By Akpan, Nsikan Paul
The research centers on incidence of oil spill in Delta State. We considered
each spill as failure and the time between two successive spills as survival time.
Thus, survival analysis was the statistical tool used in the analysis of our data.
Analysis of failure time was done with respect to competing risks. Parametric and
semi-parametric methods were applied in the analysis. A two-state-first and
second order markov chain was used to obtain the probability of occurrences
within the period of the study.
Following from the analysis, the spillage pattern with respect to the
competing risks was obtained as follows: for smaller time lags, say less than eight
days, risk 2 competes higher than risk 1 while for higher number of days, say
greater than eight days , risk 1 competes higher than risk 2. For time lag of about
eight days , their rate of competition are the same. Also the first order two state
rnarkov chain showed the probabilities of a non spill day to anon spill day, non
spill day to a spill day, spill day to a non spill day and a spill day to a spill day as
being very high, approximately equal and quite low respectively. Equally, the
second order markov chain showed that the probabilities of a non spill day
following two non spill days, a spilled day following two spill days and a spill day
following two non spilled days as being very high, almost zero and low.