Multi-temporal variation in water consumption of summer maize as determined by the Water Transformation Dynamical Processes Experimental Device

As water scarcity is one of the most urgent food security issues in China, identifying and utilizing the most effective water use management methods for improving water productivity is important. The further increase in maize production is limited by several factors, among which water management is possibly the most important, especially in North China (Liu et al. 2002; Li et al. 2003). For maximizing crop water productivity, water use efficiency needs to be increased by minimizing non-beneficial water uses (Suyker & Verma 2009; Liu et al. 2010; Pour-Ali Baba et al. 2013). Therefore, having detailed information on water consumption during crop growth periods is essential for water resource planning, irrigation control, and agricultural production (Li et al. 2013). Actual crop evapotranspiration (ET) involves soil evaporation and plant transpiration. ET is a comprehensive index that represents the overall crop water consumption and is among the main factors affecting crop productivity (Liu et al. 2002, 2016; Li et al. 2013). However, ET is significantly affected by many factors, including climatic factors (rainfall, air temperature (AT), and relative humidity (RH)) (Yang et al. 2013), crop characteristics (crop variety and development stage) (Liu et al. 2002), soil properties (soil texture, soil management, and soil water) (He et al. 2013), and groundwater conditions (water table, salinity and toxicity levels in ground water) (Soppe & Ayars 2003). In addition, these factors have multi-scale variations from hourly, daily, and development stage to annual and perennial. Studying ET on a single time scale does not consider the comprehensive influences of these factors on different time scales, and thus overlooking various underlying factors is easy. Accurately monitoring the multi-temporal variation of ET and its influencing factors plays an increasingly important role in revealing the main driving factors affecting various others. Moreover, it may be useful in planning, designing, and operating irrigation systems and crop planning in the future.

Many studies have focused on the variation of ET on perennial (Jun et al. 2012; Yang et al. 2013), annual (López-Urrea et al. 2006; Li et al. 2013; Ngongondo et al. 2013; Ma et al. 2015; Koedyk & Kingston 2016), seasonal (Tyagi et al. 2000; Kashyap & Panda 2001; Kang et al. 2003; Rana et al. 2005; Benli et al. 2006; Zhang et al. 2007; Suyker & Verma 2009; Liu et al. 2010, 2011, 2016; Beeson 2011; Hao et al. 2015; Li et al. 2017), and daily scales (Pour-Ali Baba et al. 2013). Using different methods, these studies established that the variations of climatic factors have complicated effects on ET change in different regions of the world. Nevertheless, most of these studies on the perennial or annual scale variation of ET overlooked the influence of crop characteristics and soil properties. In addition, some of these studies on the seasonal or daily scale variation of ET did not consider soil factors. Several studies that focused on the variation of ET on more than one time scale proved that considering the meteorological factors and crop characteristics in studying the multi-scale variation of ET is necessary (Liu et al. 2002; Li et al. 2003, 2008, 2013). However, previous studies paid little attention to the combined effects of climatic factors, groundwater, and soil structure on temporal variation of ET.

The multi-temporal variation of ET can be measured (directly or indirectly) or estimated. There are various methods for measuring crop ET, and they are classified into hydrological, micrometeorological, and plant physiology approaches (Rana & Katerji 2000; Wang & Dickinson 2012). Hydrological approaches are based on the principle of soil–water balance (Kashyap & Panda 2001; Li et al. 2003; Rana et al. 2005; Liu et al. 2011) and the weight of lysimeters (Tyagi et al. 2000; Liu et al. 2002; Kang et al. 2003; Benli et al. 2006; López-Urrea et al. 2006; Beeson 2011). Most micrometeorological approaches (Zhang et al. 2007; Li et al. 2008) are determined by energy balance, the Bowen ratio, aerodynamic estimates, and the eddy-covariance method. Most plant physiology approaches use the sap-flow method and chamber systems (Rana & Katerji 2000). Most of these studies used field experiments or field lysimeter experiments to focus on the effects of different crop varieties, irrigation practices, and climatic conditions on crop-water consumption at a specific location and time. Several studies adopted laboratory soil columns to investigate crop water use (Stauffer & Dracos 1986; Krzyszowska et al. 1994; Ma et al. 2010). However, several of these studies did not consider the multi-temporal variation of ET and some of them did not factor in the combined effects of soil structure, groundwater, and climatic factors on ET. Therefore, monitoring the multi-temporal variation of ET accurately is difficult with respect to the effects of soil structure, groundwater, and climatic factors. A method that estimates ET, instead of directly measuring it, was developed, and involves an analytical approach (Penman–Monteith model) (Jun et al. 2012; Li et al. 2013, 2017; Ngongondo et al. 2013; Yang et al. 2013; Ma et al. 2015) and an empirical approach (Ngongondo et al. 2013; Hao et al. 2015; Ma et al. 2015; Koedyk & Kingston 2016; Liu et al. 2016). The method regulates the environmental variables and boundary conditions of models by estimating ET. However, selecting reasonable parameters was difficult because of their complexity.

Soil texture is one of the most important soil factors affecting the multi-temporal variation of ET. The texture, thickness, and sequence arrangement of soil layers have significant effects on water movement in the vadose zone (Stauffer & Dracos 1986; Ma et al. 2010; He et al. 2013). He et al. (2013) concluded that water movement and crop growth varied among soil profiles because of the differences in the thickness and position of soil layers. The ET during the growing period of wheat–maize accounted for 72% of the total water input in silt loam–clay layered soil, thus causing relatively low water losses. Stauffer & Dracos (1986) also showed that water transport was strongly affected by the layers of packed sand. Therefore, comparing water transport in layered and homogeneous soils is necessary.

The Water Transformation Dynamical Processes Experimental Device (WTDPED) was developed by the Key Laboratory of Water Cycle and Related Land Surface Processes, Institute of Geographic Sciences and Natural Resources Research (IGSNRR), University of Chinese Academy of Sciences (UCAS), Beijing, China. WTDPED can be used to monitor the variation of ET and other water balance components on different temporal scales simultaneously and precisely under controlled climatic factors and groundwater conditions. Therefore, it is useful for studying the relationship of ET and its main influencing factors at different temporal scales under controlled climatic factors and groundwater level conditions with high precision. This study aimed to investigate the multi-temporal variation in ET throughout the growing season of summer maize in homogeneous and layered soils under combined climatic factors and groundwater conditions. It also explained the combined effects of the main environmental factors, groundwater, and soil texture on crop ET to provide valuable information for irrigation system management and new insights into the mechanisms of water transmission.

The soil moisture dynamic and transformation observation subsystem (Figure 2(b)) is composed of two large-scale weighing lysimeters with a surface area of 6 m² and a height of 3 m, which quantified the input and output of water. Each lysimeter was filled with either homogeneous or layered soil. The 5TE soil moisture, temperature, and electrical conductivity sensor (Decagon Devices, Pullman, WA, USA), the MPS-2 soil–water matrix potential sensor (Decagon Devices), and the DLS-II-type tensiometer (Institute of Geographical Sciences and Natural Resources Research, Beijing, China) were installed into each lysimeter at depths of 20, 43, 53, 63, 73, 83, 92.5, 110.5, 130.5, 150.5, 180.5, 210.5, 240.5, and 275.5 cm to monitor the moisture dynamics and transformation of the soil. The 5TE soil moisture, temperature, and electrical conductivity sensor monitored the volumetric water content, temperature, and bulk electrical conductivity of the soil; the MPS-2 soil–water matrix potential sensor monitored the water potential and temperature of the soil; and the DLS-II-type tensiometer device was used to extract the soil solution from soil layers through a vacuum. The maximal load of each weighing lysimeter was 81 t, and the resolution of the weighing lysimeter system was 180 g.

The groundwater control subsystem (Figure 2(b)) was composed of a water storage tank, water pipes, and a water supply or drainage device. The controlled electric circuit for the entire water supply and drainage process was made up of a fixed storage tube, a needle for measuring the liquid level, and the corresponding controlled software. Related control information was transferred to a computer monitoring system to check the data at any time. The groundwater level was controlled with a precision of 0.1 cm.

The average values of 10-year (2000–2010) meteorological data (AT, RH, CO₂ concentration, photosynthetically active radiation, etc.) collected in Beijing were used to simulate the whole growing period of summer maize. AT was modified every 3 h based on the average values at 01:00, 04:00, 07:00, 10:00, 13:00, 16:00, 19:00, and 22:00 to fully reflect the normal diurnal variation. RH was set as the average value of daily RH over the 10-year period. We measured the photosynthetically active radiation at different heights in the experimental facility, and the lamps were arranged according to the average value of photosynthetically active radiation over the 10-year period. Lights were turned on from 06:00 to 18:00 with three groups on. CO₂ concentration was set to 376 ppm, which was the average value over the 10-year period. The groundwater table depth was set to 2.5 m. The measurement frequency of meteorological data, soil volumetric water content, soil temperature, and soil bulk electrical conductivity was set to 1 h.

Experimental soil samples were collected from a common farmland in Huangcun Town, Daxing District, Beijing, China. Soil type was fluvo-aquic, which is the representative soil type in Beijing. Soil samples were backfilled with homogeneous soil and layered soils. The physical properties of soil in the 0–300 cm layer are listed in Table 1.

Maize was planted on May 17, 2014 and harvested on September 19, 2014. The whole growing period of summer maize (126 d) was divided into six stages: planting time to third leaf stage (PT–V3), third leaf to sixth leaf stage (V3–V6), sixth leaf to 12th leaf stage (V6–V12), 12th leaf to tasseling-silking stage (V12–VT), tasseling-silking stage to milk stage (VT–R3), and milk stage to physiological maturity stage (R3–R6).

One-way analysis of variance (ANOVA) in conjunction with the least significant difference method was used to compare the daily ET between homogeneous and layered soils at p < 0.05. Correlations between ET and AT or RH were tested using Spearman's rank correlation at p < 0.01 and p< 0.05 (two-tailed). All analyses were conducted using SPSS 16.0 (IBM Corp., Chicago, Illinois, USA).

The total ET during the whole growing period was 321.16 mm in homogeneous soil and 344.09 mm in layered soil. These results were similar to those in previous studies. For example, Liu et al. (2010) reported that the total ET of maize was 319–501 mm in the Loess Plateau, China, and Yang et al. (2013) reported the total ET of maize was 335.6 mm in the 1960s and 311.4 mm in the 2000s. Kang et al. (2003) found that the 10-year average seasonal ET was 424 mm in Northwest China.

The ET is significantly affected by crop characteristics (crop type, genotype, growth stage, and plant density), soil parameters (salinity, pH, fertility, and management), and environmental conditions (AT, RH, precipitation, solar radiation, sunshine hours, and wind speed) (Jun et al. 2012; Li et al. 2013; Yang et al. 2013). Many studies have investigated the relationship between the ET of summer maize and environmental factors, and suggested that AT, solar radiation, sunshine hours, and wind speed were positively correlated with ET, and RH and precipitation were negatively correlated with ET (Li et al. 2013; Yang et al. 2013).

To elucidate the effect of environmental factors on ET during the entire growing period of summer maize, the correlation coefficients between ET and AT or RH were calculated in homogeneous and layered soils (Table 3). ET was positively correlated with AT and negatively with RH. These results are consistent with previous studies (Li et al. 2013; Yang et al. 2013). The highest correlation coefficient between ET and AT was 0.70 (p < 0.01) in homogeneous and layered soils at PT–V3. At the beginning of the crop growth period, soil evaporation accounted for the larger part of ET, and it was the reason for the highest correlation coefficient between ET and AT occurring at PT–V3 in our study. The transpiration then increased, accompanying the growth of the leaf area (Liu et al. 2002; Kang et al. 2003), during which the minimum correlation coefficient occurred between ET and AT. Therefore, the lowest correlation coefficient observed at V12–VT was 0.53 (p < 0.01) and 0.55 (p < 0.01), respectively. RH had a negative effect on soil evaporation and transpiration. The highest correlation coefficient between ET and RH was −0.50 (p < 0.01) in homogeneous soil and −0.51 (p < 0.01) in layered soil at V12–VT, where ET maintained a relatively high value while RH decreased (Figures 3 and 4). The lowest correlation coefficient observed at V6–V12 was at −0.11 (p < 0.05) and −0.14 (p < 0.01), respectively. The correlation coefficient between ET and RH resulted from the integrated effects of AT, leaf area, and their interactions in this experiment.

Daily recharge was positively correlated to daily ET both in homogeneous and layered soils along the linear fitting lines (Figure 8(b)). The variability between daily recharge and ET was about 45% and 16% by the linear regression in homogeneous and layered soil, respectively. Five-day recharge was also positively correlated to 5-day ET in both homogeneous and layered soils, and the variability between 5-day recharge and ET was interpreted as about 46% and 29% by the linear regression in homogeneous and layered soil, respectively (Figure 8(d)). During the whole growth stages, recharge was significantly correlated to ET, and the variability between recharge and ET was interpreted as about 80% and 98% along the linear regression in homogeneous and layered soil, respectively (Figure 8(f)).

The contribution of daily recharge to daily ET varied from 6.14% to 10.14% with an average of 7.83% in homogeneous soil, while it ranged from 5.88% to 10.00% with an average of 7.63% in layered soil (Figures 5 and 8(a)). The average percentage of 5-day recharge was 7.45% in homogeneous soil and 7.09% in layered soil (Figures 6 and 8(c)). The contribution of groundwater recharge to ET at V6–V12 was 4.39% in homogeneous soil and 5.76% in layered soil, that at V12–VT was 8.67% and 6.64%, respectively, that at VT–R3, was 7.58% and 7.68%, respectively, and it was 6.64% and 6.89%, respectively, at R3–R6 (Figures 7 and 8(e)). Soppe & Ayars (2003) pointed out that the contribution of daily groundwater to daily crop water use was 40%. The reasons for the difference in our study may be the different levels of capillary upflow from the water table due to the different soil texture and crop root system or depletion levels of soil water (Soppe & Ayars 2003; Yang et al. 2007; Luo & Sophocleous 2010).

The average daily ET of summer maize in our study during the entire growing period was 2.57 mm in homogeneous soil and 2.75 mm in layered soil. During the entire growing period, the ET peaked daily at 14:00–15:00 for both homogeneous and layered soils. These results are consistent with previous studies. Li et al. (2008) found that the average daily ET value was 2.96 mm in maize and the corresponding peak time was 13:00, 14:00, 14:00, 14:00, and 13:00 at the seedling, shooting, heading, filling, and maturity stages, respectively, in a semi-arid region of North China under field conditions.

ANOVA results showed that the difference in daily ET between homogeneous and layered soils was significant (Table 4). The daily ET was significantly (p < 0.05) lower in homogeneous soil than in layered soil, and the differences in initial soil moisture content was insignificant. Therefore, the differences in the daily ET of summer maize were mainly due to differences in leaf transpiration area, thus suggesting that the latter was greater in layered soil. Therefore, maize growth in the silt sandy loam–loam layered soil was better than in the silt sandy–loam homogeneous soil because layered soil has higher water and nutrient-holding capacities (He et al. 2013).

A better understanding of the multi-temporal variation in ET during different crop growth stages is important for determining a reasonable irrigation schedule. In this study, we used WTDPED to simulate and monitor the growth and water consumption of summer maize in Beijing based on the average meteorological values taken over a 10-year period. Our objective was to illustrate the multi-temporal variation in ET and the effect of environmental and groundwater conditions on it both in homogeneous and layered soils. The main findings are as follows:

• (1) During the whole growing period, the ET per hour peaked at 14:00–15:00 on a daily basis and reached a maximum value of 0.32 mm in homogeneous soil and 0.33 mm in layered soil at V12–VT.

• (2) The variation in daily ET followed a bell curve during the entire growing period. The average value was 2.57 mm in homogeneous soil and 2.75 mm in layered soil, thus reaching the maximum value of 3.26 mm at 75 days post-planting and 3.43 mm at 61 days post-planting, respectively.

• (3) The variation in 5-day ET also followed a bell curve, and it reached the maximum value of 14.98 mm in homogeneous soil and 16.26 mm in layered soil at V12–R3.

• (4) The ET at VT–R6 accounted for approximately 50% of the total ET during the entire growing period. The ET reached the maximum value of 80.86 mm in homogeneous soil and 87.73 mm in layered soil at VT–R3. The total ET was 321.16 mm and 344.09 mm in homogeneous and layered soils, respectively.

• (5) During the entire growing period of summer maize, ET was positively correlated with AT and negatively with RH. The daily ET was significantly (p < 0.05) lower in homogeneous soil than in layered soil because of the larger maize leaf transpiration area in the latter.

• (6) Daily groundwater recharge showed a sharp increase at about 40 days post-planting, then was maintained at a relatively high level at about 40–110 days post-planting, and finally declined gradually until the end of the season. The average value was 0.22 mm. Recharge was positively correlated to ET both in homogeneous and layered soils. The contribution was up to 10.07% for daily recharge and 7.27% for 5-day recharge.

In this study, WTDPED was used to monitor the multi-temporal variation in water consumption considering the combined effect of climatic factors and groundwater under a controlled standard year, and the results are consistent with several previous studies conducted under field conditions. Some differences between the studies could be attributed to soil conditions, environmental conditions, and growing periods. Further studies are required to investigate the differences in water consumption of summer maize under different controlled environmental conditions (AT, RH, and CO₂ concentrations) and groundwater level conditions in homogeneous and layered soils. Overall, studies on the effect of environmental factors and soil structure on multi-temporal ET are important for region-sustainable water planning during crop growing seasons.

This study was supported by the Key Research Program of the Chinese Academy of Sciences (Grant number: KJZD-EW-TZ-G10) and the National Natural Science Foundation of China (Grant number: 41101029).