Aust. J. Plant Physiol., 1978, 5, 169-77

Agronomic and Physiological Responses of Soybean and Sorghum Crops to Water Deficits 11." Crop Evaporation, Soil Water Depletion and Root Distribution

G. J. B ~ r c h , * ~ R. C. G. SmithA and W. K. as on* A Department of Agronomy and Soil Science, University of New England, Armidale, N.S.W. 2351.

Present address: Division of Land Use Research, CSIRO, P.O. Box 1666, Canberra City, A.C.T. 2601.

Abstract

The effects of soil water depletion on crop evaporation and root absorption of water were studied in soybean and sorghum crops. Sorghum did not deplete the maximum soil water store by more than 100 mm, whereas rainfed crops of soybeans, cvv. Ruse and Bragg, depleted the soil water store by 130 and 170 mm, respectively. This was sufficient to reduce soybean yields by 35% and hasten maturity in both cultivars when compared with irrigated crops. The post-flowering efficiency of water use by rainfed crops of soybeans was about one-third that of sorghum.

The root distribution of Ruse and its pattern of soil water extraction indicated that during bean- fill it was unable to exploit water from much below 80 cm depth, but this effect was offset by its reaching maturity before yield was severely affected by water stress. As Ruse approached maturity, its root densities decreased in soil layers below 10 cm depth, whereas Bragg, which matured 2 weeks later than Ruse, maintained a deep root system and continued to deplete water down to 120 cm. The contrast in root distribution between soybean cultivars also influenced the level of soil water depletion at which crop evaporation fell below the potential rate. Soil and root resistances to water absorption were used to interpret the effects of root density and soil water depletion on water uptake.

The regional implications of the results were examined using a water balance model to analyse historical rainfall records. It was concluded that similar soil moisture conditions could be expected about 1 year in 5, indicating that these results have a ready application for irrigation scheduling in this area.

Introduction

Crop growth in a semi-arid environment is influenced by the diurnal water status of plants (Hsiao and Acevedo 1974), which is in turn influenced by soil water status and evaporative demand (Kaufmann and Hall 1974; Ritchie 1974). However, little is known of what effect the root distribution of a crop has on water uptake and plant water status (Ritchie et al. 1972). A comparison of wheat varieties by Hurd (1974) indicated that plants with a more extensive root system could exploit a larger soil volume, thereby making more effective use of soil water and producing higher grain yields. Few studies have tested this effect in other crops or in the field.

This paper constitutes part I1 of a series of four on the responses of soybean and sorghum crops to water deficits in the Namoi Valley, New South Wales. Crop producers in this area often experience difficulty in achieving an efficient use of irriga- tion water because of the unpredictable incidence of summer rainfall. To prevent yield depression may require only intermittant irrigation, depending on the rate of crop water use and the depth of soil from which their roots can readily extract water.

* Part I, Aust. J. Plant Physiol., 1978, 5, 159-67.

G. J. Burch, R. C. G . Smith and W. K. Mason

Therefore, the interaction between soil water depletion, crop evaporation and root absorption of water needs to be investigated for irrigated crops in this region. Other pilpers in the series describe growth and yield (Constable and Hearn 1978, part I); plant water status and stornatal behaviour (Turner et al. 1978, part 111); and photo- synthesis and water use efficiency (Rawson et al. 1978, part IV).

Materials and Methods Field Experiment

Soybeans [Glycine max (L.) Merrill] cvv. Ruse and Bragg, and sorghum [Sorghum bicolor (L.) Moench] cv. TX 610 were sown in a field trial on 4 December 1975 at the Narrabri Agricultural Resesrch Station. P!ots were watered the following day to ensure uniform germination and thereafter five irrigation treatments were maintained (part I). The measurements reported in this paper were confined to the most frequently irrigated treatment (irrigated) and the unirrigated treatment (rainfed) for soybeans. Owing to heavy rainfall in January and February (part I), no irrigation treatment was imposed on sorghum.

Soil Water

Soil water was measured throughout, but more frequently from flowering to maturity, when intensive measurements of plant water status and photosynthesis (parts I11 and IV) were made. Soil water was measured gravimetrically to 30 cm and from 30 to 140 cm by a field-calibrated model 225 Wallingford neutron moisture meter. Extraction of soil water was not observed below 120 cm and all results will refer to the top 120 cm of soil. The water retention curve of the soil was determined by the filter paper method of Fawcett and Collis-George (1967). Available soil water was considered to be that held by a soil profile 48 h after irrigation and above a soil potential of - 1.5 MPa, at which tension the soil water content was 0.18 g g-'. The maximum store of available soil water for the 120-cm rooting depth equalled 210 mm. Soil water potentials were measured under rainfed soybeans cv. Ruse and sorghum on 11, 13 and 17 March 1976 at 5-cm intervals from 10 to 35 cm, using model PT-51 Wescor soil hygrometers calibrated by vapour phase equilibration over NaCl solutions, and read with a Wescor HR-33T dew point microvoltmeter. Evaporation from the crop was determined by the soil water balance method for periods when there was no observed runoff. Over the post-flowering period, a measure was made of the efficiency of water use in producing crop yield.

Roots

Root densities were measured for sorghum and rainfed soybean cv. Ruse on 18 February, during flowering, and for both rainfed soybean cultivars and irrigated cv. Bragg on 12 March, during bean-fill.

Roots were sampled between plants within a row. Ten cores were taken from each treatment using a steel tube of 4.5 cm diameter driven into the ground by an electric hammer and retrieved by a hand winch. Cores were sectioned into 10-cm lengths and soaked for 24 h in a detergent (Calgon) to disperse the soil. Roots were recovered by washing off loose soil over a 1-mm screen and then floating roots off the remaining soil using a water bath. Root samples were photographed and their lengths determined using the intercept technique of Newman (1966), by projecting the negative onto a screen etched with a standard grid (Marsh 1971).

The root-length data from each core were smoothed by regression: this was done by fitting a polynomial to root length as a function of depth and then using the fitted value. The standard errors of the estimated root lengths at each depth were equal to or less than 0.2 cm cm-3.

Water Uptake

The rates of water absorption per unit length of root were calculated from measured root densities and the changes in soil water content over the periods 3-10 March and 14-18 March for sorghum and soybeans respectively, periods of no rainfall. No account was taken of water loss from the soil surface or vertical water flux within the profile, both of which were likely to have been small (Ritchie et al. 1972).

Response of Soybean and Sorghum to Water Deficits. I1

Analysis of Rainfall Data

Meteorological data were recorded at the Myall Vale meteoroIogicaI station, 1 .5 km from the experimental site. Rainfall records since 1935 were available from the station, and these were analysed using the water balance approach of McAlpine (1970) to give a broader perspective to the 1975-76 season. For this analysis, a water balance model was developed to predict weekly changes in the soil water store. Evaporation was predicted from mean weekly pan evaporation (class A) scaled by two functions to account for the effects of leaf area index and soil water store. The model was developed to predict changes in the available soil water under rainfed soybeans, cv. Ruse, and it accounted for a significant proportion (P < 0.05) of the observed variation in Fig. 1. The model was then used to simulate the seasonal water status for 40 past crop seasons between 1935 and 1976, assuming similar rainfed crops had been grown, and with provision for a single irrigation following seeding. From this simulation, the frequency of occurrence of different levels of available soil water for each week of crop growth was calculated and the relative frequency of each week's soil water level being exceeded was also determined.

Fig. 1. Curves: Available soil water to 120 cm depth for sorghum (A), rainfed Ruse (a) and Bragg (B)

soybean, and for irrigated Ruse ( 0 )

soybean at various times from sowing. Rainfall is also shown (histogram). The arrows denote the times of irrigation.

February I March I April

Time from sowing (days)

Results

Soil Water

Above-average rainfall in December, January and February (part I) maintained adequate soil water for crop growth in all treatments until March. In March and April, between the end of flowering and crop maturity, soil water was rapidly depleted in rainfed soybean treatments, with dissimilar rates of depletion by the two cultivars after 8 March (Fig. 1). In this period, irrigation was necessary to avoid water stress on soybeans. The leaf area of sorghum was much less than that of soybeans (part I), and sorghum reached maturity earlier. Both of these factors contributed to lower water use by sorghum in comparison with soybeans. Under rainfed conditions, the soybean variety Ruse reached maturity earlier than Bragg and used less water (Fig. 1). At maturity, Ruse had depleted only 62% of the maximum available water store, whereas Bragg had depleted 85 %.

Comparison of Crop Evaporation in Irrigated and Rainfed Soybeans

The ratio of actual crop evaporation to pan evaporation was examined in irrigated and rainfed soybean treatments for periods when there was no observed runoff (Fig. 2). The ratios for irrigated Ruse and Bragg were similar; hence, only data for Ruse are presented. Sorghum has not been considered because its depletion of soil water was insufficient to depress evaporation. In soybeans, the ratio increased from about 0.3 after sowing to about 1.2 following complete canopy closure above an

G. J. Burch, R. C. G. Smith and W. K. Mason

LA1 of 2.7 (Fig. 2a). During complete canopy closure, the ratio decreased in rainfed treatments and this was associated with low levels of available soil water (Fig. 2b) whereas, in irrigated treatments, no decline occurred. There were insufficient data at the point of decline to indicate whether this decline occurred at different levels of available soil water for the two varieties.

lrn ated Ruse

I Ramfed

d Ramfed 'J Bragg

0.2

0 20 40 60 80 100 120 140

Dec. l Jan I Feb I Mar. I Apr.

Time from sowing (days) Available soil water (mm)

Fig. 2. Relationships between t%e ratio of actual evaporation (E,) to pan evaporation (E,,,), (a) as a function of time in irrigated Ruse (o), rainfed Ruse (e) and rainfed Bragg (M) soybean, and (b) as a function of the available soil water in Ruse ( e ) and Bragg (M) soybean. The arrows denote the times of irrigation.

Root densty (cm ~ m - ~ )

Fig. 3. Root densities on (a) 18 February 1976 for rainfed Ruse soybean (e) and sorghum (A), and on (b) 12 March 1976 for rainfed Bragg (M), rainfed Ruse (o), and irrigated Bragg (0) soybean.

Root Distribution

Sorghum had significantly higher (P < 0.05) root densities than had Ruse soybeans in soil layers above 50 cm depth on 18 February. Sorghum root densities steadily decreased down to 100 cm (Fig. 3a). In contrast, rainfed soybeans had a more uniform distribution of root over'depth. Except in the surface layer, root densities in rainfed Ruse decreased over the period from 18 February to 12 March (P < 0.05), indicating that it lost roots as it approached maturity (Fig. 3a). On 12 March, rainfed Bragg had twice the root density of Ruse at depths below 40 cm (Fig. 3b). Irrigation of Bragg soybean appeared to encourage root growth in soil layers above the 50 cm level (Fig. 3b), though this effect was only significant at a depth of 15 cm.

Response of Soybean and Sorghum to Water Deficits. I1

Water Extraction

During grain-fill, sorghum roots maintained similar rates of water extraction per centimetre of root at all depths (Table 1). Soybean rates were more variable but also showed no trend over depth (Table 1). Soil water contents measured on 3 March and 14 March (Fig. 4 ) indicate that Bragg extracted water down to 120 cm, whereas Ruse failed to exploit soil water much below 80 cm, probably because of declining root densities at these depths (Fig. 3a).

Table 1. Water uptake at various depths for rainfed and irrigated Bragg and for rainfed Ruse soybean and sorghum

The period of uptake was from 14 to 18 March for Ruse and Bragg and from 3 10 March far sorgkciii. Zeros indicate n= measwzb!e wzter uptake fmrr?

that soil layer

Depth Water uptake [ lob2 cm3 (cm root)-' day-'] (cm) Rainfed Rainfed Irrigated Sorghum

Ruse Bragg B r a g

Fig. 4. Soil water content at intervals of depth in the soil to 140 cm, for sorghum (A), for rainfed Ruse (e) and Bragg (m) soybean, and of a soil profile 48 h after irrigation for Ruse ( 0 ) on (a) 3 March and (b) 14 March 1976.

0.3 0 4 0.5 0.2 0.3 0.4

Soil water content (cm3

Measured soil water potentials under sorghum never approached -1.5 MPa whereas, under Ruse soybeans, the potential of the top 10 cm of soil approached - 1.5 MPa by 1 1 March and had decreased below - 1 -7 MPa by 13 March (Table 2). In rainfed crops, the post-flowering efficiency of water use by sorghum was about three times that of soybeans (Table 3).

Water Balance Analysis

The seasonal soiI water status predicted by the water balance model for the 1975-76 season was compared with the available soil water that was expected to be exceeded

G. J. Burch, R. C. G. Smith and W. K. Mason

in 20, 50 and 80% of the previous 40 seasons analysed (Fig. 5). During the period from seeding to first flower in soybean (c. 56 days from sowing), the 1975-76 season was in fact drier than in 50 % of the years examined, and the relative frequency of the

Table 2. Soil water potentials measured by dew point hygrometers at various depths in rainfed soybeans (cv. Ruse) and sorghum on three dates

Depth Soil water potential (MPa) (cm) 11 Mar. 76 13 Mar. 76 17 Mar. 76

Soybeans cv. Ruse 10 15 20 25 3 5

Sorghum

Table 3. Total water use, yield and efficiency of water use by irrigated or rainfed soybeans and sorghum during bean and grain fill - - -

Species and Treatment Total water use Yield Water use efficiency cultivar during bean or (kg ha - l) (kg ha-' mm-')

grain fill (mm)

Soybeans cv. Bragg Irrigated 301 2847

Rainfed 192 1910 cv. Ruse Irrigated 258 2699

Rainfed 138 1707 Sorghum Rainfed 132 5405

0 I I I I I I I I I 14 28 42 56 70 84 98 112 126

I January I February I March I

Time from sowing (days)

Fig. 5. Available soil water predicted by the water balance model for the 1975-76 season (0) and curves showing the soil water status exceeded:

In 50 % of all years; A In 20% of all years; and M In 80% of all years. These curves are based on a frequency analysis over a 40-year period.

mean soil water content in this period of 1975-76 being exceeded was 0.57. During flowering (c. 56-84 days from sowing), availability of soil water was well above average and the relative frequency of the mean soil water content being exceeded in

Response of Soybean and Sorghum to Water Deficits. I1

this period was less than 0.1. During the final period of crop growth in soybean, from the end of flowering to crop maturity, the relative frequency of the mean soil water content being exceeded was 0.33.

Discussion

Water Uptake and Root Distribution-a Conceptual Model

To interpret the relationships between crop evaporation, soil water content and root distribution presented in this paper, it is first necessary to understand the role of roots in maintaining water uptake by plants to meet the evaporative demand.

Water uptake has been described as a catenary process (Huber 1924) of the form:

where Q is the transpiration flux (cm3 day-'); $,, $, and $, are the water potentials in the bulk soil, at the root surface and in the leaves, respectively; and R, and R, are the resistances to water movement in the soil and in the plant. Early studies of water uptake (Philip 1957; Gardner 1964; Cowan 1965) placed most emphasis on the soil resistance to water absorption, presumably because of the large decrease in hydraulic conductivity (K) as soil dries. More recent evidence (Hansen 1974; Taylor and Klepper 1975; Reicosky and Ritchie 1976) now indicates that R, may be the dominant resistance to water uptake while soils remain wetter than c. - 1.0 MPa. The evidence also suggests that much of the resistance is located radially in the root, and therefore the total root resistance (R,,,,) can be envisaged as the summation in parallel of a radial root resistance per unit length of root (R,,,), giving a relationship of the form R,,,, = R,,,/L, where L is the length of root contained in a specified volume of soil. This relationship indicates the likely importance of root density in determining R,,,, and consequently R,. In addition, soil and root resistances can be considered to operate in series. This organization of resistances in a complex series- parallel arrangement has important implications in determining the absorption of water by root systems.

Water Extraction by Sorghum

Sorghum had a dense root system in the upper soil layers (Fig. 3a), which is comparable with previous reports on the sorghum root system (Teare et al. 1973; Mayaki et al. 1976). Because of the low resistance to water uptake associated with high root densities, sorghum extracted proportionately more water from these soil layers (Fig. 4). However, there was sufficient water remaining within the profile, particularly below 40 cm, to maintain unrestricted crop evaporation. This was a result of its early maturity, low leaf area index (part I) and low ratio of transpiration per unit area of leaf (part IV). The uniform rates of water absorption per unit length of root by sorghum (Table I), despite a wide variation in root density, suggest that the soil was never dry enough to cause a significant resistance to water uptake. This conclusion was also supported by the measurements of soil water potential (Table 2).

Water Extraction by Soybeans

There were differences between soybean varieties in the concentration of roots in the surface soil layers, with 58 % of the roots in Ruse and 35 % in Bragg found in the

G. J. Burch, R. C . G. Smith and W. K. Mason

top 30 cm of soil (Fig. 3). This resulted in the majority of water being extracted from the top 50 cm of soil. As the soil dried in the top 50 cm, soil potentials decreased (Table 2) and the soil resistance to water absorption increased until the cumulative resistance to water uptake by surface roots (R,,,,+ R,,,,) would have exceeded the resistance incurred by deeper roots. The location of maximum water uptake therefore shifted down the profile as the balance of resistances changed. However, this down- ward process was eventually limited by a lack of roots at depth, particularly in Ruse. This would have caused a combination of root resistance at depth and soil resistance in dry surface layers to reduce water uptake, with the result that the crop was unable to meet the evaporative demand. This could provide an explanation for what appeared to be an earlier or more rapid decline in the evaporation rate for Ruse compared with Bragg (Fig. 2), which was accompanied by other physiological signs of water stress (parts I11 and IV). The decrease in evaporation by rainfed Ruse commenced when the canopy was completely closed. Subsequently there was rapid loss of leaf, probably due to the combined influence of water stress and approaching maturity (part I).

General Implications

Rapid depletion of water in the top soil by species with dense surface rooting, such as sorghum, will increase the importance of deep roots for maintaining a supply of water to the plant. Nevertheless, under drought conditions the advantage of an extensive root system can only be realized if the plant has additional means of avoiding excessive water stress at critical stages of development. Species and varieties having such characteristics are able to exploit a high proportion of the available soil water without incurring severe plant water deficits and, requiring irrigation only during extended drought periods, could be used in semi-arid areas to achieve a more efficient use of irrigation water.

The greater efficiency of post-flowering water use in producing grain yield by sorghum in comparison with soybeans indicates its better adaptation to rainfed cropping practice in dry areas. In addition, following crop maturity, more soil water remained under sorghum than under soybeans, which could be important for growing following crops without irrigation. An analysis of the rainfall and soil water balance suggested that the type of season experienced in 1975-76 could be expected with a frequency of between 1 year in 4 and 1 year in 5. This indicates that, at least in this proportion of years, sorghum and soybeans could be grown successfully with minimal irrigation, the water saved being used, possibly, to grow a subsequent winter crop. This fact is well established with sorghum, as it is widely grown as a dryland crop, but it is often assumed that soybeans need frequent irrigation for adequate yield. Our results suggest that soybeans have the ability to exploit a substantial proportion of the available soil water, thereby allowing considerable flexibility in irrigation scheduling. Irrigation can minimize the soil resistance to water uptake, but the limitation imposed by root resistance can only be overcome by selecting varieties with a deep and extensive root system. Regional irrigation practice can probably be improved by a better definition of the extractable water available to a crop at various growth stages, which is dependent on the penetration and density of roots within the soil profile. By obtaining measurements, or good estimates, of root density and water content over depth it should be possible to determine the current availability of soil water and predict the water content a t which soil and root resistances will

Response of Soybean and Sorghum to Water Deficits. I1

restrict water uptake. This could provide a basis for making more rational irrigation decisions.

Acknowledgments

Calibration of the neutron moisture meter by Mr P. 0. Cull is gratefully acknowledged. We thank the staff of the Narrabri Agricultural Research Station for their help and cooperation.

References Constable, G. A,, and Hearn, A. B. (1978). Agronomic and physiological responses of soybean and

sorgh_nr?l crops to water deficits. I. Growth, deve!opment and yie!d. Am? -1 l n n t Phyriol 5 , 159-67.

Cowan, I. R. (1965). Transport of water in the soil-plant-atmosphere system. J. Appl. Ecol. 2, 221-39.

Fawcett, R. G., and Collis-George, N. (1967). A filter-paper method for determining the moisture characteristics of soil. Aust. J. Exp. Agvic. Anim. Husb. 7, 162-7.

Gardner, W. R. (1964). Relation of root distribut~on to water uptake and availability. Agvon. J. 56, 41-5.

Hansen, G. K. (1974). Resistance to water transport in soil and young wheat plants. Acta Agvic. Scand. 24, 37-48.

Hsiao, T. C., and Acevedo, E. (1974). Plant responses to water deficits, water-use efficiency, and drought resistance. Agvic. Meteovol. 14, 59-84.

Huber, B. (1924). Die Beurteilung des Wasserhaushaltes der Pflanze. Jahrb. Wiss. Bot. 64, 1-120. Hurd, E. A. (1974). Phenotype and drought tolerance in wheat. Agvic. Meteorol. 14, 39-55. Kaufmann, M. R., and Hall, A. E. (1974). Plant water balance-its relationship to atmospheric

and edaphic conditions. Agvic. Meteorol. 14, 85-98. McAlpine, J. R. (1970). Estimating pasture growth periods and droughts from simple water balance

models. Proc. 10th Int. Grassl. Congr., Surfers Paradise, 1970, pp. 484-7. Marsh, B. a'B. (1971). Measurement of length in random arrangements of lines. J. Appl. Ecol. 8,

265-7. Mayaki, W. C., Stone, L. R., and Teare, I. D. (1976). Irrigated and non-irrigated soybean, corn,

and grain sorghum root systems. Agvon. J. 68, 532-4. Newman, E. I. (1966). A method of estimating the total length of root in a sample. J. Appl. Ecol. 3,

139-45. Philip, J. R. (1957). The physical principles of water movement during the irrigation cycle. Proc.

3rd Int. Congr. Irrig. Drain. Vol. 8, pp. 125-8. Rawson, H. M., Turner, N. C., and Begg, J. E. (1978). Agronomic and physiological responses of

soybean and sorghum crops to water deficits. IV. Photosynthesis, transpiration and water use efficiency of leaves. Aust. J. Plant Physiol. 5, 195-209.

Reicosky, D. C., and Ritchie, J. T. (1976). Relative importance of soil resistance and plant resistance in root water absorption. Soil Sci. Soc. Am. J. 40, 293-7.

Ritchie, J. T. (1974). Atmospheric and soil water influence on the plant water balance. Agric. Meteovol. 14, 183-98.

Ritchie, J. T., Burnett, E., and Henderson, R. C. (1972). Dryland evaporative flux in subhumid climate. 111. Soil water influence. Agvon. J. 64, 168-73.

Taylor, H. M., and Klepper, B. (1975). Water uptake by cotton root systems: an examination of assumptions in the single root model. Soil Sci. 120, 57-67.

Teare, I. D., Kanemasu, E. T., Powers, W. L., and Jacobs, H. S. (1973). Water use efficiency and its relation to crop canopy area, stomata1 regulation and root distribution. Agvon. J. 65, 207-11.

Turner, N. C., Begg, J. E., Rawson, H. M., English, S. D., and Hearn, A. B. (1978). Agronomic and physiological responses of soybean and sorghum crops to water deficits. 111. Components of leaf water potential, leaf conductance, 14C0, photosynthesis, and adaptation to water deficits. Aust. J. Plant Physiol. 5, 179-94.

Manuscript received 5 May 1977, revised 15 November 1977