Published online 4 April 2007
Published in Agron J 99:665-672 (2007)
DOI: 10.2134/agronj2006.0092
© 2007 American Society of Agronomy
677 S. Segoe Rd., Madison, WI 53711 USA
Sorghum
Sorghum Management Practices Suited to Varying Irrigation Strategies
A Simulation Analysis
R. L. Baumhardta,*,
J. A. Tolka,
T. A. Howella and
W. D. Rosenthalb
a USDA-ARS, Conservation and Production Res. Lab., P.O. Drawer 10, Bushland, TX 79012-0010
b Texas Agric. Exp. Stn., Blackland Res. Ctr., 720 E. Blackland Rd., Temple, TX 76502
* Corresponding author (rlbaumhardt{at}cprl.ars.usda.gov)
Received for publication March 28, 2006.
 |
ABSTRACT
|
|---|
Increasing pumping costs and declining well capacities in regions like the Southern High Plains of Texas are requiring producers to adapt cropping practices for use with irrigation levels that vary between complete replacement of crop evapotranspiration (ET) to none (i.e., dryland production). Grain sorghum [Sorghum bicolor (L.) Moench] is a crop suited to both dryland and various deficit irrigation production systems. Our objectives were to (i) identify cultural practices (planting date, population, and cultivar maturity) that maximize sorghum grain yield for widely varying irrigation strategies; and (ii) consider effective means to allocate available water among irrigation strategies that maximizes the ratio of yield to ET, that is, water use efficiency (WUE). Using the model SORKAM and long-term (1958–1999) weather records from Bushland, TX, we simulated sorghum grain yields on a Pullman soil (fine, mixed, superactive, thermic Torrertic Paleustoll) under dryland (rain) and three deficit irrigation levels (rain + irrigation = 2.5, 3.75, or 5.0 mm d–1) for all combinations of planting date (mid-May, 15 May; early June, 5 June; and late June, 25 June), cultivar maturity (early, 95 d; medium, 105 d; late, 120 d), and plant density (12 and 16 plants m–2). For 2.5 mm d–1 irrigation level, simulated grain yields were maximized with either early or medium-maturing cultivars planted in early June. In contrast, simulated sorghum yield and WUE increased with a mid-May planting date and later-maturing cultivars for irrigation levels of 3.75 and 5.0 mm d–1. We conclude that spreading water to uniformly irrigate a field with 2.5 mm d–1 produces 
16% less grain than concentrating the same water resources to variably irrigate a field at 3.75 or 5.0 mm d–1 with complementary (2:1 and 1:1) dryland areas.
Abbreviations: ET, evapotranspiration • I-WUE, irrigated water use efficiency (kg m–3) • WUE, water use efficiency (kg m–3)
|
INTRODUCTION
|
|---|
GRAIN SORGHUM is a feed grain grown on the Southern High Plains of Texas under both dryland and irrigated conditions either as a primary cash crop or late-planted alternative crop that may become increasingly important as a future feed stock for producing ethanol. The introduction of improved hybrids and water-conserving residue management practices have increased dryland grain sorghum yields at the USDA-ARS Conservation and Production Research Laboratory, Bushland, TX, by 139% between 1956 and 1997 (Unger and Baumhardt, 1999). Likewise, improved water application and irrigation scheduling practices typically increased crop yields where irrigation fully replaced ET (Howell, 2001). The groundwater level beneath the Southern High Plains is declining and, consequently, irrigation well capacity has decreased (Stewart, 2003). As a result, deficit irrigation strategies are being implemented to provide a gradual transition to alternative dryland production systems (Baumhardt et al., 1985; Norwood, 1995). Allen and Musick (1993) observed that optimizing planting date, row spacing, crop maturity, and irrigation amount to achieve maximum stable grain yield is the primary challenge for irrigated grain sorghum management.
Where irrigation well capacity is adequate to replace peak sorghum ET demand, maximum grain yields are obtained with late-maturing cultivars that have been planted sufficiently early to mature before a killing frost (Vanderlip et al., 1998). Alternatively, Stichler et al. (1997) and Vanderlip et al. (1998) recommend irrigation during critical growth stages, typically the "boot" (panicle exsertion) stage, when irrigation capacity is not adequate to meet crop ET. In a 3-yr study at Bushland, Allen and Musick (1993) concluded that medium-maturity hybrids were better adapted than late-maturity hybrids for limited irrigation timed to meet crop water requirements during key growth stages. Baumhardt and Howell (2006) pointed out, however, that deficit irrigation is often applied uniformly during the growing season because of the practical limitation imposed by irrigation well capacity that precludes large single irrigation applications during critical growth stages. Precipitation during critical growth stages can affect grain sorghum yield dramatically, when grown under deficit irrigation. At Bushland, TX, erratic growing season precipitation that varies in frequency and amount from 89 to 580 mm yr–1 can render results of short duration field experiments, to identify the best combination of cultural practices for the long-term, unreliable. However, using computer crop growth simulation with long-term weather records is an effective method to expand the basis for comparing cultural practices used in producing grain sorghum despite climatic variability.
Based on the grain sorghum simulation model SORKAM (Rosenthal et al., 1989) and using more than 40 yr of recorded weather at Bushland, TX, Baumhardt and Howell (2006) concluded that plant populations of 12 and 16 plants m–2 did not affect grain yield of sorghum receiving rain + irrigation <5.0 mm d–1. Similarly, dryland sorghum grain yield was also insensitive to the simulated plant populations (Baumhardt et al., 2005). In these studies, narrow, 0.38-m row spacing increased mean simulated grain yields 7% for dryland and 9% for irrigated sorghum compared with wide, 0.76-m row spacing on a Southern High Plains clay loam soil. The simulated dryland grain sorghum yields were greatest for early and medium-maturity cultivars planted 5 June to avoid late summer heat or water deficit stress while maturing before freezing weather (Baumhardt et al., 2005). For subtropical environments in Australia with mean annual precipitation of 620 to 675 mm (Commonwealth of Australia Bureau of Meteorology, 2006), the simulated sorghum grain yields were maximized using medium- or late-maturing cultivars when water resources were not limiting (Muchow et al., 1994). Depending on the available irrigation capacity, results from Baumhardt and Howell (2006) suggested two alternative management practices to maximize grain yield for the Southern High Plains. Where rain plus supplemental irrigation was 
2.5 mm d–1, they recommended planting early maturing cultivars during early June; and where rain plus supplemental irrigation was 5.0 mm d–1, they recommended planting late-maturing cultivars during 15 May. Crop response to an intermediate 3.75 mm d–1 irrigation level was not addressed nor was how to allocate diminishing water resources to maximize farm grain sorghum yield, that is, should fixed irrigation water resources be spread over a large area for deficit irrigation or should irrigation be concentrated on a smaller area with a complementary dryland component area. Either strategy will depend on growing season duration and variable rainfall conditions.
In this study our first objective was to identify planting practices and cultivar maturity combinations that would maximize sorghum grain yield for various deficit irrigation levels. To achieve this, we simulated sorghum grain yield and water use for all combinations of selected planting dates, plant populations, and cultivar maturity grown under four irrigation levels for each year of the (1958–1999) weather record at Bushland, TX. Our second objective was to compare the effects of various deficit irrigation strategies for a fixed water resource on overall sorghum grain yield. To achieve this we computed the weighted mean grain yield associated with either spreading available water resources to deficit irrigate a large area or with concentrating water to irrigate a smaller area combined with a complementary dryland area.
|
MATERIALS AND METHODS
|
|---|
We simulated grain sorghum growth and yield using SORKAM Version 2000 (W.D. Rosenthal and R.L. Vanderlip, personal communication, 2000) according to the methods of Baumhardt et al. (2005). Briefly, long-term (1958–1999) weather records of daily solar irradiance (MJ m–2), maximum and minimum air temperature (°C), and precipitation (mm) from the USDA-ARS, Conservation and Production Research Laboratory, Bushland, TX (35°11' N, 102°5' W; and 1170 m msl elevation), were used as input. The simulations were performed for a 1.8-m deep Pullman clay loam (fine, mixed, superactive, thermic Torrertic Paleustoll) profile divided into nine layers having similar available water and porosity characteristics (Baumhardt et al., 2005). For comparison among treatment management practices, all simulations were reset each year to a common initial plant available soil profile water content of 200 mm. This represents an optimum soil water storage normally observed for no-tillage fallow after wheat (Triticum aestivum L.) in a wheat–sorghum–fallow rotation (Jones and Popham, 1997). Because SORKAM does not include nutrient effects on sorghum growth and yield, soil fertility was assumed to be adequate to meet sorghum needs for all simulations. Maximum sorghum rooting depth was not restricted within the 1.8-m soil profile depth. Moroke et al. (2005) observed rooting to depths >1.5 m for dryland sorghum grown on a Pullman soil, which is consistent with other detailed root studies (Robertson et al., 1993). Soil water evaporation was calculated by the Priestley-Taylor method with an overall 1.45 scale factor, after Howell et al. (1989), using constants of 0.19 for albedo (Howell et al., 1989), 9.9 mm for U (Stage 1) and 7.8 mm d–1 for C (Stage 2), the two stages of soil water evaporation as reported by Steiner (1989). Runoff was calculated using the measured NRCS curve number of 82 for sorghum reported by Hauser and Jones (1991).
Crop Simulations
Grain sorghum growth and yield was simulated for all possible combinations of selected cultivar maturities (3 levels), planting dates (3 levels), populations (2 levels), and irrigation (4 levels) beginning 2 wk before the treatment planting date. From available cultivar types grown in the region, we selected three generic early, medium-, and late-maturing cultivar entries that produce 15, 17, or 19 leaves and require 95, 105, and 120 d to reach maturity, respectively, for evaluating a broad range of growing season durations. Growth and grain yields of these cultivars were simulated using planting densities of 12 and 16 plants m–2 to represent intermediate transition populations between dryland and irrigated conditions, planted on 15 May (early), 5 June (normal), and 25 June (late) in single rows 0.76 m apart. Test irrigation levels of dryland (rain only) and rain + deficit irrigation totaling 2.5, 3.75, and 5.0 mm d–1 were applied, independently of crop growth stage, using a minimum 25-, 37.5-, and 50-mm application depth for a minimum interval between irrigations of 10 d. Test irrigation treatments represent a range of well pumping capacities of approximately 3, 4.6, and 6.1 L min–1 ha–1, which are consistent with declining (weak) to adequate well capacities. The resulting 72 treatment combinations were evaluated for each of the 42 yr of weather records for a total of 3024 simulations. The SORKAM simulations continued until physiological maturity or a killing freeze, when grain yield and water use estimated from growing season rain, irrigation, and soil water were determined.
Model Uncertainty
The validity of SORKAM to simulate plant responses under variable growing conditions is critical for interpreting treatment cultural practice effects on grain sorghum growth and yield and consequent water use efficiency. Validation of SORKAM simulated grain sorghum growth and yield was done by Rosenthal and Gerik (1990), Heiniger et al. (1997) and Xie et al. (2001). Baumhardt and Howell (2006) further validated SORKAM using recorded weather and cropping conditions (e.g., planting date) for a Pullman soil. They noted that grain yields simulated using experimental weather and soil water data as input averaged 4% greater than observed yields (r2 = 0.70, RMSE = 903.5 kg ha–1) for the corresponding late- and medium-maturity hybrids planted in rows spaced 0.76 m apart at 8 seed m–2 during a 1984 to 1998 wheat–sorghum–fallow rotation study (Jones and Popham, 1997). Additionally, Tolk et al. (2003) concluded that SORKAM produced similar row width (0.38 and 0.76 m) and plant density (3.1, 6.5, 13.0 plants m–2) effects on grain yield and ET measured during a 3-yr lysimeter study.
Simulated crop growth is fundamentally connected to water use (Rosenthal et al., 1987), that is, factors controlling plant available water determine a plant stress index that regulates calculated assimilate production and distribution. Seasonal ET is limited to the sum of the soil water, treatment irrigation, and precipitation minus losses such as stormwater runoff. During the Jones and Popham (1997) 1984 to 1998 wheat–sorghum–fallow rotation study, experimentally estimated ET averaged 470 mm compared with the corresponding SORKAM simulated ET of 507 mm. The 37-mm mean difference between measured and simulated water use during this 14-yr comparison was not significant (P = 0.90). The corresponding measured stormwater runoff averaged 9 mm more than simulated growing season runoff, based on preliminary, 1999 validations assessing curve number selection. For these data, we calculated the correlation between simulated and measured yield differences compared with the corresponding simulated and measured ET differences to identify any systematic yield and ET bias. The resulting correlation, r = 0.04, was negligible, indicating errors for yield estimates were independent of the errors for water use. Based on these results we determined that SORKAM adequately simulates crop growth, grain yields, and water use throughout a broad range of climate stress conditions for various growing conditions.
Analyses
Sorghum growth, grain yield, water use, and the WUE, calculated as the ratio of yield to total rain, irrigation, and soil water used, were simulated for each of the 72 treatment combinations or experimental units. For unique soil and weather input conditions specific to a simulated year, SORKAM produces unique crop growth results for that experimental unit. We, therefore, utilized the observed climatic variability including rainfall and temperature of 42 growing seasons (1958–1999) as a source of random experimental error to test variable crop response to treatment combinations. Descriptive univariate statistics and Pearson correlation were used to identify overlapping correlation among the treatment cultural practices, recorded growing season precipitation, and all dependent simulated growth parameters (SAS Institute, 1988). For example, the planting date and cultivar maturity treatments determine the potential growing season length and, consequently, were correlated with cumulative precipitation, which precluded using precipitation amount as a covariant in subsequent analyses. Additionally, the experimental error was nonhomogeneous across irrigation levels because irrigation that increased simulated yield also decreased variability; therefore, we compared the cultural practice treatment effects within irrigation levels (Baumhardt and Howell, 2006). Our analyses were according to a factorial arrangement of a completely randomized design replicated with years using SAS general linear models ANOVA procedures.
|
RESULTS AND DISCUSSION
|
|---|
Grain Yield and Water Use
The SORKAM-simulated sorghum grain yields from 1958 to 1999 ranged from 0 to 9674 kg ha–1 and averaged 5671 kg ha–1 across all planting dates, populations, cultivar maturities, and irrigation combinations. The broad range in grain yield reflects the effects of erratic precipitation amount (47–532 mm) and growing season length (144–220 d) that is characteristic of the semiarid weather for the Southern High Plains in Texas. Supplementing rain with deficit irrigation eliminated many crop failures and decreased grain yield variability compared with dryland. Total treatment irrigation amounts varied with seasonal rain occurrence and amount. Although variability of simulated grain yields decreased as irrigation increased, annual yield differences during 1958 to 1999 reflect the usually variable growing season weather conditions including crop growth terminating freeze dates. These dates ranged from later than average, 14 November to as early as 21 September, which shortened the growing season for late maturity or late planting date treatments.
The effects of cultivar maturity, planting date, and irrigation level on simulated grain yields are shown in Fig. 1
. Simulated mean grain sorghum yields increased 30% from 3780 kg ha–1 for dryland (rain only) to 4940 kg ha–1 by supplementing rain with deficit irrigation to replace ET up to 2.5 mm d–1. Mean seasonal irrigation for this treatment was 65 mm and ranged from 52 mm for early planted, early maturing sorghum to 79 mm for late-maturing cultivars that were planted late. Increasing the supplemental irrigation to replace ET up to 3.75 and 5.0 mm d–1 resulted in mean simulated grain yields of 6790 and 7170 kg ha–1, respectively, that were 80 to 90% above dryland yields. The corresponding cumulative irrigation averaged 171 mm for the 3.75 mm d–1 irrigation level compared with 302 mm for 5.0 mm d–1, which was similar to the average "full" irrigation amount recorded for 35 producer fields during a 1998 to 2003 demonstration trial (New, 2004). The 42-yr mean irrigation applications shown in Fig. 2
and grouped by treatment irrigation levels illustrate the consistent effect of later planting dates and later maturing cultivars to increase the amount of irrigation. That is, medium- and late-maturing sorghum hybrids extended the growing season and, consequently, increased the average amount of irrigation applied 6 and 13%, respectively, compared with early hybrids. Delaying sorghum planting to 5 June and 25 June increased mean irrigation applied 7 and 12%, respectively, compared with 15 May-planted sorghum possibly because of warmer growing season temperatures and decreased benefit from timely June precipitation.
View larger version (52K):
[in this window]
[in a new window]
|
Fig. 1. Simulated (1958–1999) grain sorghum yields averaged across plant population (P) for early, medium-, and late-maturing cultivars planted on 15 May, 5 June, and 25 June for dryland, rain only, and rain + irrigation (I) for 2.5, 3.75, and 5.0 mm d–1. Bars represent standard error.
|
|
View larger version (42K):
[in this window]
[in a new window]
|
Fig. 2. Average (1958–1999) cumulative irrigation applied for rain + irrigation treatments of 2.5, 3.75, 5.0 mm d–1 to the early, medium-, and late-maturing grain sorghum cultivars planted on 15 May, 5 June, and 25 June. Bars represent standard error.
|
|
Cultivar maturity, and planting date and plant population represent flexible management practices that can be adjusted as irrigation capacity changes and, in our simulations, produced a variable effect on grain yield. For both planting densities of 12 and 16 plants m–2, dryland grain sorghum yields were greatest with earlier maturing cultivars, which limited crop exposure to water deficit stress (Baumhardt et al., 2005). Rain plus supplemental irrigation of 2.5 mm d–1 eliminated all but the planting date treatment effect (Table 1) on grain yield, that is, 15 May-planted sorghum yielded 13% less than the corresponding June-planted sorghum (Fig. 1). For the of 3.75 and 5.0 mm d–1 supplemental irrigation levels, the water deficit stress was reduced or eliminated and, consequently, simulated grain yields increased 7% for medium and 15% for late cultivars compared with the early maturing cultivar (Fig. 1). Also for the 3.75 and 5.0 mm d–1 irrigation levels, medium- and late-maturing cultivars planted 15 May and 5 June increased yield up to 15% compared with 25 June-planted sorghum because of the potentially longer, 20- to 40-d growing season. Similarly, when rain plus irrigation was 3.75 or 5.0 mm d–1, sorghum grain yield increased 2 and 4%, respectively, with the higher plant population. These results show cultivar maturity and planting date treatments with the potential to extend the growing season, interact with the irrigation capacity for variable yield responses and are contrary to using uniform crop management decisions that limit crop water deficit and maximize grain yield.
View this table:
[in this window]
[in a new window]
|
Table 1. Analysis-of-variance for planting date (D), maturity (M), and population (P) treatment effects by irrigation (I) level on simulated sorghum grain yield. The ANOVA results designate the probability of significant treatment effects.
|
|
Our simulated dryland grain sorghum water use averaged 505 mm from rain plus full initial soil profiles during the 42-yr test period. This was consistent with the 490-mm water use from 1989 to 1991 observations by Allen and Musick (1993) for sorghum planted in May on full soil profiles, but greater than the 375-mm mean water use reported in a 6-yr test by Baumhardt and Jones (2002) for mid-June planted dryland sorghum. Compared with dryland, the simulated sorghum water use increased with increasing irrigation. For example, by supplementing rain with irrigation to supply 2.5 mm d–1 to the crop, the mean water use, ET, increased 14% to 576 mm or a difference approximating the 65 mm of irrigation. The 3.75-mm d–1 irrigation treatment supplemented rain with 171 mm of irrigation and increased mean water use 144 to 649 mm, an increase of 29% above dryland. Supplementing rain with 302 mm of irrigation for the 5.0-mm d–1 irrigation treatment increased mean water use 164 to 669 mm, or a 32% increase above dryland. Although not available from the model output, the difference between applied irrigation and the water use was probably due to drainage and stormwater runoff. Of the increased simulated water use, treatment irrigations accounted for 26% of the water use in the 3.75-mm d–1 irrigation level and 45% of the water use for the 5.0-mm d–1 irrigation level treatment.
Simulated water use for treatment planting date and grain sorghum cultivar maturity are shown by irrigation level in Fig. 3
with the corresponding statistical analyses summarized in Table 2. Generally, water use decreased an average of 6% when planting was delayed from 15 May to the 5 June and another 11% with a 5 June to 25 June delay. This significant reduction in crop water use by later-planted grain sorghum (Fig. 3) could be a result of the necessarily shorter growing season. These results are also consistent with a 3-yr study by Allen and Musick (1993), who attributed the decrease in water use by later-planted sorghum to lower evaporative demand during September grain filling. Medium- and late-maturing cultivars extend the growing season 10 to 20 d compared with early cultivars and increase the potential water use; however, simulated water use was only significantly greater for medium- and late-maturing cultivars where rain plus supplemental irrigation was 3.75 and 5.0 mm d–1. This reversed the trend established for the lower irrigation treatments. Compared with early maturing sorghum, the simulated water use increased 40 to 45 mm for medium and late cultivars when planted 15 May or 5 June and irrigated at 5.0 mm d–1, but this difference decreased for late-planted sorghum and resulted in a significant interaction (Table 2). Increasing planting densities from 12 to 16 plants m–2 increased water use 10% under the 5.0 mm d–1 irrigation level. Our results suggest that under conditions where rain plus supplemental irrigation >2.5 mm d–1, management practices that extend the growing season also increased water use and, consequently, grain yield potential.
View larger version (52K):
[in this window]
[in a new window]
|
Fig. 3. Simulated (1958–1999) water use averaged across plant population for early, medium-, and late-maturing cultivars planted on 15 May, 5 June, and 25 June for dryland, rain only, and rain + irrigation for 2.5, 3.75, 5.0 mm d–1. Bars represent standard error.
|
|
View this table:
[in this window]
[in a new window]
|
Table 2. Analysis-of-variance for planting date (D), maturity (M), and population (P) treatment effects by irrigation level on simulated water use and water use efficiency (WUE). The ANOVA results designate the probability of significant treatment effects.
|
|
Grain yield and water use can indicate management practices that maximize crop productivity or minimize water consumption. Our goal, however, is to identify management practices that maximize the ratio of simulated grain yield (kg ha–1) to the corresponding water use (m3 ha–1), that is, the WUE (kg m–3). In general, simulated WUE averaged 0.71 kg m–3 for dryland, 0.85 kg m–3 for rain plus supplemental irrigation of 2.5 mm d–1, and 1.05 and 1.07 kg m–3 for 3.75 and 5.0 mm d–1, respectively. The ANOVA (Table 2) indicated that mean WUE varied significantly with planting date, cultivar maturity, and plant population depending on the irrigation levels. For example, WUE increased with progressively later planting dates from 0.84 kg m–3 for 15 May to 0.94 kg m–3 for 5 June and 0.99 kg m–3 for 25 June. Comparison of the WUE plotted in Fig. 4
illustrates a declining benefit with delayed planting as the irrigation level increases. No consistent overall trend in WUE emerged for the sorghum maturity treatments tested, averaging 0.92 kg m–3; however, cultivar effects on WUE varied with irrigation. Under dryland conditions simulated WUE decreased significantly (Table 2) from 0.79 kg m–3 for early maturing cultivars to 0.70, and 0.63 kg m–3 for medium- and late-maturity cultivars (Fig. 4). In contrast, for rain plus supplemental irrigation of 2.5 mm d–1 the cultivar maturity effects were negated and, subsequently reversed for the 3.75 and 5.0 mm d–1 irrigation levels where later-maturing cultivars significantly increased mean WUE 5 to 10% above the early cultivars. For the rain plus irrigation levels of 3.75 and 5.0 mm d–1, simulated WUE increased 2% with the higher plant density from 1.04 to 1.06 kg m–3 and from 1.06 to 1.09 kg m–3, respectively. These results further illustrate that management practices for maximizing WUE vary with irrigation level.
View larger version (53K):
[in this window]
[in a new window]
|
Fig. 4. Water use efficiency (WUE) averaged across plant population from the ratio of simulated (1958–1999) sorghum grain yield and water use for early, medium-, and late-maturing cultivars planted on 15 May, 5 June, and 25 June for dryland, rain only, and rain + irrigation (I) for 2.5, 3.75, and 5.0 mm d–1. Bars represent standard error.
|
|
Differential Yield Increase and I-WUE
To compare the incremental effects of our 2.5-, 3.75-, and 5.0-mm d–1 treatment irrigation levels, we calculated the differential grain yield attributed to the irrigation increment (Fig. 5
). Compared with dryland sorghum, the 2.5 mm d–1 irrigation level added an average 65 mm of irrigation water that increased simulated grain yields 950 kg ha–1 for the early maturing cultivar, 1150 kg ha–1 for the medium-maturing cultivar, and 1380 kg ha–1 for the late-maturing cultivar. Later-maturing cultivars extended the growing season and significantly increased yield potential with increasing irrigation (Table 3). The simulated grain yield increase for the 2.5 mm d–1 irrigation were significantly (P > 0.95) greater with later-maturing cultivars. For the 3.75 mm d–1 irrigation, the additional 106 mm of water increased overall simulated grain yield another 1850 kg ha–1 compared with the 2.5 mm d–1 irrigation. Simulated grain yields, however, decreased significantly with delayed planting from 2320 kg ha–1 for 15 May, to 1850 kg ha–1 for 5 June and 1370 kg ha–1 for 25 June, thus reversing a trend (P = 0.93) observed for 2.5 mm d–1 irrigation (Fig. 5). The differential grain yields also increased as cultivar maturity increased from 1410 kg ha–1 for early to 2050 kg ha–1 for medium and 2080 kg ha–1 for late. For the 5.0 mm d–1 rain and supplemental irrigation, the additional 131 mm irrigation resulted in an average 390 kg ha–1 grain yield increase compared with the 3.75-mm d–1 irrigation level. Simulated grain yields of the 5.0 mm d–1 rain + irrigation treatment increased >500 kg ha–1 with later-maturing cultivars, but decreased with later (June) planting dates. The smaller differential yield increase (ranging from 25 to 1010 kg ha–1) plotted by cultivar and planting date treatments (Fig. 5.) indicate diminishing grain yield returns with the additional water for the 5.0-mm d–1 irrigation level. Our results also show that factors contributing to an expanded growing season can increase yield for grain sorghum receiving adequate irrigation; however, those differential grain yields decreased with additional irrigation above 3.75 mm d–1, possibly due to reduced WUE.
View larger version (37K):
[in this window]
[in a new window]
|
Fig. 5. Simulated (1958–1999) grain sorghum yield increase relative to the adjacent lower irrigation treatment level. Values for early, medium-, and late-maturing cultivars planted on 15 May, 5 June, and 25 June that received rain and rain + irrigation for 2.5, 3.75, and 5.0 mm d–1 are averaged across plant population (P). Bars represent standard error.
|
|
View this table:
[in this window]
[in a new window]
|
Table 3. Analysis-of-variance for planting date (D), maturity (M), and population (P) treatment effects by irrigation level on simulated yield increases due to irrigation and the corresponding irrigation water use efficiency (I-WUE). The ANOVA results designate the probability of significant treatment effects.
|
|
The effects of planting date, cultivar maturity, and population on simulated sorghum grain yield and WUE were compared for each irrigation increment relative to the adjacent lower irrigation treatment level. We adapted Eq. [3] of Howell (2001) to calculate the incremental irrigation water use efficiency, I-WUE, as the ratio of the differential grain yield and applied irrigation depth for the treatment irrigation levels. The simulated I-WUE shown in Fig. 6
averaged 1.73 kg m–3 for the 2.5-mm d–1 irrigation level, 1.72 kg m–3 for 3.75 mm d–1, and 0.29 kg m–3 for 5.0 mm d–1 irrigation; thus indicating limited benefit from the additional applied water at the higher irrigation level. In contrast to WUE, the average I-WUE decreased with progressively later planting dates from 1.53 kg m–3 for 15 May to 1.25 kg m–3 for 5 June and 0.84 kg m–3 for 25 June, which we attributed to the decreasing growing season length. The decrease in I-WUE with later planting dates was significant for all irrigation levels, although less consistent for the lower 2.5 mm d–1 irrigation. Generally, early maturing cultivars with shorter growing seasons decreased mean I-WUE to 0.96 kg m–3 compared with medium- and late-maturing cultivars that averaged 1.3 kg m–3. The cultivar effects on I-WUE, however, were only significant (Table 3) for the 3.75 and 5.0 mm d–1 irrigation levels because of consistently lower I-WUE with early maturing cultivars. An interaction between planting date and cultivar maturity was attributed to the declining cultivar maturity effect on I-WUE for later planting dates. As plant density increased the I-WUE significantly increased from 1.17 kg m–3 for 12 plants m–2 to 1.28 kg m–3 for 16 plants m–2 under adequate 3.75-mm d–1 and 5.0-mm d–1 irrigation levels. Our results show management practices that increase I-WUE, although variable with irrigation level >2.5 mm d–1, generally extend the growing season by using earlier planting dates and later-maturing cultivars.
View larger version (39K):
[in this window]
[in a new window]
|
Fig. 6. Irrigated water use efficiency (I-WUE) calculated as the ratio of the simulated (1958–1999) grain yield increase and the corresponding irrigation increment relative to the adjacent lower irrigation treatment level. Values for early, medium-, and late-maturing cultivars planted on 15 May, 5 June, and 25 June that received rain + irrigation for 2.5, 3.75, and 5.0 mm d–1 are averaged across plant population (P). Bars represent standard error.
|
|
Irrigation Management Strategies
We applied these simulation results to address two practical management questions on: (i) identifying common cultural practices for variably irrigated grain sorghum, and (ii) maximizing irrigated sorghum grain yield either by spreading water resources over a large area using deficit irrigation and relying on rain, or by concentrating water resources on a smaller adequately irrigated area. Our simulations indicated that late- and medium-maturity cultivars should be planted early to extend the growing season and take greater advantage of adequate irrigation. However, these practices are poorly suited to dryland production, which benefits from planting earlier maturing cultivars on 5 June to limit crop exposure to water deficit stress. This divergence in cultural practices for optimum sorghum production indicated almost incompatible management strategies for the low to high water application levels. Possible choices for a common planting date and sorghum cultivar combination varied from a medium-maturity cultivar planted on 5 June to an early maturing cultivar planted 15 May depending on if the selection criterion were based on yield or I-WUE results, respectively.
Simple contrasts of irrigation strategies for spreading or concentrating water resources were made using simulated grain yield or I-WUE for three irrigation scenarios. Using simulation results, we selected the best planting date and cultivar combination unique to each irrigation level and compared the resulting spatially weighted yields with those calculated from the best common planting practice, based on yield or WUE (Table 4). Grain yields were calculated for an area that was uniformly irrigated at 2.5 mm d–1 and compared with weighted average yields from an equivalent area irrigated at 3.75 mm d–1 and dryland divided 2:1, and 5.0 mm d–1 and dryland divided 1:1. Compared with the yield determined for a uniform irrigation, both alternative irrigation scenarios of type 2:1 and 1:1 increased the weighted average yields approximately 20% when using the independently assigned best cultivar and planting date combinations. Using a common, yield-maximizing cultivar and planting date combination, the weighted average yield for the 3.75 mm d–1 ET and dryland divided 2:1 increased 18% compared with a 12% increase for the 5.0 mm d–1 ET and dryland divided 1:1. For the common best cultivar and planting date combination based on calculated I-WUE, the corresponding yield increases were 13 and 11% for 2:1 and 1:1 weighted averages. We conclude that concentrating water resources from an area that is uniformly deficit-irrigated at <2.5 mm d–1 onto smaller component areas that are irrigated at 3.75 or 5.00 mm d–1 with a complementary (2:1 and 1:1) dryland area increases sorghum grain yield. Using a common best cultivar and planting date combination depressed the yield benefit when concentrating water resources. Although not tested, producers that choose to concentrate irrigation resources on constituent field partitions will likely, preferentially, select more productive field areas having desirable soil properties that may enhance yield response.
View this table:
[in this window]
[in a new window]
|
Table 4. Weighted average yield calculated for uniform and nonuniform irrigation levels with a fixed total capacity. Unique or common best cultivar maturity and planting date combinations were selected based on either simulated yield or I-WUE. Base yield of 5182 kg ha–1 was for uniform irrigation at 2.5 mm d–1 of early maturing sorghum planted 5 June.
|
|
|
CONCLUSIONS
|
|---|
Our results show that cultivar maturity, planting date and population management practices to maximize yield and WUE for the Southern High Plains in Texas varied with irrigation level. Simulated grain yield consistently increased with increasing irrigation, but overall WUE did not increase appreciably when rain + irrigation > 3.75 mm d–1. Simulated grain yield differences relative to the adjacent lower irrigation treatment level increased for the 2.5 mm d–1 irrigation increment, peaked at the 3.75-mm d–1 irrigation increment, and decreased for the 5.0-mm d–1 irrigation increment. Thus, calculated I-WUE for 2.5- and 3.75- mm d–1 irrigation levels were higher than for the 5.0-mm d–1 level. The 2.5-mm d–1 irrigation level stabilized and moderately increased grain yields, but the 3.75-mm d–1 level increased both grain yields and WUE. For dryland and the 2.5-mm d–1 irrigation level, a 5 June-planted, early maturing cultivar optimized simulated grain yield regardless of plant density. For the 3.75- and 5.0-mm d–1 irrigation levels, later-maturing cultivar and earlier planting date combinations tended to increase simulated grain yields by extending the growing season to utilize the increased irrigation.
Based on those management practices that maximized either yield or I-WUE, we calculated a spatially weighted-average grain yield for uniform and variable irrigation strategies using either the best planting date and cultivar combination unique to each irrigation level or a common best cultivar maturity and planting date combination. Spreading water to uniformly deficit irrigate a field with 2.5 mm d–1 produced less grain than concentrating the same water resources to variably irrigate a field at 3.75 or 5.0 mm d–1with complementary dryland areas. We conclude that, as water resources diminish, producers should adopt the hypothesized irrigation strategy of concentrating water resources on a smaller area rather than spreading water on larger irrigated areas.
|
REFERENCES
|
|---|
- Allen, R.R., and J.T. Musick. 1993. Planting date, water management, and maturity length relations for irrigated grain sorghum. Trans. ASAE 36:1123–1129.[ISI]
- Baumhardt, R.L., and T.A. Howell. 2006. Seeding practices, cultivar maturity, and irrigation effects on simulated grain sorghum yield. Agron. J. 98:462–470.[Abstract/Free Full Text]
- Baumhardt, R.L., and O.R. Jones. 2002. Residue management and tillage effects on soil–water storage and grain yield of dryland wheat and sorghum for a clay loam in Texas. Soil Tillage Res. 68:71–82.
- Baumhardt, R.L., J.A. Tolk, and S.R. Winter. 2005. Seeding practices, and cultivar maturity effects on simulated grain sorghum yield. Agron. J. 97:935–942.[Abstract/Free Full Text]
- Baumhardt, R.L., R.E. Zartman, and P.W. Unger. 1985. Grain sorghum response to tillage method used during fallow and to limited irrigation. Agron. J. 77:643–646.[Abstract/Free Full Text]
- Commonwealth of Australia Bureau of Meteorology. 2006. Climate averages for Australian sites. Available at www.bom.gov.au/climate/averages/tables/ca_qld_names.shtml (verified 6 Dec. 2006). Bureau of Meteorology, Melbourne, VIC, Australia.
- Hauser, V.L., and O.R. Jones. 1991. Runoff curve numbers for the Southern High Plains. Trans. ASAE 34:142–148.[ISI]
- Heiniger, R.W., R.L. Vanderlip, S.M. Welch, and R.C. Muchow. 1997. Developing guidelines for replanting grain sorghum: II. Improved methods of simulating caryopsis weight and tiller number. Agron. J. 89:84–92.[Abstract/Free Full Text]
- Howell, T.A. 2001. Enhancing water use efficiency in irrigated agriculture. Agron. J. 93:281–289.[Abstract/Free Full Text]
- Howell, T.A., K.S. Copeland, A.D. Schneider, and D.A. Dusek. 1989. Sprinkler irrigation management for corn: Southern Great Plains. Trans. ASAE 32:147–154, 160.[ISI]
- Jones, O.R., and T.W. Popham. 1997. Cropping and tillage systems for dryland grain production in the Southern High Plains. Agron. J. 89:222–232.[Abstract/Free Full Text]
- Moroke, T.S., R.C. Schwartz, K.W. Brown, and A.S.R. Juo. 2005. Soil water depletion and root distribution of three dryland crops. Soil Sci. Soc. Am. J. 69:197–205.[Abstract/Free Full Text]
- Muchow, R.C., G.L. Hammer, and R.L. Vanderlip. 1994. Assessing climatic risk to sorghum production in water limited subtropical environments: II. Effects of planting date, soil water at planting, and cultivar phenology. Field Crops Res. 36:235–246.[Medline]
- New, L. 2004. Grain sorghum irrigation. B-6152. Texas Coop. Ext., College Station, TX.
- Norwood, C.A. 1995. Comparison of limited irrigated vs. dryland cropping systems in the U.S. Great Plains. Agron. J. 87:737–743.[Abstract/Free Full Text]
- Robertson, M.J., S. Fukai, M.M. Ludlow, and G.L. Hammer. 1993. Water extraction by grain sorghum in a sub-humid environment: I. Analysis of the water extraction pattern. Field Crops Res. 33:81–97.
- Rosenthal, W.D., G.F. Arkin, P.J. Shouse, and W.R. Jordan. 1987. Water deficit effects on transpiration and leaf growth. Agron. J. 79:1019–1026.[Abstract/Free Full Text]
- Rosenthal, W.D., and T.J. Gerik. 1990. Application of a crop growth model to evaluate cultural practices and optimize dryland grain sorghum yield. J. Prod. Agric. 3:124–131.
- Rosenthal, W.D., R.L. Vanderlip, G.F. Arkin, and B.S. Jackson. 1989. SORKAM: A grain sorghum crop growth model. Mp-1699. Texas Agric. Exp. Stn., College Station.
- SAS Institute. 1988. SAS/STAT users guide. Release 6.03 ed. SAS Inst. Cary, NC.
- Steiner, J.L. 1989. Tillage and surface residue effects on evaporation from soils. Soil Sci. Soc. Am. J. 53:911–916.[Abstract/Free Full Text]
- Stewart, B.A. 2003. Aquifers, Ogallala. p. 43–44. In B.A. Stewart and T.A. Howell (ed.) Encyclopedia of water science. Marcel-Dekker, New York.
- Stichler, C., M. McFarland, and C. Coffman. 1997. Irrigated and dryland grain sorghum production south and southwest Texas. B-6048. Texas Coop. Ext., College Station, TX.
- Tolk, J.A., R.L. Baumhardt, and T.A. Howell. 2003. Measured and simulated grain sorghum yield response to row spacing. In Annual meetings abstracts [CD]. ASA, CSSA, and SSSA, Madison, WI.
- Unger, P.W., and R.L. Baumhardt. 1999. Factors related to dryland grain sorghum yield increases, 1939 through 1997. Agron. J. 91:870–875.[Abstract/Free Full Text]
- Vanderlip, R., K. Roozeboom, D. Fjell, J. Shroyer, H. Kok, D. Regehr, D. Whintey, D.H. Rogers, M. Alam, D. Jardine, H.L. Brooks, R.K. Taylor, J.P. Harner, and L.N. Langemeier. 1998. Grain sorghum production handbook. C-687. Kansas State Univ. Agric. Exp. Stn. and Coop Ext. Serv., Manhattan, KS.
- Xie, Y., J.R. Kiniry, V. Nedbalek, and W.D. Rosenthal. 2001. Maize and sorghum simulations with CERES-Maize, SORKAM, and ALMANAC under water-limiting conditions. Agron. J. 93:1148–1155.[Abstract/Free Full Text]
TOP
ABSTRACT
INTRODUCTION
