Analysis of Trend and Periodicity in Long-Term Annual Rainfall Time Series of Nigeria - Hydrologic Time Series Analysis: Theory and Practice - page 251

gauges contributed to the dataset, which by 1981 has increased to 14,579
gauges. The CRU inserts synthetic zero anomaly values in regions that are
“too far” from observations (i.e., farther than 450 km), while the other schemes
simply interpolate over the entire distance. The annual rainfall time series data
were checked for consistency and normality prior to testing for trends and
periodicities.

12.3.2 Data Analysis

In this study, first of all, descriptive statistics of annual rainfall time series
including minimum, maximum, range, mean, standard deviation, variance,
standard error of the mean, kurtosis and skewness with their standard errors
were computed by using SPSS 15.0 software. These statistics help to provide
a preliminary overview of the dispersion and distribution of the data series.
Details about these descriptive statistics can be found in Chapter 2 of this
topic.
Normality tests are used to determine whether a dataset can be described
by a normal distribution or not, or to compute how likely an underlying
random variable is to be normally distributed. There are many reasons for
applying the normality tests to a hydrologic time series, which include data
screening, outlier identification, description, assumption checking, and
characterizing differences among sub-populations (groups of cases). Data
screening may show that you have unusual values, extreme values, gaps in the
data, or other peculiarities. Exploring the data this way may help to determine
whether the statistical techniques that are intended for data analysis are
appropriate. There exist different approaches to normality testing as mentioned
in Chapter 3. In this study, histogram, and box and whisker plot were used as
graphical methods, while the Kolmogorov-Smirnov test and the Shapiro-Wilk
test were used to check the presence of normality in the rainfall datasets.
Rainfall variability index is usually computed as the standardized
precipitation departure and helps to separate the available rainfall time series
into different climatic regimes such as 'very dry climatic year', 'normal climatic
year', 'wet climatic year', 'very wet climatic year', etc. (Lamb, 1982; L'Hóte
et al., 2002). Rainfall variability index (G) was calculated as (L'Hóte et al.,
2002; Oguntunde et al., 2006):

G PV (

P

)/ (1)

i

where G i is rainfall variability index for year i , P i is annual rainfall for year i ,
and P and V are the mean and standard deviation of annual rainfalls for the
1901-2000 period. When G is within ±0.5, the year is characterized as a
'normal year'; when G is between +0.5 and +1, it is characterized as a 'wet
year'; when G > +1, it is characterized as a 'very wet year'. Similarly, when
G is between -0.5 and -1, the year is characterized as a 'dry year', when G <
-1, it is characterized as a 'very dry year' (Lamb, 1982; L'Hóte et al., 2002).