Relationship between global temperature, precipitation and atmospheric circulation patterns.

MOIRA DOYLE; VICENTE BARROS

Conferencia; 11th International Conference on Southern Hemisphere Meteorology and Oceanography; 2015

During the second half of the twentieth centuryhas undergone a continuous increase in many areas of Argentina. It has beenshown that in most of subtropical SouthAmerica, east of the Andes, precipitation trends during the extreme phases ofthe ENSO constitute only a small part of the trends. In particular, in northeastern Argentina, southern Brazil andParaguay, half of the annual rainfall trend comes from El Niño phase with a minor contribution from LaNiña and the neutral phase. Another factor that contributes to increase rainfall is the change in thepredominant sea level pressure circulation  pattern over south America and theAtlantic ocean. The mode that advects humid air fromthe Atlantic Ocean over most of Argentina and Uruguay has grown at the expenseof the one that favors the flow from the tropical forest over the eastern partof the continent.However these changes do not completely explain the observed precipitationtrend in the region. In this work we extend the period of analysis to includethe first decade of the 21st century and examine the predominant patterns ofthe atmospheric circulation as described by the 925hPa stream function andtheir relation with precipitation in southeastern South America. However, sincechanges in global temperature have effects on the atmosphere´s circulation aswell as rainfall regimes the relationship between these three variables isexplored. In this sense, a principal component analysis is applied to the925hPa stream function over south America between 10-40°S and 70-30°W duringthe warm months of the southern hemisphere, i.e. October -March. The first 6components, which explain 98% of the variance are rotated using normalizedvarimax to maintain orthogonality among the new variables. The first twocomponents are consistent with the main sea level patterns described by otherauthors and described above and are correlated with local precipitation.However patterns 3 and 4 are those which present the highest significantcorrelations with the annual global temperature. In order to determine how muchof the rainfall variance is explained by the circulation patterns and by globaltemperature linear regression models are fitted to each rainfall gridpoint andtrends for the estimated and observed time series are calculated and analyzed.