Description of the study sites

The research sites, Ziway and Melkassa, are situated at 7°9′N and 38°43′E and 8°4′N and 39°31′E respectively, at 1642 and 1550 m.a.s.l. in the CRV of Ethiopia. They are characterized as semi-arid agro-ecological zones (Garedew et al., Reference Garedew, Sandewall, Söderberg and Campbell2009; Meshesha et al., Reference Meshesha, Tsunekawa and Tsubo2012). The soil in the CRV of Ethiopia is poor in fertility (Biazin and Stroosnijder, Reference Biazin and Stroosnijder2012; Temesgen et al., Reference Temesgen, Rockstrom, Savenije, Hoogmoed and Alemu2008) and is shallow soil with frequent crusting and plow pans (Biazin and Stroosnijder, Reference Biazin and Stroosnijder2012). The rainfall condition exhibits high intra-seasonal variability with a coefficient of variation of 15–40%, and temperature increased significantly (0.12–0.54°C per decade) over the past 30 years (Kassie et al., Reference Kassie, Rötter, Hengsdijk, Asseng, van Ittersum, Kahiluto and van Keulen2014).

Ziway has a well-drained clay loam soil (Sime et al., Reference Sime, Aune and Mohammed2015). The total amount of rainfall received over 2011 and 2012 was 598 and 857 mm respectively. The total amount of annual rainfall received during the experimental period of the 2011/2012 and 2012/2013 cropping seasons (June to October) was 434 and 732 mm, which is 73 and 85% of the total annual rainfall received, respectively. Melkassa has a well-drained loam soil (Sime et al., Reference Sime, Aune and Mohammed2015). The total amount of annual rainfall received over 2011 and 2012 was 923 and 924 mm, respectively. The total amount of rainfall received during the experimental period (June to October) for the same years was 685 and 822 mm, which is 74 and 89% of the total annual rainfall received respectively.

The rainfall data were obtained from meteorological stations at the Adami Tullu (in Ziway) and Melkassa Agricultural Research Centers for the Ziway and Melkassa experiments, respectively.

Crop production, mainly rainfed, cereal-based production systems and modest livestock rearing are the mainstays of livelihoods for farmers in the CRV. Maize is the major crop cultivated with significant role in the livelihoods of smallholders in the CRV (Biazin and Stroosnijder, Reference Biazin and Stroosnijder2012). There are two overlapping seasons for crop production. The first season usually extends from April to September depending on the onset of rain while the second season, which is the main rainy season, extends from June to October. July and August are the wettest months. Mid-maturing maize (Zea mays L.) is the main crop for the first (and longer) season while haricot bean (Phaseolus vulgaris L.), wheat (Triticum aestivum L.) and teff (Eragrostis tef (Zucc.) Trotter) are the main crops for the second season. Early-maturing maize is cultivated during both seasons.

Treatments, experimental design and procedures

On-station and on-farm experiments of maize were conducted for two successive years. The on-station experiments were undertaken in Ziway and Melkassa for the cropping seasons of 2011/2012 and 2012/2013. During the same cropping seasons, the on-farm experiment was conducted at Ziway in three villages with nine farmers.

There were five steps, levels or treatments both in the on-station and on-farm experiments:

1. 1. Farmers practice as a control (CTc): CT (4 passes with plough) + non-primed maize seed + no fertilizer.

2. 2. Step 1 (MP): Minimum tillage (1 pass with plough) + primed maize seed + no fertilizer.

3. 3. Step 2 (MPM): Minimum tillage + primed maize seed + 53 kg ha⁻¹ P fertilizer applied at planting.

4. 4. Step 3 (MPMM): Minimum tillage + primed maize seed + 53 kg ha⁻¹ P fertilizer applied at planting + 4 ton ha⁻¹ mulch.

5. 5. Step 4 (MPMMU): Minimum tillage + primed maize seed + 53 kg ha⁻¹ P fertilizer applied at planting + 4 ton ha⁻¹ mulch + 53 kg ha⁻¹ N fertilizer applied at maize knee height.

The sources of P and N fertilizer were diammonium phosphate (DAP) and urea respectively.

The first level in the ladder represented farmers practice, whereas the other steps were used to give farmers different technology options in maize production (Figure 1). Most of these technologies are cost-saving technologies besides their potential to mitigate the effects of seasonal rainfall variability on maize production. Except seed priming, the individual effects of these technologies on maize production were studied in the same sites during the same period (Sime and Aune, Reference Sime and Aune2014; Sime et al., Reference Sime, Aune and Mohammed2015). At the same time, the same research team was interested to investigate the sequential effects of these technologies as a package on the agronomic and economic responses in maize. The choice of technologies was based on the major production constraints in the study area. They were packaged and sequenced in the order of the production costs, importance of inputs and level of availability of inputs. Minimum tillage was introduced as an option to conventional tillage with more tillage requirement and rental costs (Aune et al., Reference Aune, Bussa, Asfaw and Ayele2001; Sime et al., Reference Sime, Aune and Mohammed2015). In Ethiopia, fertilizer is expensive to most farmers, despite its unavailability and traditional way of application (CSA, 2008; Spielman et al., Reference Spielman, Byerlee, Alemu and Kelemework2010). Due to high prices, the use of urea is limited as a result, there is a failure to use DAP and urea in proper combinations (Endale, Reference Endale, Gebre-ab and Dorosh2010). This was the reason for the introduction of the fertilizer microdosing method, the pocket application of fertilizers beside seeds, which saves fertilizers besides increasing crop fertilizer use efficiency (Aune and Bationo, Reference Aune and Bationo2008; Sime and Aune, Reference Sime and Aune2014). The other constraint to maize production in relation to the impact of rainfall variability include poor seed germination, seedling establishment and partial or complete crop failure, that was the reason for the introduction of on-farm seed priming as an option for dry seed sowing. Seed priming is cheap and does not add any expenditure to farmers (Aune and Bationo, Reference Aune and Bationo2008) and improves maize establishment and yields in semiarid agriculture (Harris et al., Reference Harris, Joshi, Khan, Gothkar and Sodhi1999). Yield losses or no yield benefits are likely in the short-term practice of conservation tillage, which may need up to 10 years (Giller et al., Reference Giller, Witter, Corbeels and Tittonell2009). This was the reason for cofounding minimum tillage with seed priming. Mulching was used earlier than urea, because it is cheaper and more available to most farmers which otherwise would have been brought to the end of the ladder. It was introduced as part of the technologies to mitigate the effect of seasonal rainfall variability on maize production. In this regard, mulching was reported to give a promising significance to maize production in the region (Sime et al., Reference Sime, Aune and Mohammed2015).

Figure 1.

A ladder showing different level of inputs as package of technologies.

Key: CTc: conventional tillage (CT) as a control (c); MP: minimum tillage (M) and seed priming (P); MD: microdosing of DAP fertilizer; M: surface mulching with maize stover; and U: microdosing of urea fertilizer.

The crop tested was drought tolerant early maturing variety of maize (Z. mays L. var. Melkassa-2). Maize seed was primed by soaking in water for 12 hours, then surface drying for 2 hours on dry cloth and being kept under shade and sowed immediately. Fertilizer and seeds were applied in two separate operations.

The plot size was 3 m× 4.5 m (13.5 m²). The spacing was 75 cm between rows and 25 cm between plants. The seed rate was 27 kg ha⁻¹. The treatments were arranged in a randomized complete block design with four replications. The blocks were separated by a 1.5 m wide open space. Hand weeding was undertaken using a local hand hoe three and six weeks after planting. Permanent plots were used for the entire period of the experiments. Mulched treatments retained their maize stalks as mulching. A 4 ton ha⁻¹ mulch was applied based on Matsumoto et al. (Reference Matsumoto, Plucknett and Mohamed2004) to control weed and increase soil moisture in the drylands of Ethiopia.

The different treatments presented above can be considered as different levels in the ladder approach (Aune and Bationo, Reference Aune and Bationo2008). There were four levels in the ladder to intensify the productivity of maize. The inputs were successively added to the first, second, third and fourth level in the ladder (Figure 2).

Figure 2.

Cumulative rainfall (mm) at Ziway (left) and Melkassa (right) during the experimental period from first of June to November of 2011 and 2012.

Source: The sources of the rainfall data for Ziway and Melkassa sites are Adami Tulu and Melkassa Agricultural Research Centers repectively.

Farmers at the study sites actively participated in the planting, weeding and harvesting; providing the main source of labour. The on-farm experiments were managed mainly by farmers and partly by researchers, following jointly agreed protocols.

Agronomic data measurements

The major agronomic data collected include days to emergence, pocket seed germination (50%), seedling vigour (rated 1–5 where: 1 = poor, 2 = low, 3 = moderate, 4 = vigorous, 5 = very vigorous), days to tassel (50%) and to physiological maturity (75%), lodging count, plant height (cm), and grain and stover yield (kg ha⁻¹). Plants fallen, inclined or with broken stalk were considered as lodged. Plant height was measured from the ground level to the base of the tassel for five randomly selected plants per plot. Biomass weight was taken after sun drying for 8 days. Threshing was done manually. To avoid border effects, data of all the parameters were taken from the four central rows, thus, the net plot size was 9 m². After harvesting cobs, for each treatment, cobs were shelled, weighed and grain moisture content was measured immediately by a multigrain digital moisture meter. Eventually, the grain yield weight was adjusted at 12.5% moisture level. Yield was extrapolated and then reported on a hectare basis.

Economic data measurements

The agronomic efficiency, risk and profitability of using fertilizers were assessed by calculating the fertilizer use efficiency (FUE), value cost ratio (VCR) and gross margin (GM).

Agronomic fertilizer use efficiency

The agronomic FUE of each treatment was computed as the difference in yield (kg ha⁻¹) between each treatment and control divided by the amount of fertilizer applied (kg ha⁻¹).

\begin{equation*} FUE = \frac{{Yt \times Ct}}{{Ft}}, \end{equation*}

where FUE is the agronomic fertilizer use efficiency of treatment t; Yt is the grain yield of treatment t; Ct the grain yield of the control treatment; and Ft is the rate of fertilizer used for treatment t.

Economics of fertilizer use

Standard enterprise budgeting techniques were used to estimate production costs and profitability. GM and VCR were used for the analysis of profitability of each treatment. Average labour cost for the various farm activities of planting, fertilizer application, mulching, weeding and harvesting was estimated as person⁻¹ day⁻¹. Oxen rental cost for one pass was estimated at 11 US$ ha⁻¹. The input costs of maize seeds and fertilizer DAP and urea kg⁻¹ was estimated at 1.14, and 0.82 and 0.63 US$ respectively at the local market. The price of grain, averaged over 2011 and 2012, was estimated at 0.23 US$ kg⁻¹ at local markets. Gross income (GI) was calculated from grain price. Labour cost was estimated at 1.64 US$ person⁻¹ day⁻¹ across the sites (the average rate paid by research institutions and private organizations). The study sites have similar input and output markets. Total variable cost (TVC) was calculated as the sum of labour and input costs. Monetary values related to costs and incomes were converted from Ethiopian Birr at the exchange rate of one US$ to 18.24 Ethiopian Birr.

For each treatment, the GM was calculated as follows:

\begin{equation*} {\rm GM} = {\rm GI} - {\rm TVC} \end{equation*}

The VCR was calculated as follows:

\begin{equation*} {\rm VCR} = \frac{{\Delta Y \times p}}{{Cf}}, \end{equation*}

Where ΔY denotes incremental grain yield resulting from treatment, p denotes grain yield price kg^–1, and Cf denotes fertilizer cost ha^–1.

The VCR for each treatment measures the increase in revenue relative to the increased cost of fertilizer compared to the control treatment.

Statistical analyses

The SAS System (Version 9.3) from SAS Institute Inc. (SAS, 2011) was used for the statistical evaluation of treatment effects on maize yield and yield characteristics compared to the control. To determine statistical significances between differences in means of the treatments, analyses of variance were carried out. Wherever there were significant differences, mean separations were carried out using least significant difference (LSD). Significant differences between means of treatments were determined at 5% significance level (p value < 0.05). Unless mentioned in the text, only significant differences between means of treatments were reported.