Description of studied area

Field experiments were conducted during the dry cropping seasons of 2016 and 2017 at Melkassa Agricultural Research Center (8°24′N latitude and 39°21′E longitude at altitude of 1,550 masl). The center is located in the semi-arid part of Awash basin of Ethiopia, downstream of Koka dam.

The study area receives about 818 mm rainfall annually. On the other hand, the evapotranspiration demand of the area is about 2,567 mm/year. Greater percentage of the rainfall (67%) occurs from mid-June to mid-September and the maximum is in July (Figure 1). The water balance in the study area indicates the need for irrigation throughout the year except during the months of July and August (Figure 1). The rainfall occurred during the study period was greater than the long-term average rainfall of the area, while the reference evapotranspiration was less than the average of the long-term average.

Figure 1.

Relationship between reference evaporation (ETo) and rainfall (RF) during the study period and the long-term averages at Melkassa, Ethiopia.

The mean monthly temperature recorded during the study period was greater than the long-term mean monthly temperature recorded in the study area (Figure 2).

Figure 2.

Average long-term and study period mean temperature comparison at Melkassa, Ethiopia.

The textural class of the soil of the study area was clay loam. Table 1 shows some of the physiochemical properties of soils of the experimental site.

Table 1.

Physical and Chemical Characteristics of the Study Site Soil at Melkassa, Ethiopia.

Treatment and design

The experiment consisted three furrow irrigation systems (CFI, AFI, and FFI) and three levels of irrigation applications (100%ETc, 75%ETc, and 50%ETc). The 100%ETc application is a control treatment based on allowable soil moisture depletion (p). The experimental treatments had a split plot arrangement in a randomized complete block design with three blocks in which furrow irrigation system was assigned to main plots while irrigation levels were assigned to sub-plots.

Maize agronomy

Maize seed cultivar Melkassa-II, suitable for dry areas, was sown on clay loam soil. Plots were 4.5 m × 5 m. The seed was sown at plant and row spacing of 25 and 75 cm, respectively, with a total plant population of 100 per plot. The spacing between plots of the same block was 1.5 m and the spacing between blocks was 3 m. Planting date was January 14 and January 23 for the 2016 and 2017 experiments, respectively. The treatments received di-ammonium phosphate, DAP, ((NH₄)₂HPO₄) at a rate of 100 kg ha⁻¹ during planting and two times split application of urea (CO(NH₂)₂) at planting and knee height (40 days after planting) at rate of 50 kg ha⁻¹. The experimental treatments received two common irrigations one at planting and the other after germination for better crop establishment. The plant was harvested on May 23 and May 29 for the 2016 and 2017 experiments, respectively. Both years the experiment was conducted at the same site following main season haricot bean harvesting.

Irrigation water management

Irrigation applications to meet treatment levels were based on allowable soil moisture depletion (p = .50) of the total available water in the crop root depth (Allen et al., 1998), which for maize in the present study was 90 cm. Daily climatic data (maximum and minimum temperatures, humidity, wind speed and actual sunshine hours) were used in CROPWAT 8.0 for Windows (Food and Agriculture Organization, 1992) to compute reference evapotranspiration, ETo. The crop evapotranspiration, ETc was estimated using “Kc ETo” approach.

The sum of daily ETc was added between two irrigation events following the water balance equation (1).

(1)

Where, ETc is crop evapotranspiration (mm), P is precipitation (mm), I is irrigation (mm), D is deep percolation (mm), and ΔS is change in soil water storage (mm).

Irrigation was provided for the control treatment when 50% of the total available water within the root depth was depleted. The remaining treatments received the allocated percentage of control treatment in the same date. The gross irrigation application was estimated by using average application efficiency of 60% for furrow irrigation (Raine & Bakker, 1996). The required depth of irrigation water was applied using 3-inch Parshall flume.

Data collection

The crop was harvested after physiological maturity, when relatively low grain moisture content was attained (18%–25%) (SESI Technologies, 2020). Data on aboveground biomass and grain yield were collected from each plot. Stover weight was measured after oven drying at 70°C temperature for 48 hours. Maize cobs were harvested and shelled. The grain was weighed, grain moisture measured using digital moisture meter and eventually corrected for moisture content at 12.5%. Yield was extrapolated and then reported on a hectare basis. To avoid border effects, both stover and cobs were collected from the three central rows, and two plants were left from both ends of the furrow length, with a net plot size of 9 m².

Crop WUE was computed as a ratio of grain yield (kg/ha) to seasonal crop evapotranspiration (mm ha/ha). Harvest index based on maize grain yield and biomass yield was calculated using equation (2) (Du et al., 2010):

(2)

Where, HI is the harvest index, Y_g (kg ha⁻¹) is the grain yield, and Y_b (kg ha⁻¹) is the biomass yield.

The yield response factor, Ky, was estimated following equation (3) (Doorenbos & Kassam, 1979).

(3)

where, Ya and ETa are, respectively, the actual yield and actual ET for the deficit treatments, Ym and ETm are the maximum yield and maximum ET, respectively; obtained from the fully irrigated treatment, and (1–Ya/Ym) is the relative yield decrease to the corresponding relative ET deficit (1–ETa/ETm).

Data analysis

The obtained data were subjected to analysis of variance suitable for split plot design and year considered to be random effect while furrow method and irrigation level were considered to be fixed effects. Analysis of variance tests were performed for the model using the Mixed procedure (SAS version 9.00, SAS Institute) with block considered to be a random effect (KLI, 2007). Mean separations were conducted using the SAS macro pdmix800 (Saxton, 1998) with Fisher’s LSD at an error rate of 5% (p = .05).

Depth of irrigation water applied

Irrigation water depth of 594.10 mm was given to the maize for its entire growing period under conventional furrow irrigation with full irrigation (CFI 100%ETc). For CFI 75%ETc, 75% of full irrigation (445.58 mm) was applied, while for CFI 50%ETc irrigation depth of 297.05 mm was applied in CFI. In both AFI 100%ETc and FFI 100%ETc treatments, total irrigation depth of 297.05 mm was given under full irrigation treatment, because water was applied to half of the plot area through alternate and specific furrows, respectively. For AFI 75%ETc, AFI 50%ETc, FFI 75%ETc, and FFI 50%ETc irrigation depth of 222.79 mm and 148.53 mm of water was applied respectively.

Maize grain yield and water use efficiency

Maize grain yield and WUE were not significantly different between years (Table 2). No significant interaction between years and furrow method, and year × furrow method × irrigation level was observed for grain yield and WUE (Table 2). The analysis of variance showed significant variation on maize grain yield and WUE among the different furrow irrigation methods and deficit irrigation levels (Table 2). Grain yield obtained under the main plot factor was significantly greater for CFI than AFI, while the lesser was from FFI (Table 3). However, no significance difference in WUE was observed due to furrow method. The difference in grain yield recorded due to difference in subplot factor, irrigation level, was significant. Greater grain yield was obtained for 100%ETc followed by 75%ETc, and lesser value was obtained from 50%ETc. WUE was statistically similar for 75 and 100%ETc, but lesser for 50%ETc (Table 3). Among the furrow methods, FFI had the least performance in grain yield and WUE compared to CFI and AFI (Table 3).

Table 2.

Significance of Furrow Method, Irrigation Level, Furrow Method × Irrigation Level, Replication, Replication × Furrow Method, Year, Year × Furrow Method, and Year × Furrow Method × Irrigation Level on Maize Parameters Over Two Years and Three Randomized Complete Blocks at Melkassa, Ethiopia..

Table 3.

Grain Yield, WUE, Biomass and Harvest Index for Furrow Method and Irrigation Level on Maize at Melkassa, Ethiopia. Values are the Means of Two Years and Three Replicates.

Significance differences were observed on maize grain yield and WUE by the interaction effect of furrow methods and irrigation levels (Table 2). The greatest grain yields were measured for CFI 100%ETc and CFI 75%ETc, which were not significantly different from each other (Table 4). Application of CFI 75%ETc gave the highest maize yield from the other deficit irrigation levels and showed no difference with AFI 100%ETc application. The lowest maize grain yield in CFI was obtained from 50%ETc irrigation application, which is also statistically similar with result of AFI at 75%ETc irrigation (Table 4). A similar study by Kang et al. (2000) confirmed that reduction in yield due to AFI application was statistically similar to FFI.

Table 4.

The Furrow Irrigation Method × Irrigation Level Interaction on Grain Yield, WUE, Biomass, and Harvest Index (HI) of Maize in 2016 and 2017 at Melkassa, Ethiopia. Values are the Means of Two Years and Three Replicates.

The WUE of AFI 75%ETc and AFI 100%ETc was greater than the other treatments, however the difference between them was not significant. Kang et al. (1998) and Mehari et al. (2020) showed that AFI is most effective to improve WUE. The lowest WUE was obtained from FFI 50%ETc irrigation and it was statistically similar result with CFI 100%ETc application (Table 4). Better WUE was also observed from deficit application of CFI 50%ETc and CFI 75%ETc. However, in terms of crop yield, the CFI 75%ETc had comparable yield with AFI 100%ETc. This could be because of better application efficiency and less evapotranspiration related to AFI and deficit irrigation (Hassene & Seid, 2017; Subhan et al., 2021). On the other hand, the obtained WUE values agreed with the globally measured average WUE for maize which ranged from 11 kg ha-mm⁻¹ to 27 kg ha-mm⁻¹ (Yazar et al., 2009).

The values estimated for WUE have some very important implications. Under a limited water supply situation where the goal may be to achieve the highest possible WUE, using AFI 100%ETc and AFI 75%ETc offers opportunities for water savings. Otherwise, if the objective is to maximize yield, the use of CFI with 100 or 75%ETc would be better.

Biomass and harvest index

No significant difference was observed over years, by interaction of year and furrow method, and year × furrow method × irrigation level for biomass and HI (Table 2). Biomass and harvest index were reduced significantly from CFI to AFI and FFI in furrow method treatments and from 100%ETc to 75%ETc and 50%ETc (Table 3). The result was in agreement with those reported by Bryant et al. (1992) and Farré and Faci (2009). Bryant et al. (1992) reported that water stress reduced yield by reducing both accumulated biomass and the harvest index. Farré and Faci (2009) also reported a significant effect of limited irrigation on HI.

Higher biomass yield was observed from the 100%ETc application in CFI and shows no significant difference with deficit irrigation application of 75%ETc and 100%ETc under CFI and AFI, respectively (Table 4). The FFI performed low in biomass (BM) yield and shows no significant differences with CFI at 50%ETc and AFI at 75%ETc and 50%ETc applications. Low harvest index, HI for maize was recorded from FFI and similar response was obtained from AFI with 50%ETc application. Highest HI of 0.71 was recorded from CFI with 75%ETc application and showed no difference with 100%ETc application under AFI. The AFI with 100%ETc application saved 50% water compared to the CFI furrow irrigation with 100%ETc application and could be considered best water saving technology in water limited area (Table 4).

Yield response to water deficit

Maize grain yield response to water deficit was obtained from the relationship between seasonal yield reduction and the corresponding seasonal ET deficit (Table 5). The relative yield reduction corresponding to relative water deficit was noticed to increase with increase in irrigation deficit in all the three furrow methods. For the same amount of water deficit, the yield reduction incurred was greater for FFI than AFI (Table 5). The greatest yield response factor was obtained from FFI 50%ETc, while it was less than 1 for all the other treatment combinations (Table 5). As Ky represents a measure of the relative sensitivity of a crop to drought in a particular environment (Ferreira & Gonçalves, 2007), the obtained result indicates that using combined furrow method and irrigation level as a management strategy could be employed in semi-arid regions as effective scarce water utilization practices. As noted by Araya et al. (2011), values for Ky above 1 indicate that the crop is sensitive to moisture stress, whereas values below 1 indicate that the crop can tolerate some levels of moisture stress in its growing environment.

Table 5.

Relative Yield Reduction, Relative ETc Deficit and Yield Response Factor of Maize Using Different Irrigation Methods and Deficit Irrigation at Melkassa, Ethiopia. Values are the Means of Two Years and Three Replicates.