The impact of irrigation on yield of alfalfa and soil chemical properties of saline-sodic soils

Study site and plant material

The experiment site (N 45°34′36″–45°34′29″, E 123°2′24″–123°2′47″) was located 15 km east of Baicheng City, Jilin Province, China. The climate of the study site is a temperate continental monsoon climate with an annual precipitation of 370–400 mm and the annual evaporation more than 1,500 mm (Figure 1, data from the Meteorological Bureau of Baicheng City, Jilin Province).

The field experiment was conducted from September 2015 to July 2018 on a saline-alkali artificial grassland, which was established in the spring of 2009 by incorporating 20 cm of sandy soil into a degraded natural grassland. The Alfalfa cultivar of Gongnong No. 1 of M. sativa L. cv. was used as the plant material in this experiment and the alfalfa was harvested in July and October every year. The irrigation water was supplied from a local well (Depth > 100 m, pH = 7.5, electrical conductivity (EC) = 50 μS cm⁻¹).

The backgrounds of the physico-chemical properties of the soil (at depth of 0–100 cm) are shown in Table 1.

The average soil bulk density (0–100) was 1.59 g cm⁻³ and there was a wide range for the soil pH (8.52–10.45). The soil EC averaged ≥200 μS cm⁻¹. The soil at the depth of 0–20 cm and 20–100 cm were classified as sandy soil and loam, respectively, according to the international soil system.

Experimental design

The spatial distribution of the experiment plots followed a randomized block design (Fig. 2). Each plot was 5 × 4 m = 20 m².

The alfalfa on the western Songnen Plain was harvested at flower stage, thus, there were eight irrigation treatments based on the three growing stages of alfalfa with twelve replications (Table 2).

The irrigation water amount (I, mm) was calculated as following: (1) $$I=C\times {\text{ET}}_{\text{c}}$$ (2) $${\text{ET}}_{\text{c}}{\text{ = ET}}_{\text{0}}\text{ × }{K}_{\text{c}}\text{ = }P\text{ + }I\text{ – }D\text{ – }R\text{ ± Δ }W$$ (3) $${\text{ET}}_{\text{0}}=\frac{0.408\text{Δ}(Rn-G)+\text{γ}\frac{900}{T+273}{\text{μ}}_2(e_s-e_a)}{\text{Δ}+\text{γ}(1+0.34{\text{μ}}_2)}$$

ET_c, alfalfa evapotranspiration or crop water use, mm; C, design irrigation coefficient, 30%, 60%, 90%, and 0% in this study; K_c, crop coefficient, 0.83 (Van der Gulik, 2001); ET₀, reference ET for alfalfa (mm), calculated by the Penman–Monteith method (Allen et al., 1998); Rn, net radiation, MJ m⁻² d⁻¹; G, soil heat flux density, MJ m⁻² d⁻¹; T, mean daily air temperature at two m height, °C; μ₂, wind speed at two m height, m s⁻¹; e_s, saturator vapor pressure, kPa; e_a, the actual vapor pressure, kPa; Δ, slope of the vapor pressure curve, kPa °C⁻¹; γ, psychometric constant, kPa °C⁻¹.

Data collection

The relative chlorophyll content (SPAD) in each plant was determined using a SPAD-502 Chlorophyll Meter (Minolta Co. Ltd., Osska, Japan). The SPAD was measured every week following the beginning of the regreen stage of alfalfa. The SPAD of each alfalfa was measured five times and then the mean values of SPAD were calculated. Five alfalfa stems were selected randomly from each plot for measuring the shoot height (SH) using a ruler from the base of stem to the end.

Dry yields (DM) were measured by harvesting a 1.00 × 1.00 m quadrat from the center of each plot when the alfalfa reached 50-flower. Samples were air-dried in a repository. The DM was calculated by the 1st and the 2nd harvest of alfalfa DM.

Each subsample of alfalfa was separated into the leaf and stem fraction for measuring the ratio of stem to leaves (SLR). The SLR was calculated as the average of the 1st and 2nd harvest of alfalfa.

The soil water balance and ET_c were calculated using following equation (Carter & Sheaffer, 1983): (4) $${\text{ET}}_{\text{c}}\text{ = }P\text{ + }I\text{ – }D\text{ – }R\text{ ±Δ }W$$

ET_c, evapotranspiration, mm; P, precipitation, mm; I, irrigation amount, mm; D, downward drainage out of root zone, mm; R, surface runoff, mm; ΔW, change in soil water storage, mm.

Drainage (D) below the root zone was assumed to be zero since water applied with each irrigation was less than or equal to the water deficit in the one m soil profile of the fully irrigated treatment. Runoff (R) was considered zero because the experimental plots were surrounded by 0.15 m high and 20 m long plastic levees around its perimeter and the basin were meticulously prepared to be level.

The meteorological data from 2015 to 2018, including precipitation (P), were collected by the meteorological station (HOBO U30 Weather station; Onset Computer Corporation, Bourne, MA, USA). The actual amount of irrigation (I, mm) was recorded by the water meter set up in each plot. Neutron probe tubes were installed for measuring the ΔW weekly at the depth of 0–100 cm.

The irrigation should be slight and applied frequently at every growth stage to avoid water-logging the plot. WUE (kg m⁻³) was calculated according to Eq. (2) (Ojedaa et al., 2018): (5) $$\text{WUE\hspace{0.17em}= }\frac{Y}{{\text{ET}}_{\text{c}}}$$

Y, yield, g m⁻².

The water table significantly declined following the extreme drought from 2015 to 2017, and large areas of new paddy fields were developed nearby. Thus, hydraulic pressure was too low to support all the plots equally. This may have caused insignificant differences in the actual water consumption among the different irrigation treatments.

The alfalfa water sensitive indexes of different growth stages were calculated as (Jensen, 1968): (6) $$\frac{Y}{Y_m}={{\displaystyle \prod _{i=1}^n\left(\frac{{\text{ET}}_i}{{\text{ET}}_{mi}}\right)}}^{{\lambda }_i}$$

n, stage number of alfalfa; i, serial number of alfalfa growth stages; Y, actual yield, kg ha⁻¹; Y_m, maximum or potential grain yield with water not limiting production, kg ha⁻¹; ET_i, actual evapotranspiration in growth stage i, mm; ET_mi, maximum or potential evapotranspiration in growth stage i, mm; λ_i, water sensitive index (WSI) in growth stage i.

Soil samples were collected with a soil auger at a depth of 0–100 cm in September 2015 and October 2017 in order to measure for soil chemical properties, such as soil pH, soil EC, sodium absorption ratio (SAR), and total alkalization (TA).

The collected soil samples were air-dried, mixed, sieved through a two mm sieve, and then analyzed according to the methods described by the United States Department of Agriculture (USDA) (1954). Soil pH and EC were measured in a soil-water suspension with a 1:5 soil/water ratio, using a pH and EC meter (LeiCi Co. Ltd., Shanghai, China, and specifications of the glass electrodes were DJS-1C and PHSJ-3F, respectively). The K⁺ and Na⁺ concentrations from the soil-water extracts were analyzed by using flame photometry (Shanghai Precision Co. Ltd., Shanghai, China), whereas the concentrations of Ca²⁺ and Mg²⁺ were determined by atomic absorption spectrometry (Haiguang GGX605; Shanghai, China). The SAR and TA were calculated by the equations below (Wang et al., 2018).

The differences of soil pH, EC, SAR, and TA at 2015 and 2017 were calculated for representing the changes in the soil chemical properties.

Data analysis

For each soil sample, one-way ANOVA was performed on SLR, DM, SH, SPAD, and WUE using SPSS (P < 0.05) (Version 20.0; IBM, Armonk, NY, USA). Multiple linear regression was conducted to determine the relationship between the actual alfalfa yield and the predictable alfalfa yield.

Origin 8.0 (OriginLab Corporation, Northampton, MA, USA) was employed for the figuring.

Precipitation and evaporation during the growing season

The results of the monthly precipitation and evaporation during the alfalfa growth periods from 2015 to 2018 (during the April and October) were shown in Fig. 2.

The total precipitation and evaporation during the alfalfa growth season were 399.25 and 1,020.01 mm, respectively (Fig. 3). The highest monthly precipitation was 91.75 mm, which was observed in July, and the lowest monthly precipitation was 7.15 mm, which was observed in April. The highest monthly evaporation was 205.60 mm (April) which was double the monthly evaporation of October (101.67 mm).

The bio-characteristics of alfalfa under different treatments

The SH (cm), SPAD, DM (g m⁻²), and the SLR of different treatments were shown in the Table 3.

It was clear that the alfalfa DM, SLR, and SH were significantly (P < 0.05) influenced by irrigation in the western Songnen Plain. The DMs of the irrigated plots were 455.26–1,393.86% heavier than the DM of CK. The SH of irrigation treatments were 79.04–124.61% higher than the SH of CK. In addition, SLRs of irrigation treatments were 1.87% higher than the SLR of CK. Furthermore, irrigation treatment enhanced the chlorophyll concentration more than the CK; however, there were no significant differences (P > 0.05).

The DMs of alfalfa irrigated at only one growth stage was significantly (P < 0.05) lower than the DMs of alfalfa irrigated at two growth stages. The DM of I2 reached up to 980 g m⁻², which was the highest yield among the deficit irrigation treatments, which was followed by I1 (953.33 g m⁻²). The highest DM of alfalfa irrigated at one growth stage was observed at the I5 treatment (786.67 g m⁻²). Evidence observed in this study showed that the effect of irrigation treatments on SLR was not significant (P > 0.05). The SLRs of I1, I4, and I5 were decreased and the SLRs of I2, I3, I6, and I7 were increased by irrigation. The highest SLR was 2.62, which was observed at full irrigation treatment (I7) and then followed by SLR of I2 (2.24) and I2 (2.2.4). The SHs of eight irrigation treatments ranged from 55.11 to 123.78 cm. Results of SH showed that I7, which was irrigated at regreen, branch, and flower stages, performed the best on the increase of the alfalfa SH, followed by the I2 and I6 treatments. The values of SPAD increased from 53.33 to 64.87 among the eight irrigation treatments. Maximum SPAD (64.87) was observed from the I7 treatment, and the minimum SPAD (53.33) from CK.

The actual water consumption and water use efficiency

The total amount of water applied to alfalfa ranged from 85.00 to 755.00 mm in different irrigation treatments (Fig. 4). The actual water consumption of I1, I2, and I3 were 390.18, 410.82, and 386.59, respectively. Additionally, the actual water consumption of I4, I5, and I6 was 305.46, 314.48, and 288.08 mm, respectively. Furthermore, the I7 consumed 733.00 mm of water. The lowest water consumption was observed in the CK treatment (85.49 mm).

The WUE of seven treatments with irrigation ranged from 1.33 to 2.50 kg m⁻³ and were significantly higher than the WUE of CK (P < 0.05), however, no differences among the WUEs of irrigation treatments were found (P < 0.05). The lowest WUE (1.33 kg m⁻³) was observed at CK. There were no significant differences in WUE between the I1 (2.47 kg m⁻³), I2 (2.42 kg m⁻³), and I3 (2.28 kg m⁻³), which were irrigated at two growth stages of alfalfa. Compared to I4 and I6, which were irrigated at one growth stage of alfalfa, I5 has reached the highest WUE (2.50 kg m⁻³). This might imply that irrigation at the branch stage has an important influence on the WUE of alfalfa.

Irrigation schedule for western Songnen Plain

The water insensitive indexes of alfalfa at the regreen, branch, and flower stages were shown in Fig. 5 below.

The Fig. 5 showed that the WSI of the branch period was significantly higher (P < 0.05) than the WSIs of the regreen and flower periods. In addition, there was no significant (P < 0.05) difference between the WSI of regreen and flower.

The water production function of alfalfa in the western Songnen Plain was determined with λ₁, λ₂ and λ₃ these was shown in Fig. 5. Then the water production function of alfalfa in the Songnen Plain was displayed in Eq. (13).

The determination correlation coefficient (r = 0.71) between the simulated alfalfa DM and the amount of irrigation water was 0.71, which was in the critical coefficient between the simulated alfalfa DM and the amount of irrigation water (0.30–0.93). Thus, the Jensen model was reliable for the western Songnen Plain, Northeast China for irrigation scheduling. The simulated yield obtain from the Jensen model of the western Songnen Plain was lower than the actual yield (Fig. 6). However, the simulated alfalfa DM would be closer to the actual alfalfa DM when the amount of irrigation water reached at 150.00–350.00 mm.

The relationship between the alfalfa yield and the amount of irrigation on the artificial grassland in the western Songnen Plain can be simulated with the cubic curve (y = 125.188 + 3.449 x + 0.002 x² − 10⁻⁶ x³, 650 > x > 0, R² = 0.91, F = 65.61 > F_0.01 (3, 20) = 4.94) (Fig. 7). Alfalfa yield could rapid increase coupled with the amount of irrigation increase when less than 650 mm water was supplied.

Based on the data above, 236.50 mm of irrigation water was recommended at the branch stage of alfalfa for the eastern Songnen Plain, Northeast China.

Effects of irrigation on soil chemical properties

Results of changes in soil chemical properties after 2-years of irrigation were shown in the figures below.

Irrigation had significant effects on the chemical properties of the soil at different soil depths (0–100 cm, Figs. 8–11). For example, the soil EC, SAR, and TA of irrigation treatments showed greater decreasing effects than CK. During the growing season, the average soil EC of different irrigation treatments decreased 210.21–287.24 μS cm⁻¹, and the average soil EC of CK decreased 46.35 μS cm⁻¹. The results of soil SAR (average of 0–100 cm) revealed that irrigation reduced 8.10–9.00 (mmol_c/L)^1/2, and on the contrast, the average soil SAR of CK decreased only 0.93 (mmol_c/L)^1/2 from September 2015 to October 2017.

The greatest decrease of soil TA was observed with the I7 irrigation treatment (4.64 mmol_c L⁻¹), while the lowest decrease of soil TA was found in the I3 treatment (3.29 mmol_c L⁻¹), and the soil TA without irrigation decreased 2.88 mmol_c L⁻¹. The effects of irrigation on soil pH were complicated. For instance, the soil pH at a depth of 0–40 cm and 80–100 cm decreased at 0.003–0.80, and a 0.13–0.56 increase of soil pH at the depth 40–60 cm was found.

As shown in Fig. 8, I2, I3, I4, and I5 showed a greater decrease on soil pH (0–00 cm, ΔpH = 0.14–0.30) than the other treatments. The soil EC of I4 and I5 decreased 283.89 and 287.24 μS cm⁻¹, which were larger than other treatments (Fig. 9). The soil SAR decreases 8.96 and 9.00 (mmol_c/L)^1/2, which were observed with the I7 and I5 treatments (Fig. 10); the soil TA under the treatments of I5, I6, and I7 (ΔTA = 4.15−4.65 mmol_c L⁻¹) showed a greater decrease than the other treatments (Fig. 11).