Pesquisa Agropecuária Brasileira
versão On-line ISSN 1678-3921
LYRA, Gustavo Bastos et al. Homogeneous regions and probability density functions of pluvial precipitation in Táchira State, Venezuela. Pesq. agropec. bras. [online]. 2006, vol.41, n.2, pp. 205-215. ISSN 1678-3921. http://dx.doi.org/10.1590/S0100-204X2006000200004.
The objective of this work was to determine regions of monthly pluvial precipitation based on the seasonal and the probability distribution that best fit to precipitation of those areas in the State of Táchira, Venezuela. Long-term series of 24 throughout 62 years of monthly precipitation data of 25 climatological stations were used. The Ward's clustering methods of analyses was used to group the months with similar monthly pluvial precipitation and also the climatological locations with similar precipitation (homogeneous regions). The adjust of the probability density functions exponential, Gamma, Gumbel, normal, log-normal were evaluated with three parameters as well as Weibull distributions in order to observe monthly precipitation data. The precipitation seasonal variation at the State of Táchira presents three periods statistically defined as: dry, transition, and wet. For the dry and wet periods, four homogeneous regions of locations with similar monthly precipitation could be identified and, in the transition period, three regions. In the dry period, the recommended probability distribution is the exponential, except for the homogeneous regions with the largest rainfall values during that period, when Gamma distribution is better. In the wet period, in all regions, normal distribution prevails, except for August, when Gamma prevails. As far as transition periods are concerned, Gamma distributions are better in April and normal distribution, in November.
Palavras-chave : rainfall; probability density functions; cluster analysis; Ward.