By Samuel Njoroge and Shamie Zingore
Smallholder maize productions systems in western Kenya exhibited wide spatial and temporal variability in yield responses to N, P and K application at the field level. Nutrient omission trials provide evidence of strong and ubiquitous N limitations and high vulnerability for rapid decline in P and K under continuous cropping without balanced nutrient applications.
Crop production in smallholder systems in sub-Saharan Africa (SSA) is strongly limited by poor soil fertility that results from continuous cropping with low nutrient applications (Kihara et al., 2015). Subsequently, crop yields in smallholder farming systems in SSA are far below potential yields (Van Ittersum et al., 2016). The yields of cereal crops in SSA can be readily raised (Mueller et al., 2012) when nutrient deficiencies are addressed (Adediran and Banjoko, 1995). However, profound variability of crop yield responses to nutrient applications across and within fields presents challenges for effective nutrient management.
There is limited information on how crop yield responses to nutrient applications vary over time, as most on-farm experiments are only conducted for a limited number of growing seasons. Such information is necessary to fine-tune site-specific fertilizer recommendations for improving crop productivity, fertilizer use efficiency, and profitability in the long-term. This is particularly important considering the ongoing fertilizer crisis which is severely limiting the ability of smallholder farmers in SSA to access and afford fertilizer inputs. This article outlines results from a series of on-farm nutrient omission trials conducted in the same experimental plots over six consecutive seasons to quantify variations in maize yield response to varied N, P and K applications over space and time.
Field trial description
The study was conducted within a 10 km x 10 km study site in Sidindi, western Kenya. Annual rainfall ranges between 1,600 to 2,000 mm and is distributed over two distinct seasons with a long rainy (LR) season from March to July, and a short rainy (SR) season from September to December. On-farm nutrient omission trials with maize were established in 24 sites representative of major soil units in the study area. The experiment included a set of five treatments including a control (no nutrients applied), P+K, N+K, N+P, and N+P+K established in plots measuring 10 m x 10 m, with each farm serving as a complete block. Throughout the experimental period, short-season maize variety DK8031 was planted at the recommended spacing. Trial plot locations and allocated treatments remained the same throughout the study period.
At physiological maturity, all maize plants were harvested within a 2.25 m x 3 m net plot. Total cob weight and grain moisture content were determined, and grain yield calculated and expressed as 88% dry matter. The effect of treatment on grain yield was analyzed at a seasonal level using a generalized linear model, and differences in treatment means evaluated for significance using a Tukey HSD test in R software. Differences in yield variation between treatments were assessed by calculating coefficient of variation (CV) values for each treatment in each season. Scatter plots of CV values and seasons were then constructed, and regression lines fit for trend assessment. Relative yield (RY) was used as a measure of the yield responses to N, P and K and was determined as the ratio between nutrient limited yield and yield in the NPK plot (Equation 1). Relative yield values < 1 indicate a response to the applied nutrient, while values ≥ 1 indicate no response to the applied nutrient.
Eq. 1: RYi,j,s = GYi,j,s / GYnpk,j,s
Where:
RYi,j,s = Relative yield in treatment plot i at field j in season s
GYi,j,s = Grain yield in treatment plot i at field j in season s
GYnpk,j,s = Grain yield in the NPK treatment plot at field j in season s
Effect of imbalanced and balanced nutrient applications on maize yields
Mean maize yields were consistently largest in the balanced NPK treatment, and smallest in the control treatment (Table 1). Yield losses resulting from nutrient omission were large and increased over time as indicated by yield differences between the NPK and control treatments of 2.7, 2.8, 3.4, 3.4, 3.5, and 4.2 t ha-1 in the six consecutive cropping seasons, respectively. Imbalanced nutrient applications also resulted in substantial yield losses with mean yields in the PK treatment being significantly smaller than with NPK in all seasons. While yields in the NK and NP treatments were not significantly different from NPK yields in the first season, significantly smaller yields were observed in the NK treatment in all five subsequent seasons, and in the third and last season for the NP treatment (Table 1). In the last season, sustained N, P and K omission resulted in mean yield losses of 3.9, 2.4 and 1.7 t ha-1, respectively.
Table 1. Mean maize grain yields in t ha-1 at 88% dry matter for on-farm (n=24) nutrient omission trials conducted over six consecutive seasons in Sidindi, western Kenya.
‡ LR and SR refer to long and short rains seasons respectively.
HSD refers to the honest significant differences between means as per the Tukey test.
Imbalanced nutrient applications were characterized by larger variations in maize yield responses (Fig. 1). On average, variability was greatest in the control treatment followed by the NK treatment and was lowest in the NPK treatment. Variability remained constant for the NPK treatments but increased significantly (P<0.05) for the control and NP treatments.
Variations in yield response to N, P and K over space and time
Cumulative frequency plots of relative yields depict the spatial and temporal variations in yield responses to N, P and K (Fig. 2). Responses to N, P and K differed between fields within a season (spatial variation) as did responses to a particular nutrient across seasons (temporal variation). The range of relative yield values for each cumulative frequency line indicates the variability in nutrient responses between fields, while a shift in of the cumulative frequency lines from one season to another indicates temporal variations in observed responses (Fig. 2).
In the first season, strong responses to N (RYPK < 0.5) were observed in only 29% of fields. In the subsequent five seasons, the percentage of fields strongly responsive to N (RYPK < 0.5) increased to 48, 57, 57, 61, and 96%, respectively. For P, only 4% of fields showed a strong response to P (RYNK < 0.5) in the first season. In the subsequent five seasons, 22, 30, 35, 26, and 43% of fields were strongly responsive to P (RYNK < 0.5), respectively. Lastly for K, the first season indicated that only 4% of fields where strongly responsive to K (RYPK < 0.5), while in subsequent seasons this increased to 17, 13, 9, 13, and 30%. Although the proportion of fields responsive to P and K were smaller than those responsive to N, the effects of P and K omission in deficient fields were drastic, with yields losses of up to 80% relative to the NPK treatment in some of these farms, particularly from the second cropping season onwards (Fig. 2b and 2c).
Findings from this study confirm the strong effects of soil fertility variability on maize productivity and nutrient requirements. Variability was largest in imbalanced treatments and least with balanced NPK applications, reflecting the potential of balanced fertilization in managing variability in crop yield responses in such smallholder farming systems. Cropping with no nutrient applications resulted in large and substantially increasing yield losses, while cropping with imbalanced nutrient applications also resulted in substantial yield losses.
Nitrogen was deficient on most farms, although the observed responses differed between farms in a season. However, these spatial differences in response to N decreased over time as illustrated by the narrower range of relative yields in the last seasons. Temporal differences in response to N were weak as illustrated by the minimal change in mean RYPK over time, and by the minimal shift in the location and spread of the RYPK cumulative frequency lines particularly in the first five cropping seasons.
Smaller spatial-temporal differences in response to N will limit the scope for improving nitrogen use efficiency (NUE) through targeted N application based on the spatial-temporal N response patterns. Yet N application is still critical across all farms in all seasons, and NUE can be improved by focusing on better timing and placement of fertilizer N to match crop requirements.
Large mean RYNK yields indicated that maize yield response to applied P was not significant in the first season. This was likely due to residual effects of P applied in previous
seasons (Janssen et al., 1987; Kifuko et al., 2007). However, the large variability in RYNK observed between farms in the first season, with about half of fields showing no response to P, indicates strong spatial variation. Differences in yield response to P can be linked to differences in the P fertility status of the soil, reflecting differences in historical field management (Vanlauwe et al., 2006). Omitting P for more than one season resulted in significant reductions in yield as indicated by the significantly smaller mean RYNK values, and an increasing proportion of fields responsive to P. This indicates the need for regular application of P to sustain productivity.
Strong spatial-temporal patterns in response to K were also observed. Two out of 24 fields showed very strong response to K, while a decline in mean RYNP was observed over time. The strong K deficiencies in a limited number of sites could be due to the presence of localized K deficiency hotspots (Kihara et al., 2016), and continuous removal of harvest products without application of K fertilizer (Chianu and Mairura, 2012; Zörb et al., 2014). Further, K deficiencies are expected to become more pronounced at higher N and P application rates. Fertilizer recommendations should therefore account for the need to supply K in combination with N and P, particularly in K deficiency hotspots (Kihara et al., 2016).
Given the prevailing fertilizer crisis that is limiting the ability of farmers to access and afford fertilizer applications required to attain and sustain high yields, findings from this study indicate that farmers in such smallholder settings will face yield losses of up to 50% if they drastically reduce fertilizer applications. Such yield reductions would substantially impact crop productivity, food security and farmer incomes.
Findings from this study however provide potential short-term nutrient management options that farmers can apply to mitigate severe yield losses. In low fertility soils, yield losses can be reduced by applying reduced quantities of balanced NPK applications as such soils are expected to be deficient in N, P and K. In moderate to high fertility soils that have had previous large applications of organic or inorganic
fertilizers, severe yield losses can be mitigated by committing available resources to accessing and applying N fertilizers plus modest applications of P fertilizers, as such soils are expected to be mostly deficient in N. It should however be noted that these are short term (1 to 2 season) measures, and that balanced fertilization is required for long-term sustainability.
Summary
Strong spatial-temporal differences in maize yield responses to N, P and K were observed in smallholder farming systems in western Kenya. Fertilizer recommendations must account for such spatial and temporal differences for efficient and effective nutrient management. Improved fertilizer recommendations are critical for buffering the impact of the prevailing fertilizer crisis by improving crop productivity, nutrient use efficiency and farmer incomes, and reducing the risk associated with uncertainties in crop yield responses to fertilizer applications. Short-term measures involving reduced rates of balanced NPK applications in low fertility soils, and a focus on N and modest P application rates in moderate fertility soils can help mitigate against severe yield losses resulting from limitations in access to and affordability of fertilizer among smallholder farmers due to the on-going fertilizer crisis.
Dr. Njoroge is Program Coordinator at the African Plant Nutrition Institute (APNI), Nairobi, Kenya; e-mail: s.njoroge@apni.net. Dr. Zingore is Director of Research & Development at APNI, Benguérir, Morocco.
Acknowledgement
This article is adapted from Njoroge et al. 2017. Strong spatial-temporal patterns in maize yield response to nutrient additions in African smallholder farms. Field Crops Research 214 (Supplement C) 321 – 330. https://doi.org/10.1016/j.fcr.2017.09.026.
Cite this article
Njoroge, S. and Zingore, S. 2022. Variability in yield response strongly affects maize productivity and nutrient requirements, Growing Africa 1(2), 15-19. https://doi.org/10.55693/ga12.EWJD1663
REFERENCES
Adediran, J.A. and Banjoko, V.A. 1995. Response of maize to nitrogen, phosphorus, and potassium fertilizers in the savanna zones of Nigeria. Comm. Soil Sci. Plant Analy., 26(3-4), 593-606.
Chianu, J.N. and Mairura, F. 2012. Mineral fertilizers in the farming systems of sub-Saharan Africa. A review. Agron. Sust. Dev., 32(2), 545-566.
Janssen, B.H., Lathwell, D.J., and Wolf, J. 1987. Modeling long-term crop response to fertilizer phosphorus. II. Comparison with field results. Agron. J., 79(3), 452-458.
Kifuko, M.N. et al. 2007. Effect of combining organic residues with Minjingu phosphate rock on sorption and availability of phosphorus and maize production in acid soils of western Kenya. Exp. Agric., 43(1), 51-66.
Kihara, J. et al. 2015. Maize response to macronutrients and potential for profitability in sub-Saharan Africa. Nutr. Cycl. Agroecosys., 171-181.
Kihara, J. et al. 2016. Understanding variability in crop response to fertilizer and amendments in sub-Saharan Africa. Agri., Ecosys. and Env. 229, 1-12.
Mueller, N.D. et al. 2012. Closing yield gaps through nutrient and water management. Nature, 490(7419), 254-257.
Van Ittersum, M.K. et al. 2016. Can sub-Saharan Africa feed itself? Proc. Nat. Acad. Sci., 113(52), 14964-14969.
Vanlauwe, B., Tittonell, P., and Mukalama, J. 2006. Within-farm soil fertility gradients affect response of maize to fertiliser application in western Kenya. Nutr. Cyc. Agroecosys., 76(2-3), 171-182.
Zörb, C., Senbayram, M., and Peiter, E. 2014. Potassium in agriculture – Status and perspectives. J. Plant Physio., 171(9), 656-669.
