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

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 Kc values under CA as compared to CT. The result depicted that CA with a drip irrigation system significantly (α = 0.05) reduced Kc 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.


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 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.
Water 2020, 12, x FOR PEER REVIEW 3 of 20 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.

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 m 2 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. 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.

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 m 2 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. 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.
Water 2020, 12, x FOR PEER REVIEW 4 of 20 Figure 2. Gravity based drip irrigation system on a paired 't' experimental design; half of the plot was used for conservation practice and the other half is the conventional tillage practice.

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 m 2 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 (Figures 2 and 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.
CA 500 liter Tanker CT Figure 2. Gravity based drip irrigation system on a paired 't' experimental design; half of the plot was used for conservation practice and the other half is the conventional tillage practice.

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 m 2 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 (Figures 2 and 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.

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].

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).

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

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].

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).

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.

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.
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).
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].

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

Experimental
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) 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).

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 B1) 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 Section (Table B1).

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) 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).  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). 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).

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.
Water 2020, 12, x FOR PEER REVIEW 12 of 20

Effect on Crop Coefficient (Kc)
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 and Table B2, 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). 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 (Ke) 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.  Tables 3 and A4, 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). 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.

As shown in
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). 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.  [64] and cabbage compared with Gulik and Nyvall [65] for CA management and CT practice.
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 Kc (10%) was observed for pepper. The variations in Kc 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 Kc 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 (Kc 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 Kc 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.  [64] and cabbage compared with Gulik and Nyvall [65] for CA management and CT practice.
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.  [58] and pepper compared with Asres [66] for CA management and CT practice.
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].

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 ( = 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.  [58] and pepper compared with Asres [66] for CA management and CT practice.
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].

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.
Author Contributions: A.Y.Y. contributed to the experimental design, data analysis and interpretation, and drafted the manuscript; T.T.A. contributed to the experimental design, data acquisition and analysis, and revised the manuscript for the scientific content; N.F.A. contributed to data collection and analysis; S.A.T. contributed to the experiment, data analysis and interpretation, and revised the manuscript; M.K.J. contributed to data analysis and revised the manuscript; M.R.R. contributed to data acquisition and revised the manuscript. All authors have read and agreed to the published version of the manuscript. Acknowledgments: We would like to acknowledge the Ethiopian National Meteorological Agency (ENMA) for providing quality data for this research.