Factors influencing Maize yield Gaps on Smallholder farms in Vihiga and Kakamega Counties of Western Kenya

Abstract

Despite decades of investment in new agricultural technologies, crop yields of main staple crops such as maize (Zea mays), rice (Oryza sativa) and wheat (Triticum aestivum) continue to stagnate in many parts of Sub Saharan Africa. As a result, there have been large yield gaps; the difference between potential and actual yield >50%. Research has shown that factors causing yield gaps ranging from biophysical, field management and socio-economic are studied in isolation. An inclusive approach is needed where biophysical, socio-economic and field management factors are studied to enhance yields. The overall objective was to provide insights into factors influencing yield gaps by studying socio-economic, management and biophysical variables on maize fields at a farmer level as a solution to enhancing yields. The specific objectives were: To assess biophysical factors influencing maize yield gaps; To determine the effect of farmer derived management on maize yield variability; To analyze the effect of the interrelationship between socio-economic, management and biophysical factors on maize yield gaps; To determine the effect of spatial arrangements (fields) differentiated by distance from the homestead on maize yield gaps. The study was conducted in two contrasting sites; Mukuyu and Shikomoli of Western Kenya for a period of two years; 2016 and 2017. The sites contrast in agro-ecology, market access and population density. Multi-stage sampling design was adopted to select regions and villages followed by random selection of 70 households; 35 households in Mukuyu and 35 in Shikomoli. A total number of 170 maize fields which were the study units were identified and georeferenced from the 70 households. In the year 2016, soil sampling and analyses were done to characterize soil properties. Field measurements to determine within season biophysical variables were done at two key maize development stages; ear initiation (stage 1) and tasseling and silking (stage 3). Maize output was also collected and yield determined per hectare. Yield gaps were then computed at a farmer level by comparing the determined yields at the 90th percentile to other yields. Household surveys were conducted to collect field management and socio-economic factors. Satellite imagery was acquired and processed to map yield gap variability at different spatial arrangements with respect to distance from the homestead. In the year 2017, on farm trial plots with best and average farmer derived management practices were laid out on 33 smallholder farms using the randomized complete block design. The management practices were based on survey findings from 2016 and included; nutrient supply, weed management and plant density. An integrated analysis comprising the Generalized Linear Mixed Model (GLMM), Classification and Regression Tree analysis (CART), Factor Analysis (FA), Linear Mixed Effects Model analysis (LMER) and Spatial Analysis Techniques was used to analyze the collected data. Results showed that the average measured maize yield and yield gaps for Mukuyu were 3.8 t ha-1 and 1.8 t ha-1 while for Shikomoli they were 2.7 t ha-1 and 2.6 ha-1 respectively. This represented 35% and 54% of unachieved yields for Mukuyu and Shikomoli respectively. Factor Analysis showed socio-economic variables as the overarching factor influencing maize yield gaps over biophysical and management across the two sites. The GLMM identified education, age, membership to groups, access to markets, family labour, gender, credit facility, maize variety, crop residue utilization insitu, quantity of organic and inorganic fertilizer use, while CART identified maize density, chlorophyll values, maize height, and depth to compact layer as consistent factors affecting yield at both sites. Also, according to CART weed cover at early stages and maize density at late stages were the most limiting factors in maize production in Mukuyu and Shikomoli, respectively. The GLMM analysis also showed a two-way significant interaction effect between socio-economic, management and biophysical factors on maize yield gaps which was agro-ecology specific. In Mukuyu inorganic fertilizer use and gender of operator as female, weed coverage at early maize stages and crop residue utilization as animal feed, positively interacted to influence maize yield gaps. While low weed coverage at early maize stages and phosphorus, depth of compaction and crop residue use insitu, number of organic fertilizer and cation exchange capacity, negatively interacted to influence maize yield gaps. In Shikomoli, membership to groups and timeliness in execution of agronomic activities such as land preparation, planting and weeding negatively interacted to influence maize yield gaps. The LMER analysis on on-farm trial data revealed that highest yields were recorded on the best farmer derived management treatments and averaged 7.8 and 6.6 t ha-1 for Mukuyu and Shikomoli. These yields represented 45 and 35% between farm and inter-annual yield variation when compared to average-derived farmer practices and best farmer management practices from past surveys respectively. Spatial analysis techniques demonstrated that heterogeneous patterns of high, average and low yield gaps were found on fields closer to the homestead. While nearly homogenous yield gap patterns were found on fields further from the homestead. Factors such as inorganic fertilizer use, weed control, early land preparation, hired and family labour use and large land sizes were utilized on spatial arrangements further the homestead. Organic fertilizer and family labour use was utilized on fields closer to the homestead. The findings indicate that large yield gaps > 30% exist on smallholder farms showing a scope for farmers to exploit the gap. The findings also demonstrate that an integrated approach can result in consistent, agroecology specific and interacting factors influencing yield gaps applicable at different scales of decision making; farmer, local, county, national and regional in improving yields. The high yields from the on-farm trial research (best management plots) demonstrate the potential to reduce maize yield gaps on smallholdings. Since maize is a staple crop in Kenya and in most parts across the globe, policy measures aimed at improving general soil fertility, market accessibility, relaying agricultural information and encouraging family involvement in agronomic activities are needed. Agro-ecology and field specific measures focused on improving particular soil nutrient types and levels, weed management and plant density are also required. Delineating management zones based on yield gap patterns will also help promote field-specific land management to enhance yields. Further research in yield gap studies could focus on the effect of post-harvest handling practices in reducing yields and in using crowdsourcing methods via innovatively developed mobile applications to collect data.