Developing an integrated soil fertility management decision support tool for arabica coffee (coffea arabica) in selected areas of northern Tanzania

PhD Thesis

Coffee is one of the major export crops in Tanzania, contributing 24% to the agricultural gross domestic product (AGDP). The crop contributes directly to the livelihoods of over 420 000 farm families and indirectly to over 2 million people employed in the coffee value-chain. The Tanzanian average coffee production is variably pegged at 45 000 – 52 000 metric tons annually, while smallholder coffee productivity per tree ranges between 250 and 300g of parchment which is low compared to the world average of 500 – 600 g per tree. In the northern zone, for instance, annual coffee production trend indicates a decline over years as from 1980. During the first coffee stakeholders’ conference in 2009, soil fertility decline was pointed out by representatives of coffee growers as one of the most limiting factors for coffee productivity and sustainability. In the absence of a clear soil fertility intervention strategy in the coffee growing areas, with scanty and incoherent soil fertility data and limitations in their reliability and usability, it would seem impossible to verify the farmers’ claims or devise an intervention pathway. This formed the rationale of this work, whose objective was to develop a system that will make the soil analytical data useful for coffee farming. A model was required to quantitatively translate the soil data into estimated coffee yield, and also to recommend nutrient input application for best returns. This study was undertaken in Hai and Lushoto Districts, Northern Tanzania. The two districts were picked on merit of both growing coffee (thus experiencing the problem of coffee productivity decline) and each belonging to different geological origins (volcanic and metamorphic-gneissic parent materials, respectively). The first task was to establish the farmers’ perception of soil fertility decline as a problem and their attitudes towards integrated soil fertility management (ISFM) for coffee, thereby identifying the appropriate intervention strategies. Based on questionnaire data involving 126 respondents, both farmers’ awareness of the problem iii and their attitude were highly significant (at p<0.01). Age, total land area under coffee and total off-farm income negatively affected farmers’ attitude. As farmers get older, they tend to refrain from innovation. For the two districts, ISFM interventions will make a better impact to younger and more energetic farmers with enough land for coffee production and who depend largely on this crop for their livelihood. It was therefore concluded that the interviewed farmers echoed the concern that their representatives made in 2009. Another study was conducted to assess the soil fertility status against the soil fertility requirements for Arabica coffee, thus scientifically verifying the above concern. A total of 116 soil augerings and 10 soil profiles were described, and soil samples analyzed for the key soil fertility parameters. Soil fertility was assessed qualitatively, quantitatively and spatially. It proved to be considerably low in the study areas, and much lower in Lushoto than in Hai. Immediate recommendations to address the declining soil fertility were given, which include integrated farm management, adequate supply of essential nutrients and building capacity to produce high quality soil data and to interprete them in terms of coffee productivity. Following the two earlier studies, a review of existing crop models was made and a Wageningen model called QUEFTS was picked as a benchmark. It was recalibrated with the coffee crop in mind, and a new model called SAFERNAC (Soil Analysis for Fertility Evaluation and Recommendation on Nutrient Application to Coffee) was developed. The model was checked for accuracy and applicability and found to be capable of reproducing the actual yields by 80-100%. The new model, tested with soils of Hai and Lushoto Districts, proved to be a useful tool for coffee land evaluation and ISFM planning. Through a screen house experiment, different organic materials were tested for nutrient release potential and the data used as inputs to the model for yield estimation under different nutrient management options. Mineral N, P and K release varied significantly (P<0.001) among the organic materials and between the two soil types representative of the study areas. The model demonstrated its potential in iv suggesting appropriate nutrient management options for both organic and conventional farmers, and showed that green manure plants have great potential in coffee ISFM. The model was further expanded to involve prices of inputs and outputs for the determination of net returns and coffee profitability. It was used to obtain yields from a soil of known properties receiving different levels of input N, P and K from both organic and inorganic sources (ISFM). The costs of inputs were derived from experiences in Northern Tanzania, while coffee prices were estimated to range between 1250 and 2500 TZS kg-1. The economically optimum input applications (in equivalent terms) that gave highest net returns and value: cost ratios were found to be 401, 332 and 418 kE ha-1 for soils of low, medium and high fertility, respectively. The recommendations emanating from this study are outlined hereunder: 1. ISFM efforts should focus on younger and more energetic farmers with enough land and who depend largely on coffee for their livelihood. TaCRI and other coffee stakeholders should devise a programme to encourage young people to take coffee farming as a viable business. 2. Factors affecting farmers’ decisions on fertilizer use should be taken into consideration in devising an ISFM strategy for the coffee farmers. 3. TaCRI and the coffee extension machinery at district level should continue to promote the right kind of nutrient management strategy to the farmers. Also, promotion of the improved coffee varieties among farmers should continue. 4. Farmers should be encouraged to come forward and pre-test the model under TaCRI guidance. The pre-testing should include interested organic and conventional farmers around Lyamungu and Yoghoi, specifically to test the target yield and respective ISFM options suggested. v The following activities are envisaged for future perfection of this work: 1. To continue research on the four green manure plants (Mucuna, Lupine, Canavalia and Crotalaria) as to their appropriate application methods in a farm, and notify coffee farmers as soon as the results becomes available. 2. To include in the coffee ISFM programme other plants with nutritive value, such as “tughutu” (Adhatoda engleriana), wild sunflower (Tithonia diversifolia) and fishbean (Tephrosia vogelii). 3. To carry out more research on the applicability of the model to all categories of coffee growers, with issues of shade-grown coffee, intercropped with staple food crops, etc. 4. To integrate the model to more generic agrometeorological models as climate change becomes more and more important in the coffee growing areas. 5. To establish the appropriate entry points in the inclusion of secondary macronutrients and micronutrients to the model and the way they influence the availability of N, P and K to coffee. 6. To collaborate with computer programming specialists and develop a full-fledged software for SAFERNAC.

Prof. James Teri, the Chief Executive Director of Tanzania Coffee Research Institute