This is the first study to show the global Cut-off Low (COL) activity in 46 models participating in the Coupled Model Intercomparison Project Phase 5 (CMIP5) and Phase 6 (CMIP6). The COL historical simulations for the period 1979–2005 obtained from the CMIP5 and CMIP6 models and their ensembles are compared with the ERA5 reanalysis using an objective feature-tracking algorithm. The results show that the CMIP6 models simulate the spatial distribution of COLs more realistically than the CMIP5 models. Some improvements include reduced equatorward bias and underestimation over regions of high COL density. Reduced biases in CMIP6 are mainly attributed to the improved representation of the zonal wind due to the poleward shift of the subtropical jet streams. The CMIP5 models systematically underestimate the COL intensity as measured by the T42 vorticity at 250 hPa. In CMIP6, the intensity is still underestimated in summer, but overestimated in winter in part due to increased westerlies. The overestimation is enhanced by the finer spatial resolution models that identify more of the strong systems compared to coarser resolution models. Other aspects of COLs such as their temporal and lifetime distributions are modestly improved in CMIP6 compared to CMIP5. Finally, the predictive skill of climate models is evaluated using five variables and the Taylor diagram. We find that 15 out of the 20 (75%) best coupled models belong to CMIP6, and this highlights the overall improvement compared to its predecessor CMIP5. Despite this, the use of the multi-model ensemble average seems to be better in simulating COLs than individual models.

Abstract. Quality control of climate data obtained from weather stations is essential to ensure reliability of research and services based on this data. One way to perform this control is to compare data received from one station with data from other stations which somehow are expected to show similar behavior. The purpose of this work is to evaluate some visual data mining techniques to identify groupings (and outliers of these groupings) of weather stations using historical precipitation data in a specific time interval. We present and discuss the techniques' details, variants, results and applicability on this type of problem.

Resumo O objetivo deste estudo é avaliar os termos da equação da tendência da energia cinética do distúrbio (K’) nas previsões de 24 h e 48 h dos modelos regionais BRAMS, WRF e ETA. Estes modelos foram integrados operacionalmente no Centro de Previsão de Tempo e Estudos Climáticos (CPTEC). O período de avaliação foi de 1 de junho a 31 de agosto de 2016, sendo que o domínio utilizado foi a região da América do Sul. A resolução horizontal dos modelos é de 5 km e a temporal é de 6 horas. As condições iniciais e de contorno foram obtidas do Global Forecast System (GFS) do National Centers for Environmental Prediction (NCEP), com resolução horizontal de 0,25°. Os modelos foram inicializados apenas com a análise das 1200 UTC. Os modelos foram integrados com a versão não-hidrostática, sendo que o WRF utilizou a parametrização cumulus de Kain-Frisch, o ETA a parametrização culumus de Betts-Miller e o BRAMS Grell e Freitas. Em geral, os três modelos subestimam K’ nas latitudes médias, principalmente sobre os oceanos, porém o ETA é o que menos subestima e o WRF o que tem o maior viés negativo. Com relação aos termos de conversão baroclínica e barotrópica, além do termo de convergência do fluxo ageostrófico, os três modelos apresentam um padrão semelhante aos das análises numéricas do GFS.