Evapotranspiration and Yield Impact Tools for More Water-Use Efficient Alfalfa Production in Desert Environments

Abstract

Drought and climate change have decreased water availability for agriculture, especially in the desert of southwestern USA. Efficiency enhancements in irrigation management aimed at conserving water are key to adjust to limits in water supply, improve profitability and sustainability of alfalfa production in arid and semiarid areas. This study intended to conduct a field-scale analysis to develop yield and ET estimation tools for the effective use of irrigation water in a desert alfalfa production system. Extensive data collection and trials were carried out over three years in nine fully irrigated commercial alfalfa fields in the low deserts of California. The seasonal crop water consumption measured using the residual of energy balance method varied from 1381 mm to 1596 mm across the experimental sites and crop seasons. Variable seasonal dry mater (DM) yields ranged from 23.01 Mg ha⁻¹ to 29.90 Mg ha⁻¹. The results indicated that the first five cuttings each year were the most productive cuttings with a mean DM value ranged between 3.29 (cut 1) and 4.21 (cut 4) Mg ha⁻¹ but declined in later cuttings. An average annual water productivity (WP) value of 17.0 kg ha⁻¹ mm⁻¹ was determined across the sites varying from 15.5 to 18.9 kg ha⁻¹ mm⁻¹. The findings suggested that one may lose up to 1.44 Mg ha⁻¹ alfalfa yields with moderate summer deficit irrigation strategies, using 40% less water applied than full irrigation practices over the summer period of July–September. A more severe summer water deficit, with no irrigation event over the summer period of July–September may result in a potential water savings of 0.234–0.246 (ha·m) ha⁻¹ and 19–21% seasonal yield losses in the desert environment. This study describes the seasonal yield pattern, the crop water use-production function, and the crop coefficient values for various harvest cycles over the crop season. These tools may assist farmers to quantify water savings and estimate yield losses for more accurate and effective irrigation management strategies to meet water conservation objectives and for the resiliency of alfalfa production in the desert region.

1. Introduction

In the low desert of California, alfalfa (Medicago sativa L.) is the dominant water user due to its large-cultivated area and long growing season. Alfalfa has a total hectarage of nearly 94,000 in the low deserts of the Palo Verde and Imperial Valley agricultural areas of California, accounting for more than 30% of the crop area [1,2,3]. This irrigated region is a part of the lower Colorado watershed where alfalfa and other forages account for a large component of agricultural water use. Due to recurring long-term droughts and water shortages in the Colorado River Basin, the main source of irrigation water, there are growing concerns on the yearly shortfall between water supply and demand. Limited water supplies impact alfalfa production as well as other crops and urban use. Therefore, implementing water conservation tools and practices, and efficient use of irrigation water is necessary to sustain agricultural system in the region.

While nearly 95% of the desert alfalfa is currently irrigated by surface irrigation systems [1], various irrigation strategies and on-farm water conservation practices have already been adapted by the local farmers to enhance the efficiency of water use. This includes utilizing subsurface drip irrigation (SDI), overhead linear move irrigation, and automated surface irrigation, and management strategies of tailwater recovery systems and deficit irrigation strategies [1].

Accurate estimate of alfalfa water-use as well as yield impacts of water deficits are critical for efficient irrigation water management, drought strategies, water transfer planning, and hydrological analyses. Worldwide, alfalfa seasonal evapotranspiration (ETc) is estimated between 800 and 1600 mm depending on climate and length of growing period [4,5].

According to a two-year measurement using eddy covariance (EC) method at a commercial field located in Northern California, a range of 1249 to 1381 mm was reported as alfalfa seasonal water use under surface irrigation over six harvest periods [6]. Alfalfa actual ET was estimated 625 mm at the windy arid environment of Curlew Valley, Utah using the residual of energy balance calculations [7]. The average ET for 751 alfalfa fields in the Lower Rio Grande Valley, New Mexico, using three OPEC (One Propeller Eddy Covariance) towers was determined 901 mm, which ranged from 386 to 1241 mm [8].

The average crop coefficient values of alfalfa varied from 0.80 to 1.09 in the Sacramento Valley, with a range of 0.45–0.60 just after harvest and 1.10–1.20 just before harvest [6]. The experimental fields were under common farmer irrigation practices with border check surface irrigation. The range of alfalfa crop coefficient (fields under border check surface irrigation) was found 0.5 to 1.3 in saline and shallow water tables area of the Imperial Valley, California based on measurements of applied water and soil moisture depletion data [9]. A wide range of the seasonal alfalfa actual crop coefficient values for the 751 fields investigated in the Lower Rio Grande Valley, New Mexico was reported (varied from 0.2 to 0.8, with an average value of 0.59) [8], in which irrigation methods, improper irrigation scheduling, limited water supply, interference of harvesting schedule with the irrigation, and cultural practices were recognized as the reasons of crop coefficient variability. Thus, these ET measurements likely reflected under-irrigated crops. An average seasonal actual crop coefficient value of 0.84 was found using the surface renewal (SR) and eddy covariance measurements in four commercial alfalfa fields in the Palo Verde Valley [1]. These alfalfa fields were under border check surface irrigation and furrow irrigation.

Alfalfa is an herbaceous crop with rapid growth characteristics, and its yield is considered to be linearly related to evapotranspiration [10], particularly within a growth period. A wide range of 10–34 kg ha⁻¹ mm⁻¹ alfalfa water productivity has been reported in earlier studies [11,12,13,14,15,16,17,18,19,20,21]. The review of nine studies conducted in the west and Midwest of the U.S. showed that averagely 188 mm is required to produce 1 Mg of alfalfa [22]. Some of these studies used older alfalfa varieties and had less frequent cutting schedules than present practices.

Although alfalfa has a high-water demand footprint, it has a range of water advantages during times of water scarcity [23]. These include its deep rooting patterns to utilize residual moisture during drought, high water use efficiency compared with other crops, and perenniality which enable quick re-growth without stand establishment. However, alfalfa is particularly valuable for its flexibility to be ‘deficit irrigated’ when water is unavailable, or partial-season water applications. The fact that multiple harvests of alfalfa are possible enables partial season irrigations to produce some harvest, while conserving water for other purposes, unlike annual crops with a single harvest. The ability of alfalfa to sustain temporary droughts, maintain yields and survive severe irrigation deficits is somewhat unique among productive crops.

Three basic deficit irrigation strategies have been proposed for drought mitigation in alfalfa fields [24], including: (1) triage (cease irrigation of some alfalfa fields while fully irrigating others), (2) starvation diet (deficit-irrigate the entire alfalfa field during the whole crop season), and (3) partial-season irrigation (fully irrigate fields the early cuttings, then cease irrigation pathway through the rest of the season). Moderate deficit irrigation (MDI), skipping some of the irrigation events over the summer period, has been proposed as a water conservation tool in California low desert [1]. However, estimates of yield impacts are often lacking for various of these strategies. Yield impact tools, and water conservation estimates are necessary for recommending various strategies for farmers.

Deficit irrigation was investigated as a sustainable crop production strategy to maximize water productivity and to stabilize alfalfa yields while conserving irrigation water [24,25]. A recent study reported that deficit irrigation could be applied to alfalfa at the bud stage when it less affects the yield than deficit irrigations at the regrowth or branching stage in the arid and semi-arid areas [25]. The result of a study conducted in the arid area of Northwest China suggested that the partial root-zone drying could be a promising technique for alfalfa production with improved water productivity and positive effect on quality characteristics [26].

At the field scale, alfalfa crop water use, yield, and water productivity are variable. These variabilities may be attributable to inappropriate irrigation management due to the lack of sufficient irrigation scheduling knowledge and tools, spatial and temporal variations in water availability, agronomic constraints, and irrigation systems and infrastructure limitations. Therefore, knowledge of alfalfa crop water-use, crop coefficients and water deficit impacts may assist in promoting effective irrigation management.

Since agriculture is the major water consumer in many regions, adjustments in cropping patterns and practices are likely to be necessary when confronted with genuine water scarcity, as is currently the case in the Colorado watershed. Although wholesale changes in crops and extensive fallowing have been proposed, partial-season deficit irrigation of alfalfa is an important strategy to consider, which could maintain forage production and farm profitability while conserving scarce water resources. The tools that are needed to understand the ability of agriculture to respond to water scarcity are (1) accurate on-farm estimates of total water use and water application practices, (2) yield estimates across soil types that are sensitive to seasonal yield patterns and potential yield penalties from water deficits, and (3) value of improved water application methods that will further improve water productivity under water scarcity.

The objective of this research is to characterize season-long evapotranspiration and seasonal yield patterns of low-desert alfalfa production using current farm practices across various soil types and conditions. It is aimed at developing applied tools to design deficit irrigation management strategies for profitable and sustainable alfalfa production in an era of water constraints.

2. Materials and Methods

2.1. Experimental Sites

The experiments were conducted in nine commercial alfalfa fields (designated “site A” through “site I”) over a three-year period (the 2019 season to the season 2021) in the Palo Verde and Imperial Valleys, California (

Table 1. Irrigation methods and soil texture of the experimental sites. “A” through “I” represent alfalfa experimental sites. SDI, LMI, and ASI abbreviate subsurface drip irrigation, linear move sprinkler irrigation, and automated surface irrigation, respectively.

Experimental Site	Irrigation Method	Soil Texture
		Sand (%)				Silt (%)				Clay (%)
		Generic Horizon (m)				Generic Horizon (m)				Generic Horizon (m)
		0–0.3	0.3–0.6	0.6–0.9	0.9–1.2	0–0.3	0.3–0.6	0.6–0.9	0.9–1.2	0–0.3	0.3–0.6	0.6–0.9	0.9–1.2
A	Furrow	44	47	42	48	45	45	51	46	11	8	7	6
B	Furrow	40	50	75	84	40	38	20	12	20	12	5	4
C	Border	32	26	88	95	55	55	9	4	13	19	3	1
D	Border	69	92	82	86	18	5	10	8	13	3	8	6
E	SDI	54	40	28	31	28	38	45	37	18	22	27	32
F	SDI	22	19	33	63	45	46	45	32	33	35	22	5
G	ASI	26	22	23	17	39	49	54	54	35	29	23	29
H	LMI	14	11	10	10	19	6	32	20	67	83	58	70
I	LMI	35	44	51	30	31	25	28	53	34	31	21	17

Dominant soil textures of the sites at the topsoil (0–0.6 m) are categorized as: loam (A–B), clay (C), sandy loam (D and F), silty clay loam (E and G), silty clay (H), and silty loam (I).

). The area has a true desert environment with an annual average air temperature, total annual precipitation, and potential evapotranspiration (ETo) of 21.4 °C, 78 mm, and 1782 mm, respectively [1]. The average air temperature of June and July is nearly 32.2 °C [1].

Dominant soil textures ranged from sandy loam at sites D and E, to loam at sites A and B, and silty clay loam to clay at sites C, F–I (Table 1). The experimental fields represent various irrigation management practices in the region consisted of conventional straight border check (flood) irrigation (sites C and D) and graded furrow irrigation (‘bedded alfalfa’, sites A and B), subsurface drip irrigation (SDI, sites E and F), linear move sprinkler irrigation (sites H and I), and automated border irrigation (site G). During the time of this study, farmers utilized ‘normal’ full irrigation (not deficit) practices common to the region.

The Colorado River water was the source of irrigation water for all sites with an average pH of 8.1 and an average electrical conductivity of 1.17 dS m⁻¹. All fields were planted in October 2018, except site H and I that were planted in October 2019. Alfalfa fields were typically harvested in a 28 day to 33 day cycle during spring and summer, with a total of 8–9 cuttings per year during the study period.

2.2. Evapotranspiration Measurements

The actual evapotranspiration (ETa) was measured using the residual of energy balance (REB) method with a combination of surface renewal and eddy covariance techniques. The sensible heat flux density (H) was estimated by the SR and EC methods to calculate the latent heat flux density (LE) using the REB approach [1,27,28,29,30,31,32].

A full flux density tower was set up at the experimental sites (Figure 1). Two chromel-constantan thermocouples model FW3 with 76.2 µm diameter (Campbell Scientific, Inc., Logan, UT, USA) were used to measure high frequency temperature data for computing uncalibrated sensible heat flux (H0) using the SR technique. An RM Young Model 81000RE sonic anemometer (RM Young Inc., Traverse City, MI, USA) was used to collect high frequency wind velocities in three orthogonal directions at 10 Hz to estimate H for the latent heat flux density calculations using the EC technique.

The measurements of soil heat storage at three different locations using three HFT3 heat flux plates (REBS Inc., Bellevue, WA, USA) inserted at a 0.05 m depth below the soil surface, and the measurements of soil temperature using three 107 thermistor probes (Campbell Scientific, Inc., Logan, UT, USA) at three depths in the soil layer above the heat flux plates were used to estimate ground heat flux (G). An NR LITE 2 net radiometer (Kipp & Zonen, Ltd., Delft, The Netherlands) was used to measure net radiation (Rn). The data were recorded using a combination of a Campbell Scientific CR1000X data logger (Campbell Scientific, Inc., Logan, UT, USA) and a CDM-A116 analog input module (Campbell Scientific, Inc., Logan, UT, USA)). Except for the soil sensors, all other sensors were set up at 1.8 m above the ground surface.

As an additional sensor, Tule sensor (Tule Technologies, Inc., Oakland, CA, USA) was used in some of the monitoring sites to estimate ETa using a simple surface renewal technique. The estimated daily ETa from the Tule sensors at those sites was verified by comparing with the ETa measured from the full flux density towers and the results were adopted to fill the gaps during some periods that other sensors needed time-consuming maintenance.

2.3. Normalized Difference Vegetation Index Data

The METER Spectral Reflectance Sensor was used to measure Normalized Difference Vegetation Index (NDVI) (METER Groups Inc., Pullman, WA, USA). The sensor has two versions: (1) cosine-corrected sensor (SRS-Ni Hemispherical Sensor) for incident radiation and (2) a field stop lens sensor (SRS-Nr Field Stop Sensor) for reflected radiation from the canopy. METER ZL6 data logger (METER Groups Inc., Pullman, WA, USA) was used to record the data on a 30 min basis. The SRS-NDVI was mounted 1.6 m above the ground surface.

2.4. Soil Moisture Monitoring

Watermark Granular Matrix Sensor (Irrometer Company, Inc., Riverside, CA, USA) was used to measure soil water potential at multiple depths of 15–120 cm, on a continuous basis. The data of soil moisture sensors were recorded by a 900M Monitor data logger (Irrometer Company, Inc., Riverside, CA, USA)) on a 30 min basis.

2.5. Soil Salinity Survey

Soil properties and salinities were surveyed and characterized within a footprint area of 30 m × 30 m around the ET monitoring stations in each site. At each site, soil cores at four distinct depth ranges (0–30, 30–60, 60–90, and 90–120 cm) were taken from five sampling locations. A comprehensive laboratory analysis was conducted on the composite soil samples for each depth range.

2.6. Alfalfa Dry Matter Yields

In each site, yields were measured from 12 sub-plots (replicates) at each site with 1.5 m wide and 2.0 m long PVC quadrats, hand, harvested at 6–8 cm cutting height. This is the most common cutting height used by commercial harvesting equipment. Fresh weights of plants harvested within the quadrats were recorded, then samples were dried for three days in conventional oven at 60 °C, weighed and recorded to adjust alfalfa dry matter (DM) content.

2.7. Moderate Summer Deficit Irrigation Trials

On four selected farmer’s fields (sites A–D), two moderate deficit irrigation strategies were implemented on each field, in addition to the fully irrigated alfalfa. Deficit strategies were (1) elimination of two irrigation events during July–August, and (2) elimination of three irrigation events during the period July–September. The fields were divided into three large plots to allocate an average are of approximately 7 ha to each of the three strategies (Full, elimination of two, and elimination of three irrigations during the summer period).

2.8. Data Set and Analysis

The extensive data set collected from the alfalfa sites under fully irrigated practices (site A–I) and moderate summer deficit irrigation strategies (site A–D) were used for this analysis. The data set includes three-year data (2019–2021) for the sites A through D and two-year data (2019–2020) for the sites E through I.

Both the EC and SR techniques were individually employed to determine sensible heat flux density. The advantage of using both the SR and EC methods was that they are independent and similar results provide a high level of confidence in the data used [32,33]. Latent heat flux density was calculated as LE = Rn − G − H. After determining LE, ETa in mm d⁻¹ was calculated by dividing the LE in MJ m⁻² d⁻¹ by 2.45 to obtain the ET values in mm d⁻¹. Water productivity (WP) was calculated as the ratio of alfalfa dry matter (kg ha⁻¹) to ETa (mm). Fisher’s protected least significant difference (LSD) test was used to evaluate significant differences in DM yield and WP values between the experimental sites.

Using the daily ETa determined in each monitoring site and the daily ETo from the spatial CIMIS (California Irrigation Management Information System) data [34] for the coordinates of the monitoring site, the daily actual crop coefficient (Ka) was calculated as the ratio of ETa to ETo. CIMIS uses the Penman–Monteith equation, and a version of Penman’s equation modified by Pruitt–Doorenbos [35]. Spatial CIMIS combines remotely sensed satellite data with traditional CIMIS stations data to develop site specific ETo on a 2000 m grid.

3. Results

3.1. Weather Parameters

While trivial differences were found in the amounts of average daily solar radiation, air temperature, and dew point temperature among the study seasons, differences of average daily wind speed were significant (Figure 2). The average daily wind speed was nearly 20% greater in the 2019 season than the season 2021 (2.3 vs. 1.9 m s⁻¹), mostly occurred during the first six months of the year. A higher wind speed increases crop water use and consequently impacts irrigation management in terms of irrigation amounts and frequency. Average daily solar radiation (241 W m⁻²) and air temperature (22 °C) were slightly higher, and average dew point temperature (6.6 °C) was slightly lower in the 2020 season than the other two crop seasons. During the 2020 season, the month of June had the highest average daily solar radiation of 346 W m⁻², followed by May (341 W m⁻²) and July (336 W m⁻²). Greater solar radiation and air temperature increase potential evapotranspiration and as a result irrigation water needs in alfalfa fields. There was higher than normal rainfall in the fall of 2019 season and the winter of 2020 season (Figure 2). Average annual rainfall for this region is 75 mm while the total precipitation was 110 mm in a six-month period of October 2019 through March 2020.

3.2. Crop Water Use

Variable daily crop water consumption (ETa) was mostly affected by harvest schedule (over each growth period), time of the year, and irrigation management. For instance, the ETa at site A over the 2019 season ranged between 2.8 mm d⁻¹ after the June 29 cutting, increasing to 10.4 mm d⁻¹ at midseason full crop canopy on July 20 (Figure 3a). The maximum and minimum ETa during the harvest cycle of November–December 2019 at this site were 2.7 mm d⁻¹ and 0.2 mm d⁻¹, respectively. The maximum daily ETa was 9.1 mm d⁻¹ in the 2020 season and 8.6 mm d⁻¹ in the season 2021.

The seasonal ETa varied widely for each crop harvest cycle and throughout the study seasons in each site. For instance, the value varied from 241.6 mm for the fifth harvest cycle to 159.3 mm for the seventh harvest cycle of the 2019 season at site A, while it ranged between 228.5 mm and 125.7 mm for the eighth harvest cycle of the 2020 season at this site (

Table 2. Cumulative actual ET (ETa) per harvest cycles of the 2019 and 2020 seasons at site A. The site had eight cuttings in each season.

Study Season	Harvest Cycle Number
	1	2	3	4	5	6	7	8
ETa (mm)
2019 season	194 (301) *	207.0	218	234	242	181	159	166
2020 season	134	218	203	214	229	192	166	126

* This harvest cycle includes the entire period from the first irrigation event after planting seeds on 14 October 2018, through 23 March 2019 (first harvest of the 2019 season at this site). The value was 195 mm from 1 January 2019, by the first harvest on 23 March 2019.

).

Alfalfa season-long crop water consumption may be affected considerably by irrigation schedule and other management practices, soil types and conditions, and weather parameters. These results demonstrated that the seasonal crop water consumption varied from 1381 mm at site F in the 2019 season to 1596 mm at site A in 2019 (Figure 4). An average seasonal crop water use of 1510 mm was determined for alfalfa across all experimental sites.

3.3. Salinity

The mean EC_e values at the effective crop root zone (30–120 cm) demonstrated that salinity level was in the ‘acceptable’ range for alfalfa crop at the majority of experimental sites except sites H and I, both fields under linear mover sprinkler irrigation, in the 2019 season (H-19 and I-19) (Figure 5). The average ECe values at the depths of 30–90 cm for sites H and I were 4.9 and 4.2 dS m⁻¹, respectively. A higher level of salt accumulation was also observed at site F, field with a silty clay-loam dominant soil texture under subsurface drip irrigation. While likely no significant salinity stress influenced the amounts of crop water consumption across the experimental sites, the ETa could be affected by higher level of salinity at sites F, H, and I. The results suggested that advanced irrigation delivery systems such as SDI and linear move need to be effectively managed to maintain low soil salinity in the desert environment over time. This is not a challenge observed at the surface irrigated fields where flood or sprinkler irrigation techniques may provide excessive water to leach salts.

3.4. Crop Coefficient Values

The number of harvests per season was 8–9 at each site over the study period. Due to frequent harvesting events, actual crop coefficient oscillates over the harvest cycles (Figure 3b). The crop coefficient values depend greatly on the growth stage, ranging from lowest during initial growth stage (stubble), just after each harvest, and reaching the greatest at full canopy development, attained prior to each harvest. The K_a ranged from nearly 0.3 to 0.5 after hay is cut, to a maximum of nearly 1.24 at full canopy over each harvest cycle. Changes also occur over the season, with lower values later in the season (Figure 3).

Alfalfa demonstrated seasonality values of crop coefficients in which the average value of harvest cycles may vary over the season. Lower crop coefficient values were obtained at the early and late season harvest cycles and higher values in mid-season harvest cycles. For instance, the average Ka for the harvest cycles of April (harvest cycle 2) and May (harvest cycle 3) was 0.96 and 0.87, respectively, at site A in the 2019 season (Figure 3). However, these values declined to 0.76 and 0.81 for the harvest cycle August (harvest cycle 6) and September (harvest cycle 7) at this site in the same season.

3.5. NDVI

Lower daily NDVI values were detected for the July–September harvest cycles when compared with the May–June harvest cycles reflecting poorer growth during this period (Figure 6). Alfalfa plants regrowth after harvest was slower during the summer period than the May–June months. The heat could be the main explanation for this result. The results illustrated that the average peak daily NDVI values at full canopy could be a good indicator to represent average crop coefficient values of alfalfa fields in each harvest cycle. For instance, the average maximum daily NDVI at site D was 0.93, 0.91, 0.85, and 0.83 during the harvest cycles of May through August, respectively. These values are relatively similar (±2.3% points difference) to the average crop coefficient values determined for the corresponding harvest cycles at this site.

3.6. Dry Matter Yields

Alfalfa dry matter measurements indicated that the first five cuttings were the most productive cuttings across the experimental sites (Figure 7). Early season DM yields varied from 3.29 (cut 1) to 4.21 (cut 4) Mg ha⁻¹ over the first five cuttings. However, early-season yields were more variable across the sites and years (Figure 7). The standard deviation value of DM yield was 1.24, 0.81, 0.80 Mg ha⁻¹ for the first through third cuttings, respectively. The high SD value of the first cutting is due to a greater alfalfa yield in the first harvest after establishing alfalfa fields. Alfalfa fields often have a higher yield at the first harvest after planting, that typically occurs 5–6 months after planting. This is not something that happens during the following seasons and consequently a greater standard deviation was observed for the first cutting.

Lower mean DM values were obtained for the last four cuttings, varying between 2.87 (cut 6) and 1.48 Mg ha⁻¹ (cut 9). The maximum and minimum DM yield observed per cut was 5.43 and 1.24 Mg ha⁻¹ across the sites and the seasons, respectively.

Seasonal alfalfa DM yields varied across the experimental sites. The mean seasonal DM value varied from 23.01 Mg ha⁻¹ at site E (the 2020 season) to 29.90 Mg ha⁻¹ at site B (the 2019 season) (Figure 8). Overall, alfalfa yields were greater in the first year after the establishment than the following years, and generally decline over the subsequent years (Figure 8). Plant stand loss over the time is likely one of the main reasons of yield reduction. An average of 26.25 Mg ha⁻¹ was determined as the seasonal alfalfa yield of the experimental sites.

While we expected greater crop water use at the fields under SDI due to better water distribution uniformity over time and space, the findings of this study and an earlier study [1] clearly revealed that this expectation depended upon how irrigation water was managed. For instance, the soil moisture and applied water data verified that sites C and G in the 2019 season likely experienced moderate water stress even though farmers had no plan for deficit irrigation at these sites. The soil moisture tension data clearly demonstrated that site C could have benefited from one extra irrigation event between mid-March and late-April 2019 (Figure 9a). The cumulative ETa was 163 mm at this site for that harvest cycle while the total water applied was 148 mm, indicating deficit irrigation practices at that site.

3.7. Water Productivity

An average seasonal water productivity of 17.0 kg ha⁻¹ mm⁻¹ was obtained across all experimental sites (Figure 10). The seasonal WP value varied from 15.5 kg ha⁻¹ mm⁻¹ at site C, field under furrow irrigation, in the 2021 season to 18.9 kg ha⁻¹ mm⁻¹ at site B, field under border irrigation, in the 2019 season.

The WP values per each harvest cycle for sites B, C, and D in the 2019 season are given in

Table 3. Water productivity per each harvest cycle at sites B, C, and D over the 2019 season.

Cutting No.	1	2	3	4	5	6	7	8
		Water productivity (kg ha−1 mm−1)
Site B	18.1	21.8	27.1	21.5	17.0	17.8	17.4	9.4
Site C	13.6	24.4	22.1	19.4	14.4	14.3	12.7	10.5
Site D	19.6	24.5	25.0	24.3	13.2	12.6	11.4	10.2

. In general, the WP values were greater at cuts 2–4 than later cuts. For instance, the third cutting had the greatest WP at the site B (27.1 kg ha⁻¹ mm⁻¹) while the greatest WP was occurred in the second cutting at site C (24.4 kg ha⁻¹ mm⁻¹). These values are near to the maximum values reported for alfalfa WP by earlier studies, 10–34 kg ha⁻¹ mm⁻¹ [11,12,13,14,15,16,17,18,19,20,21].

These data show that alfalfa WP in the region is very high during the harvest cycles of April–June and declines considerably during the harvest cycles of July through the late fall (the second half of crop season). The findings revealed that the average WP of the second half of alfalfa crop season may be reduced by nearly 40% (site B) to more than 90% (site D) in compared with the average WP of the first half of crop season. Site D has a sandy loam soil texture which exhibited moderate water stress around the harvest cycles even during the first half of crop season (Figure 9b). However, this water stress continued and was even more severe during the late summer period.

3.8. Impact of Less Water Applied than Regular Farmer Practice

The imposition of deliberate moderate summer water deficits over three growing seasons resulted in negligible yield reductions (less than 5.7%), ranging from 0.47 Mg ha⁻¹ at site B in the 2021 season (D2) to 1.44 Mg ha⁻¹ at site A in the 2019 season (D1) (Figure 11). These differences were not significant statistically. The seasonal DM yields of site B and A in the corresponding seasons for the farmer irrigation practice were 26.2 and 27.7 Mg ha⁻¹, respectively. The average seasonal alfalfa DM yield of sites A–D for the plots under farmer irrigation practice was 26.8, 25.1, and 24.2 Mg ha⁻¹ in the 2019, 2020, and 2021 seasons, respectively.

The modest yield reductions observed could be a consequence of reduced water distribution uniformity caused by deficit irrigation scenarios. Large-scale farming systems (furrow and borders with a longer length than 370 m) along with the use of surface irrigation methods could be the primary reasons for lower water distribution uniformity values over time and space while implementing water deficits practice. Farmers generally apply more water to achieve higher water distribution uniformity under check flood irrigation; however, even a moderate deficit irrigation strategy may reduce distribution uniformity, affecting alfalfa yields during the course of water deficit practice. A similar trend was found across all four experimental sites over the study seasons.

The results demonstrated that 4 to 10% (0.038–0.124 (ha·m) ha⁻¹) of the seasonal irrigation water delivered to the alfalfa sites can be conserved by adopting moderate summer deficit irrigation strategies (Figure 12). 0.124 (ha·m) ha⁻¹ is nearly 40% of the average water applied in irrigation farmer practices over the summer period. The average seasonal water applied at the sites A–D for the plots under farmer irrigation practice was 1.196, 1.168, and 1.150 (ha·m) ha⁻¹ in the 2019, 2020, and 2021 seasons, respectively.

The findings suggested that implementing the MDI strategies could serve as an effective water conservation tool and provide a reliable source of available water as well as sustain the economic viability of alfalfa production. Following this deficit irrigation strategy, one may lose up to 1.44 Mg ha⁻¹ DM yields due to conserving up to 0.124 (ha·m) ha⁻¹ water. It needs to be noted that insignificant soil water depletion, non-crucial salt accumulation, and minor impacts on plant stands and hay quality could be expected due to adapting this moderate deficit irrigation practice in the region [1].

4. Discussion

4.1. Effects of Irrigation Methods and Management Practices

Better irrigation management is key to increasing alfalfa water productivity in the desert environment. The data presented here provide ET estimates, the possibilities of water savings at different times of year, and potential yield impacts of deficit strategies targeted to conserve water.

The results showed that changes in irrigation technology itself is not a fix-all solution to improving hay yields and water productivity. For instance, the seasonal ETa was 1450 mm at site E and 1481 mm at site F (both sites under subsurface drip irrigation) in the 2020 season. These amounts were lower than the measured ETa at the surface irrigated sites over this season (Figure 4). The seasonal WP values were also lower, 15.9 kg ha⁻¹ mm⁻¹ at site E and 15.5 kg ha⁻¹ mm⁻¹ at site F, than all of the surface irrigated sites in the 2020 season (Figure 10). Maintenance and gopher strikes were among the major challenges at sites E and F which impacted irrigation management and as a result influenced the total hay yields per season. This has been commonly observed in production alfalfa fields utilizing SDI [36]. Eventually, the farmer reseeded these two fields and converted them to alfalfa seeds from alfalfa hay in the 2021 season to reduce the number of irrigation events per season and to mitigate the negative impacts of drip tape damages and leaks from rodents.

In this study, the greatest alfalfa crop coefficient values were obtained during the harvest cycles of April–June, the period that alfalfa yield productivity is considerably higher than the remainder of crop season. According to the seasonal trends of alfalfa crop coefficient values (Figure 13), the average crop coefficient values for the various harvest cycles are as follows: 0.82 for the March harvest cycle, 0.95 for the harvest cycles of April–May, 0.89 for the harvest cycles of June–July, 0.84 for the harvest cycles August–September, and 0.75 for the fall harvest cycles. Thus, more efficient irrigation systems could incorporate different Kc values for the different times of year to reflect differences in water demand.

These crop coefficient values reflect local desert environments in terms of climate, soil, and water and crop management and may assist farmers in more accurate and effective irrigation scheduling. It is particularly important that farmers ensure that fields have a full profile of water at the beginning of the season and irrigate fields fully by early- to mid-July following the proposed crop coefficient values for the individual harvest cycles. Early season irrigation to fill the crop root zone will produce healthy roots in the lower root zone to take full advantage of the soils water holding capacity and the plants long root system. This could be one reason why site C had a lower seasonal crop water use (1433 mm) and DM yield (23.5 Mg ha⁻¹) in the 2019 season (Figure 4 and Figure 8). Alfalfa DM yield at this site was nearly 9% lower than the average DM yield achieved at all experimental sites during the same crop season.

4.2. Irrigation Strategies to Reduce Water Use

Deficit irrigation strategies in alfalfa are primarily feasible due to the seasonal yield pattern (Figure 7 and Figure 14). The cumulative percentage of alfalfa yields at the sites A–D over three growing seasons suggested that approximately 73–74% of desert alfalfa production occurred by mid-July (Figure 14). The fact that substantial yields are achieved by mid-summer enables decisions to be made to either water fully, or conserve water for the remainder of the season. Deficit irrigation strategies could affect production over three harvest cycles of late July through late September–mid October, the timeframe that desert alfalfa typically produces about 21–22% of seasonal yields. This provides some guidance to estimate yield reductions should more severe deficit irrigation strategies be required.

A combination of triage (eliminating non-productive fields), and partial-season production is recommended when faced with curtailment of water supplies. In this low desert environment, the MDI strategies are likely to be the most viable approach of deficit irrigation of alfalfa fields. The findings of the deficit irrigation treatments suggested an average of 4–6% seasonal yield reductions resulting in up to 40% less water applied (0.124 (ha·m) ha⁻¹) than standard farmer practice (Figure 11 and Figure 12).

Severe deficit irrigation strategy with no irrigation events over a three-month summer period of July–September, may cause greater yield loss. A recent three-year study at the University of California Desert Research and Extension Center (alfalfa trials under SDI) suggested that cutting 75% crop water demands during the summer (a 90 day period) may reduce an average of 13% yields and about 9% plant stand densities [37]. Yield and plant stand losses, soil water depletion, and salt build-up due to the summer water deficit were meaningful at this research trial. Salinity build-up is less likely to occur if the field was a surface irrigated field.

Considering the seasonal yield pattern (Figure 14), a severe water deficit (likely 0.234–0.246 (ha·m) ha⁻¹ water cuts during the summer period) may cause 19–21% seasonal yield losses (nearly 5.1 Mg ha⁻¹) in the desert region. This prediction is in a good agreement with the results of a one-year study conducted in the Palo Verde Valley in 2010 [38]. It is important to state, however, that a portion of the water saved may need to be applied in the fall-winter to refill the soil profile and maintain salts accumulated in the soil. The amount of this excessive water depends on soil type and condition, and contribution to the crop water needs from the water table (as capillary movement of water upward from the water table). This amount could be nearly 25% (or even more) of the total water cuts. Although policy is not always clear on this, salinity management with good drainage and excess water applications in desert conditions is a key beneficial use to maintain soil health and crop productivity in desert environments.

The developed alfalfa crop water use-production function across the experimental fields (Figure 15) indicates that it takes about 59 mm of ETa to make 1 Mg ha⁻¹ of alfalfa dry matter (range 1433 ≤ seasonal ETa ≤ 1595). An earlier study found that it takes about 53 mm of ETc (crop evapotranspiration under standard conditions) to make 1 Mg ha⁻¹ of alfalfa DM yield in the San Joaquin Valley, California [39].

An average seasonal alfalfa dry matter yield of 25.7 Mg ha⁻¹ was determined using the alfalfa production function and the average seasonal crop water use (1510 mm) across the experimental sites (Figure 15). Assuming 5.1 Mg ha⁻¹ (20% of the seasonal yield) as alfalfa DM yield losses caused by a severe summer deficit irrigation (0.24 (ha·m) ha⁻¹ water savings over the summer period), an average seasonal crop water consumption of 1390 mm was predicted using the alfalfa water-production function. In other words, nearly 120 mm reduction in crop water consumption should be expected from implementing a severe summer deficit irrigation practice in desert alfalfa. This is about 20% of the total irrigation water cuts from the severe deficit irrigation strategy. It is important to note that more severe summer cutoffs in previous research have resulted in stand losses on these heavy clay-loam soils with excessive heat in low desert environments [11].

Alfalfa fields with a predominant soil texture of sandy-to-sandy loam are not likely to be appropriate fields for severe water deficit strategies. These fields often can routinely exhibit moderate water stress around harvests, even under full irrigation practices, particularly over the summer season (Figure 9b).

The optimal deficit irrigation strategy depends on water availability, water conservation incentives programs, alfalfa hay price, and individual farming operations. Both severe and moderate summer deficit irrigation strategies are likely to be feasible and effective water conservation tools in the region. It enables growers to choose the optimal practice considering their farming operations in line with incentives, hay prices, and other factors. For instance, a farming operation that has already adapted multiple water conservation technologies or has more efficient on-farm water management systems might decide to integrate those practices with the MDI strategies. A farmer could adapt severe summer deficit irrigation for the fields that are not equipped with any other irrigation technologies and/or a higher water table is available.

5. Conclusions

The Colorado River Basin is facing increasing uncertainty concerning water supplies and these limits are likely to continue to challenge farmers and policy makers in the future. Water management strategies to address this problem include improved irrigation technology, careful monitoring, but also deficit irrigation strategies to conserve substantial amount of water while sustaining agriculture. This field-scale analysis of alfalfa crop water use and crop coefficients, seasonal yield pattern, crop water use-production function, water productivity, and feasibility of moderate and severe summer deficit irrigation strategies provide tools for estimation of water conservation and yield impacts for desert regions of the southwestern United States. The crop coefficient values developed in this study may assist alfalfa farmers in more accurate and effective irrigation scheduling. As an irrigation practice, it is important to ensure that alfalfa fields have a full profile of water at the beginning of the season and irrigate fields fully by early- to mid-July following the proposed harvest cycle crop coefficient values. Deficit irrigation practices should only be implemented late in the season when ET levels are high, and production is lower. The optimal deficit irrigation strategy in the desert region depends on water availability, water conservation incentives programs, alfalfa hay price, and individual farming operations. Both severe and moderate summer deficit irrigation strategies are likely to be feasible and effective water conservation tools for alfalfa. The findings of the alfalfa trials suggested an average of 4–6% seasonal yield reductions resulting in up to 40% summer water savings compared with standard farmer practice over the summer period. More severe summer water deficit, likely 0.234–0.246 (ha·m) ha⁻¹ water cuts during the summer period, may cause up to 5.1 Mg ha⁻¹ yield losses in the region.