, using the information on dust episodes provided by Ministerio para la Transición Ecológica y el Reto Demográfico, the monthly mean column concentration, the number of days above the 60th percentile and the number of days above the 95th percentile were analysed for two periods: the historical period and the future period. The historical period is from 1980 to 2009 and the future period is divided into two, mid-century (2030-2059) and late century (2070-2099). In addition, the SSP (Shared Socioeconomic Pathways) scenarios from CMIP6 describing CO2 concentrations in the future will be used for the future period. So first of all, the monthly averages of dust column concentrations in the historical period for the three CMIP6 models and for the MERRA2 measurements are compared, which allows us to discard the IPSL model for future simulations, as its behaviour is quite far from the observed one. Then the monthly averages of dust column concentration in the future are analysed for the GFDL models in the SSP585 and SSP245 scenarios and MIROC6 in the SSP126, SSP245, SSP370 and SSP585 scenarios. Since a general increase in the monthly mean dust column concentrations is observed, the number of days above the 60th and 95th percentile is studied to determine whether this increase is due to the increased intensity of the episodes or the duration of these intrusions. For the historical period, the IPSL model does not represent the stationarity of the observations and the number of days above the 60th percentile is much higher than for the 95th percentile. Therefore, it can be said that the calima episodes were not too intense in the past. For the future, an increase in the number of days with dust intrusions is generally observed for the two selected CMIP6 models, which indicates that the increase in the monthly mean is due to the longer duration of the dust intrusions and, to a lesser extent, to the intensity. So, in order to obtain more information on this matter, this work was finalised by studying the future trends for MIROC and GFDL. A study of the trends in annual dust column concentrations shows a gradual increase, which can be associated with more dust episodes as well as with an increase in intensity. Consequently, the trend in the number of annual days of extreme events (95th percentile) was analysed for both models and an increasing behaviour was observed. However, although the increase in desert aerosol concentrations can be related to the increase in the number of these episodes, a study of the average dust concentrations for the events in each year has been carried out. From this study, which turned out not to be statistically significant, it is possible to conclude that it cannot be considered an important cause for the growth of dust concentrations. Finally, it can be concluded, in first place, that the potential of the MIROC and GFDL models is favourable and consequently they postulate to be good simulators for the future. Furthermore, the future increase in frequency and intensity of desert dust intrusions is evident, in particular for the worst-case scenario concerning CO2 concentrations. Therefore, under the initial conditions and assumptions proposed, this work reflects the worsening of the calima episodes in the Canary Islands and stimulates contributing to the slowing down of climate change.