2.2.1. Mann-Kendal Test

It is always essential to work out monotonic trends in the time series of any geophysical data before any further use. In this study, the Mann-Kendall test [30–32] was employed to detect the trends that exist in both the temperature and rainfall time series. This method is defined as a non-parametric, rank-based method which is commonly used to extract monotonic trends in the time series of climate data, environmental data or hydrological data. The Mann-Kendall test statistic gives information about the trend of the total time series and its significance. However, it is important to investigate how the trend varies with respect to time. Therefore, the calculation of the forward/progressive (*u*(*t*)) and backward/retrograde (*u* (*t*)) values of the Mann-Kendal test statistic is essential in order to investigate both the potential trend turning points and the general variability of trends in respect to time. This method is called the Sequential Mann-Kendall (SQ-MK), and it is well explained by [33] and other authors [31,34]. This method has been found to perform very well in trend analyses of stream flow and precipitation [35] and also in the field of earth remote sensing [36].

#### 2.2.2. Multilinear Regression

One of the primaries aims of this study is to identify the relationship between the studied time-series (temperature and rainfall at Conakry station in the present paper) and climate indices such as TNA, Niño3.4, AMM and AN. The multi-linear regression (MLR) is a method that is frequently used to explain the relationship between one continuous dependent and two or more independent variables (climatic indexes in this case). The MLR model output *yi* based on a number "*n*" of observations can be expressed as follow:

$$y\_i = \beta\_0 + \beta\_1 \mathbf{x}\_{i2} + \dots + \beta\_p \mathbf{x}\_{ip} + \varepsilon\_i. \tag{2}$$

where

$$\mathbf{i} = 1, 2, 3, \dots, n,\tag{3}$$

where in this case, *yi* is the dependent variable *xip* represents the independent variables, β<sup>0</sup> is the intercept, and β1, β2, ... β*<sup>p</sup>* are the *x*'s coefficients. The final term (ε*i*) represents the residual term which the model should always keep its contribution as minimum as possible.
