Managing transitions in smallholder coffee agroforestry systems of Mount Kenya
Coffee farming has been a major foundation of Kenya’s rural highland economy for the last four decades or so. Over 600,000 smallholder farmers organized in 579 cooperatives are engaged in the subsector. Coffee was a major source of income, employment and food security until the late 1980’s. Though Kenya produces some of the finest world coffee, the collapse of the International Commodity Agreement (ICA) on coffee and entry into the world market by major producers like Vietnam marked a near collapse of Kenya’s coffee. Exports fell by over 50% between the year 2000 and 2010. This was accompanied by significant loss of productivity (declined to a meagre 200 kg/ha from 600 kg/ha). The situation has contributed to poor living standards in coffee growing areas. Interestingly, there are no credible alternative investments to merit the allocation of constrained farm resources to replace coffee growing. In addition, there are concerns that the current resource base can no longer support enhanced productivity. This study used several research designs to investigate the performance of smallholder coffee agroforestry systems around Mount Kenya. More specifically, enterprise adoption and adaptation practices in the event of increased or decreased coffee production were researched. The evolution of coffee agroforestry systems was also evaluated and management of soil fertility determined. Using coffee yields data obtained from 180 smallholder coffee farmers by stratified random sampling techniques, coffee farm typologies were identified. These farm typologies/categories were labeled as increasing, decreasing and constant - representing their historical trends in coffee production. These farms were then used to investigate current productivity behavior. Simple descriptive statistics such as means, range, counts, enterprise scoring, diversity analysis pair wise correlations and regressions were used to compare farmer enterprise intensification strategies. Results have showed that farms that are decreasing coffee production, though had smaller land sizes are not significantly different from those in the coffee increasing category. Further results showed similarities in farmer enterprise diversification strategies. Coffee was nonetheless declining in smaller farms compared to farm sizes where it was increasing. Results also showed that farms with increasing coffee yields are associated with productive milk enterprises. These farms appear to afford and benefit from larger amounts of fertilizer and manure application. Coffee declining farms view banana and maize as likely alternatives to coffee, perhaps in a strategy to secure household food security. The study has showed that land size, coffee production (number of bushes, cherry yields/Ha), livestock units, agroforestry trees, banana, maize value and nutrient inputs (manure and fertilizer) and labour costs are important factors to assess coffee farms productivity and distinguish farm types. Results have showed the importance of creating more awareness among policy makers in order to promote enterprises that are of interest to farmers. This research also investigated tree diversity presently maintained by smallholders showing a shift in coffee cultivation practices. Trees on farm are traditionally appreciated for product benefits such as timber, fuel wood and food. They are also important for enhanced farm biodiversity and environmental services such as enhanced nutrient cycling. This study applied diversity analysis techniques such as species accumulation curves, rènyi diversity profiles and species rank abundance, to investigate farm tree diversity. At least 190 species were recorded from 180 coffee farms. For all the species enumerated, alpha diversity (H0) = 5.25 and H∞ = 0.89. Results showed that the 10 and 25 most abundant species comprise 75% and 91% of tree individuals present on farm, respectively. Results suggest that, though there is high abundance of tree individuals on farms they are of less richness and evenness. Species richness per farm was calculated at 17 species (15- 19.2, P = 0.95). Grevillea robusta was highly ranked in terms of relative density and dominance across surveyed farms at proportions of 41- 42%. Tree species basal area distribution showed that fruit trees such as, Persea americana, Mangifera indica and timber species such as, Cordia africana, Vitex keniensis and Croton macrostachyus are the most dominant but are of lower relative density. Species diversity analysis by coffee agro-ecological zones revealed that the upper-midland (UM) 3 is ranked significantly higher than UM2 and UM1. Results have implied that farmers with larger quantities of coffee (Coffea arabica L.) also retain more species diversity than farmers with stagnated production even though this evidence was inconclusive. Skewed patterns of species heterogeneity and structure among smallholder coffee plots provide indicators of divergent species cultivation. Tree species richness distribution between farms is strongly influenced by agro-ecological zones and presence of coffee cultivation. Only 22.5% of agroforestry tree abundance on farm was categorized as indigenous. Tree basal area ranking implied that fruit and native timber species are retained longer on coffee farms. Finally, this study assessed the implications of recent changes in coffee cultivation on soil fertility management. It was hypothesized that significant soil nutrient exports have occurred from coffee systems and that present nutrient prevalence are unknown and likely to be poorly managed. The purpose of this research was to inform concerns that with poor soil fertility prevalence, coffee systems face a danger to deteriorate to low production systems. Near-infrared (NIR) spectroscopy was used to analyse soil constituent properties for some 189 soil samples collected on 94 farms (within coffee plots). One third of the samples were used to build calibration models giving correlation coefficients between measured and partial least square (PLS) predicted soil properties. Correlations were strong (r > 0.70) except for P, Zn and Na demonstrating the potential of NIR to accurately predict soil constituents. Principal component analysis (PCA) was then used to develop soil nutrient indices (principal components scores) to serve as representative soil nutrient prevalence indicators. PC scores were also used as dependent variables in regression analysis. Collected data is robust to show that soil organic C, total N and probably P were most deficient across the coffee sites surveyed. Farmer nutrient application practices showed wide variability of fertilizer and manure use. Manure application is less than fertilizer and negatively correlated to farm size. Estimation of manure use per household was however challenging due to quantification and timing aspects of application. Collated evidence showed that farmers with increasing coffee production were more likely to afford larger fertilizer and manure application. Overall results point out that smallholders deliberately concentrate nutrient application on farm enterprises with good market performance. Coffee cultivation has in the past benefited from fertilizer credit facilities from farmer cooperative movements and government bilateral programmes. Declined coffee production is therefore seriously jeopardizing the amount of fertilizer that can be loaned to farmers. In conclusion, this study has identified a number of factors associated with smallholder decision making, resource use and enterprise adoption and adaptation behavior within coffee agroforestry systems of Mount Kenya. Research findings have allowed recommendations to be made on how best to promote farmer resource use, understand farmer decision making and enterprise choices that are of interest to farmers. The study has contributed to knowledge of farmer livelihood strategies when managing coffee farms in conditions of reduced profitability.