EFFECT OF DEFICIT IRRIGATION REGIMES ON GROWTH, YIELD, AND WATER USE EFFICIENCY OF MAIZE ( Zea mays ) IN THE SEMIARID AREA OF KIBOKO, KENYA

Background: irrigated of 70% of all freshwater scarcity worldwide, the effectiveness of Methodology: A study was carried out for two seasons in Kiboko, Makindu Sub-County during 2018 and 2019 short and long rains, respectively to evaluate the response of maize growth, yield, and water use efficiency to deficit irrigation in the semi-arid area. The experiment was arranged in Randomized Complete Block Design with three replicates. The irrigation regimes were T1 (100 % field capacity), T2 (75 % field capacity, T3 (50% field capacity), T4 (25 % field capacity), and T5 (rain-fed) were evaluated. Results: In season I, there was a difference (P≤0.05) on Plant height, leaf area, and leaf area index, in T1 compared to T5. Plants in T1 were higher (308.1cm) than those in T5 (263cm) (control). Irrigation deficit showed an effect (P≤0.05) on maize growth in season II, with plant height of 270.3cm in T1 compared to 95.6cm in T5. The yield components showed a difference (P≤0.05) on cob-size, 100grains weight, aboveground biomass and harvest index in both seasons. The highest yield of 10.9 and 10.2 t ha -1 was obtained in T1in Season I and II, respectively and lowest in T5 (8.8 t ha -1 and 3.0 t ha -1 ) in the season I and season II, respectively. Higher aboveground biomass and yield were obtained under full irrigation, and declined under varied deficit irrigation regimes. Water use efficiency had no significant difference at the different treatments in the season I, since rains were moderately reliable, thus allowing pausing of irrigation with little water stress. However, in season II, a difference (P≤0.05) in water use efficiency (WUE) was observed. Generally, water use efficiency ranged from 19.6 to 22.kg ha -1 mm -1 in season I and 16.6 to 24.8 kg ha -1 mm -1 in season II. Implication: Irrigating maize at 50% water deficit increases the WUE with minimal yield decline, hence a better deficit irrigation strategy in water conservation under scarcity situation. Conclusion: Growth and yield of maize increased with increased amount of irrigation water and decreased under reduced irrigation while WUE increased with reduced irrigation and decreased under sever water stress.


INTRODUCTION
Regulated deficit irrigation (RDI), a concept coined in the 1970s, controls soil water deficit at certain times in a season to reduce irrigation water requirements (Stewart and Steiner, 1990). This practice has shown grain yield substantially increased during the last decade. The rapid decline of water resources in recent years, however, has led to an urgent need for a reduction of irrigation to make agriculture sustainable in Kiboko semi-arid area (Kipkorir et al., 2001).
Global cereal use is projected to increase by 14% by 2027, mainly due to higher food and feed use in developing countries OECD/FAO, (2018). Maize consumption is expected to increase by 16% by 2027, with maize used for animal feed increasing its overall share of total use to 58% in 2027, largely due fast expanding livestock sectors as well as rising incomes in most developing countries and the consequent growth in meat and poultry consumption. This would translate to an increase in maize demand as livestock feed especially for poultry and pigs (Locke et al, 2013). Maize for human consumption will increase mainly in developing countries, especially those in Sub-Saharan Africa where populations are growing rapidly and white maize is an important staple for several countries OECD/FAO, (2018) as well as in fuel industries and breweries in the production of ethylic alcohol (Dabija et al., 2021).
Deficit irrigation (DI) systems are among the management systems that have been successfully implemented in various crops (Tari, 2016;Zhang et al., 2016). In DI, crops are exposed to a certain level of drought stress by withholding irrigation at specific growth stages or reducing the amount of irrigation water, either during a particular period or throughout the growing season. Therefore, crops under DI receive an amount of water below their full requirement, which, under optimal conditions increases irrigation water use efficiency (IWUE) in exchange for an acceptable yield penalty (Chen et al., 2018;Huang et al., 2005). This yield penalty can be economically tolerable compared with the cost or value of water saved in water-limited environments. DI has been successfully implemented to maximize IWUE and increase yield per unit water used in various crops (Chuanjie et al., 2015), including maize (¸Cakir 2004;Payero et al., 2006;Aguilar et al., 2007;Jahansouz et al., 2014;Domínguez et al., 2012).
In Kenya, DI has been brought about by the increase in population, migration into the Arid and Semi-Arid Lands (ASALs), and climate change (Kinama et al., 2007) and variability. In the ASALs of Kenya, rainfall variability across and within seasons has resulted in moisture deficits. Climate change enhances soil evaporation and reduces water available to crops due to the expected temperature increases. Indeed, soil evaporation takes up to 50% of the total rainfall in the soil water balance in semi-arid areas (Kinama et al., 2005).
This experiment was designed and conducted to evaluate maize crop response to regulated deficit irrigation in the Kiboko area.

Study site
The study was carried out at KALRO Kiboko Research Centre, latitude 02° 127 S, longitude 37° 437 E, elevation 975 m above Sea level, and approximately 160 km southeast of Nairobi, the capital city of Kenya (Maingi et al., 2001).
The soils of the area comprise of well-drained Fluvisols, Ferralsols, and Luvisols, according to the (USDA, 1997) soil classification. The soil texture is sandy loams that have very high drainage (Wiesman et al., 2000). Rainfall is bimodal, with the short rain in October-December and long rains start from March-June (Wiesman et al., 2000). The mean annual rainfall is less than 500 mm (Juma, 2012).
The relief of the area is flat to gently undulating linear with a slope of 2%. The land use is a research site with a border cultivated area and abandoned trial site. The land is cultivated for field crops such as sorghum (Sorghum bicolor), maize (Zea mays), beans (Phaseolus vulgaris), cowpea (Vigna unguiculata) and pigeon pea (Cajanus cajan).

Experimental Design
The experiment was laid out in a Randomized Complete Block Design (RCBD) with five treatments replicated three times. Four soil water deficit irrigation regimes were applied throughout the growing period, and rain-fed treatment acted as control. Irrigation water was applied at different regimes i.e. T1 (100% FC), T2 (75% FC), T3 (50% FC) T4 (25%FC) and T5 (rain-fed) only rain-fed with no irrigation. Duma 43 maize variety was used as a test.
The water was applied by drip irrigation system and amount applied at each treatment was calculated from the full irrigation treatment (100%) using the maize crop water requirement (CWR) at 100 cm rooting depth.

Drip irrigation installation
The system consisted of one filter, seven valves, Tjoints, start connectors, Polyvinyl chloride (PVC) pipes, drip lines, end lines, and L-bow.
The treatments were irrigated individually and the water controlled by the use of valves in the system. The main valve controlled T1 since it was the last to go off during irrigation; while T2, T3, and T4 were controlled by individual valves. The duration of irrigation for each treatment was calculated from the system discharge per hour.

Christiansen's Coefficient of Uniformity (CCU)
Christiansen (1942) "defined" the coefficient of uniformity (CCU) as the ratio of absolute difference of each value from the mean and the mean of means. The Christiansen's Coefficient of Uniformity (CCU) can be expressed as in Eq. 1 Where, n -Number of the depth measurements of the water applied, each representing an equal irrigated area. Ximeasured application depth in liters (L). µ mean application depths in liters (L). CUcoefficient of uniformity (%) The uniformity test was taken from 12 plots after the complete installation of the drip irrigation system. Three drip laterals were selected in each plot from the edges and middle of the plot. Graded beakers in mm were placed in all the selected drip laterals in each plot to collect water during the testing process. The drip irrigation system was open to run for 10 minutes and stop, the water collected in the beaker was recorded, a mean value was obtained in each plot (Xi) and mean of means (µ) was obtained by the means (∑Xi/n) got from the 12 plots.
The coefficient of uniformity (CCU) was 96% which indicates almost equal distribution of all discharges from the emitters. Ascough and Kiker (2002) reported that the CU values (in %) for various irrigation systems varied from 17.4 to 95.2 per cent.

Coefficient of variation (CV)
This is the ratio of actual emitter discharge to the design emitter discharge in litres per hour (L h -1 ).

CV= Q_act/Q_design ________________________2
The coefficient of variation (CV) was 0.93 which indicate high accuracy of the emitters discharge efficiency, thus the variation between the system discharge and actual emitters was 7%. Similar coefficient of variation has been reported (Solomon, 1984;Burt et al., 1997;Ascough and Kiker, 2002).

Irrigation water
Irrigation amounts were calculated according to evaporation pan records (Allen et al. 1998), using the equation given below (Equ): Where, I = amount of irrigation water, A = ratio of depth of irrigation water applied to the cumulative evaporation, Ep = the cumulative evaporation amount, and Kpc = coefficient (including crop coefficient, and application efficiency).

Irrigation time
It was calculated according to following equation; t= (I x A)/q ________________________________4 Where, t = irrigation time (hrs), I = depth of applied irrigation water (mm) A = wetted area by emitters (m 2 ) and q = emitters discharge (L h -1 ).

Measurements of soil moisture and evapotranspiration
Soil moisture content was monitored at a depth of 30, 45 60, 75, 90, 105, and 120 cm weekly using the gravimetric technique.
A soil sample was collected at each plot using a soil auger, and the sample was weighed before oven drying at 105 o C for 24 hours to constant weight.

ET = (P + I + SG) -(D + R) -ΔS ______________5
Where, ET= evapotranspiration (mm), P= precipitation (mm) taken from nearby meteorological station, I = Irrigation water (mm) applied, D = deep percolation (mm), ΔS = changes in soil moisture content (mm), R = runoff and SG = the groundwater contribution to plant available water (mm). D, SG and R was found to be negligible during the experimental period, hence the equation was rewritten as below; ET = P + I -ΔS ____________________________6

Agronomics Practices
Land preparation was done by ploughing and harrowing with a tractor and then subdivided into plots of 3m x 6m with a border spacing of 1m. Two maize seeds were planted per hole. After germination, one seedling was thinned to obtain plants per hole and a population of 44444 plants ha -1 . The spacing between rows was75 cm and between plants was 30 cm.

Plant height (cm)
Plant height was recorded at 30, 60, and 90 days after emergency and at harvest for each treatment. The height readings were taken from the soil surface to the leave base of highest fully expanded leaf. Measurements were taken from five tagged plants per treatment using a meter ruler.

Leaf area index (LAI)
The leave length and width for the five tagged plants in each plot at the different water levels (T1, T2, T3, T4, and T5) was measured at the central part of the leaf at 50% heading, the leaf length and width were obtained for each plot, and the leaf area was calculated using Watson (1947).
Where; LA= leaf area, L is the length 0.75 is the maize correction factor.
The leaf area index (LAI) was estimated from leaf area per plant (A) divided by land area per plant (p). Where, LAI = leaf area index, A = leaf area per plant (cm 2 ) and P = land area per plant (cm 2 ).

Total dry matter weight (tha -1 )
Total dry matter weight was recorded at harvest from five randomly selected plants per plot. The plant was separated from the plant's root portion, and then it was labelled and partially dried in before oven drying it at 60 o C.

Seed grain weight (g)
One hundred grains weight was recorded from each plot from five randomly selected plants, and an average for the treatments. This measurement was done using a weighing machine.

Grain yield (t ha -1 )
Grain yield in tha -1 from each plot was recorded from air-dried cob, separated and cleaned before drying it to 14% moisture content. The grains were weighed and recorded in kilo grams (kg) before it was converted to tha -1 .

Harvest Index (%)
This refers to the crop's economic yield divided by total dry weight, as Donald (1992)

Water Use Efficiency (WUE)
WUE was estimated from the yield in kilogram (kg) and actual crop evapotranspiration ETc (mm) with the equation given below (Karuku et al., 2014, Araya et al., 2011, Song et al., 2019. Where, WUE is water use efficiency (kgm -3 ), Y is the yield (kg ha -1 ), and ETc is the crop reference evapotranspiration.

Data analysis
The data analysis was done with the aid of GenStat 19 th edition (Lane and Payne, 1997) and subjected to analysis of variance (ANOVA) with means differences separated by Duncan's multiple range test at 95% confidence level (P≤0.05 level of significance).

Chemical soil properties
The soil chemical properties of the experimental site are presented in Table 1. The soil pH obtained was 7.15, which is within the required pH for effective maize growth that ranges from 5.0 to 7.0 (FAO, 2012).

Physical soil properties
The soil physical properties of the study site are shown in Table 2. The particle size distribution showed that sandy texture was dominant.
The sandy soil content was 68.5, clay 26.8 and the silt soil content shows low soil contain of 4.6%, thus the textural class of the soil was sandy clay loam according to the textural triangle.
The bulk density indicated a slight variation with depth and ranged from 1.30 g cm-3 at the depth of 0 -15 cm to 1.55 gcm-3 at the depth of 90 -105 cm. This could be because of decrease in organic matter content with depth and compaction due to the weight of the overlying soil layer (Brady and Weil, 2002). The soil moisture content at field capacity and permanent wilting point was at PF 4.2 and PF 2.0 respectively and the hydraulic conductivity (Ksat mmday-1) was high which indicated high permeability of the soil.

Climatic data
Climatic data are shown in Table 3. Maximum and minimum air temperature ( o C), rainfall (mm), relative humidity (%), wind speed (ms -1 ) at screen height (2m about the ground) and sunshine hours were obtained from Kiboko research station.   growth (Sanchez and porter, 2014). The average rainfall in season I was 3.9 mm with the highest rainfall of 10.4 mm recorded in December and the lowest of 1.4mm in October. Season II had 0.37mm as its average rainfall which indicate a low rainfall in both seasons though higher in season I than in season II. Rainfall occurrence depends greatly on the temperature and weather conditions (Trenberth, 2011, Mawonike andMandonga, 2017).
A high temperature increases the rate of potential evaporation which would deplete the soil moisture content (Nkuna and Odiyo, 2016). Relative humidity (RH) on average was 82 and 78% in season I and II, respectively which moderately high. Relative humidity (RH) directly influences the water relations of plant and indirectly affects leaf growth, photosynthesis, pollination, occurrence of diseases and finally economic yield (Hoogenboom, 2000). The dryness of the atmosphere as represented by saturation deficit (100-RH) reduces dry matter production through stomatal control and leaf water potential (Grange and Hand, 1987). The wind speed was 192 and 163 ms -1 in season I and II, respectively whereas sunshine recorded an average of 6.9 hours in both seasons.

Plant height (cm)
Maize height was not significantly affected by deficit irrigation (T1, T2, T3 and T4) in season I. However, there was a significant difference (P≤0.005) observed between T1 (100% FC) at 308 cm and T5 with 263 cm plant height at the maturity stage. The finding is in agreement with Rosadi et al. (2005) who found out that a small difference in moisture deficit levels did not affect plant height. In season II, plant height had highly significantly (P≤0.005) difference between deficit irrigation regimes and rain-fed, with a maximum maize height of 306 cm obtained in T1 followed by 262 in T2, 225 cm in T3, 197cm in T4 and the least plant height of 96 cm was recorded under in T5. Water is an important component of plant cell and raw material for photosynthesis. Carbohydrates are manufactured from water combine with carbon dioxide (CO2) in the presence of sunlight. Water keeps the plant turgid and erect; moisture deficiencies in plant result in cell flaccidity and the plant drops and wilt. Tari (2016) and Jia et al. (2017) found out that maize plant grown under sufficient moisture content produce high plant height while water stressed condition produces dwarf maize plant.

Leaf area and leaf area index
Leaf area and leaf area index was recorded during the growing period and the data obtained in season I revealed non-significant difference among the deficit irrigation treatments. However significant (P≤0.005) difference was noted between fully irrigated (T1) treatment that recorded 718 cm and 4.8 leaf area and leaf area index and 661cm and 4.4 obtained under T5. Pandey et al. (2000) recorded the highest value of leaf area index for corn that was obtained under the conditions of full irrigation (without stress). In season II deficit irrigation has high significant (P≤0.005) effect on the leaf area and leaf area index, a maximum leaf area and leaf area index of 700 and 4.6 recorded in T1 followed by 591 and 3.8 in T2, 540 and 3.3 in T3, 525 and 2.9 in T4 and the least leaf area and leaf area index of 242 and 1.2 was observed under rain-fed (T5). The findings agree with (Bouazzama et al., 2010) who found out low leaf area index in the treatments under more water stress.

Maize yield components
Yield attributes of maize measured during the harvesting time includes; cob size (cm), grain weight per 100 seeds, aboveground biomass, yield and harvest index (HI) are shown in Table 5.

Cob size (cm)
Cob size was recorded at the harvesting stage and data obtained in season I shown a significant effect of deficit irrigation on maize cob size. Among the deficit irrigation regimes, the maximum cob size of 19.6cm was observed in T1 which was no significantly (P≤0.05) difference from its immediate irrigation regimes of T2, T3 and T4. Whereas

rain-fed)PH (plant hight),LA (Leaf area) and LAI (leaf area index). Mean followed by the same letter in a column
are not significantly different from each other at (P≤0.05) level. the minimum cob size of 18.1cm was obtained in T5, this was significantly (P≤0.05) difference compare to the irrigated treatments. In season II deficit irrigation had a high significant effect on cob size, with highest cob size of 19.9cm obtained in full irrigated treatment (T1) followed by T2 (75% FC), T3 (50% FC) , T4 (25% FC) and least cob size of 13.2 cm was recorded under rain-fed (T5).

Grain weight per 100 seeds
A deficit irrigation regime had significant (P≤0.05) effect on grain weight (g). In season I, the maximum grain weight of 39.6g was recorded in T1 which has no significant effect from T2 while T3, T4 and the rain-fed that recorded the least grain weight of 33.6 g had a significant (P≤0.05) difference. In season II grain weight shown a high significance difference, among the deficit irrigation regimes, a maximum grain weight of 41.3g obtained under full irrigation (T1) which has no significance deference from T2 and T3, however there was a significance (P≤0.05) difference noted in T4 and T5 that obtained 35.1g and 18.1g respectively compared to full irrigated treatment. Grain filling stage requires adequate moisture content to facilitate the assimilation of dry matter to the grains, hence water stress at this stage will reduce the assimilation of dry matter to the grain as well as cause the production of sterile pollen grains thus low grain weight (Du et al., 2015 andLi et al., 2018) found that water stress in reproductive stage reduces grain weight of maize.

Above ground biomass (t/ha)
The above ground biomass (t/ha) was found to be linear with deficit irrigation. The data collected in season I, revealed a significant effect (P≤0.05) of deficit irrigation on above biomass, a maximum above ground biomass of 35.2 tha -1 was recorded in T1 which was no significance difference from to 33.9tha -1 obtained from T2, but significance difference to T3, T4 and T5 that obtained the minimum above ground biomass of 28.1tha -1 . In season II deficit irrigation had high significant effect on biomass accumulation, with a maximum of 33.8tha -1 recorded in in T1 followed by T2, T3, T4 and T5 that obtained the least biomass of 14tha -1 . Generally, accumulation of above ground biomass of maize depends on the level of deficit irrigation regime and it reduces significantly with decrease in deficit irrigation. The findings are in agreement with Igbadu et al. (2008) who reported that deficit irrigation at any growth stage resulted in decrease in both biomass and grain yield. Yazar et al. (1999) and Pandey et al. (2000), who reported that deficit irrigation definitely reduces yield of maize crop, and that maize dry matter and grain yield increased significantly with irrigation.

Grain yield (t/ha)
Grain yield of maize was significantly (P≤0.05) affected by deficit irrigation regimes. In season I the data collected revealed a maximum grain yield of 10.9t/ha obtained in T1 which was no significantly difference from T2 but significantly (P≤0.05) different T3, T4 and T5 that record the lowest grain yield of 8.4tha -1 . In season II maximum yield 10.2tha -1 was obtained in full irrigation (T1), 9.1t ha -1 in T2, 8.5t ha -1 in T3, 6.0 t ha -1 in T4 and lowest yield of 3.0tha -1 was obtained in rain-fed. Season I has low yield variation between deficit irrigation and rain-fed condition whereas season II has high yield variation between irrigated and rain-fed, theses could be as result of rainfall pattern between the two seasons. Season I slightly moderate rainfall that had added significant moisture content to the soil compared to season II that received very little rainfall (Table 5), hence the crop was mostly depending on irrigation thus the effect of deficit irrigation and water stress cause the yield variation in season II. The result clearly shows that maize yield is linear with deficit irrigation regimes, and this agrees with the findings of (Naescu, 2000, Karam et al., 2003;Farre et al., 2006;Mengü and Ozgurel, 2008, Oktem, 2008, Golzardi et al., 2017, who reported that deficit irrigation reduces the yield of maize crop, and maize dry matter increases significantly with irrigation. The findings are also agreeing with Rhoades and Bennett (1990) and Lamm et al. (1995), who reported that it is difficult to plan deficit irrigation for maize without causing yield reduction.

Harvest index (HI)
The harvest index of maize was almost the same in season I Table 6. However, in season II harvest index revealed high significant (P≤0.05) difference, with high harvest index obtained in T1 and T2 which was highly significant (P≤0.05) to 0.29 recorded in T3, 0.25 in T4 and 0.2 in T5 as the least harvest index. Yield and above ground biomass in season I were moderately high along with a small variation among all the treatment, which results to low variation in harvest index in season I. Golzardi et al. (2017), Mohammadi et al. (2018), and Xue et al. (2018) reported that maximum harvest index of maize was produced when filed was well irrigated. Bryant et al. (1992) indicated that water stress reduces yield by reducing accumulated biomass and the harvest index. However, (Traore et al., 2000) found that the harvest index was affected by water deficit only when stress was imposed during anthesis.

Maize water use efficiency (WUE)
The effects of deficit irrigation on water use efficiency of maize are shown in Table 6. Water use efficiency of maize was found to be significantly (P≤0.05) different and varies with seasons and irrigation level. The values recorded for water use efficiency of maize ranges from 16.6 to 24kgha -1 mm -1 . In season I, the maximum water use efficiency of 22kgha -1 mm -1 obtained under rain-fed (T5), which was significantly (P≤0.05) difference compare to 19.7 and 19.6kgm -1 mm -1 obtained in T1 and T2 respectively.
In season II, water use efficiency was highly significantly (P≤0.05) difference, with a maximum WUE of 24.8 kg ha -1 mm -1 obtained in T3 followed by T1 and T2 that recorded the same water use efficiency of 23.7 kgha -1 mm -1 , 19.8 obtained in T4 and lowest was 16.6 kg ha -1 mm -1 recorded under rain-fed (T5).
Rainfalls were insufficient and unreliable in season II; as a result, the corps was entirely dependent on irrigation, which results in high water use efficiency. Regulated deficit irrigation (RDI), as described and used by Rawson and Turner, (1983), and Fabeiro et al., (2002), can further improve WUE. The RDI maintains crop plants under water deficit stress during some of the growth stages by controlling irrigation amounts. Fully irrigated plants usually have widely opened stomata. Plants open their stomata for CO2 uptake and carbon gain but will lose significant quantities of water at the same time (Kang and Zhang, 2004). A small narrowing of the stomatal opening can reduce water loss substantially with little effect on the photosynthesis rate. Earlier research predicted that plants generally should have the capability to increase their WUE in this way, thereby maximizing their chance of surviving a period of drought, potentially without a great reduction in carbon gain and biomass accumulation; however, this may occur only when crops are aerodynamically well coupled to the atmosphere (Grieu et al., 1988). However, during critical growth stages, it is particularly important to maintain plant water supply and status (McLaughlin and Boyer, 2004).

CONCLUSIONS
Generally, irrigating maize (Zea mays) under deficit irrigation in the study area will have the following effect on its productivity; • Irrigating maize at 50% water deficit would improve water use efficiency without much reduction in yield in the study area.
• High maize yield performance along all the treatments in season I was due to moderate rainfall received, while yield variations in season II were due low and unreliable rainfall hence crops were entirely grown under irrigation as such deficit irrigation effect were observed all the treatments.
• Deficit irrigation under 25% field capacity (FC) reduces yield and water use efficiency with 41% and 14%, respectively.
Funding statement. This work was fully funded by Borlaug Higher Education for Agricultural Research and Development (BHEARD).

Conflict of interest.
The authors confirm that there no known conflict of interest associated with this publication.
Compliance with ethical standards. Do not apply. No human participants or animals were used in the studies undertaken in this article by any of the authors.
Data availability. Data is available with Lubajo Bosco Wani (lubajobosco402@gmail.com) upon reasonable request.