Study area

The study was conducted at Lôh Djiboua and Gôh in Centre-West, Côte d'Ivoire. The main environmental characteristics are shown in Table 1. The two regions are composed of Gôh-Loh-Djiboua district (Figure 1). This district is a post-pioneer zone with an orchard of 22 years old (Deheuvels, 2007). These regions are under the influence of tropical humid climate with four-season, including two dry seasons and two rainy seasons. Since the late 1970s, the decline in annual rainfall amounts in the West African sub-region, particularly in Côte d'Ivoire, also affected this area (Brou et al., 2005; Kassin et al., 2008). The average humidity of 85% has a strong seasonal variation. The minima are obtained between November and March. The average temperature is 27°C and varies annually between 19 and 33°C.

 

 

 

METHODS

The rainfall data analyzed were collected at the meteorological stations of CNRA both based on Divo and Gagnoa. These data covered a period of 30 years, namely from 1986 to 2015. The two-rainfall series was complete. The methodology used for the analysis of the data consisted of first studying the evolution and distribution of rainfall over the whole period per decade. Then some rainfall parameters, such as annual rainfall heights and cumulative rainfall of a rainy season were calculated and compared to the thresholds required for cocoa (Mian, 2007).

Evolution of annual rainfall

The study of the variation of annual rainfall was done by calculating the rainfall index (Equation 1), which allows the determining of the excessive and deficit years relatively to each meteorological station (Servat et al., 1998). Thereafter, the frequency  of  years  with  annual rainfall less than 1200 mm,  the minimum threshold required for cacao tree was determined.

Where,= quantity of rainfall during year I;Average annual rainfall over the study period; S = standard deviation of the annual rainfall amounts during this period

A year is humid, when the rainfall index is positive; otherwise it is dried when this one is negative (Soro et al., 2011). Thereafter, the proportion of years with annual rainfall less than 1200 mm, the minimum threshold required for cacao was determined.

The rainfall index was then ranked according to the range of values (Table 1) to allow an assessment of the extent of drought or moisture for each year of the rainfall series (Wu et al., 2005).

Monthly rainfall distribution

The study of the monthly rainfall distribution was based on rainfall quantity criteria (IFCC, 1978). Thus, a month is dry when it records less than 50 mm of rain; a month is considered as deficit when it records a rainfall amount between 50 and 100 mm; a month is wet when it records a rainfall greater than 100 mm.

The occurrence probability of humid, deficit and dry months per year led to determine the wettest and driest months over the observation period.

Climatic characterization of rainy season

The characterization of the growing seasons consisted of determining the beginning and end of  the  rainy  seasons  by  the rainfall balance method (Cocheme et Franquin, 1976). According to this method, the rainy period is defined by the position of the 10 days rainfall curve above the 10 days ETP curve of the latest normal (1971 to 2000). The beginning and the end of the rainy season are thus at dates corresponding to the intersections of the two curves. Subsequently, the probability of rainy seasons was calculated by 10 days by comparing the rainfall totals of the major rainy seasons per year at the threshold of 700 mm required for cocoa farming.

 

 

Variation of annual rainfall

The study showed a secession of wet and dry periods (Figures 2 and 3) in the two stations. At Divo, the three decades were moderately dry ranging from 60 to 80% of years characterized by rainfall deficits (Table 2). The extreme dry and wet years were almost non-existent in this station.

 

 

At Gagnoa, the decade (1986-1995) was moderately dry with 60% years characterized by a rainfall deficit (Table 3). On the other hand, the decades of 1996 to 2005 and 2006 to 2015 were moderately humid from 60 to 80% of the years having recorded rainfall surpluses. Only 10% of the years were extremely wet in this station. Furthermore, taking into account the minimum threshold of 1200 mm of annual rains, required for cocoa farming, over the three decades 50 to 60% of the years at Divo and 20 to 30% at Gagnoa were deficient (Table 4).

 

 

Monthly rain distribution

During 30 years of observations, March, April, May, June, September and October were the rainiest months, with probabilities of occurrence of events P> 100 mm greater than 50%, on one hand (Table 5). On the other hand, January and December were the driest months.

 

 

Evolution on the beginning and the end of the rainy seasons

Divo station at the Lôh Djiboua region

Figure 4 indicated that the beginning of the rainy seasons varied between the 1st 10 days of January (1996 to 2005) and the 2nd 10 days of March  (2006  to 2015) with the frequency of the events P> ETP of 50 and 40% (Table 6). However, over the 30 years of observations (1986-2015), the rainy season started at the 1st 10 days of March with a probability of P> ETP of 50%. Over the past three decades, the beginning of rainy seasons varied between 30 and 70 days with early rains during the 1996 to 2005 decade.

 

 

 

The end of the rainy seasons also fluctuated from the 1st 10 days of May (1986-2006) to the 2nd 10 days of July (2006-2015), that is, an offset of 10 to 50 days with P> ETP event probabilities of 0% at the 2nd 10 days of May and 10% at the 3rd 10 days of July. Nevertheless, over all the 30 years, the rainy seasons ended from the 2nd 10 days of July with a probability of the P> ETP event of 43% in the 3rd 10 days of July (Table 7). So the length of the rainy seasons was thus 70 days during the decade (1986-1995) and 130 days during the decades 1996-2005 and 2006-2015. As a result, the duration of the rainy season was increased with 60-days from 1986-1995 to 1996-2005 and 2006-2015.

 

 

Gagnoa Station in the Gôh region

The beginning of the rainy seasons in Gagnoa fluctuated between the 2nd 10 days of March (1986 to 1995, 1996 to 2005) and the 3rd 10 days of February (2006-2015) during the three decades (Figure 5). However, the probability of occurrence of the P> ETP events (Table 8) was higher (60%) in the 3rd 10 days of March than in the 2nd 10 days (50%) during the decade (1986 to 1995). The beginning of the rainy season therefore varied between 10 and 20 days from decade to decade. In addition, over the 30 years, the probability of rains settling during the 3rd 10 days of March was greater (60%) than that of the 2nd 10 days (47%).

 

 

 

Moreover, in the past 30 years, 90% of the years (1986-1995) and 40% during the decades 1996 to 2005 and 2006 to 2015 recorded cumulative rainfall less than 700 mm, the minimum required for the cacao cultivation during the great rainy season.

 

 

 

 

 

 

Evolution of annual rainfall

The study showed a fluctuation of the annual rainfall during the decades over the 30 years (1986-2015) in the two regions. The average annual rainfall  calculated over this period is between 1200 and 1400 mm, instead of 1500 and 1700 mm, during the period 1951-1960 (Dabin, 1973). This downward trend in rainfall was already highlighted by several other authors (Servat et al., 1995; Brou et al., 1998; Bari and al., 2016). However, the inter-annual average rainfall recorded remains above 1200 mm, the minimum required for cocoa farming, on one hand.  On the other hand, at the level of each weather station, the study showed the existence of years with a rainfall of less than 1200 mm. This average is 56% at Divo and 23% at Gagnoa. This reflects a gradual drying up of the climate in the study area and especially at Divo; where the three decades (1986-1995, 1996-2005 and 2006-2015) were moderately dry. The existence of these drier years with a rainfall of less than 1200 mm is a constraint for cocoa farming.

Rainfall distribution

The study of rainfall distribution revealed that the rainfall regime at the two stations was bimodal. The rainiest months are March, April, May and June for the great rainy season and September and October for  the short season. However, the beginning and the end of the seasons fluctuated from decade to decade. The shift of the beginning of the seasons from decade to decade at Divo is between 30 and 70 days and in Gagnoa between 10 to 20 days with a high probability of installation of the rains from the 2nd 10 days of March in the two stations. The study also revealed an increase in the rainy seasons during the last two decades, namely 1996-2005 and 2006-2015 for the two stations. However, this development has not affected much rainfall, especially at the Divo station since more than 70% of the years during the decade 2006-2015 recorded less than 700 mm of rain required for cocoa farming. Rainfall deficit recorded during this period exposes cocoa trees to water deficit.

The water deficit is a constraint for all crops and more particularly for the cocoa tree, a very sensitive plant. Several studies indeed showed the negative influence of the water deficit on the establishment of a young cacao and on the productivity and sustainability of a mature cacao. Indeed, for a cocoa seedling in the establishment phase some authors (Boyer, 1973; Freud et al., 2000; Petithuguenin, 1995) showed that the water deficit stops the growth, then leads to the death of young cocoa trees with a high mortality rate during the extremely dry years. For example, the mortality rate observed in Togo fluctuates from 15 to 53%  depending on the type of soil (Jagoret and Jadin, 1993). In Côte d'Ivoire; the mortality rate observed in a sample of 50 plots in the East is 39% (Petithuguenin, 1995). The water deficit also delays the development and entry into production of young cocoa trees. This study confirms that, in West Africa, the survival of young cocoa trees, with a shallow root system, is particularly conditioned by a good distribution of rains, on one hand. On the other hand, adult cocoa trees can withstand periods of water deficit greater than 3 months, due to their deep root system; but, their production potential is affected in different ways. When the water deficit occurs a few weeks before flowering, it leads to an increase in the final yield by stimulating flowering. A few weeks after fruit set, the same deficiency causes wilting of the young fruits. If it occurs 3 to 4 months after setting, it causes the reduction of the size of the cocoa beans. This explains the small size of the cocoa beans during the intermediate harvest in West Africa. Similarly, the size of the beans in the main crop is decreasing more and more. Such a phenomenon is linked to the water deficit recorded during the great rainy season (Brou et al., 2003). When the water deficit occurs over several months and is repeated from year to year, it weakens and reduces the longevity of the cocoa trees. This explains the shortenedness of the mature phase and the lateness of the senescent one (Petithuguenin, 1995)

Impact of disturbance of the rainy seasons on cocoa farming

The study of the distribution of rainfall between 1986-1995, 1996-2005 and 2006-2015 revealed a shift in the dates of  the  beginning  and  end  of  the  growing seasons. The time lag is about 30 days between the was observed between the 1986-1995 decade and the decade 1986-1995 and the decade 1996-2005 and 70 days between the decade 1996 to 2005 and the decade 2006 to 20015.However, at Gagnoa, the lag is  less  pronounced.  It  is  30  days  between  the decade 1996-2005 and the decade 2006-20015. Any time lag1996-2005 decade. But, generally, over the period (1986 to 20015), the main rainy season was established at Divo from the 1st 10 days of March to the 2nd 10 days of July and at Gagnoa from the 3rd 10 days of March to the 1st 10 days of  July,  while  in practice the planting of cacao is usually made in May or June. That could partly explain the high mortality of seedlings or even the failure of replanting during the first two years of replantation. Climate change not only affects the agricultural calendar, but also the cacao growth and production cycle. In fact, at eastern Côte d'Ivoire, the first cocoa production zone, the main harvest is shifted from October to December (Brou et al., 2003). Many other studies have revealed that cacao productivity, development and quality are strongly affected by the amount, distribution and duration of rainfall (Almeida and Valle, 2007; Balasimha and Bhat, 1991; Moser et al., 2010). Consequently, cacao is not generally considered resilient in the face of extreme weather conditions, particularly during prolonged periods of drought (Alvim and Kozlowski, 1977; Almeida and Valle, 2007). Cacao’s noted low tolerance to prolonged water limitation is concerning given that climate models predict reduced rainfall, and increased temperatures will become significant limitations to production in the near future, with some areas of West Africa already being affected (Agbongiarhuoyi et al., 2013; Bari et al., 2016; Hutchins et al., 2015).

Solutions for sustainable cocoa production in the two regions

The study revealed a variation in rainfall over the years causing the shifts of the beginning and the end of the main rainy season in the two regions. Under these conditions, producers no longer know when to plant cocoa and associated food crops. This situation increases the vulnerability of rural populations in these two regions to face the adverse effect of climate change. Although the study showed a high probability to have the beginning of the main rainy season by the 2nd 10 days of March in both regions, it would be advantageous to make rainfall predictions and make these information available to farmers on time to help them install their crops at the good time. Similar studies have been done elsewhere (Cetin, 2015a; Cetin et al., 2018; Cetin et al., 2010; Cetin and al., 2016). In addition, the use of climate prediction models in combination with GIS will help define areas suitable for cocoa farming both regionally and nationally (Cetin, 2015b;Cetin,2015c;Kaya et al.,2018;Cetin et al., 2018b; Cetin et al., 2018c). In addition to deforestation, current cocoa full sun production systems have led to land degradation (Tondo et al., 2015; N'guessan et al., 2017; Kone et al., 2012).The promotion of agroforestry systems based on cocoa could not only contribute to restoring soil fertility but also diversify the income sources of cocoa farmers to make them less dependent on cocoa and vulnerable (Koko et al., 2013).

This system also provides stress alleviating services (Jacobi et al., 2015; Verchot et al., 2007; Tscharntke et al., 2010; Beer et al., 1998). In coffee, shade trees protect the plants from drought by reducing the evaporative transpiration demand, and increasing the infiltration capacity of the soil (Lin, 2007). Shade can affect cacao stomatal conductance, ameliorating the effects of drought especially during seedling establishment. Such interaction was observed by Frimpong et al. (1999). Currently, chocolate makers have initiated in several countries the promotion of agroforestry in cocoa farming through numerous projects including REDD+. Drought-resistant plant material for cocoa could also be used in replantation or rehabilitation of farms (Medina and Laliberte, 2017; Smith, 2014). A review was done on the effects of drought and temperature stress and increased CO2 in Theobroma cacao Costa Rica: Bioversity International) and short cycle varieties for associated food crops. However, based on the concept that a healthy cacao tree under balanced nutrition is more likely to be resilient to abiotic stresses, potassium amendment was highlighted for its potential to mitigate the effect of drought in cacao orchards. Interactions between potassium nutrition and drought stress have been observed in other crops – e.g. sorghum (Asgharipour and Heidari, 2011); cassava (Ezui et al., 2017); olive (Erel et al., 2014); and highland banana (Taulya, 2013) – and could be a management opportunity towards drought resilience adaptability in cacao. Aside from depending on potassium to regulate intrinsic osmotic functions, cacao has a specifically high demand for potassium, particularly for pod structure, and so it could benefit from supplemental potassium. Therefore, the best period to apply potassium is thought to be during pod set and development (Almeida and Valle, 2007). Irrigation can also be adopted as a solution to provide cacao with additional water they need to express their potential.

 

The current evolution and distribution of the rains  show a preponderance of moderately dry years in the Lôh Djiboua region and moderately humid in the Gôh region during the three decades studied. In addition, the study showed that the beginning of the main rainy season is between the 1st and 2nd 10 days of March in both regions. However, a high probability of rainfall deficit during rainy seasons, unfavorable to cacao was also revealed. So at the short term planting in both areas could be done earlier in March at the beginning of the rainy season to allow the young plants to benefit from all the rains so as to reduce their mortality rate. But tests will have to be carried out in real environment to confirm these periods. Agroforestry could also be a solution to mitigate the negative effects of climate on cocoa farming. Special attention should also be given to the choice of the soils. These soils should have a good water retention capacity in order to compensate for the rainfall deficit during the rainy seasons.

 

The authors have not declared any conflict of interests.