The Relationship between Temperature and Rainfall Variability in the Levubu Sub-Catchment, South Africa

Abstract: A study has been done in the Levubu sub-catchment with the aim of investigating the relationship between temperature and rainfall variability. Monthly rainfall and temperature data for the period 1964/65-2009/2010 were used for this study. Four rainfall stations located in the Levubu sub-catchment and one temperature station were used in the study. The Standardization Procedure, Pearson’s correlation analysis and Cross correlation analysis techniques were employed. The latter two methods were used to establish relationships between rainfall and temperature in the study area. The study revealed the occurrence of an extremely wet period (1999/2000) and two moderately dry periods (1982/83 and 1992/93) within the study period. Wet events were predominant during the first half of the study period and mostly dry events followed thereafter. Correlation analysis was positive when rainfall and temperature were analyzed on a monthly time-scale but negative on an annual time-scale. Annual rainfall was correlated with annual maximum and minimum temperatures and the results were moderately negative r<-0.5. monthly="" rainfall="" was="" correlated="" with="" maximum="" and="" minimum="" temperatures="" the="" results="" were="" positive="" r="">0.8. Annual rainfall was cross correlated with annual maximum and minimum temperatures and the results were moderately positive 0.8 at 95% confidence limits. Monthly rainfall was cross correlated with monthly maximum and minimum temperatures and the results were greater than 0.85 at 95% confidence limits. A pattern in the behavior of rainfall and temperature with respect to one another was also established. During cool years, rainfall was seen to be high, while high temperatures were associated with below average rainfall. The study recommends that further studies on rainfall and temperature with more data sets need to be done before future climate predictions can be made using high resolution time scales such as daily data analysis.