Precipitation variability in the inter-tropical Andes - Climatological studies for paleoclimatic reconstruction




Swiss-NSF 8220050401 P.I.: M. Vuille
NSF ATM-9707698, P.I.: R.S. Bradley

Project summary

Introduction
        This project focused on interannual to interdecadal climate variability in the tropical Andes where important ice core records have been recovered (Quelccaya and Huascaran in Peru, Sajama and Illimani in Bolivia) and where other new coring projects are planned (i.e. Chimborazo, Ecuador). The overall goal was to provide a better understanding of the circulation patterns and related oceanic forcing mechanisms, which are associated with precipitation and temperature variations (and associated isotopic changes) to help in interpreting the Holocene climatic record of tropical ice cores. Since the data situation is rather poor over this region, we put a lot of effort in establishing a new and dense climatic database of monthly temperature and precipitation data. This was crucial for the success of the project, because only a dense station network can provide sufficient information to resolve the spatial variability within the Andes, which separates two completely different climatic regimes (tropical Pacific and desert coastal areas to the west; moist tropical lowland areas of the Amazon and Chaco basins to the east). Scientists and Institutions from Ecuador, Peru, Bolivia, Chile, USA, France, Germany and Switzerland contributed their data to establish a dense database, which now consists of 411 (268) precipitation (temperature) records between 0° and 23°S along the Andean range.
Our studies were primarily based on Varimax rotated Principal Component Analysis (PCA) and simple Correlation and Compositing techniques. The PCA analysis was used to extract the main modes of interannual temperature and precipitation variability and to show the regional domains of the different modes. Correlation and compositing techniques were used to relate those modes to oceanic forcing in both the tropical Pacific and Atlantic domain and to verify the established relationships during extreme warm and cold phases of the respective oceanic basins. The following is a discussion of the main results from the two subareas that were analyzed in detail, Ecuador and the Altiplano, and from an analysis of temperature trends over the last six decades.

Altiplano
  The main spatiotemporal modes of interannual temperature and austral summer (DJF) precipitation variability in the Central Andes were identified based on a two-way principal component analysis (PCA) of 30-year (1961-1990) monthly station data and related to contemporaneous tropical Pacific and Atlantic sea surface temperature anomalies (SSTAs). In addition, various meteorological fields, based on National Centers for Environmental Prediction / National Center for Atmospheric Research (NCEP/NCAR) reanalysis, NOAA-Outgoing Longwave Radiation (OLR), Andean radiosonde and station data, were analyzed during periods of strong positive and negative SSTA and the respective composites tested for local significance using a Student’s t-test approach. In addition we showed how the atmospheric circulation over the Bolivian Altiplano changes during periods of extreme WET and DRY phases and during HIGH and LOW index periods of the Southern Oscillation, which is of special importance for the interpretation of the Nevado Sajama ice core.
Temperature variability in the Central Andes is primarily related to El Niño – Southern Oscillation (ENSO) and closely follows SSTA in the central equatorial Pacific with a lag of 1-2 months. In the southern Altiplano, temperatures have significantly increased since the late 1970s. DJF precipitation is also primarily related to ENSO, featuring below (above) average precipitation during El Niño (La Niña). Precipitation over the dry western part of the Altiplano shows the closest relationship with ENSO, due to ENSO-induced atmospheric circulation anomalies. Precipitation variability over the western Altiplano features a decadal-scale oscillation, related to a similar climatic shift in the tropical Pacific domain in the late 1970s. Over the northern Altiplano the precipitation signal is reversed in the austral summer following the peak phase of ENSO, presumably due to the temporal evolution of tropical Pacific SSTA, rapidly switching from one state to the other. No evidence for a tropical Atlantic influence on DJF precipitation was found. SSTAs in the tropical NE Atlantic, however, presumably are influenced by heating and convection over the Altiplano through an upper air monsoon return flow, altering the strength of the NE trades that emanate from the Sahara High.

Ecuador
        In the Andes of Ecuador (1°N - 4°S) the main spatiotemporal modes of seasonal precipitation and temperature variability were identified in a similar way, based on a PCA of monthly station data. The score time series of the main modes were correlated with tropical Pacific and Atlantic Sea Surface Temperature Anomalies (SSTA) to detect areas of significant oceanic forcing. To confirm the results, a reverse procedure was applied, by extracting the main modes of tropical Pacific and Atlantic SSTA, again by means of PCA. The score time series of these main SSTA modes were then correlated against station precipitation and temperature anomalies to see whether similar coherency patterns emerged. In most cases, this two-way approach strengthened the findings obtained and the results compared very favorably. A prime result of our analysis is, that despite the close proximity to the Pacific, precipitation variability in the Andes of Ecuador is not related to SSTA in the tropical Pacific domain alone. The El Niño-Southern Oscillation (ENSO) influence is most dominant in the northwestern part of the Andes and associated with below (above) average precipitation during El Niño (La Niña) years. However, precipitation along the eastern Andean slope is related to a dipole-like correlation structure in the tropical Atlantic, featuring positive correlations with SSTA to the south of the ITCZ, and negative correlations to the north. The proposed mechanism involves positive SSTA in the tropical South Atlantic and contemporaneous negative SSTA in the tropical North Atlantic, resulting in increased rainfall on the eastern Andean slopes. The only region with slightly increased precipitation during El Niño events is confined to a narrow area along the western Andean slope between 1°-3°S in close proximity to the Pacific. Temperature variability in the Andes can largely be explained by SSTA in the Pacific domain. The temperature response closely follows SSTA in the NINO3 and NINO3,4 regions with approximately one month lag.

Temperature trends
    Temperature trends in the tropical Andes are determined over the last six decades (1939-1998) based on a newly established station database including 268 stations between 1°N and 23°S and from 0 m to 5000 m a.s.l. Annual average temperature time series are determined using the First Difference method and trends are estimated using both an ordinary least squares (OLS) and a more robust least absolute residuals (LAR) approach. Temperatures in the tropical Andes have increased by 0.10 - 0.11°C/decade since 1939. The rate of warming has more than tripled over the last 25 years (0.32 - 0.34°C/decade). Of the last 13 years only one (1996) was below average, and the last two years of the series, associated with the 1997/98 El Niño, were the warmest of the last six decades. There is a general decrease of the observed warming trend with increasing elevation, except for the lower eastern slopes of the Andes, where the trend is insignificant. On the Pacific side, below 1000 m, temperatures have increased by 0.39 - 0.40°C/decade since 1959, while temperatures have increased only by 0.09 - 0.16°C/decade above 4000 m. However, despite the lower rate of warming, the trend toward increased temperatures is still significant at the 95% confidence level, even at the highest elevations.
 

Related Publications


 
 

University of Massachusetts | Department of Geosciences | Climate System Research Center | Mathias' Homepage | Current research