RESUMEN La ocurrencia de eventos secos en la región Sudeste de Brasil (SEB, por sus siglas en inglés) durante el verano (temporada de lluvias) se ha evidenciado en los últimos años, principalmente debido a eventos extremos previos en las temporadas 2013/14 y 2014/15. Los análisis de sequía se suelen realizar con datos mensuales. Aquí nuestra metodología aborda el tema con datos diarios para generar un análisis exhaustivo. Los eventos secos se evaluaron en diferentes subregiones de precipitación homogénea dentro del SEB, durante 37 temporadas diciembre-febrero (DJF, por sus siglas en inglés) y con dos escalas de tiempo diferentes de duración: sinóptica (5-9 días) e intraestacional (≥10 días). Se encontraron dos patrones dinámicos principales distintos para los eventos secos en las partes sur y centro-norte de SEB, respectivamente, pero no se identificaron diferencias importantes en las diferentes escalas de tiempo de ocurrencia. Los eventos del sur se caracterizaron por una cresta estacionaria que actuaba sobre todo el sur de América del Sur, lo que dificultó la aproximación de los sistemas transitorios al SEB del sur. Al mismo tiempo, este patrón mostró una configuración de Zona de Convergencia del Atlántico Sur (SACZ, por sus siglas en inglés) desplazada hacia el norte. En los eventos centro-norte, una alta presión centrada entre las regiones sur y sureste de Brasil se asoció con las condiciones de sequía. También se verificó un desplazamiento anómalo hacia el sur de los sistemas meteorológicos característico del verano sudamericano para estos eventos. Sobre el Atlántico Sur, se identificó una configuración de anomalía SST opuesta entre los eventos del sur y del centro-norte.

ABSTRACT The occurrence of dry events in the Southeast Region of Brazil (SEB) during summer (rainfall season) has been in evidence in the last years, mainly due to previous extreme events in the 2013/14 and 2014/15 seasons. Drought analyses are usually carried out with monthly data. Here our methodology addresses the issue with daily data to generate a thorough analysis. Dry events were evaluated in different homogeneous precipitation sub-regions within the SEB, over 37 December-February (DJF) seasons and with two different timescales of duration: synoptic (5-9 days) and intraseasonal (≥10 days). Two main distinct dynamic patterns were found for dry events in southern and central-northern parts of SEB, respectively, but no significant differences were identified in the different timescales of occurrence. Southern events were characterized by a stationary ridge acting over the whole of southern South America, making the approximation of the transient system to southern SEB difficult. At the same time, this pattern showed a northern-shifted South Atlantic Convergence Zone (SACZ) configuration. In the central-northern events, a high pressure centered between Brazil’s South and Southeast regions was associated with the dryness conditions. An anomalous southward shift of meteorological systems characteristic of the South American summer was also verified for these events. Over the South Atlantic, an opposite SST anomaly configuration was identified between southern and central-northern events.

Resumo A rede de estações meteorológicas no Brasil é muito recente, com poucas estações no país com mais de 100 anos de dados. Além disso, muitas das séries oriundas destas estações apresentam falhas, seja por falta de medição (no caso das estações convencionais), seja por falha nos equipamentos (no caso das estações automáticas). Encontram-se diferentes métodos na literatura para o preenchimento destas falhas. Neste trabalho são comparados seis métodos, os quais foram aplicados a séries de dados mensais conhecidas de temperatura média compensada e precipitação para estações meteorológicas localizadas no estado do Rio Grande do Sul, região sul do Brasil. Os resultados para temperatura média compensada indicaram os métodos de regressão linear múltipla e ponderação regional como os mais adequados para estimativas de dados faltantes enquanto que para precipitação não houve um método a se sobressair.

Abstract The network of weather stations in Brazil is very recent, with few stations in the country possessing data over more than 100 years. In addition, many of the series from these stations present missing data, either due to lack of measurement (in the case of conventional stations) or due to equipment failures (in the case of automatic stations). There are different methods in the literature for filling these gaps. In this work, six methods are compared and applied to series of known monthly data for mean compensated temperature and precipitation for weather stations located in the State of Rio Grande do Sul, in Southern Brazil. The results for mean compensated temperature suggest that multiple linear regression methods and regional weighing are best suited for estimating missing data, while for precipitation there was no superior method.