Impact of Climate Variability on Crop Yields in Southern Togo

Rain fed agriculture constitutes the prime sector in Togo that governs the rural economy, given its contribution to gross domestic product (GDP), employment, and food supply. It represents about 38.8% of GDP in 2013 and 41.7% of GDP in 2014, sustains two thirds of the population and provides more than 20% of export earnings [1]. The sector plays critical roles in the national strategy for hunger reduction and poverty alleviation. Food crops are the main agricultural economy subsector in Togo and represents 68% of the country's agricultural GDP [2]. Togolese diet is based mainly on corn (Zea mays), rice (Oryza sativa), cassava (Manihot esculenta) and cowpea (Vigna unguiculata). Climate is the most important factor that governs food production and causes interannual variability in socio economic and environmental systems related to the availability of water resources [3]. Climate change is referred as a change in average weather conditions or long-term variation in climate variables. According to the Intergovernmental Panel on Climate Change (IPCC) Fifth Assessment Report (AR5), the global surface mean temperature (Tmean) rose by 0.85°C during the 1880-2012 period, which is likely due to the observed increase in anthropogenic greenhouse gas concentrations [4]. In fact, global near-surface air temperatures have increased at a rate unequal to any other periods on record including paleo periods [5,6]. Such increases have substantial consequences on precipitation and its variability, especially drought and flood episodes in both the tropics and the subtropics [7,8]. The developing countries rely more on agriculture. They are more likely to be affected by climate change making the climate variability a developing countries issue in recent time because of its largely harmful impacts on agriculture. The main climate risks facing West Africa are combination of flooding, drought, poor distribution of rain, late rains, violent winds and coastal erosion. High population growth, deforestation, very traditional agriculture techniques and improper use of land have resulted in massive land degradation with losses of fertile soil. Today, many rural families can barely make their living from agriculture. The basic climatic elements such as precipitation and temperature show seasonal or annual fluctuations somewhat different from expected climatic conditions, which are very important to agriculture production. Precipitation fluctuation has significant long and short term impacts on natural resources particularly forests, lakes, wetlands and rivers. Regardless of the presence of surface and groundwater resources, the low precipitation over the growing season seriously affects the region's agricultural activities that lead to food insecurity.


Introduction
Rain fed agriculture constitutes the prime sector in Togo that governs the rural economy, given its contribution to gross domestic product (GDP), employment, and food supply. It represents about 38.8% of GDP in 2013 and 41.7% of GDP in 2014, sustains two thirds of the population and provides more than 20% of export earnings [1]. The sector plays critical roles in the national strategy for hunger reduction and poverty alleviation. Food crops are the main agricultural economy subsector in Togo and represents 68% of the country's agricultural GDP [2]. Togolese diet is based mainly on corn (Zea mays), rice (Oryza sativa), cassava (Manihot esculenta) and cowpea (Vigna unguiculata). Climate is the most important factor that governs food production and causes interannual variability in socio economic and environmental systems related to the availability of water resources [3]. Climate change is referred as a change in average weather conditions or long-term variation in climate variables. According to the Intergovernmental Panel on Climate Change (IPCC) Fifth Assessment Report (AR5), the global surface mean temperature (T mean ) rose by 0.85°C during the 1880-2012 period, which is likely due to the observed increase in anthropogenic greenhouse gas concentrations [4]. In fact, global near-surface air temperatures have increased at a rate unequal to any other periods on record including paleo periods [5,6]. Such increases have substantial consequences on precipitation and its variability, especially drought and flood episodes in both the tropics and the subtropics [7,8]. The developing countries rely more on agriculture. They are more likely to be affected by climate change making the climate variability a developing countries issue in recent time because of its largely harmful impacts on agriculture. The main climate risks facing West Africa are combination of flooding, drought, poor distribution of rain, late rains, violent winds and coastal erosion. High population growth, deforestation, very traditional agriculture techniques and improper use of land have resulted in massive land degradation with losses of fertile soil. Today, many rural families can barely make their living from agriculture. The basic climatic elements such as precipitation and temperature show seasonal or annual fluctuations somewhat different from expected climatic conditions, which are very important to agriculture production. Precipitation fluctuation has significant long and short term impacts on natural resources particularly forests, lakes, wetlands and rivers. Regardless of the presence of surface and groundwater resources, the low precipitation over the growing season seriously affects the region's agricultural activities that lead to food insecurity.
In Togo, the analysis of farmers' perception to climate change by revealed that farmers perceived high increase in temperature [9]. Some studies in Togo revealed a rainfall deficit since 1970 [10][11][12][13][14][15]. Sogbedji [16] also found that the decrease in seasonal rainfall amount represents a serious threat to maize growth during the second growing season. Poudel and Shaw [17] used multiple regression analysis to determine the relationship between climate variables such precipitation and temperature on crop yields in Nepal. Regression models based on historical yield data and climate data are relatively accurate for the prediction of changes in the yield due to climate change [18]. To analyse the effects of climatic variables on crop production in Patigi, Nigeria, Tunde [19] used the correlation and regressions analyses and concluded that the climatic variables such as precipitation and temperature have impact to certain limit on the selected crop yield. Adamgbe  Nigeria reported that rainy days and rainfall amount had strong positive relationship with maize yield [20] (r=0.747 and r=0.599, respectively). Several other studies have discussed long-term linear tendencies of climate variables such as precipitation, temperature and others climatic parameters [3,[21][22][23][24][25]. However, modest information data between climatic variables and crop yields exists on Togo, particularly in southern Togo which plays an important role in Togo's economy. The objective of this study was (i) to investigate the change in precipitation and temperature, (ii) evaluate the impact of climatic variability on crop yields. The results of this study will help different actors such as farmers, universities, crop consultants, NGO, hydrologist, and environmentalists to develop and adopt adaptation strategies for agriculture development and sustainable environment.

Site description and data collection
Togo is a small West African nation with estimated population of about 6,191,155 inhabitants [26]. It borders in the south by Atlantic Ocean; Ghana lies to the west; Benin to the east; and the north of Togo is bound by Burkina Faso. Togo lies mostly between Latitudes 6° and 11°N and Longitudes 0° and 2°E ( Figure 1). Monthly average air maximum and minimum temperature and precipitation obtained from daily data series for four weather stations such as Lomé, Tabligbo, Atakpamé and Kouma-konda were obtained from the Meteorological Department of Togo covering the period 1970 to 2014. The geographical coordinates, the long-term mean climatic variables and standard deviation under study are summarized in Table 1. Annual means of air maximum and minimum temperature and precipitation were estimated and used in trend analysis. Crops yields data were provided by Ministère de l' Agriculture, de l'Elevage et de l'hydraulique (MAEH) Togo. Amongst the selected crops, only rice is irrigated, the other such as maize, cassava, cowpea and groundnut are not irrigated.

Trend analysis
For the analysis of trend in annual and monthly maximum and minimum air temperature and annual and monthly precipitation, the Mann-Kendall test [27][28][29], a nonparametric method for trend analysis, was used. It should be noted that the Mann-Kendall test is a statistical test widely used for the analysis of trend in climatologic [30,31] and hydrologic time series [32,33]. There are two advantages of using this test: it is a nonparametric test and does not require the data to be normally distributed and the test has low sensitivity to abrupt breaks due to inhomogeneous time series [34]. According to this test, the null hypothesis (H0) assumes that there is no trend (the data is independent and randomly ordered) and this is tested against the alternative hypothesis (H1), which assumes that there is a trend. The Mann-Kendall test statistic S is given as follows: where xi is the data value at time i, n is the length of the dataset and sign (x i -x j ) 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 [35]. The Theil-Sen estimator (TSE) is an unbiased estimator of the true slope in simple linear regression [36]. If there is a trend in the data, the magnitude of the change in any variable can be denoted by trend slope ß [36,37]: where x i and x j are data values at time t i and t j (i> j), respectively.

Linear regression
Linear regression analysis was applied for analyzing trends in the time series. The main statistical parameter drawn from regression analysis is the slope that indicates the mean temporal change in the variable under study. Positive values of the slope show increasing trends. The coefficient of determination R square drawn from regression analysis is also used. The total change during a decade was obtained by multiplying the slope by 10 years [34].

Impact of variation in climatic variables on crop yields
The precipitation and temperature data were tested against crop yields in southern part of Togo. Yield records of 5 major crops commonly grown in southern of Togo was collected at the Ministry of agriculture, livestock and hydraulic from 1990 to 2014. Lome is not an agriculture production area so; no yield record has been obtained. The crops for which data are available are maize, rice, cassava, cowpea and groundnut. Amongst these crops, only rice is irrigated in southern Togo. Multiple regression analysis was carried out on the crop yield using growing season precipitation and temperature values covering 1990 to 2014 period. The Statistical analysis software STATA was used to perform the Student t-test. This had enabled the possible determination of relationships between the precipitation, temperature and crop yield in southern of Togo. When the probability associated with t-student is less than 5%, then the independent variable has a significant influence or contributes to the variation in crop yield. The equation associated with the change in yield as a function of precipitation and temperature is: Where γ is the yield (kg), a temperature-related coefficient, T temperature data (°C), b precipitation coefficient, P the precipitation data (mm) and ε error term or other variables which have an influence on the yields, C the constant.

Trend in climatic variables
Trend in annual precipitation: Figure 2 presented the time series annual precipitation. Annual precipitation varied with location and ranged from a maximum of 2344 mm registered in 1985 at Koumakonda to a minimum of 423 mm at Lomé in 2000. Annual precipitation had relatively low coefficient of determination R2 with years ranging   [15,38] indicated also an increase of temperature in these locations. In Kouma-konda, T min rose by 0.4°C and T max decreased by 0.3 per decade. It is found that minimum temperature increased faster than maximum temperature in southern Togo ( Table 4). The same result was obtained in Senegal by Djaman et al. [3] who revealed that T min increased more rapidly than T max (0.    and Caesar et al. [40]. Our results are consistent with Vose et al. [41] who reported that T min increased more rapidly than T max (0.204 vs. 0.14°C decade -1 ) from 1950 to 2004. It is also found that amongst study stations, the coastal area (Lomé) had the highest increase in annual T min and T max with Sen's slope 0.048, 0.040, respectively. The increasing of temperature in southern Togo is not surprising but it only proves that global warming revealed for the whole world by Vose et al. [41] and Weart [42], can be identified even at local scales (southern Togo). The finding is in agreement with IPCC prediction which states in 2007 that, the Earth could warm by 3°C this century.
The trend analysis in monthly T min is presented in Table 4 shows an increasing trend in T min at all four locations (P<0.0001). For the four locations, the trend in T min revealed the lowest slopes from April to October which is the growing season and the highest slope from November to March which coincide with the dry season in southern Togo. A non-significant trend (P>0.05) of T max was observed in January, March and August in Tabligbo, in January to July, October and November at Kouma-Konda. The highest increased in T max at all locations was observed in December (Table 4). It can be deducted that decadal change increase in T min varied from 0.1 to 0.7 with more warming tendency in November to March at all locations which corresponds to the dry period in southern Togo. Overall, the increase in annual and monthly average T min and T max from 1971 to 2014 could have a serious implication in agricultural productions, reduced crop yields in tropical areas leading to increased risk of hunger [43]. Overall, the coastal area (Lomé) has the highest increase in monthly average T min and T max . This could be explained by urban heat island. Lomé is capital city and is the most populated city in Togo (40% of the population and over 90% of domestic industries). The results of this study are in agreement with Gadédjisso [9] who revealed an increase of air temperature throughout Togo. Similar trends in air temperature were reported in West Africa by Sylla et al. [22].

Impact of variation in climatic variables on crop yield:
The results of multiple regressions of precipitation and temperature on crop yields are presented in Table 5. In Atakpame, the results revealed that the precipitation and temperature have non-significant effect on the yields record. In Kouma-konda, the results revealed that the temperature has significant influence on the variation of maize yields (p<0.1) and on cowpea (p<0.01). For other selected crops in Kouma-konda, it is showed that the variation of precipitation and temperature have nonsignificant effect on them. In Tabligbo, for all the selected crops, the variation of precipitation and temperature have non-significant effect on the crop yields. The non-effect of precipitation and temperature variability on the selected crop yields obtained in this study is also obtained by in Northern part of Nigeria [44]. This is an indication that precipitation and temperature variability may not necessary affect some selected crops. Despite the climate variability, farmers used drought resistant seed to adapt their production system to climate change. Our results indicated a slight increase in growing season precipitation with non-significant effect on crop yields, which is in contradiction n.s.: Non-Significant + Significant at 10% * Significant at 5% ** Significant at 1% ***Significant at 0.1% Table 4: Summary of the trends analysis in monthly minimum and maximum temperature at Atakpamé, Kouma-konda, Lomé and Tabligbo.
with Sogbedji [16] who revealed that in Togo a decrease in seasonal rainfall amount represents a serious threat to maize growth. It has also been reported by Adamgbe and Ujoh [20] in Nigeria that rainy days and rainfall amount had strong positive relationship with maize yield (r=0.747 and r=0.599, respectively). In addition, Adamgbe and Ujoh [20] observed that rainfall characteristics jointly contributed 67.4% in explaining the variations in the yield of maize per hectare in Nigeria. The effect of water deficiency depended on growth stages, deficiency level and environmental changes during drought conditions. During later productive stages, yield losses from drought decreases as crop is approaching physiological maturity.
Morover, Kouma-Konda is situated in forest zone and deforestation during the last decades may have contributed to the slight decreasing of precipitation and increasing of temperature. Maize and bean are affected by the increasing of temperature. The same result is also obtained by Amouzou et al. [15] who revealed that climate variability minimally affected the yield of maize in maritime region of Togo. Serious attention need to be paid to the affected crops by climate variability in some locations, otherwise, the effect will increase and affect the other crops. With changes in climate parameters, mostly an increasing trend in air temperature combined with precipitation variability, several actions have been taken by agricultural and water resources actors to contrast the effects of climate change on Togo's agriculture. As climate conditions change, crop growers are increasingly receptive to new agricultural technologies offered by extension services [33]. These farmers and extension service capacity to adapt to climate  variability may explain the non effect of precipitation and temperature on crop yields. The use of drought-resistant seeds is becoming more common, which minimizes the risk of crop failure due to prolonged drought. The PNIASA program initiated by the government of Togo in 2010 is consistent with the National Poverty Reduction Strategy, and has a balanced focus of investment in support of improved agricultural production, improving the institutional framework, and physical infrastructure for higher productivity in the agricultural sector. Several other actions are being taken by the government of Togo, such as rehabilitation of water reservoirs, rehabilitation of irrigation schemes, and agricultural land development of 666 hectares at Mission-Tove and 585.5 hectares at Agome Glozou and implementing the national environmental policy and disaster management policy. Despite the adoption of strategic options, there is a need to continue implementing them and more participative decision-making is necessary for the durability of the innovation. Moreover, policy-makers should be involved at every step from the introduction to the adoption process. In addition, more research should be done on all of the strategic options to strengthen the system.

Conclusion
The objective of this study was to analyze the trend in climatic variables such as annual and monthly precipitation, annual and monthly T min , T max using historical data from four weather stations through the year 1970 to 2014 using the Mann-Kendall test and Sen's slope estimator and to evaluate the impact of climatic variability on crop yields in southern Togo using multiple regression analysis. Results showed a non-significant increasing trend (P>0.05) in annual precipitation at Atakpamé, Lomé, Tabligbo and non-significant decreasing trend (P>0.05) at Kouma-konda from 1970 to 2014. There was a significant warming trend (P<0.0001) in T min , T max in Atakpamé, Kouma-konda, Lomé and Tabligbo during 1971-2014 period. T min increased by 0.3, 0.5 and 0.3°C per decade at Atakpamé, Lomé and Tabligbo respectively and T max increased by 0.3, 0.4 and 0.4°C per decade at Atakpamé, Lomé and Tabligbo, respectively. In Kouma-konda, T min rose by 0.4°C and T max decreased by 0.3 per decade. T min is found to increase faster than T max . The coastal area (Lomé) has the highest increase in annual and monthly T min and T max . Results of multiple regression analysis revealed that in Atakpame and Tabligbo, neither the precipitation nor the temperature have effect on the selected crop yields. In Kouma-konda, it revealed that the increased of temperature has significant effect on maize and bean. The increasing trend in T min and T max imply an increase in evapotranspitation which is of great concern to crop producers. Therefore, specific actions such as water management practices, crop diversification, new drought-resistant seed etc. are needed for resource planning and management for the sustainability of agriculture for future generations.