Conditional probability , dry and wet periods weather forecasts recovery course using Markov chain model (Case study: Khuzestan province, Iran)

Authors
I. Salehvand
A. Gandomkar
E. Fattahi
Abstract
The aim of this study is the spatial analysis of rainfall in wet and dry months of the year in south-west
Iran especially khouzestan. To this end, the model of Markov chain is used. In Markov chain, it is assumed
that the climate has only two states: dry or wet. For instance, rainfall has only two states: it rains (1) or it
will not rain (0).There is not a third state. The statistical years are not the same in this research and change
from 10 to 50 years. Twenty stations are selected in Khuzestan and the surrounding provinces (synoptic
and climatology).The three months of the summer are not included in calculations because the precipitation
was zero or near zero. As the result of that, the reoccurrence of dry periods is generally longer than that
of the wet ones. In the other words, dry periods last a long time in south-west Iran. Regarding area, long
dry periods cover a wide area in the southwest while the area covered by short-term periods is less. When
the length of time of periods increases to possibly 8 days, the changes are reversed and the longer time the
probability of 8-day aridness takes, the more area it covers and the shorter time it takes, the less area it
covers.
Keywords
drought
Terry
Khuzestan
Markov chain