Reference evapotranspiration is a key parameter in hydrological and meteorological studies and used to determine the actual water use rate for various crops. The objectives of this study were to explore trend in the grass-reference evapotranspiration (ETo) through years 1961–2011 and to identify trend in the aridity index as an indicator of change in climate in Togo. ETo was calculated using the FAO-56 Penman–Monteith method, and trends analyses were performed with non-parametric statistics proposed by Mann–Kendall and the Sen slope estimator. Results showed that annual ETo varied from 1,440 to 1,690 mm at Lomé, from 1,761 to 1,905 mm at Tabligbo, and from 1,839 to 1,990 mm at Sokode. The Mann–Kendall test revealed significant increase in annual ETo at Tabligbo (Z = 2.89) and Sokode (Z = 2.29). Annual ETo is much more stable at Lomé, with non-significant decrease. In Togo, according to the three study sites, the 1961–2011 period annual aridity index varied from 0.26 to 0.99 at Lomé, 0.38 to 0.98 at Tabligbo, and 0.45 to 1.08 at Sokode. The Mann–Kendall test revealed a declining trend in the ratio of precipitation/ETo which adversely implies an increasing severity of the aridity index at all the sites, prejudicial to rainfed agriculture practiced by about 90% of Togolese crop growers.
Increasing interest has been shown in the study of change in climate parameters and its effects on the hydrological cycle and water supply. Research has been conducted to detect climate changes, trends, and variability in various parts of the world considering some climate parameters such as air temperature, rainfall depth, reference evapotranspiration, and pan evapotranspiration (Schwartz & Randall 2003; Garbrecht et al. 2004; Hegerl et al. 2007; Fu et al. 2009; Saghravani et al. 2009; Hakan et al. 2010). Climate change is expected to intensify the hydrological cycle and to alter one of its important components, evapotranspiration (Huntington 2006). A linear warming trend over the last 50 years has been recorded at a rate of 0.13 °C per decade (IPCC 2007a). In particular, all of Africa is highly likely to experience warming during this century, with the warming expected to exceed the global average (IPCC 2007b). Therefore, analyzing how climate change affects reference evapotranspiration is critical for understanding the impact of climate change on the hydrological cycle. To evaluate evapotranspiration in the context of climate change, Zheng et al. (2009) characterized the cause of the decreased pan evaporation during 1957–2001 in the Hai River Basin, and found that the declining wind speed is the climate variable that impacted ETo. Goyal (2004) studied the sensitivity of evapotranspiration in terms of change in temperature, solar radiation, wind speed, and vapor pressure for a 32 year (1971–2002) period from the arid zone of Rajasthan, India, and the results showed that the calculated ETo was most sensitive (14.8%) to temperature. Porter et al. (2012) reported grass-reference evapotranspiration (ETo) and alfalfa reference evapotranspiration (ETr) sensitivity to measurement errors in wind speed and air temperature followed by incoming shortwave (solar) radiation. Bandyopadhyay et al. (2009) also studied reference evapotranspiration trends in India and reported decreases in ETo all over India. On the other hand, several researchers also reported increases in ETo trends. Yu et al. (2002) observed increasing trends in reference evapotranspiration at Kao-Hsiung, south Taiwan, using 48 years of data. Hess (1998) reported an increasing trend in reference evapotranspiration in the northeast arid zone of Nigeria, due to the increases in wind speed. Myneni et al. (1997) and Milly & Dunne (2001) reported that the accelerated reference evapotranspiration over North America is assumed to be due to a rise in temperature over the past century. Dinpashoh et al. (2011) reported more pronounced temporally increasing trends in ET than the decreasing trends over Iran, with the wind speed found to be the most dominant variable influencing ETo in all months except the winter months. Other studies reported changes in pan evaporation, which is related to reference evapotranspiration through appropriate and locally developed pan coefficients. Liu et al. (2004) found that the decrease in solar radiance was most likely the driving force of the trend in pan evaporation from 1955 to 2000 in China. In the southern and eastern parts of the Hai River Basin, the annual reference evapotranspiration was dominated by the decreasing trends, and the reference evapotranspiration was more sensitive in decreasing gradient to relativity humidity, temperature, shortwave radiation, and wind speed (Zhao et al. 2014).
Precipitation is one other key component in the hydrologic cycle that affects numerous locations of the world. Reduction in seasonal precipitation is becoming recurrent, and many countries are concerned by the concept of climate change. Climate change has resulted in extreme drought conditions in some parts of the world and flooding in other parts (van de Giesien et al. 2010), and according to a modeling study, anthropogenic climate change may soon yield increases in the frequency and severity of droughts and the expansion of deserts (Manabe et al. 2004). In particular, environmental changes in Africa have been mostly directly related to rainfall (Zheng et al. 1997). Many studies revealed a drastic decrease in precipitation in Africa (Hubert et al. 1989; Mahé & Olivry 1995; Bricquet et al. 1997; Servat et al. 1999; Balme-Debionne 2004; Van Vyve 2006). Nicholson & Grist (2001) identified several changes in the general atmospheric circulation that accompanied a shift to drier conditions in the West African Sahel. This atmospheric circulation is believed to generate and maintain wave disturbances that modulate the rainfall field. Rotstayn & Lohmann (2002) showed that a prominent feature is the drying of the Sahel in North Africa, and suggested that the indirect effects of anthropogenic sulfate may have contributed to the Sahelian drying trend. A few studies in Togo revealed a rainfall deficit since 1970 (Klassou 1996; Badameli 1996, 1998; Adewi 2002; Adewi et al. 2010). The decreasing trend in precipitation against the unbalanced evapotranspiration reduces water availability that can be expressed in terms of aridity.
Aridity is usually expressed as a function of rainfall and temperature. The long-term difference (or ratio) between ETo and precipitation (P) has been considered a measure of aridity used in several climate classiﬁcation schemes (Köppen 1936; Thornthwaite 1948; Prentice 1990). UNESCO (1979) applied an aridity/humidity classification system based on the average annual precipitation divided by the average annual potential evapotranspiration (PET). Aridity results from the presence of dry, descending air. Therefore, aridity is mostly found in regions with anticyclonic conditions, such as the subtropical area where Togo is located. Climatological aridity is a critical environmental factor that helps to determine the character and sustainability of natural vegetation, rainfed agriculture, and terrestrial ecosystems. The aridity index is qualified by the index of drought by Rind et al. (1990).
To achieve water conservation and sustainability, climate interactions with various aspects of the water cycle should be a research priority. However, despite the broad studies in several countries and regions on the trends in reference evapotranspiration and precipitation, very little information is available on the temporal trend analysis of ETo over Togo. Therefore, the present study is undertaken with three objectives, which are as follows: (1) to detect the monotonic linear trends in the monthly and annual ETo time series using the Mann–Kendall non-parametric test; (2) to estimate the slopes of trend lines of ETo times series using the Theil–Sen's estimator method; and (3) to estimate the trend in the aridity index in Togo.
MATERIALS AND METHODS
The study was conducted in Togo, where three weather stations at Lomé (6 °9′56″ N, 1 °15′16.24″E, elevation 22 m), Tabligbo (6 °34′59″ N, 1 °30′00″ E, elevation: 76 m), and Sokode (8 °58′59″ N, 1 °07′59″ E, elevation: 417 m) were selected for reliability and the long-term dataset without missing data covering the period 1961 to 2011. A record of monthly average climatic parameters including air maximum and minimum temperature, minimum and maximum relative humidity, wind speed, solar radiation, and precipitation were used to estimate monthly evapotranspiration.
Reference evapotranspiration estimation model
The aridity index was defined by UNESCO (1979) as the ratio of the average annual precipitation divided by the average annual PET. Monthly and annual aridity indexes during the 1961–2011 period were calculated using the estimated ETo and the precipitation data.
Temporal trend analysis
A positive value of Z indicates that there is an increasing trend, and a negative value indicates a decreasing trend. The null hypothesis, H0, that there is no trend in the records, is either accepted or rejected depending on whether the computed Z statistics are less than or more than the critical value of Z statistics obtained from the normal distribution table at the 5% significance level. If |Z| > Z(1−α/2), the null hypothesis of no autocorrelation and trend in dataset is rejected, in which Z(1−α/2) is corresponding to the normal distribution with α being the signiﬁcance level.
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, while negative values of the slope indicate decreasing trends. The total change during the period under observation was obtained by multiplying the slope by the number of years (Tabari & Marofi 2011).
RESULTS AND DISCUSSION
Trends in annual reference evapotranspiration
Trends in monthly reference evapotranspiration
Trends in aridity index
In Togo, the increase in annual evapotranspiration is associated with decreasing trends in annual precipitation. Under rainfed and irrigated agriculture, the difference between precipitation and ETo, or the ratio of precipitation over ETo named the aridity index, is very important in water resources planning, complementating of irrigation management, and conservation agriculture. In Togo, according to the three sites in this study, the annual aridity index varied from 0.26 to 0.99 at Lomé, 0.38 to 0.98 at Tabligbo, and from 0.45 to 1.08 at Sokode (Figure 7). The 1961–2011 period mean aridity indexes were 0.55 at Lomé, 0.63 at Tabligbo, and 0.70 at Sokode. Therefore, Lomé and Tabligbo are in the dry sub-humid zone and Sokode is in the wet humid zone according to the classification of UNESCO. Increased aridity is a robust proximate cause of desertification, both indirectly through greater rainfall variability and directly through prolonged droughts (Costa & Soares 2012; Hrnjak et al. 2013). Years or months when the aridity index is greater that unity are broadly classified as wet since precipitation met evaporative demand. Similarly, years or months with an aridity index less than unity are broadly classified as dry. The Mann–Kendall test revealed a declining trend in the ratio of precipitation/ETo, which adversely implies an increase in the severity of the aridity index at all the sites. However, the trend was significant at Tabligbo and Sokode at a 95% confidence interval while it is not significant at Lomé, although it had a deceasing trend at that site similar to the others. The monthly aridity index varied from 0.07 to 1.9 at Lomé, 0.06 to 1.15 at Tabligbo, and from 0.02 to 1.9 at Sokode (Figure 8). The aridity index showed two patterns at Lomé and Tabligbo corresponding to the two cropping seasons in south Togo, while at Sokode it presented the pattern of the long crop growing season in the central region of Togo (Figure 9). The Mann–Kendall test revealed different trends in the monthly aridity index depending on the month and location. A significant increase in aridity severity was observed in December at Lomé and a significant reduction in the severity of the aridity index was noted in August and October at Lomé. Similar to the results of this study, Some'e et al. (2013) reported a mean annual P/ETo that varied from 0.06 to 1.0 in Iran. During the 1951–2010 period, the aridity index over Italy varied from 0.22 to 3.08 with a national territory average of 0.9 (Salvati et al. 2013). For production seasonality, the monthly aridity index should be a better indicator of the onset and establishment of the crop growing season, and the choice of appropriate crop species within rainfed agriculture systems. The increasing trend in the aridity index is an expression of the lack of moisture or water for crop production. Some studies in Togo revealed a rainfall deficit since 1970 (Klassou 1996; Badameli 1996, 1998; Adewi 2002; Adewi et al. 2010). Adewi et al. (2009) found that there is a relation between the rainfall disturbance and water stress at the critical period of maize growth and development in the country. Precipitation will be able to cover crop evapotranspiration demands when the aridity index is equal to or greater than 1. Crop actual evapotranspiration is estimated by the product (kc*ETo) (Allen et al. 1998; Djaman & Irmak 2013), with crop coefficient kc that varies with crop growth, crop development, watering regimes, and the environment (Djaman & Irmak 2013). Thus, the long-term monthly average aridity index could be used for setting the planting period under rainfed production. With regard to the monthly aridity index, the period November–March is considered dry throughout the country, and the month of August is considered dry only in the maritime region. With regard to the aridity index, in the maritime region of Togo where Lomé and Tabligbo are located, late March–early April is well indicated for the planting period during the first growing season and early September is the appropriate time of planting during the second growing season. Late April is indicated for the beginning of the crop growing season in the central region of Togo where Sokode is located. From May to October, the aridity index was close to or greater than unity except in August at Lomé and Tabligbo. This explicitly expresses more abundant rainfall, such as to counterbalance or exceed what is required by PET. The onset of the cropping season, using the aridity index, coincided with the results of Adewi et al. (2010), who delimited the beginning and the end of the crop growing season across Togo based on historical climatic data of the 1950–2000 period. Sogbedji (1999) found that the decrease in seasonal rainfall amount represents a serious threat to maize growth during the second growing season. In Togo, where crop production is essentially rainfed, water management requires more attention to better planting date choice and the crop species with regard to water availability and crop evapotranspiration, as proposed by Adewi et al. (2010). The increasing trend toward aridity in recent times has been reported by Dregne & Chou (1992), Nicholson (2003), Hanafi & Jauffret (2008), and Gaughan & Waylen (2012). Amégadjé (2007) reported an aridity index of less than 0.75 with diminishing precipitation and the number of rainy days in the savannah region of Togo. Over the past 60 years, Togo has experienced three major droughts, in 1942–1943, 1976–1977, and 1982–1983, leading to severe famines. Predictions for 2025 by the SCENGEN model showed that the declining trend in rainfall is set to continue and the country is expected to be 10–30% drier than the previous 50 years (CNI 2001). Rainfall amount and temporal distribution is primordial to crop production. Whenever the aridity index is high, uniform distribution of rainfall may meet crop evapotranspiration and reduce water loss through excessive runoff and deep percolation.
The objective of this study was to analyze the trend in the grass-reference evapotranspiration (ETo) calculated by FAO-56 Penman–Monteith method using historical data from three weather stations through the years 1961–2011 and to identify trend in the aridity index as an indicator of change in climate in Togo. Results showed that annual reference evapotranspiration varied from 1,440 to 1,690 mm at Lomé, from 1,761 to 1,905 mm at Tabligbo, and from 1,839 to 1,990 mm at Sokode. The Mann–Kendall test revealed a significant increase in annual ETo at Tabligbo and Sokode. Annual ETo was more stable in Lomé, with a non-significant decrease. In Togo, according to the three sites under study, the annual aridity index varied from 0.26 to 0.99 at Lomé, 0.38 to 0.98 at Tabligbo, and from 0.45 to 1.08 at Sokode. The 1961–2011 period mean aridity indexes were 0.55 at Lomé, 0.63 at Tabligbo, and 0.70 at Sokode. The Mann–Kendall test revealed a declining trend in the ratio of precipitation/ETo, which adversely implies an increase in the severity of the aridity at all the sites in this study. Information regarding trends in the aridity index as a result of climate change is necessary for policy-makers and water resources managers within the context of water resources management, hydrology, agriculture, and the environment. These results may constitute some useful information required by different levels of actors in development, particularly in reducing the vulnerability of rainfed crop growers in Togo. The findings of this research suggest the need to consider ETo changes in planning for agricultural and water resources projects in the dynamics of climate change. Most of the rural poor households in Togo rely for their livelihood and food security on highly climate sensitive rainfed subsistence or small-scale farming, pastoral herding and direct harvesting of natural services of ecosystems such as forests and wetlands (CNI 2001). Therefore, adaptation efforts to climate change should target conservative water resources management technologies, crop breeding for more drought and heat tolerant varieties to increase crop productivity under adverse weather, and promote and support surface runoff water harvesting and the development of small-scale irrigation in upland and lowland areas.