Temperature in Latin America varies depending on the different subregions. For tropical Latin America, temperature depends on cloud cover and altitude; for other subregions, altitude, advection, and, at the southern cone, sea-surface temperature (SST) play more predominant roles. The following description is based on observed records; where possible, paleoclimatic information has been included to present a wider view.
Central America shows different signs for temperature trends, according to the specific area under analysis. For example, in Costa Rica, Alfaro (1993) identifies a positive trend in daily maximum temperature. Gómez and Fernández (1996) and OCCH (1999) have identified negative trends for large areas of Costa Rica and Honduras. MARENA's (2000) analyses of time series for Nicaragua find only a small increase in mean temperature for Managua, which might be associated with growth in urbanization.
For northwestern South America, monthly mean air temperature records show a warming of 0.5-0.8°C for the last decade of the 20th century (Pabón, 1995a; Pabón et al., 1999; Quintana-Gomez, 1999). Colombia also presents increasing trends in the time series for the daily series of daily mean and minimum temperature for the past 30-40 years. Similar patterns have been observed in average monthly dew point and relative humidity (Mesa et al., 1997; Pérez et al., 1998). Coastal cities from northern Peru presented increases in air temperature since 1940, where 16 El Niño events were reported (Jaimes, 1997; SENAMHI, 1999).
In several cities in southern and southeastern Brazil, studies on long-term tendencies for air temperature, from the beginning of the 20th century, have indicated warming tendencies (Sansigolo et al., 1992). This could be attributable to urbanization effects or to systematic warming observed in the south Atlantic since the beginning of the 1950 s (Venegas et al., 1996, 1998). In the Amazon region, Victoria et al. (1998) have detected a significant warming trend of +0.63°C per 100 years.
Data since the beginning of the 20th century do not show a clear tendency in mean temperature in the southern cone, but there is a decrease in the thermal range. Moreover, south of 50°S there are indications of a positive tendency (Hoffman et al., 1997). However, when a shorter record is used for the analysis, Argentina and Chile show a large warming rate of 1.2-3.0°C per 100 years (Rosenblüth et al., 1997).
In south tropical Argentina, warming is observed only during the austral autumn season (Bejarán and Barros, 1998). The Argentina humid pampa, represented by Buenos Aires, presents a warming as a result of urban effects (Camilloni and Barros, 1997). Intensity and persistence of heat and cold waves present tendencies in which the sign depends on the region (Rusticucci and Vargas, 1998).
In extra-tropical west South America (Chile), surface air temperature has varied differently during the 20th century. South of approximately 45°S, temperatures have been increasing in stepwise fashion (Aceituno et al., 1993). In the area spanning about 35°S to 45°S, the most significant feature is a well-defined cooling of 1-2°C from the 1950s to the mid-1970s.
At decadal scales, multiple climate records throughout Latin America consistently exhibit a shift in the mean during the mid-1970s. This could be a climatic consequence of sudden changes in the climatology of the Pacific Ocean (Trenberth, 1990).
Using tree-ring and glacial evidence, summer temperatures in northern Patagonia show distinct periods of higher and lower temperatures during the past 1,000 years (Villalba, 1994; Villalba et al., 1997). For instance, there was a cold interval from AD 900 to 1070, followed by a warm period from AD 1080 to 1250 (coincident with the Medieval warm period). Warm climatic episodes similar to that observed during the 1980s may have occurred in the recent past under preindustrial CO2 levels in northern Patagonia (Chile and Argentina) (Villalba et al., 1997).
Latin America, in general, shows important variations in temperature, some of which might be connected to change in climate. At the same time, these variations might depend on the origin and quality of the data as well as the record periods used for the studies.
Other reports in this collection