^{1}

^{*}

^{2}

^{3}

^{3}

^{4}

Rainfall and temperature are climatic variables mostly affected by global warming. This study aimed to investigate the temporal trend analysis in annual temperature and rainfall in the Southern Togo for the 1970-2014 period. Daily and annual rainfall and temperature were collected from four weather stations at Atakpame, Kouma-Konda, Lome, and Tabligbo. The temperature variability was determined by the Standardized Anomaly Index (SAI) and the annual rainfall variability was determined using the Standardized Precipitation Index (SPI). The Mann-Kendall test was used for trend analysis. Mann-Kendall statistical test for the mean annual, mean annual minimum and maximum temperature from 1970 to 2014 showed significant warming trends for all stations except Kouma-Konda where mean annual maximum temperature had exhibited non significant cooling trend (
*P* = 0.01). For Standardized Precipitation Index in the 12-month time scale, dry tendency dominates Atakpamé (55.7%) and Kouma-Konda (55.5%) while wet tendency dominates slightly Lomé (50.9%) and Tabligbo (51.4%). The Mann-Kendall test revealed an increasing trend in standardized anomaly index at all the sites, prejudicial to rainfed agriculture practiced by about 90% of Togolese crop growers. The trend analysis in the climate variables indicated a change in climate that necessitates some specific actions for resources management sustainability and conservation.

Climate is the most important factor that governs food production and causes inter-annual variability in socioeconomic and environmental systems related to the availability of water resources [

An increasing trend in air temperature has been reported in different parts of the world including the tropical region in Africa. [

Precipitation is a key component in the hydrologic cycle that is affected in several parts of the world. Reduction in precipitation is becoming recurrent and many countries are concerned by the concept of climate change, which is becoming a national primordial challenge. Climate change has resulted in extreme drought conditions in some parts of the world and flooding in other parts [

The need for reliable statistics to assess the past and plan for the future is becoming increasingly important in Togo. To better understand climate change in Togo, its manifestations and to further monitor flood, drought occurrences and make agricultural planning easier, it is important to study trend analysis in standardized precipitation index (SPI) and standardized anomaly index (SAI). The present study aimed to evaluate: (i) the variability of annual rainfall using Standardized Precipitation Index (SPI), (ii) the variability in temperature using Standardized Anomaly Index (SAI), (iii) the trend of wet and dry periods using the Man-Kendall statistical test, (iv) the probability of occurrence of different categories of wet and dry periods.

Togo is a small West African nation with estimated population of about 6,191,155 inhabitants [

The study was conducted in Southern Togo, where four weather stations Lome (6˚9'56''N, 1˚15'16.24''E, elevation 22 m), Tabligbo (6˚34'59''N, 1˚30'00''E, elevation: 76 m), Atakpame (7˚31'37''N, 1˚7'36''E elevation 250 m) and Kouma-konda (6˚95'N, 0˚58'E, elevation 643 m) were selected for reliability and the long-term daily dataset without missing data covering the period 1970 to 2014. A record of monthly average climatic parameters including air mean, maximum and minimum temperature, were used to estimate the Standardized Anomaly Index and precipitation was used to estimate the Standardized Precipitation Index. All the data were collected from the National Meteorological Department of Togo.

Two homogeneity tests were applied on the data and the Skewness and Kurtosis coefficients were determined (Appendix A,

the 04 stations are approximately symmetric then homogenous. The Kurtosis test indicated that both temperature and precipitation are flat distributed. Moreover, the statistical package of standardized Precipitation index used, transformed the data into normal distribution before estimating the SPI values.

For each of the stations, annual mean temperature, mean annual minimum temperature and mean annual maximum temperature series were analyzed for fluctuation using Standardized Anomaly Index (SAI) which is a commonly used index for regional climate change studies [

where r is the mean temperature of the year, r_{i }is the long-term mean, and σ is the standard deviation of annual mean temperature for the long-term.

A period when below long-term average was dominated is considered as cooling period and a period when above long-term average was most persistent is a warming period.

The concept was developed by [

where

where

The parameters α and β are estimated using the following formulae

where

and

When the probability density function is integrated with respect to x using the estimates of α and β, a cumulative probability

Substituting t for

which is the incomplete gamma function. However, the gamma distribution function is undefined for

where

Finally, the cumulative probability distribution

However, due to the complexity of following these steps to compute SPI manually, the United States National Drought Mitigation Centre developed a program that computes SPI from monthly precipitation data at required time scales.

This SPI program (SPI_SL_6.exe) available at:

(http://drought.unl.edu/MonitoringTools/DownloadableSPIProgram.aspx) was downloaded and used in this research.

All negative SPI values indicate the occurrence of drought, while all positive values stand for wet periods. A table of SPI magnitude as presented in

SPI value | Interpretation |
---|---|

≥2.0 | Extremely wet |

1.5 to 1.99 | Severely wet |

1.0 to 1.49 | Moderately wet |

0.99 to −0.99 | Near normal |

−1.0 to −1.49 | Moderately dry |

−1.5 to −1.99 | Severely dry |

≤−2.0 | Extremely dry |

Source: [

Various time scales are used for the computation of SPI on which changes in precipitation can affect different aspects of hydrologic cycle. These time scales include 1, 3, 6, 9, 12, 24, 36 months. The choice of these time scales depends on the interest of research. In this study, the time series of SPI 12 month time scale was utilized to compute SPI for the precipitation data from 1970 to 2014 to determine the frequency of occurrence of wet/dry conditions and the risk of flood and drought in southern part of Togo.

For the analysis of temporal trend in standardized anomaly index and standardized precipitation index, the Mann-Kendall test [

where x_{i} is the data value at time i, n is the length of the dataset and sign( ) is the sign function which can be computed as:

For n > 10, the test statistic Z approximately follows a standard normal distribution:

in which Var(S) is the variance of statistic S.

A positive value of Z indicates that there is an increasing trend, and a negative value indicates a decreasing trend. The null hypothesis, Ho, that there is no trend in the records is either accepted or rejected depending on whether the computed Z statistics is less than or more than the critical value of Z statistics obtained from the normal distribution table at the 5% significance level [

If the data has a trend, the magnitude of trend can be denoted by trend slope β [

where x_{i} and x_{j} are data values at time t_{i} and t_{j} (i > j), respectively.

The temperature anomalies which occur in Atakpamé, Kouma-konda, Lomé and Tabligbo during 1970-2014 periods are described for the mean annual; mean annual minimum and maximum temperatures. Figures 2-4 show Standardized Anomaly Index for the mean annual, minimum and maximum temperature at (A) Atakpamé, (B) Kouma-Konda, (C) Lomé and (D) Tabligbo, respectively. For mean annual temperature, all stations were marked with below long term average between 1970 and 2000 indicating a period of cooling (

The anomalies in the mean annual minimum temperature (

Mean annual maximum temperature at all stations (

four (04) remarkable warming years during the first decades (1970-1980) followed by alternate warming and cooling years from 1980 to 1990. Thereafter, there was a remarkable cooling period until 2006. After 2006, there was warm period till 2014. The year to year departures were not pronounced, indicating the slight variation from the mean values [

The Standardized Anomaly index for min, max and mean temperature were further subjected for trend analysis using Mann-Kendall statistical test. Figures 5-7 show the trend analysis of Standardized Anomaly Index for the mean annual, minimum and maximum temperature at (A) Atakpamé, (B) Kouma-Konda, (C) Lomé and (D) Tabligbo, respectively.

Results in

Station | Temperature | Z statistic | Significance | Slope (Q) | Constant (B) |
---|---|---|---|---|---|

Atakpamé | Mean annual | 6.65 | *** | 0.064 | −1.34 |

Minimum | 6.39 | *** | 0.062 | −1.29 | |

Maximum | 6.04 | *** | 0.062 | −1.35 | |

Kouma-Konda | Mean annual | 4.07 | *** | 0.050 | −1.20 |

Minimum | 5.42 | *** | 0.063 | −1.37 | |

Maximum | -0.46 | n.s. | -0.006 | 0.061 | |

Lomé | Mean annual | 7.09 | *** | 0.071 | −1.52 |

Minimum | 6.26 | *** | 0.063 | −1.26 | |

Maximum | 5.99 | *** | 0.065 | −1.42 | |

Tabligbo | Mean annual | 5.63 | *** | 0.064 | −1.39 |

Minimum | 5.17 | *** | 0.057 | −1.20 | |

Maximum | 5.18 | *** | 0.064 | −1.44 |

Mann-Kendall test of significance Levels 90% (*), 95% (**), 99% (***), n.s., Non-significant.

rose by 1.1˚C, 1.6˚C, 2.2˚C and 1.2˚C at Atakpamé, Kouma-konda Lomé and Tabligbo respectively. The evolution of the maximum temperature is also important. From 1970 to 2014, the Tmax rose by 1.3˚C at Atakpamé, 1.7˚C at Lomé, 1.7˚C at Tabligbo and decrease by 0.13˚C at Kouma-konda. The warmest years started from 2000 in all locations. Moreover, the 1976-1977 and 1982-1983 years were also warm which confirms the droughts occurrence in these years leading to severe famine as revealed by [

Our results confirm the findings of [

perceived an increase of temperature in Togo from 1961 to 2013. Similar results were also found by [

The results of the SPI for 12-month time scale of Atakpamé, Kouma-konda, Lomé and Tabligbo are shown in

2007 was extended to 2010. At Kouma-konda, there were 31 cases of wet period which represents 43.7% of the study period, out of which 14 were moderately wet, 11 were severely wet and 6 were extremely wet. On the other hand, there were 40 cases of dry period which represents 56.3%, out of which 23 were moderately dry, 13 were severely dry and 4 were extremely dry.

The results of this study showed higher frequency of dry periods than wet periods. The incidence of the decrease of wet periods can be observed by the reduction of the crop growing season. Moreover, Southern Togo is the most

SPI | Category | Number of time in 45 years | |||
---|---|---|---|---|---|

Atakpamé | Kouma-konda | Lomé | Tabligbo | ||

≥2 | Extremely wet | 2 | 6 | 3 | 0 |

1.5 - 1.99 | Severely wet | 8 | 11 | 9 | 10 |

1 - 1.49 | Moderately wet | 19 | 14 | 22 | 30 |

0 - 0.99 | Mild wetness | 45 | 34 | 51 | 51 |

0 to −0.99 | Mild dryness | 47 | 41 | 42 | 44 |

−1 to −1.49 | Moderately dry | 26 | 23 | 21 | 19 |

−1.5 to −1.99 | Severely dry | 15 | 13 | 13 | 13 |

≤−2 | Extremely dry | 5 | 4 | 6 | 10 |

Total | 167 | 146 | 167 | 177 |

Locations | Number of case | Percentage of number of case | ||
---|---|---|---|---|

Wet | Dry | Wet | Dry | |

Atakpamé | 29 | 46 | 38.7 | 61.3 |

Kouma-Konda | 31 | 40 | 43.7 | 56.7 |

Lomé | 34 | 40 | 45.9 | 54.1 |

Tabligbo | 40 | 42 | 48.8 | 51.2 |

populated and where more industries are established (40% of population and 90% of industries) [

Despite the high occurrence of dry period, it appears some sporadic flooding. The highest was in 2007-2008 [

Monotonic increasing or decreasing trend in annual precipitation is tested with the non parametric Mann-Kendall test (Z-statistic) and secondly the slope of a linear trend is estimated with the non parametric Sen’s method [

The slight increasing and decreasing trend agrees with the fact that long time scales of SPI respond slowly and stably to the variation in daily rainfall [

According to the Z statistic for the four stations, it becomes evident that there is conformity in magnitude of the statistic when an altitude and ecological zone

factors are taken into consideration. The station of Lomé, Tabligbo and Atakpamé are below 400 m while Kouma-Konda is above 600 m. Morover, Kouma-Konda is situated in forest zone and deforestation during the last decades may have contributed to the slight decreasing of precipitation.

Based on the above results, it is of immense importance to discuss the ecological, economic, and social impacts that could result if increasing precipitation trends continue in these areas in the future. For coastal areas, in particular, vulnerability to loss of land, destruction of residence are major problems and therefore, coastal insurance is of great importance. The vulnerability might further be aggravated if extreme rainfall episodes continue in the future and consequently result in inland and coastal flooding. Institutional changes, coastal regulation, and management goals have to be, therefore, adapted in a timely manner. Increased precipitation can influence the water quality and possibly result in the outbreak of waterborne diseases due to sewage overflows (in case of combined sewers) and/or ineffectiveness of wastewater treatment systems to handle increased load [

The probability of occurrence of various categories of wet and dry periods at Atakpamé is shown in

SPI | Category | Number of time in 45 years | Percentage of occurrence | Severity of events |
---|---|---|---|---|

≥2 | Extremely wet | 2 | 1.2% | 1 in 22.5 years |

1.5 - 1.99 | Severely wet | 8 | 4.8% | 1 in 5.6 years |

1 - 1.49 | Moderately wet | 19 | 11.4% | 1 in 2.4 years |

0 - 0.99 | Mild wetness | 45 | 26.9% | 1 in 1 year |

0 to −0.99 | Mild dryness | 47 | 28.1% | 1 in1year |

−1 to −1.49 | Moderately dry | 26 | 15.6% | 1 in 1.7 years |

−1.5 to −1.99 | Severely dry | 15 | 9% | 1 in 3 years |

≤−2 | Extremely dry | 5 | 3% | 1 in 9 years |

Total | 167 | 100% |

SPI | Category | Number of time in 45 years | Percentage of occurrence | Severity of events |
---|---|---|---|---|

≥2 | Extremely wet | 6 | 4.1% | 1 in 7.5 years |

1.5 - 1.99 | Severely wet | 11 | 7.5% | 1 in 4.1years |

1 - 1.49 | Moderately wet | 14 | 9.6% | 1 in 3.2 years |

0 - 0.99 | Mild wetness | 34 | 23.3% | 1 in 1.3 years |

0 to −0.99 | Mild dryness | 41 | 28.1% | 1 in 1.1 years |

−1 to−1.49 | Moderately dry | 23 | 15.8% | 1 in 2.0 years |

−1.5 to −1.99 | Severely dry | 13 | 8.9% | 1 in 3.5 years |

≤−2 | Extremely dry | 4 | 2.7% | 1 in 11.3years |

Total | 146 | 100% |

years. This is followed by extremely wet year with percentage and probability of occurrence 4.1% and 1 in 7.5 years.

12-month time scale probability of occurrence of different categories of wet and dry periods at Lomé is shown in

More than half of the study period at Lomé (50.9%) and Tabligbo (51.4%) was revealed wet period while at inland stations such as Atakpamé (55.7%) and Kouma-Konda (55.5%), the dry period was dominant. The higher percentage of occurrence of mild wetness explained the wet period observed at Lomé (30.5%) and Tabligbo (28.8%) stations. The percentage of mild dryness explained also the dry period in plateau stations as Atakpamé (28.1%) and Kouma-Konda (28.1%). The climate at all stations is near normal according to the SPI classification by [

SPI | Category | Number of time in 45 years | Percentage of occurrence | Severity of events |
---|---|---|---|---|

≥2 | Extremely wet | 3 | 1.8% | 1 in 15 years |

1.5 - 1.99 | Severely wet | 9 | 5.4% | 1 in 5 years |

1 - 1.49 | Moderately wet | 22 | 13.2% | 1 in 2.1 years |

0 - 0.99 | Mild wetness | 51 | 30.5% | 1 in 0.88 year |

0 to −0.99 | Mild dryness | 42 | 25.2% | 1 in 1.1 years |

−1 to −1.49 | Moderately dry | 21 | 12.5% | 1 in 2.1 years |

−1.5 to −1.99 | Severely dry | 13 | 7.8% | 1 in 3.5 years |

≤−2 | Extremely dry | 6 | 3.6% | 1 in 7.5 years |

Total | 167 | 100% |

SPI | Category | Number of time in 44 years | Percentage of occurrence | Severity of events |
---|---|---|---|---|

≥2 | Extremely wet | 0 | 0% | -- |

1.5 - 1.99 | Severely wet | 10 | 5.7% | 1 in 4.4 years |

1 - 1.49 | Moderately wet | 30 | 16.9% | 1 in 1.5 years |

0 - 0.99 | Mild wetness | 51 | 28.8% | 1 in 0.86 year |

0 to −0.99 | Mild dryness | 44 | 24.9% | 1 in 1 year |

−1 to −1.49 | Moderately dry | 19 | 10.7% | 1 in 2.3 years |

−1.5 to −1.99 | Severely dry | 13 | 7.3% | 1 in 3.4 years |

≤−2 | Extremely dry | 10 | 5.7% | 1 in 4.4 years |

Total | 177 | 100% |

vealed by [

The trend analysis was performed in Standardized Anomaly Index and Standardized Precipitation Index for southern Togo using the Mann-Kendall test and Sen’s slope estimator. Results for standardized anomaly index indicated that there are significant warming trends for the mean annual, mean annual minimum and maximum temperature in the period 1970-2014 for all stations except in Kouma-Konda where mean annual maximum temperature exhibited non significant cooling trend (P = 0.01). Mean annual temperature, mean annual minimum and maximum temperatures increased respectively from 1.2˚C; 1.1˚C; 1.3˚C in Atakpamé to 1.9˚C; 2.2˚C; 1.7˚C in Lomé between 1970 to 2014 but drop slowly to 1.4˚C; 1.2˚C; 1.7˚C in Tabligbo during the same period. In Kouma-Konda, the mean annual temperature, mean annual minimum rose respectively by 0.7˚C and 1.6˚C. However, maximum temperatures decreased by 0.13˚C. For Standardized Precipitation Index, drying tendency dominates Atakpamé and Kouma-Konda in the 12-month time scale; mild dryness has the highest percentage of probability of occurrence in both locations. Wet tendency dominates slightly Lomé and Tabligbo in the 12-month time scale; mild wetness has the highest percentage of probability of occurrence in these locations. The increasing of temperature and dry period is of great concern; it implies an increase of evapotranspiration which affects crop yields. The information provided by this study can be to support at local level decision-makers in order to monitor flood and drought. Therefore, agricultural planning and government policies in these areas should be based on recent rainfall, temperature trends. This study should be extended to other drought and flood prone area and to all over the country at large and the impact of the climate variability on crop yields should also be investigated.

The first author designed the study and produced the results. All authors supervised the study and analyzed the results.

The authors declare no conflict of interest.

Koudahe, K., Kayode, A.J., Samson, A.O., Adebola, A.A. and Djaman, K. (2017) Trend Analysis in Standardized Precipitation Index and Standardized Anomaly Index in the Context of Climate Change in Southern Togo. Atmospheric and Climate Sciences, 7, 401-423. https://doi.org/10.4236/acs.2017.74030

Stations | Temperature | ||
---|---|---|---|

Skewness | Kurtosis | ||

Atakpamé | Minimum | 0.06 | −1.32 |

Maximum | −0.19 | −1.29 | |

Mean | −0.15 | −1.29 | |

Kouma-Konda | Minimum | 0.51 | −1.31 |

Maximum | −0.09 | −1.13 | |

Mean | 0.01 | −1.09 | |

Lome | Minimum | 0.68 | −0.95 |

Maximum | −0.63 | −1.17 | |

Mean | −0.29 | −1.16 | |

Tabligbo | Minimum | 0.54 | −0.97 |

Maximum | −0.23 | −1.38 | |

Mean | −0.03 | −1.14 |

Stations | Precipitation | |
---|---|---|

Skewness | Kurtosis | |

Atakpamé | −0.26 | −1.67 |

Kouma-Konda | −0.24 | −1.32 |

Lome | 0.97 | 0.02 |

Tabligbo | −0.10 | −1.39 |