Rainfall data for a 37 years period revealed a variation in the annual mean amount of rainfall (Fig. 2). Rather than an absolute change, results for all the three stations show rainfall variability. The rainfall in Kwara is below normal, a little above normal in Ogun, and varied in Ondo. The rainfall variability found in the data can be attributed to seasonal and interannual climatic variability, which is much more several particularly in the savannah region where Kwara is located (Adejuwon 2004; Odekunle et al. 2007; Ayanlade 2009). As depicted in Fig. 2, annual rainfall was mostly below average before 2010, but varies greatly in the last decade, with readings above average in all the stations. Rainfall peaked in 1991 (with 1,982 mm) and 1995 (1,999 mm) in Ondo state, with 2016 (2,206 mm) recording the highest rainfall during the study period (Fig. 2). In Ogun state, rainfall peaked in 1996 (2,026 mm), 2006 (2,216 mm), and 2015 (2,086 mm), with 2006 recording the highest rainfall. Finally, in Kwara, rainfall peaked in 1991 (1,468 mm), 1995 (1,409 mm), 1998 (1,596 mm), and 1999 (1,539 mm), with the highest peak recorded in 2014 (2,467 mm) with a very unusual high value. Generally, the annual average of rainfall was lowest (673 mm) in Kwara state, with similar values in Ogun (1,028 mm in 1986 and 1,037 mm in 2013) and Ondo states (1,041 mm). Values follow an expected pattern, with higher values in the rainforest ecological zone close to the ocean and lower values in inward zones closer to the savanna. Ogun state received the highest rainfall between 1995 and 2010. The highest rainfall within a year was recorded in 2014 in Kwara, the state belonging to Guinea savanna ecological zone, which recorded a value slightly above the peak of rainfall in Ondo (which belongs to the forest ecological zone).
Climatic variability in the study area can be linked to global climate oscillating systems as El Niño – Southern Oscillations, sea surface temperatures (SSTs), and Inter-Tropical Discontinuity (ITD) have been reported to be responsible for inter-annual variability in Africa climate (Stige et al. 2006). Such climate variability can cause negative departures from normal climate, as documented in a part of Nigeria and elsewhere (Ashipala 2013; Owusu et al. 2015; Ayanlade et al. 2018a). Studies have further revealed that rainfall associated with SST strongly influences rainfall variability in Nigeria in conjunction with the role of the ITD as the equatorial displacement of the Atlantic Subtropical High suppresses the northward summer migration of the ITD thereby resulting into rainfall variability (Bello 2008; Ayanlade et al. 2019).
High variability was observed in the maximum temperature of all stations between 1982 and 2017 (Fig. 4). The temperature was highest in Kwara state, the northernmost state which falls within the Guinea savanna – an area dominated with relatively low rainfall and high temperature – and lowest in Ondo state, the southernmost state. Ondo and Ogun state recorded temperatures below average between 1982 and 2010. However, temperatures have experienced an increase from 2012 till date, especially in Ogun state with the temperature higher gradually increases. As shown in Fig. 3, all the stations experienced similar minimum temperatures patterns from the early 1980s. Temperature oscillations in the maximum temperature were recorded between 1990 and 2010 in Ondo and Ogun with an upward trend observed only in Kwara, though from 2010, the temperature has generally increased in all stations (Figs. 3 and 4).
As shown in Fig. 3, the annual minimum temperature is highest in Ogun state with the lowest value in 1994 (22.8 °C), a value that coincides with the highest minimum temperature in Ondo state. Despite the southern location of Ogun state, the minimum temperature is generally above the minimum temperature for the other two stations, among which Ondo displays the lowest temperature than Kwara. These values represent the highest temperature difference in Kwara and the lowest temperature difference in Ogun state. The period of temperature increase in Kwara state (i.e., years 2000, 2001, and 2002) corresponds to the period of rainfall decrease which could be attributed to increasing temperature and decreasing rainfall as a result of the increasing evapotranspiration and desertification also reported in other areas of Nigeria (Ayoade 2003). The trend of the maximum temperature varies considerably in all stations. However, while the trend is irregular in the early years of the study period, an upward increasing trend was observed in the later years. This recent increase in temperature could be linked to the effect of changes in climate and to the rapid urbanization in the ecological zones under consideration (Mabo 2006).
Relationship Between Climate and Crop Yield
The results of the Pearson correlation analysis between annual minimum temperature and crop yields during the growing season are shown in Table 1. Annual minimum temperature had positive but weak correlation with maize in Ondo (0.142) but not significant at p <0.05, a weak and not statistically significant negative correlation of −0.078 at p <0.05 in Ogun state, and a positive correlation of 0.219 in Kwara at p <0.05. Rice, on the other hand, had a positive but weak correlation of 0.167 which is statistically not significant at p <0.05 in Ondo state while in Ogun state with a positively strong and statistically significant correlation coefficient of 0.674 at p <0.05 and a strong positive and statistically significant correlation coefficient of 0.481 in Kwara state at p <0.05. The correlation coefficient for cassava in Ondo state is a weak and statistically not significant value of 0.22 at p <0.05 while a strong positive and statistically significant correlation coefficient of 0.82 at p <0.05 in Ogun and a positive but not significant correlation of 0.25 at p <0.05. Generally, annual minimum temperature had a strong, positive, and statistically significant correlation coefficient at p <0.01 and p <0.05 in rice for Ogun and Kwara states and cassava only in Ogun state.
Table 2 shows the result of Pearson correlation analysis between annual maximum temperature and crop yields during the study period. The annual maximum temperature had a positive weak and statistically not the significant relationship with maize in Ondo (0.198) at p <0.05, a weak and not statistically significant negative correlation of −0.337 at p <0.05 in Ogun state, and a positive correlation of 0.342 in Kwara at p <0.05. Rice had a positive but weak correlation of 0.160 which is statistically not significant at p <0.05 in Ondo state while in Ogun state with a positive, very strong, and statistically significant correlation coefficient of 0.88 at p <0.05 and a very strong positive and statistically significant correlation coefficient of 0.77 in Kwara state at p <0.05. The correlation coefficient for cassava in Ondo state is a negative, weak, and statistically not significant value of −0.12 at p <0.05 while a very strong positive and statistically significant correlation coefficient of 0.87 at p <0.05 in Ogun. A positive and significant correlation of 0.63 at p <0.05. Generally, the annual maximum temperature had a strong, positive, and statistically significant association with rice and cassava yields in Ogun and Kwara states but not in Ondo state. The maximum temperature was not associated with maize yields in any of the states.
The result of the Pearson correlation analysis between annual rainfall and crop yields during the study period is shown in Table 3. Annual rainfall had a negative, weak, and statistically not a significant relationship of −0.15 with maize Ondo at p <0.05, a weak and not statistically significant positive correlation of 0.003 at p <0.05 in Ogun state and a positive correlation of 0.08 in Kwara at p <0.05. Rice showed a negative but weak correlation of −0.091 which is statistically not significant at p <0.05 in Ondo state while in Ogun state with a positive, weak, and statistically significant correlation coefficient of 0.14 at p <0.05 and a weak, negative, and statistically not significant correlation coefficient of −0.22 in Kwara state at p <0.05. The correlation coefficient for cassava in Ondo state is a negative, weak, and statistically not significant value of −0.08 at p <0.05 while a positive and statistically not significant correlation coefficient of 0.046 at p <0.05 in Ogun was observed. And a positive, weak, and not the significant correlation of 0.06 at p <0.05 was observed for Kwara state. Generally, annual rainfall had a weak and statistically not significant association with maize, rice, and cassava yields in the three studied states.
Farmer’s Adaptive Strategies
Figure 5 shows the percentage error graph of farmers’ coping or adaptive capacity toward extreme climatic conditions. According to our results, 16.7% of the respondents are engaged in agricultural diversification, 17.3% are engaged in changing the crops they cultivate, 20.2% are engaged crop rotation, and 19.6% are practicing mixed farming.
As shown in Fig. 5, the farmers are coping or adapting to extreme climatic conditions but generally with low capacity. This is because the results show that 19% of farmers had implemented changes in the size of farmland, 14.9% are engaged in different crop composition, 33.3% are engaged in agricultural intensification, 30.4% are engaged in bush fallowing, 29.8% are changing the daily working time, 47.6% are presently engaged in agroforestry, 27.4% are changing the seasonal timing of sowing, 41.1% are changing the harvest time, 35.7% are into irrigation, 31.5% are into the use of fertilizer, 32.1% are into the use of pesticide, 56.5% are into getting loan and credit facilities, 32.7% are into getting other sources of income so as not to depend on their farm only, 44% are changing the method of storing their food, 27.4% are changing the quantity of food consumed by the family, and 24.4% are changing their residence by moving to another area for farming. From Fig. 5, seeking for loans and credit facilities, changes in food storage, agroforestry, changing the time of harvest, use of irrigation, use of fertilizer, and agricultural intensification, among others, are the most practiced adaptation options.