Experimental Evaluation for the Impacts of Conservation Agriculture with Drip Irrigation on Crop Coefficient and Soil Properties in the Sub-Humid Ethiopian Highlands

Abstract

A field experiment consists of conservation agriculture (CA) and conventional tillage (CT) practices were set up in two areas, Robit and Dangishta, in sub-humid Ethiopian highlands. Irrigation water use, soil moisture, and agronomic data were monitored, and laboratory testing was conducted for soil samples, which were collected from 0 to 40 cm depth before planting and after harvest during the study period of 2015–2017. Calculation of crop coefficient (Kc) revealed a significant decrease in K_c values under CA as compared to CT. The result depicted that CA with a drip irrigation system significantly (α = 0.05) reduced K_c values of crops as compared to CT. Specifically, 20% reductions were observed for onion, cabbage, and garlic under CA whereas 10% reductions were observed for pepper throughout the crop base period. Consequently, irrigation water measurement showed that about 18% to 28% of a significant irrigation water savings were observed for the range of vegetables under CA as compared to CT. On the other hand, the results of soil measurement showed the CA practice significantly (α = 0.05) increased soil moisture (4%, 7%, 8%, and 10% increment for onion, cabbage, garlic, pepper) than CT practice even if irrigation input was small in CA practice. In addition, CA was found to improve the soil physico-chemical properties with significant improvement on organic matter (10%), field capacity (4%), and total nitrogen (10%) in the Dangishta experimental site. CA with drip irrigation is evidenced to be an efficient water-saving technology while improving soil properties to support sustainable intensification in the region.

1. Introduction

Agriculture is the core driver for the economy of Ethiopia and long-term food security which supports about 85% of the population’s livelihoods, and 43% of gross domestic product (GDP) [1,2,3]. However, the current farming practices in the nation are mainly traditional tillage using animals and hand tools; which resulted in the loss of soil fertility and crop productivity [4,5,6,7,8]. On the other hand, water is a valuable resource in food production, nevertheless, the competition for water is high among various sectors due to several reasons including population growth, climate change, and poor management of rainwater. Crop production from smallholder farmers constitutes more than 90% of the food supply in the nation [9]. This production system is highly dependent on rainfall and the system uses the traditional approach [10]. The combined effects of traditional farming practices [11], with poor rainwater management [12] caused severe soil degradation [13,14]. Consequently, food production and water uses were incompetent in the region [15,16,17]. Therefore, it is critical to advance the farming system and restore degraded soil to efficiently utilize available water resources for sustainable food production [7,18,19].

Rainfall variability and irregularity are major concerns in a rainfed production system and smallholder farmers are vulnerable in crop production. Crop failures are common in the region due to rainfall variability and frequent drought [20,21]. Consequently, food insecurity often turns into famine with the least adverse climatic incident, predominantly, affecting the livelihoods of the rural poor [22,23]. Expanding small-scale irrigated agriculture could play a significant role in sustaining the livelihood of smallholder farmers. However, only 4% to 5% of the cultivated land in Ethiopia is under irrigation [1] with poor water use efficiency, about 40% [24]. Water availability is limited in the long dry season in the nation which necessitates the use of efficient water application technology in expanding irrigated agriculture.

Ineffective on-farm irrigation management practices lead to poor water distribution, nonuniform crop growth, excessive leaching of soil nutrients all of which decrease water and crop productivity [25,26,27]. Conservation agriculture (CA) is found to be an effective strategy in the region to improve soil quality, water management, and crop yield [5,6,28]. Conservation agriculture (CA) is defined as minimal soil disturbance (no-till, NT) and year-round permanent soil cover (mulch) combined with rotation in time [29,30]. CA continued being promoted over conventional agriculture (CT) practice to enhance soil health and sustain long-term crop productivity and reduce soil erosion [29,30]. The overall goal of CA is to make better use of agricultural resources with limited external inputs by enhancing the ecosystem through integrated management of soil and water [31,32,33]. On the other hand, drip irrigation is considered as the most efficient water application technology and has the potential to improve water productivity [34,35]. Combining CA with drip irrigation is found to be an ideal approach to maximize crop production while conserving the environment [6,28,36,37]. Crop coefficient can be used as an indicator to measure the effectiveness of a particular soil and water management practices [38]. Therefore, the main aim of this study was to evaluate the effects of CA with drip irrigation on crop coefficients in the sub-humid Ethiopian highlands. The specific objectives were (1) to evaluate irrigation water savings for irrigated crops under CA and CT water management systems and (2) to evaluate the impacts of CA and CT management practices on crop coefficients both under drip irrigation. The results of this study would assist stakeholders to determine the amount of irrigation needed precisely under CA and CT practices in the regions.

2. Materials and Methods

2.1. Study Area

This study was conducted in the sub-humid Ethiopian highlands at two experimental watersheds; Dangishta and Robit (Figure 1). The Dangishta watershed is in Dangila Woreda, about 80 km northwest of Bahir Dar whereas Robit watershed is about 12 km northeast of Bahir Dar. The climate of Dangila Woreda is sub-humid with an average annual rainfall of 1600 mm and an average annual potential evaporation of 1250 mm [39]. The maximum and minimum temperature during the study period (from 2015 to 2017) in Dangishta was 33 and 25 °C, respectively. The average annual rainfall in Robit is about 1400 mm with a minimum temperature of 23 °C and a maximum temperature of 31 °C. Dangila and Robit are one of the Agricultural Growth Program (AGP) and feed the future Woredas in Amhara Regional State. Rainfed agriculture and livestock production are the main livelihood systems in these regions where the dominant crops produced in the regions are maize, millet, barley, and teff [40]. On the other hand, tomato, onion, potato, cabbage, pepper, and garlic are the dominant irrigated vegetables produced in the regions. Shallow groundwater is the source of irrigation in Dangila [39] in which manual pulley systems are widely practiced to lift water from the wells. Groundwater level ranges from 4 to 15 m below the ground surface in the dry season [41]. In the rainy season, the groundwater level for shallow wells was observed close to the ground surface (as low as 1 m in some shallow wells). Similarly, in Robit, irrigation is widely practiced using shallow groundwater with water table depth that varies from 8 to 20 m [42] using a manual pulley system as a means of water lifting technology.

2.2. Experimental Design and Field Setup Procedures

2.2.1. Experimental Design

The effects of CA and CT practice on crop coefficients was evaluated using a paired ‘t’ experimental design (Figure 2) which is mathematically powerful and allows good control of individual differences [18,43]. De Winter [43] evaluated the paired t-test for an extremely small sample size and concluded that the applicability of the t-test as low as two replicates. Detailed description of the experimental design in Dangishta and Robit study sites can be found in Assefa [5] and Assefa et al. [18]. The experimental setup was established in 2015 on a 100 m² plot size where half of this size was assigned to CA and another half to the CT practice through randomization. A total of 14 farmers participated in this experiment; seven farmers in Robit and seven farmers in Dangishita for each cropping season. Farmers have grown various vegetables in different seasons on the experimental plots; cabbage, pepper, garlic, and onion and their detailed cropping pattern are shown in

Table 1. Rotation of crops and management activities at Dangishta and Robit experimental watersheds adopted from Assefa et al. [19].

Site	Vegetable	Management Activity	Date
Dangishita	Garlic (1st cycle)	Tillage 1	13 October 2015 and 16 October 2015
		Mulch application 2	25 October 2015
		Planting	28 October 2015
		UREA application	28 November 2015
		Irrigation Application	6 November 2015–22 February 2016
		Harvesting	3–4 March 2016
	Onion (2nd cycle)	Tillage 1	14 March 2016 and 16 March 2016
		Mulch application 2	15 March 2016
		Planting	17 March 2016
		Irrigation Application	15 March–3 May 2016
		Harvesting	24–26 June 2016
	Garlic (3rd cycle)	Tillage 1	15 February 2017
		Mulch application 2	17 February 2017
		Planting	17 February 2017
		DAP 3 application	4 March 2017
		Irrigation Application	17 March–3 June 2017
		Harvesting	20–22 June 2017
	Pepper (4th cycle)	Tillage 1	12 March 2018
		Mulch application 2	14 March 2018
		Planting	15 March 2018
		Irrigation Application	15 March–9 June 2018
		Harvesting	12–20 July 2018
Robit	Tomato (1st cycle)	Tillage 1	2 September 2015
		Mulch application 2	23 October 2015
		Planting	24 October 2015
		Malathion 4 application	22 November 2015
		Irrigation Application	24 October 2015–12 March 2016
		Harvesting	1–15 March 2016
	Garlic (2nd cycle)	Tillage 1	19 March 2016
		Mulch application 2	21 March 2016
		Planting	22 March 2016
		Irrigation Application	23 March–21 June 2016
		Harvesting	10–18 July 2016
	Cabbage (3rd cycle)	Tillage 1	27 October 2016
		Mulch application 2	8 November 2016
		Planting	9 November 2016
		UREA 3 application	20 December 2016
		Dimeto 4 40% application	15 November 2016
		Irrigation Application	9 November 2016–25 February 2017
		Harvesting	15–26 February 2017

Note: ¹ Only for CT plots; ² only for CA plots; ³ Fertilizer; ⁴ Pesticide.

. The gravity based drip irrigation system was applied for all plots. Shallow groundwater was the source of irrigated agriculture for the experimental sites. Farmers use a rope and pulley system to extract water from the shallow well to an elevated water tank. Irrigation water was then applied to the plots from the water storage tanks through gravity using the drip system.

2.2.2. Field Setup Procedures

A serious discussion was conducted with the local government and community to select potential farmers for this experiment. Additionally, a focus group discussion was followed with farmers for their willingness to participate in this experiment and the availability of shallow groundwater wells close to their gardens was checked. A 10 by 10 m plot close to the household or at a walking distance was identified and prepared for CA and CT managements. A total of 10 beds were prepared on a 100 m² plot 30 cm furrows in between. The size of the beds was kept uniform for the entire plots in the experimental sites. CA and CT beds were decided randomly by tossing a coin. The drip irrigation system (main line, submain line, drip kit, sediment filter, and laterals) was installed on the beds attached to water storage tanks, 500 L size for each plot. The water storage tanks were installed about 1.5 m high from the ground for gravity flow with optimum pressures in the drip pipes (Figure 2 and Figure 3). All CA beds were then mulched using thick dried grass even if Robit farmers were a bit difficult in adopting the intervention. The farmers discuss and select the type of vegetable to be planted for each season. The same variety of seeds were given to the farmers, either direct seeded or the seedlings were grown on a nursery bed. For the case of seedlings, the vegetables were transplanted on the prepared beds. A data protocol was prepared and monitored by a local trained expert.

2.3. Data Monitoring and Collection

2.3.1. Soil Physio-Chemical Property

Soil samples were taken from both CA and CT plots before and after harvest during the 3rd cycle of vegetable production (Table 1) from Dangishta and Robit experimental sites. A total of 16 soil samples from Robit and 20 soil samples from Dangishta were taken from the surface 10 cm depth; half of the samples were taken before transplanting hereafter called the beginning of the experiment and the remainder immediately after harvest hereafter called the end of the experiment. In addition, during the 2nd cycle vegetable production in Dangishta a total of 20 samples before transplanting from two depths; 0–20 and 20–40 cm depth were collected from five plots. In both Dangishita and Robit sites, a representative soil mass of about 1 to 1.5 kg was sampled. The samples were analyzed in the Amhara Design Supervision Works Enterprise (ADSWE) soil laboratory. The samples were analyzed for field capacity, permanent wilting point, soil texture, available organic matter, pH, electrical conductivity, total nitrogen, available P, available K, iron status, and cat-ion exchange capacity. Laboratory analysis of these parameters is quite cumbersome and the details of each analysis and approaches done by ADSWE can be referred from Tesema et al. [39].

2.3.2. Meteorological Data

Climatic data (rainfall, maximum and minimum temperature, wind speed, sunshine hour, relative humidity) were collected from the nearby meteorological station, Dangila for Dangishta sites, and Bahir Dar for Robit sites. Evapotranspiration (ETo) was estimated using the Penman-Monteith method [44] as it incorporates more climatic variables than other methods.

The effective rainfall (Pe) was computed based on Bos et al. [45] as shown in Equation (1).

$$\left(\begin{array}{cc}Pe=0.8\times P-25& P>75mm/month\\ Pe=0.6\times P-10& P<75mm/month\end{array}\right)$$

(1)

2.3.3. Agronomic and Water Use Data

The length of the growing period and water requirement of crops depends on various conditions including climate, variety of crops, date of planting (growing season), water, and other agricultural inputs. Wellens et al. [46], for instance, stated that the growing period of cabbage may range from 120 to 140 days with a root depth ranging from 40 to 50 cm and its water requirement ranges from 350 to 500 mm with variable irrigation frequency, 3 to 12 days [47]. The cabbage was transplanted in Robit at a recommended spacing of 20 cm between plants and 30 cm between rows having a maximum rooting depth of 50 cm. The length of the four growing stages of cabbage was 21 days for initial, 25 days for development, 65 days for mid, and 15 days for the end stage. Crop water use increases during the growing period with a peak towards the mid-stage of the season. The rotation of crops at each study site is shown in Table 1. Garlic was directly seeded in Dangishita at a depth of 40 cm and with 20 cm spacing between rows and 30 cm spacing between plants. The length of the four growth stages of garlic were 20 days for initial, 30 days for development, 40 days for mid, and 30 days for the end stage. After garlic was harvested onion and pepper were irrigated in Dangishta with a similar spacing of garlic. Their length of growth stages for initial, development, mid, and late stages were 5-day, 25-day, 40-day, and 30-days for onion, respectively. On the other hand, 25, 35, 40, and 25 days were the growing stage of pepper for initial, development, mid, and late stages, respectively.

Irrigation water use for CA and CT was monitored separately during each irrigation period under the farmers’ practice. Participant farmers were trained on the water saving potential of conservation practice. Farmers observe soil moisture and use their intrinsic knowledge to determine if irrigation was needed. We observed that farmers used a different irrigation frequency by themselves; once in three days for CA and every day for CT management in most cases. For ease of measurement, a 500 L taker was installed at an elevated position and its outlet was attached to the drip hose for each plot. At the end of the drip, the hose flow control valve was fitted, and farmers control the flow manually during the irrigation of CA and CT plots. The number of water storage tanks used was counted separately for CA and CT plots to determine the volume of water used for irrigation. Soil moisture at the top 20 cm depth was monitored using TDR (total domain reflectometer) probes every two or three days for CA and CT. Readings were taken by trained agricultural extension agents by inserting a pair of 20 cm length TDR roads into the soil from representative points on the plot (CA and CT) and averaged for each recording. TDR probes were calibrated using gravimetric soil moisture determination techniques to increase data quality. Soil samples were collected from TDR measurement locations. The samples were weighted and put to oven at 105 °C for 24 hr. The dry soil mass was measured after oven drying. The gravimetric moisture content was determined as a ratio of weights of water to weight of dry soil. The gravimetric water content was then converted to volumetric water content and plotted against the TDR readings for calibration.

2.4. Crop Coefficients

The precision of irrigation water use for various crops depends on the proper estimation of reference evapotranspiration (ET_o) and crop coefficient (K_c) [48]. Crop coefficient (K_c) was computed by dividing actual evapotranspiration by reference evapotranspiration on a daily basis for each growing season for both CA and CT management practices, as shown in Equation (2) [18]. Average K_c for the growing period was used to evaluate the effect of CA practice over CT. A paired t-test was used to compare the two means, for CA and CT practices, at a 5% significance level and determine if changes in the mean values were significant.

$$K_C=E{T}_c/E{T}_o$$

(2)

where K_c, ET_c, and ET_o are crop coefficients, crop evapotranspiration, and reference crop evapotranspiration, respectively. Crop evapotranspiration was computed from soil water balance Equation (3) and reference crop evapotranspiration was computed by the Penman-Monteith method using meteorological data using Equation (4). The soil water balance equation [39] can be expressed as shown in Equation (3).

$$E{T}_c=P+I+U-R-D+Q_i-Q_o$$

(3)

where P, I, U, R, D, ET_c, Q_i and Q_o are effective rainfall (mm), irrigation (mm), upward capillary rise into the root zone (mm), runoff (mm), deep percolation beyond the root zone (mm), crop evapotranspiration (mm), amount of water in the root zone at time t_i (mm), and amount of water in the root zone at time t_i+1 (mm), respectively. Q_i and Q_o are determined by multiplying volumetric soil moisture content measured by TDR at time t_i and t_i+1 by plant root depth, respectively. The difference between the amount of water in the root zone (Q_i) and the amount of water in the root zone after a given time interval (Q_o) is the change in storage. Upward capillary rise into the root zone (mm), runoff (mm), and deep percolation beyond the root zone (mm) were assumed null in the drip irrigation managed system. This experiment was focused on dry season production with drip irrigation. Several studies including Bozkurt et al. [49], Sezen et al. [50], and Shedeed et al. [51] showed that deep percolation, runoff, and capillary rise under drip irrigation is very small and can be neglected. Assefa et al. [19] in the same experimental site showed that the deep percolation and runoff in the dry irrigation period under drip irrigation was minimal. The upward capillary rise depends on depth of groundwater table and the rooting depths. The root depth is dynamic throughout the crop growth stages and varies depending on crop types. However, from the field observation during harvest, the maximum rooting depths for onion, garlic, cabbage, and pepper are less than 60 cm. On the other hand, the depth of groundwater level varies spatially and temporarily; in the dry season 4–15 m in Dangishta and 8–20 m in Robit was observed [52]. Gao et al. [53] explained that the upward capillary rise is negligible when the groundwater depth is more than 2.5 m.

Reference crop evapotranspiration (ET_o) can be computed using the Penman-Monteith equation as shown in Equation (4). The Penman-Monteith method uses several climatic parameters and gives relatively more accurate results [44].

$$E{T}_{\mathrm{o}}=\left\{0.408\mathsf{\Delta }\left(R_n-G\right)+\gamma \left(900/T+273\right)U_2\left(e_s-e_a\right)\right\}/\left\{\mathsf{\Delta }+\gamma \left(1+0.34U_2\right)\right\}$$

(4)

where ET_o, R_n, G, T, U₂, e_s, e_a, Δ, γ are reference evapotranspiration (mm day⁻¹), net radiation at the crop surface (MJm⁻²day⁻¹), soil heat flux density (MJm⁻²day⁻¹), air temperature at 2 m height (°C), wind speed at 2 m height (ms⁻¹), saturation vapor pressure (KPa), actual vapor pressure (KPa), slope vapor pressure curve (KPa°C⁻¹), and psychometric constant (KPa°C⁻¹), respectively.

3. Results and Discussion

3.1. Effects of CA with Drip Irrigation on Soil Physico-Chemical Properties

The effects of CA with drip irrigation on soil physico-chemical properties were analyzed from soil samples at the experimental sites and tested at 10% significant level.

Table 2. Average soil physico-chemical properties during the 3rd cycle vegetable production (Table 1) both at the beginning and end of this experiment. Beginning stands for before transplanting, end stands for immediately after harvest during 3rd cycle vegetable production (after two years of conservation agriculture (CA) practice).

Experimental Site	Dangishta				Robit
Irrigation Season	Beginning		End		Beginning		End
Parameter\Treatment	CA	CT	CA	CT	CA	CT	CA	CT
Soil PH	5.9	5.8	6.4	6.2	5.1	5.2	5.5	5.6
Organic matter (%)	4.8	4.3	4	3.7	3.2	3	4.9	4.8
Electrical Conductivity (ds/m)	0.1	0.1	0.1	0.1	0.4	0.4	0.1	0.1
Total Nitrogen (%)	0.3	0.2	0.2	0.2	0.2	0.2	0.2	0.2
Average Potassium (ppm)	1349	1411	1118	997	305.6	496.3	454.4	415.5
Average phosphorous (ppm)	24.7	25.4	32.2	24.2	21.6	13.7	18.3	19.1
Average field capacity (%)	36.8	35.5	n.a	n.a	32.8	32.7	34.5	24.6
Sand (%)	46.4	36.4	29.2	26	19.5	19	20.5	22.5
Silt (%)	24.6	29.4	34.4	32	29.5	28.5	27.5	29
Clay (%)	29	34.2	36.4	42	51	52.5	55	45.5
Average Permanent wilting point (%)	24.2	25	n.a	n.a	22.2	22.2	26.8	24.5
Iron (%)	18.5	18.4	n.a	n.a	12.9	12	16.9	13.6
Average cation exchange Capacity (%)	21.5	17.4	25.8	23.3	25.6	27	28	32.8

Note: N.a. refers to data not available.

shows soil sample analysis collected from Robit and Dangishta sites for the impact of CA on soil physico-chemical properties.

The soil physico-chemical properties, (Table 2) vary across the management conditions. Optimum plant growth is attained for a range of pH 5.5 to 7 [54]. The result obtained from the soil test showed that the average value of pH for both experimental sites (Robit and Dangishta) is within the suitability range for plant growth. The change between treatments in Dangishta was significant whereas the reverse is true for Robit at a 10% significance level (

Table A1. A one tailed paired t-test for the average soil physico-chemical properties, where OM, EC, TN, Av. K, Av.P, Av. FC, Av.Pwp, Fe, and Av. CEC are organic matter, electrical conductivity, total nitrogen, average potassium, phosphorous, field capacity, permanent wilting point, iron, and cation exchange capacity, respectively.

Items	pH	OM (%)	EC (ds/m)	TN (%)	Av. k (ppm)	Av. P (ppm)	Av. FC (%)	Av. Pwp (%)	Fe (%)	Av. CEC (%)
CA	5.9	4.8	0.1	0.3	1349	24.7	36.8	24.2	18.5	21.5
CT	5.8	4.3	0.1	0.2	1411	25.4	35.5	25	18.4	17.4
N	5	5	5	5	5	5	5	5	5	5
p-value	0.15 **,a	0.01 *,a	0.5 **,a	0.05 **,a	0.3 **,a	0.4 **,a	0.08 **,a	0.37 **,a	0.4 **,a	0.06 **,a
CA	6.4	4	0.1	0.2	1118	32.2	n.a	n.a	n.a	25.8
CT	6.2	3.7	0.1	0.2	997	24.2	n.a	n.a	n.a	23.3
N	5	5	5	5	5	5				5
p-value	0.09 **,b	0.02 *,b	0.2 **,b	0.03 *,b	0.2 **,b	0.24 **,b				0.02 *,b
CA	5.1	3.2	0.4	0.2	305.6	21.6	32.8	22.2	12.9	25.6
CT	5.2	3	0.4	0.2	496.3	13.7	32.7	22.2	12	27
N	4	4	4	4	4	4	4	4	4	4
p-value	0.15 **,c	0.24 **,c	0.11 **,c	0.35 **,c	0.18 **,c	0.07 **,c	0.43 **,c	0.5 **,c	0.04 *,c	0.16 **,c
CA	5.5	4.9	0.1	0.2	454.4	18.3	34.5	26.8	16.9	28
CT	5.6	4.8	0.1	0.2	415.5	19.1	24.6	24.5	13.6	32.8
N	4	4	4	4	4	4	4	4	4	4
p-value	0.15 **,d	0.4 **,d	0.35 **,d	0.38 **,d	0.45 **,d	0.44 **,d	0.1 **,d	0.13 **,d	0.03 *,d	0.06 **,d

Note: N = number of replicates, ^a = Dangishta at beginning, ^b = Dangishta at end, ^c = Robit at beginning, ^d = Robit at end, * α = 0.05, ** α = 0.1.

).

Tillage practice affects organic matter availability in the soil in which its amount depends on various site-specific conditions; rainfall, air temperature, management practice, type of plant, soil temperature, and drainage condition [55] and has its consequence on soil structural stability, soil erosion, and nutrient availability in the soil [56]. As shown from Table 2, at the beginning of the experiment (3rd cycle vegetable production, after two years of mulch application), CA has a greater mean difference in organic matter (4.8) than CT (4.3) which is significantly higher at 10% significant level in Dangishta (Table A1). This is because CA improves organic matter availability in the soil as a result of the biological decomposition of organic mulch covers.

The variation in the organic matter between CA and CT in the Robit experimental site at the beginning and end of the experiment was not significant (Table A1) even if CA has a higher average organic matter than CT (Table 2). This may be due to a decrease in organic matter supply caused by a decrease in mulch cover in CA. The macronutrients which are usable for plant growth (nitrogen, potassium, and phosphorous) have a higher mean value in the CA management than CT systems. Statistically (α = 0.1), at the beginning of the experiment in Dangishta, nitrogen showed a significant difference in CA than CT, but the reverse is true for phosphorus and potassium (Table A1). The result from Robit showed there was no significant difference in micronutrients (nitrogen, potassium, and phosphorous) between the two treatments at (α = 0.1) (Table A1).

The water holding capacity of the treated soil (CA) is higher than the untreated soil (CT) and as a result, the permanent wilting point for CA is less than CT. As shown in Table 2, the average field capacity of CA (36.8%) is larger than CT (35.5%) in Dangishta experimental site which is significant at (α = 0.1) (Table A1). At the end of the experiment in Robit site, the average field capacity of the soil is significantly higher in CA (34.5%) treatment than CT (24.6%) at a 10% significant level (Table A1). In general, the soil physio-chemical properties for the successive irrigation season were variable due to change in climate, management conditions, and type of crop to be planted.

The soil physico-chemical analysis result taken from a sample of soil at 20 and 40 cm depth showed that soil properties vary with soil depth. Figure 4 showed that the surface soil or top-soil (0–20 cm depth from the surface) contains more organic matter (4.98%) in CA treatment than CT (4.15%) which is significant (

Table A2. A one tailed paired t-test for the average depth integrated soil physico-chemical properties taken from Dangishta experimental site.

Items	pH	OM (%)	EC (ds/m)	TN (%)	Av. k (ppm)	Av. P (ppm)	Av. CEC (%)	Fe (%)
CA	5.388	4.982	0.222	0.228	250.92	21.204	33.92	30.094
CT	5.268	4.156	0.28	0.238	289.34	24.684	31.88	31.186
N	5	5	5	5	5	5	5	5
P-value	0.3 **,a	0.005 *,a	0.25 **,a	0.21 **,a	0.27 **,a	0.13 **,a	0.08 **,a	0.36 **,a
CA	5.406	3.932	0.192	0.204	203.04	9.426	30.72	26.516
CT	5.582	3.42	0.142	0.17	316.76	9.432	29.76	26.592
N	5	5	5	5	5	5	5	5
P-value	0.14 **,b	0.06 **,b	0.16 **,b	0.05 **,b	0.05 **,b	0.49 **,b	0.37 **,b	0.48 **,b

Note: N = number of replicates, ^a = sample taken from 0–20 cm depth, ^b = sample taken from 20–40 cm depth, * α = 0.05, ** α = 0.1.

) at 5% significant level. The lower layer (20–40 cm depth from the surface) has an average organic matter content of 3.93% in CA and 3.42% in CT and is significant at a 10% significant level (Table A2). Electrical conductivity (EC), cation exchange capacity (CEC), soil pH, potassium (K), phosphorus (P), and total nitrogen (N) showed a variation in magnitude across depth in the two treatment conditions (CA and CT) but statistically there was no significant difference between CA and CT across soil depth at 10% level of significance (Table A2).

3.2. Amount of Water Used and Soil Moisture Content

Farmers applied a relatively small amount of irrigation water on CA as compared to CT. A one tailed paired t-test was done (

Table A3. A one tailed paired t-test for the average irrigation water use for variety of vegetables.

Items	Cabbage	Garlic	Onion	Pepper
CA	318.1	199.2	332.4	211.4
CT	385.2	275.4	406.4	245.4
N	5	5	5	5
P-value	0.006 *,a	0.0000236 *,b	0.00074 *,b	0.0074 *,b

Note: N = number of replicates, ^a = Robit experimental site, ^b = Dangishta experimental site, * α = 0.05, ** α = 0.1.

) and the average irrigation water use of cabbage, onion, garlic, and pepper was significantly (α = 0.1) higher in CT than CA. The average irrigation water used in CA showed a significant decrease (18%) as compared to CT in the cabbage experiment. About 385 and 318 mm of average irrigation water was applied to the CT and CA systems (Figure 5) for cabbage production, respectively. Allen et al. [47] explained that the average water consumption of cabbage varies from 350 to 500 mm depending on differences in the climate and soil condition. The amount of water applied to cabbage in the CA plot (318 mm) was below the lower limit suggested by Allen et al. [47] indicating significant soil moisture improvement under CA practice. CA reduces soil evaporation and preserves soil moisture due to the mulch cover and improvement of soil quality and structure through the biological decomposition of the organic mulch. The amount of water applied to cabbage in CT plots was consistent with Allen et al. [47] findings. Similarly, Tiwari et al. [57] estimated the average seasonal water requirement of cabbage to be 400 mm on clay soil and sub-humid climate. The average irrigation water applied to garlic was 275 mm in CT and 199 mm in CA plots. A significant irrigation water use was observed for garlic (28%) in CA when compared with CT. Similarly, a significant irrigation water saving was observed for onion and pepper under the CA treatment. The result showed there was 18% and 14% irrigation water savings under CA over CT treatment for onion and pepper, respectively. The average irrigation water applied in the onion experiment (Figure 5) is consistent with Bossie et al. [58] and with Tebal [59] for pepper in the CT plots. A one-tailed paired t-test analysis is presented in the Appendix B (Table A3).

The result from the soil moisture measurement showed that the average volumetric soil moisture content in the treated soil (CA) was higher than the conventional tillage (CT) practice in most of the irrigation periods. This shows conservation agriculture is a means of saving irrigation water which is essential in water scares areas and the driest season of irrigation (Figure 6).

3.3. Effect on Crop Coefficient (K_c)

The crop coefficient was determined as a function of reference and actual evapotranspiration. Reference evapotranspiration ranges from 2.8 to 5 mm per day (Figure 7) both for Dangishta and Robit watersheds which is consistent with Enku and Melesse [60] findings.

As shown in

Table 3. Maximum, minimum, and average crop evapotranspiration (ET_c) in mm/day for Robit and Dangishta watersheds.

Watershed	Treatment	Max	Min	Average
Robit	CA	4.0	2.8	3.4
	CT	4.3	3.0	3.6
Dangishta	CA	3	2.8	2.8
	CT	3.75	3.3	3.5

and

Table A4. A one tailed paired t-test of crop coefficient values of different vegetables under their respective growth stages at 5% significance level.

A, Initial stage					B, Mid-stage
Item	Cabbage	Onion	Garlic	Pepper	Item	Cabbage	Onion	Garlic	Pepper
CA	0.67	0.66	0.59	0.59	CA	0.97	0.93	0.82	1.22
CT	0.71	0.67	0.65	0.65	CT	1.19	1.24	1.07	1.31
N	4	5	5	5	N	4	5	5	5
p-value	0.13	0.44	0.26	0.12	p-value	0.04	0.003	0.02	0.27
C, Developmental stage					D, Late-stage
Item	Cabbage	Onion	Garlic	Pepper	Item	Cabbage	Onion	Garlic	Pepper
CA	0.89	0.78	0.67	1.12	CA	0.72	0.84	0.75	0.89
CT	1.02	0.93	0.82	1.28	CT	0.84	0.95	0.82	1.01
N	4	5	5	5	N	4	5	5	5
p-value	0.18	0.10	0.10	0.26	p-value	0.08	0.19	0.20	0.04

, the consumptive use of crops for the CA treatment was significantly (α = 0.05) less than the CT treatment for both experimental sites (Robit and Dangishta). This is because the conservation of irrigation water is mainly achieved through the mulch system. This indicates that the amount of stored water in the root zone for the CA treatment will be conserved than the CT practice.

A relatively higher crop coefficient result has been found under CT practice than CA (

Table 4. Average crop coefficients of cabbage in Robit, garlic, onion, and pepper in the Dangishta experimental site, respectively.

Treatment	Initial	Developmental	Mid	Late	Crop Type
CA	0.67	0.89	0.97	0.72	Cabbage
	0.59	0.67	0.82	0.75	Garlic
	0.95	1.12	1.22	0.89	Pepper
	0.66	0.78	0.93	0.84	Onion
CT	0.71	1.02	1.19	0.84	Cabbage
	0.65	0.82	1.07	0.82	Garlic
	1.19	1.28	1.31	1.01	Pepper
	0.67	0.93	1.24	0.95	Onion

). The higher crop coefficient obtained under CT management system can be associated with higher irrigation water use and losses associated with soil evaporation. The dual crop coefficient approach is ideal to understand the impacts of soil wetting by rainfall or irrigation and for estimating the effect of conservation practice in reducing soil evaporation. In this study, the single crop coefficient approach was used for determining crop coefficients at the different growing stages. The dual approach divides the crop coefficient in to crop transpiration and soil evaporation coefficients [61,62,63]. The crop transpiration coefficient “basal coefficient” needs more parameters to be determined and needs more numerical calculations. The soil evaporation coefficient (K_e) also needs estimation of constant parameters from a range of values which results in overestimation or underestimation of the coefficients. The crop height at each growing stage was not measured for basal coefficient determination and therefore the single crop coefficient was used. As the crop develops, the ground cover, crop height, and the leaf area also change which brings differences in the amount of evapotranspiration during the various growth stages, which finally brings variation in the crop coefficient.

During the initial period, the leaf area is small, and evapotranspiration is mainly in the form of soil evaporation. The crop coefficient during the initial period is high when the soil is wet during irrigation and low when the soil surface is dry. Particularly in the CT water management system, the irrigation applied to the plot is evaporated without any control and the major evapotranspiration in this stage is soil evaporation. The average crop coefficient of cabbage in developmental, mid, and late stage was 0.89, 0.97, 0.72 for CA treatment and 1.02, 1.19, 0.84 for CT practice, respectively.

The average initial crop coefficient for the cabbage experiment was 0.67 for CA and 0.71 for CT. This result showed a difference from crop coefficient developed by FAO for cabbage (i.e., K_c = 0.45), for non-stressed, well-managed crops in sub-humid climates. For the garlic experiment, a crop coefficient of 0.59 for CA and 0.65 for CT was found.

According to the previous studies conducted by Ayars [64] an initial crop coefficient of 0.53 for garlic was estimated which shows a difference from this experiment. The difference may be due to variations in climate, soil conditions, and water management practices. A coefficient of 0.67, 0.82, 0.75 for CA and 0.82, 1.07, 0.82 for CT treatment were found (Table 4) for garlic. A 20% average K_c reduction was observed for both cabbage and garlic. The results were compared with previous studies conducted by Ayars [64] for garlic and Gulik and Nyvall [65] for cabbage (Figure 8). The minimum crop coefficient value was observed in the initial stage and gets increased in the development stage. The maximum crop coefficient was observed in the mid-stage and finally decreased in the harvesting stage as shown in Figure 8. This trend was seen both in the CA and CT water management system. During the mid and late-stage plant, leaves begin to age, and K_c gets decreased at the end of the growing period.

In CA and CT water management systems, a higher crop coefficient was observed at the mid-stage because the leaf area is large and crop evapotranspiration (ET_c) is in the form of evaporation from the soil and transpiration from the plant leaf. During the late stage, the plant leaves begin to die which resulted in a decrease in K_c value. At the initial and developmental stages, the variation of crop coefficient was not statistically significant (Table A4). Whereas in the mid-stage except for pepper, the variation in crop coefficient between CA and CT is significant (α = 0.05) (Table A4). The average crop coefficient values for cabbage, garlic, and onion throughout the crop base period showed a significant (α = 0.05 and α = 0.1) increase in CT than CA (

Table A5. A one tailed paired t-test of average crop coefficient values of cabbage, onion, garlic, and pepper throughout the crop base period.

Item	Cabbage	Onion	Garlic	Pepper
CA	0.81	0.80	0.71	1.05
CT	0.94	0.95	0.84	1.20
N	16	20	20	20
p-value	0.007 *	0.005 *	0.003 *	0.054 **

Note: N = number of replicates, * α = 0.05, ** α = 0.1.

). Therefore, CA is a promising strategy to save scarce water resources through a decrease in runoff, evaporation, and an increase in soil moisture and groundwater recharge. An improvement in soil fertility was observed in CA practice from the biological processes happening in the soil environment.

Similar findings in the crop coefficient of onion and pepper showed an increasing pattern from initial to mid-stage and finally decreased at the late stage (Figure 9). For onion experiments a coefficient of 0.67, 0.93, 1.24, and 0.95 were observed in the CT plots whereas in CA plots a coefficient of 0.66, 0.78, 0.93, and 0.84 were found at the initial, developmental, mid, and late stages, respectively. A significant (α = 0.05) reduction in K_c (10%) was observed for pepper. The variations in K_c were found to be consistent with Bossie et al., [58] findings which are (0.47 at the initial stage, 0.7 at the developmental stage, 0.99 at mid-stage, and 0.46 at a late stage). The result found from the pepper experiment showed a rapid increase in K_c value when compared with other crops (Table 4). In the CT experiment, a crop coefficient of 1.19, 1.28, 1.31, and 1.01 was observed in the four growing stages which are different from Asres [66] findings (K_c value of 0.6 at initial, 0.78 at developmental, 0.95 at mid-stage, and 0.6 at harvest), respectively. The result obtained from the CA experimental showed a K_c value of 0.95 at the initial stage, 1.1 at the developmental stage, 1.2 at mid-stage, and 0.9 at the late stage.

The variation in CA and CT was also clearly seen in vegetable production (yield of vegetable per hectare of land). The CA management system increased productivity than the CT practice. The average vegetable yield in CA was 3.2, 3.05, 23.58, and 19 tha-1 for onion, garlic, cabbage, and pepper, respectively. Whereas in the CT, the average vegetable yield was 2.81, 1.96, 21.54, and 13.6 tha-1 for onion, garlic, cabbage, and pepper, respectively. The result depicts that CA increased productivity by 16%, 56%, 9%, and 40% for onion, garlic, cabbage, and pepper, respectively. This was due to the improvements in soil quality and structure [19,67].

4. Conclusions

Conservation agriculture with a drip irrigation system was found efficient for agricultural water management than the conventional tillage system. About 18% of irrigation water saving was observed for cabbage and onion experiment under conservation agriculture than conventional tillage. Similarly, 14% and 28% of water savings were also achieved in the pepper and garlic experiment in Dangishta, respectively.

Calculation of crop coefficient revealed a significant decrease under conservation agriculture as compared to conventional tillage practice. The result depicted that conservation agriculture with a drip irrigation system significantly (α = 0.05) reduced crop coefficient values of crops as compared to conventional tillage. Specifically, 20% reductions were observed for onion, cabbage, and garlic under conservation agriculture whereas 10% reductions were observed for pepper throughout the crop base period.

Similarly, the result obtained from the soil test indicated that conservation agriculture was a promising approach for soil quality improvement than the conventional tillage practice. A significant improvement was observed in organic matter (10%), field capacity (4%), and total nitrogen (10%) under conservation agriculture than conventional tillage practice at a 10% significant level corroborates this finding. Therefore, shifting the current agricultural practice is worth enough for improving water and crop productivity in the region.