Abstract:
Understanding the behavior of the climatic parameters of a geographical area is a
prerequisite for any effort towards developing the agriculture sector. Influence of stochastic
and seasonal elements driving climatic dynamics pose a challenge to the physical
geographer who is trying to build up a model of understanding as such. Of particular
concern in these regards is the need to comprehend the, rainfall patterns in the dry-zone
regions of Sri Lanka, which also happen to be extensive paddy producing regions. This
paper focuses on a case study of the Anuradhapura district, and provides a time-series
modeling approach to understand the behavior of rainfall patterns for the period of January
1989 till December 2004.
Adopting a singularly focused approach as per the Occam's Razor Test in scientific
logic, this paper explains the development of a uni-variate autoregressive model for
analyzing the temporal behavior of historical rainfall data, with the ultimate aim of
arriving at a methodology for forecasting rainfall trends to aid in agricultural-decision
making in the Anuradhapura District. This approach provides the analyst the advantage of
treating the system as black box to evaluate the apparent-predominant functional behavior
without knowing another inter-connective and circular factors which influence the system.
Key findings indicate that the rainfall patterns in the study area are non auto-regressive, as such - they do not depend on the past history of rainfall; but are
predominantly depending on nonlinear trend and a seasonal pattern of order 12. This
indicates that in order to arrive at a comprehensive forecasting model for rainfall in
Anuradhapura, the need to focus on the influences of non-endemic and regional to global
climate phenomena is apparent.
