What shapes yields of East African Highland banana? An explorative study from farmers’ fields

Context or problem East African Highland banana is a main staple and cash crop for millions of people in the Great Lakes region of Africa. Reliable methods for banana yield estimation are lacking and factors underlying yield variability are not well understood. Objective or research question We studied the interrelations and relative importance of yield components (yield per mat, mat density, bunch weight, cycle duration and the number of bunches harvested per mat per year or per hectare per year) to on-farm yields. We also explored various methods to estimate yields and tested hypotheses on influence of several biotic and abiotic factors on yields in farmers’ fields. Methods From July 2019 until June 2021, yields, bunch weights and management practices in 118 fields in West and South-West Uganda were monitored and detailed data on 10 mats per field were collected monthly. This was complemented by a one-off bunch weight survey covering 171 fields. Results Median yields in West and South-West Uganda were respectively 31 and 32 ton fresh weight (FW)/ha/year in 2019/2020 and 38 and 36 ton FW/ha/year in 2020/2021. Bunch weights and the number of bunches harvested per hectare were major determinants of yields, while mat density was of minor importance. However, higher mat density was correlated with lighter bunches, longer cycle durations and fewer bunches harvested per mat. Fields that were reportedly manured in the past produced heavier bunches, but the effect of manure applied in 2019/2020 on yields in 2020/2021 was not significant. Higher yields and heavier bunches were associated with mulching, but the underlying causality needs unravelling. Banana Xanthomonas Wilt (BXW) was a major yield-reducing factor. Farmers had to remove diseased suckers in 38% of the monitored fields. Modest amounts of mineral fertilizer had a positive effect on yields and bunch weights, but only in absence of BXW. Conclusions Bunch weights and the number of bunches harvested per hectare are major determinants of banana yields. Mineral fertilizer application increases yields within 12 months in the absence of BXW. Manure application seems to need longer duration for effect. On-farm banana yields are higher than reported in literature. Implications and significance Yield estimation methods need to be standardized. Future initiatives for improving banana yields should focus on controlling BXW prior to advocating input use. Research should be conducted to unravel the effect of mulch and different combinations and rates of manure and mineral fertilizer on banana yields.