Oceanic influence on the precipitation of the south‐east of Venezuela

Abstract

AbstractThe Caroní catchment located in the south‐east of Venezuela accounts for 70 per cent of the total hydropower energy of the country. On a year to year basis, it has been shown that low frequency large scale ocean‐atmosphere phenomena are highly coupled to the hydroclimatology of the region, El Niño‐Southern Oscillation (ENSO) being a major forcing mechanism of climatic and hydrological anomalies. Regional differences in amplitude and timing are due to complex orographic interactions, land surface‐atmosphere feedback mechanisms and the evolution of dominant synoptic meteorological conditions. A detailed analysis of the relationship between rainfall and several large scale ocean‐atmospheric variables was carried out to determine the potential use of large scale climatic information as predictors of the rainfall anomalies over the region. The problem was tackled in two ways: (a) first a seasonal dynamic rainfall model was fitted to monthly rainfall for different locations. In this case rainfall is assumed as a normal variate w which has been transformed to account for its departure from normality and truncated to account for the positive probability mass of zero values, which corresponds to negative values of the normal variable. The time series of the model parameters and the macroclimatic variables are inspected for their potential relationship with local rainfall via the stochastic model. (b) Second, dynamic linear regression models between the macroclimatic variables as predictors and the rainfall anomalies as predictant were fitted to evaluate and quantify the significance of these dependencies. Consistent patterns are observed with the Tropical Atlantic and Pacific ocean temperature anomalies, in which a significant negative relationship has been present since 1976, indicating an overall decrease (increase) in rainfall when the Pacific and the Tropical Atlantic are warmer (colder) than normal. In all cases the results suggest that the relationships between rainfall anomalies and the macroclimatic variables are not constant with time. Copyright © 2002 John Wiley & Sons, Ltd.