Introduction
Accurate and optimal water supply to cereal crops is critical in growing and producing maximum yields. Excessive soil-water can cause nutrient and oxygen deficiency (Cannel et al., 1980), limit the germination of seeds (Fausey et al., 1985), induce the dying of young plants (Lee et al., 2010), reduce root growth/plant height (Lee et al., 2010), reduce yields (Belford et al., 1985), and increase
susceptibility to disease (Belford et al., 1992). These effects are caused by the depletion of oxygen or oxygen diffusion in soils. In addition, microbial growth increases, which can cause the formation of sulfides and butyric acid that are toxic to plants, and the potential for root diseases is increased (Ashraf and Rehman, 1999). These negative effects by excessive soil-water result in a reduced yield reliability and suboptimal production.
In Korea, many paddy fields have generally high groundwater levels or greater soil-water contents compared to general uplands because rice requires flooded conditions during cultivation. The area of paddy fields in Korea is 934,000 ha, which is more than 55% of total arable lands (NSO, 2015). Most of these paddy field soils have low infiltration rates and poor drainage properties (Jung et al., 2011). Recently, the Korean government decided to support cereal crop cultivation in paddy fields to reduce overproduction of rice. Crops cultivated in paddy fields have shown physiological stress caused by excessive soil-water or high groundwater level (Jo et al., 1996; Ji et al., 2009). There is a need to overcome excessive soil-water problems for cereal crops. Therefore, it is necessary to fully understand crop responses to excessive soil-water conditions in paddy field soils.
In case of excessive soil-water conditions, Hiler (1969) introduced the stress day index (SDI) to quantify the cumulative stresses imposed on crops by water. Researchers studied yield-SDI relationships in several crops: corn (Hiler and Clark, 1971; Ahmad and Kanwar, 1989; Mallikarjunaswamy et al., 1999), soybean (Purwanto et al., 1993), and cotton/wheat (Kandil et al., 2001). This yield-SDI relationship is based on crop susceptibility (CS) factors (Lewis et al., 1974; Howell and Hiler, 1975; Hardjoamidjojo et al., 1982). CS factors describe the plant’s susceptibility (or response) to environmental stresses and depend upon the species and the stage of development of a given crop. CS factors have been investigated for various physiological growth stages of corn and soybean under excessive soil-water or flooded conditions in various studies (Evans and Skaggs, 1984; Mukhtar et al., 1990; Evans et al., 1990; Purwanto et al., 1993). Evans et al. (1990) proposed the concept of NCS (Normalized Crop Susceptibility) factors to eliminate the effects of environmental elements other than flooding, such as, genotype, soil type, fertility, and temperature. NCS values were applied to yield-SDI relationships instead of CS factors and determined the reduction degree of corn and soybean yields and effects on crop growth under excessive soil water stress (Evans et al., 1990). CS and SDI concepts have been applied to soybean cultivated in paddy fields in Korea (Jung et al., 2011), but there has been no attempt to use the yield-SDI model to predict yields under excessive soil-water conditions.
The objectives of this study were to evaluate excessive soil-water stress to cereal crops by using CS factors at various growth stages and evaluate the use of the yield-SDI model to predict cereal crop yield from paddy fields with various soil-water contents or groundwater table levels.
Materials and Methods
Yield-SDI (Stress-Day Index) model
The SDI concept is determined from the stress day (SD) factor and the CS factor. The SD factor is a measure of the intensity and duration of stress. The CS factor is a measure of the crop susceptibility to a unit of stress and is a function of the crop species and its stage of development. Hiler (1969) defined the SDI as
(1)
where n is the number of growth stages and SD and CS are the stress day and crop susceptibility factors, respectively, for period i. The SD values can be replaced by SEW30 values. SEW30 (cm-day) represents the sum of groundwater level differences when it exceeds a depth of 30 cm, because Sieben (1964) considered the depth of 30 cm to be the critical excessive groundwater level for crops.
Sieben (1964) proposed crop yields to changes of groundwater table depths as
(2)
where xi is the groundwater table depth below the soil surface on day i, and n is the number of days in the period being considered. The CSi is defined as the yield reduction ratio per unit of SDi at a given excessive soil water stress in i –the growth stage. Hiler (1969) expressed CSi as
(3)
where Xi is the yield from a treatment subjected to a unit water stress during i –the growth stage, and X is yield when a crop is kept under no water stress throughout the season. Later, Evans and Skaggs (1984) suggested the concept of NCS factor by normalizing CS factors to reduce the sensitivity of the CS factors to stress duration.
(4)
where n is the number of growth stages. The relationship between crop yield and SDI can be determined based on CS factors using regression analysis. The generalized yield-SDI relationship determined from simple linear regression by
(5)
where Yj is the absolute yield (kg·ha-1) observed in year j, Yp is the potential or base maximum yield that would occur in the absence of any soil-water related stress, a is the yield reduction per unit of SDI (slope of regression line), and SDIj is computed from eq.(1) using CS or NCS values. The relative yield (RY) was calculated by
(6)
where RYj is the relative yield, which is expressed as a percent of potential yield and b is the RY reduction per unit of SDI. Experimental site
The pot and field experiments were conducted to investigate responses of five crops to excessive soil-water conditions at National Institute of Crop Science, Crop Production Technology Research Division in Miryang, Gyeongnam, Korea. Foxtail millet (Setaria italica L.), proso millet (Panicum miliaceum L.), adzuki bean (Vigna angularis L.), and sorghum (Sorghum bicolor L.) were tested in 2013, and sesame (Sesamum indicum L.) was tested in 2015.
1) CS factor experiments
The CS experiments were conducted to investigate the response of excessive soil-water at five growth stages; tillering, booting, flowering, milk development, and ripening (Zadoks growth scale, Zadoks et al., 1974). Each crop was planted in a wagner pot (1·5000-1) filled with silt loam soils with chemical fertilizer as a basal manure. Each crop was planted in a pot with five replicates for each growth stage. A total of 30 pots for one crop were prepared including five pots with no stress treatments. All pots were managed with adequate watering, and growth stages were observed. Five pots were placed in bigger pots (1·2000-1) filled with water when the plants achieved each growth stage. Excessive soil-water treatment was applied by keeping the water level in these pots at a 10 cm depth for 10 days (Fig. 1). After 10 days, pots with crops were pulled out from water and the crops grew until harvest. The growth characteristics and yield components were measured from every pot.
2) Field experiments
The field experiments were performed in paddy fields in Miryang, Gyeonam. The soil of the experimental field was classified as Gagog series (fine silty, mixed, nonacid, mesic family of Aeric Endoaquepts). The paddy fields were divided into two plots by drainage class; poorly and somewhat poorly drained (Fig. 2). The size of each plot was 30 m by 50 m. The characteristics of the soils from the two plots are shown in Table 1 . Two different drainage systems were applied to alleviate effects of excessive soil-water in a plot: open ditch and pipe drainage system (Fig. 2). The open ditch was a 30 cm wide and 30 cm deep trench along one side of the plot and the pipe drainage was buried 50 cm deep along a side of the plot. A total of 4 subplots were placed in the site.
In 2013, sorghum, foxtail millet, proso millet, and adzuki bean were cultivated in each subplot of the experimental site as shown in Fig. 2. Sesame was cultivated in 2015 in each subplot. In each subplot, soil-water contents were measured by soil moisture probe (Easy AG, Sentek Pty. Ltd., Stepney, Australia) at 20 cm depth from the soil surface. Levels of groundwater were measured by sensors with automatic water level recorders (Remote Data Systems, Inc, NC, USA) from 0 cm to 150 cm. All measurements were applied for all four subplots and measured every hour during the growing seasons.
Statistical analyses
All data from experiments were analyzed statistically by analysis of variance (ANOVA) and Duncan’s multiple range test at 90% using SPSS v20 (SPSS Inc., Chicago, IL, USA).
Results and Discussions
CS and NCS factors
The CS and NCS factors were determined from the yields of sorghum, foxtail millet, proso millet, adzuki bean, and sesame according to the controlled excessive soil-water condition (Table 2). The crops had the greatest decreases in yields when the excessive soil-water was applied at the tillering stage and the least yield decrease when the excessive soil-water was applied at the ripening stage. The CS and NCS factors from the five crops were calculated from Equations (3) and (4) and the results are shown in Table 2 & 3. The CS and NCS factors had similar trends across growth stages for all the crops, except sesame. The yields decreased statistically less than the yields in no water stress treatments (p < 0.1) where the excessive soil-water was applied at the later growth stage. The yields and CS factors of sesame from the milk development and ripening stages showed opposite results to other crops’. However, there was no statistical difference in the values between the milk development and ripening stages (p > 0.1).
Table 3. CS and NCS factors of five crops with excessive water stress at five growth stages and no water stress.
|
| Different letters within the same column are statistically different at 90%. |
Evans et al. (1986) discussed that NCS factors are more insensitive to crop variety or to environmental factors. NCS factors reduced variations of the CS factor values across growth stages and crops. Therefore, the NCS factors were applied instead of CS factors in this study. The SDI values for five crops were calculated by Equation (1) (Table 4). The RY values were calculated using yields with no water stress and yields at each growth stage. Previous studies calculated the yield-SDI model from field data, but model accuracy was lower than 80% (Ahmad and Kanwar, 1989; Evans et al., 1991; Purwanto et al., 1993; Kandil et al., 2001). In this study, CS factors and yield-SDI models were calculated from the indoor and fully controlled experiment (pot experiments) to improve the precision of model predictions. All data from CS experiments were applied to the yield-SDI models using Equation (5) and. (6). The fit and results of model calculations are shown in Fig. 3. The r2 values of the models were greater than 80% except the one for adzuki bean. This result showed better fit results than the previous studies that calculated the yield-SDI models from actual field data.
Table 4. Stress-day indexes (SDI) of five crops with excessive water stress at five growth stages and no water stress.
|
Shaw (1974) and Sudar et al. (1979) proposed zero threshold stress levels which means no yield from crops under drought condition. Evans et al. (1991) applied this concept to yield-SDI model under excessive soil-water condition and set the SDI value at yield = 0 (SDI0) as a threshold of excessive water stress in yields. Evans et al. (1991) suggested the SDI0 value for corns was 141 cm-day and SDI0 value of soybeans was 154 cm-day. If SDI values for corn were greater than 141 cm-day, yield-SDI model were set as y = 0. In this study, SDI0 values were calculated from the yield-SDI model (Fig. 3).
The sorghum SDI0 values was 112 cm-day, proso millet was 67 cm-day, foxtail millet was 87 cm-day, adzuki bean was 78.6 cm-day, and sesame was 270 cm-day. This result meant that proso millet was more vulnerable to excessive soil-water compared to other cereal crops, followed by Foxtail millet. On the other hand, sesame was most tolerant to the excessive soil-water condition.
Evaluation of the yield-SDI model
Jung et al. (2014) reported the same results from their wet injury tests of crop cereals in paddy fields. They concluded that the degree of wet injury of five cereal crops were proso millet > foxtail millet > adzuki bean > sorghum. In this study, proso millet was most vulnerable to wet injury followed by adzuki bean.
The yield-SDI model from each crop was applied to actual yields from paddy fields for evaluation. Average values of groundwater levels and soil-water from the paddy field of all treatments are shown in Table 5. In 2013 and 2015, somewhat poorly drained paddy fields with a pipe drainage treatment had the lowest groundwater level and soil-water throughout the growing seasons. The poorly drained paddy field with an open drainage treatment had the highest groundwater level and soil-water (Fig. 4). These conditions led to the greatest stress of excessive soil-water to crops in the poorly drained paddy field with an open drainage treatment while they led to the smallest stress to crops in the somewhat poorly drained paddy field with a pipe treatment.
Table 5. Hydrological properties from paddy fields of two drainage classes and two drainage treatments.
|
| Different letters within the same column are statistically different at 90%. |
Fig. 4
Groundwater levels from paddy fields for all treatment in 2013 (a) and 2015 (b); poorly drained-open was a poorly drained field with open ditch drainage treatment, poorly drained-pipe was a poorly drained field with pipe drainage treatment, somewhat poorly drained-open was a somewhat poorly drained field with open ditch drainage treatment, and somewhat poorly drained-pipe was a somewhat poorly drained field with pipe drainage treatment.
There was no significant difference across drainage treatments in crop yields from foxtail millet, proso millet, and adzuki bean (p > 0.1, Table 6). Sorghum and sesame showed a significant difference between poorly drained fields with open drainage and somewhat poorly drained fields with pipes (p < 0.1). The average yield of each crop from no water stress cultivation in uplands was 330 kg·10 a-1 for sorghum, 298 kg·10 a-1 for foxtail millet, 257 kg·10 a-1 for proso millet, 104 kg·10 a-1 for adzuki bean, and 90 kg·10 a-1 for sesame (NSO, 2013). Foxtail millet from paddy fields resulted in a 20% to 50% reduction of yields and proso millet to a 40 % to 60% reduction compared to the average yields from uplands. However, sorghum and sesame did not show a reduction in yields compared to the average yields from uplands. As discussed above, foxtail millet and proso millet were more vulnerable to excessive soil-water conditions than sesame and sorghum. In Table 7, SDI values for each crop were calculated as per Equation (1). Proso millet and adzuki bean from all drainage classes and drainage treatments had greater SDI values than the SDI0 value. Foxtail millet had only one SDI value which was greater than SDI0. Therefore, the yield-SDI model of proso millet, foxtail millet, and adzuki bean expected no yield under this condition. Based on SDI values, proso millet, foxtail millet, and adzuki bean had the most significant damage from excessive soil-water. The yield results also confirmed that there was no difference from proso millet, foxtail millet, and adzuki bean (p > 0.1). They all had wet injury across all treatments. Sorghum and sesame yield data were applied to calculate the relative yield (actual RY) and the yield-SDI models calculated expected relative yield (expected RY).
Table 6. Average and standard deviation values of yields from five crops from all treatm.
|
| Different letters within the same column are statistically different at 90%. |
Sorghum and sesame showed a fit with r2 > 90% between actual RY and the expected RY (Fig. 5). Evans et al. (1991) and Purwanto et al. (1993) evaluated the yield-SDI models of corns and soybeans compared to actual field yields. Their best fitting results were found to range between r2 values of 60 and 80%. Both studies determined their models based on field data instead of controlled experiment data. Previous studies reported that sorghum is one of the most tolerant crops to excessive soil-water compared to other cereal crops (Jung et al., 2013). Wei et al. (2013) reported that sesame has a severe sensitivity to waterlogging condition. But in this study, sesame had a slope value of 0.0039 from yield-SDI model while sorghum had one of 0.0146. In other word, yield decrease of sesame was less than that of sorghum as SDI increased. This result meant that sesame was more tolerant to wet injury than sorghum.
Fig. 5
Relations between expected yields calculated using Eq. (6) and actual RY (Table 7) of sorghum (top) and sesame (bottom).
Conclusions
This study evaluated the effect of excessive soil-water or high groundwater level conditions on yields of five cereal crops. The yield-SDI model was adopted to evaluate the effect of excessive soil-water and predict yields under excessive soil-water or high groundwater conditions. The model was useful for certain crops such as sorghum and sesame, which had relatively greater tolerance to excessive soil-water. However, the model was unable to predict yields of proso millet, foxtail millet, and adzuki bean from paddy fields in excessive soil-water condition. These results concluded that these three crops may not be suitable for cultivation in paddy fields. On the other hand, sorghum and sesame were more suitable to paddy filed cultivation and their yields can be easily predicted under various soil-water conditions or groundwater levels. More studies will be performed to obtain additional field data for cereal crops with various groundwater tables to test yield-SDI models and improve the models for proso millet, foxtail millet, and adzuki bean. This study may contribute to more precise water management for cereal crops in paddy fields and provide a guideline to manage water conditions in paddy field soils to cultivate cereal crops.
