Effect of mustard cover crops on corn growth, yield and soil water storage in southwest Texas

Abstract Mustard cover crops can provide multiple soil health benefits but may adversely affect cash crops by depleting soil water. This was tested in a center-pivot irrigated corn cropping system in southwest Texas from 2018 to 2020. Changes in biomass growth of both the cover crops and corn, and soil water content in the root zone were monitored multiple times during the main growth periods of the crops. Upon termination, biomass production of cover crops averaged 4271 kg/ha, with equivalent nitrogen and phosphorus inputs to the soil of 92 kg/ha and 12 kg/ha, respectively. Over 2 years, water use during the main growth period of the mustard cover crops and the fallow area averaged at 1.4 mm/day and 1.1 mm/day, respectively, and water use of corn following the mustard and fallow averaged at 4.0 mm/day and 3.7 mm/day, respectively, although the differences were only marginally significant. Our data suggest that water use of cover crops was affected by their growth patterns. The mustard cover crops depleted soil water in a dry year (2019–2020), but to a less extent in a wet year (2018–2019). In either year, however, corn yield was not reduced in association with cover crop use, although the improved shoot biomass growth in corn associated with mustard cover crops was observed in the wet year but not in the dry year. Mustard cover crops tended to reduce deep drainage during the corn seasons, which has implications in preventing the leaching loss of nitrogen in the crop root zone.


Introduction
A major challenge for global agriculture in future decades is increasing crop production by 60% by 2050 (Alexandratos and Bruinsma 2012) while also ensuring long-term sustainability under projected rainfall uncertainty. This calls for significantly increased efficiency in both crop breeding and field management. Corn (Zea mays L.) is one of the world's major staple food crops, and improved and sustainable corn production relies on optimal management and utilization of the stored soil water. Yet the crops' water requirements frequently cannot be met due to widespread soil degradation, including soil compaction, erosion and reduced water infiltration and retention. One strategy to ameliorate these problems is by introducing winter cover crops into the cropping The cover crops were followed by corn in the summers. The focus of the work was on the effects of cover crops on soil water storage and the growth of corn that followed the cover crops, with a special emphasis on the comparisons among different cover crop types. We hypothesized that (a) mustard cover crops differing in growth patterns would have the same level of water consumption; and (b) growing selective mustard cover crops in the winter would not significantly affect the growth and yield of corn or deplete soil water in southwest Texas.

Descriptions of field site, layout of plots and crop management
The experiment was conducted at the Texas A&M AgriLife Research and Extension Center at Uvalde, TX (latitude 29 12 0 34 00 N, longitude 99 45 0 03 00 W; elevation 276 m). The area has a semi-arid climate with an average annual rainfall of 663 mm, annual evapotranspiration of 1506 mm, and mean May and December temperature of 24.7 and 12.2 C, respectively. The site has clay soil (Fine-silty, mixed, active, Hyperthermic Aridic Calciustolls) of the Uvalde series (USDA. 1976). Soil texture analysis indicates 57% clay, 31% silt and 12% sand. Soil bulk density averages at 1.38 g/cm 3 (0-20 cm of soil depth), 1.43 g/cm 3 (20-80 cm), and 1.58 g/cm 3 (80-120 cm). The soil volumetric water content at field capacity and wilting point are 0.35 cm 3 /cm 3 and 0.18 cm 3 /cm 3 , respectively. The average soil pH for the 0-120 cm soil depth is 8.3. The average soil organic matter content (SOM) for the 0-20 cm soil depth is 3.6%. On 1 December 2017, a mixture of Kodiak (Brassica juncea (L.) Czern) and White Gold (Sinapis alba L.) mustard seeds provided by the Pacific Northwest Farmers Cooperative (Spokane, WA) was planted through broadcasting in a 2.43-ha field at a seeding rate of 16.8 kg/ha. An adjacent 2.43-ha field was used as control (without mustard). The mustard-planted and control fields combined occupied a quarter of a 20-ha center pivot field. On 8 March 2018, the mustard cover crop was incorporated in the soil (see Supplemental Figure S1 for field condition immediately before and after incorporation). Grain sorghum (variety DEK S37-07) was planted on 10 April 2018 in 1-m row spacing in both the mustard and control areas. The sorghum grain was harvested from 27 July to 1 August 2018.
On 6 December 2018 and 21 November 2019, the same mustard area used in 2017 was planted with four types of mustard cover crops, namely, Trifecta, White Gold, Kodiak, PNW01, and a mixture of Kodiak and White Gold, using the same seeding rate as in 2017. Again, an adjacent 2.43-ha land was used as control (without mustard, see Supplemental Figure S1). Similar to 2017, in both 2019 and 2020, the mustard cover crops were terminated in early March by lightly incorporating the chopped shoot materials into the topsoil. Following the termination of mustard crops, a commercial corn crop (variety DKC64-69) was planted (see Table 1 for dates of major crop management events) in 0.8-m row spacing in both the mustard and control fields, following the initial fertilization with 73 kg/ha of N and 97 kg/ha of P. A 1.2 ha refuge corn was planted nearby in the same field. The corn crop in the center pivot field was managed conventionally and the weed control was done using glyphosate. In both years, 85 kg/ha of N was injected as liquid fertilizer at the 7-leaf stage. A second injection of 85 kg/ha of N was done at the tasseling stage but was skipped in 2019 due to equipment malfunction.
The field had been managed under strip and conventional tillage since 2009, but in this project, only the conventionally managed areas were used for field sampling. Eighteen plots, each 9 m Â 18 m in dimension, were delineated, with three plots located in each of the mustard-planted areas, including the fallow area (see Supplemental Figure S2 for layout of the plots). Due to lack of a plot planter and to avoid excessive tractor traffic in the field, the experimental plots were blocked but not truly randomized (18 of them; see Supplemental Figure S2). The lack of true replication may cause potential bias in statistical analysis, but we believe it was at a lesser extent due to the lack of a progressive change in total soil water storage across the field before the beginning of this cover crop study, as seen in Supplemental Figure S5B. The 18 plots were used for sampling mustard or corn biomass and leaf area. Twelve metal access tubes (150 cm long; 4.08 cm diameter) were installed in 12 of the 18 plots for monitoring soil water content from the 0 to 120 cm depth profile using a neutron probe. The 0-120 cm soil depth was used in this study because the presence of the pale Bk layer (USDA 1976; see Supplemental Figure S3) at 100-120 cm in the soil of this area, which typically retards root penetration. The 0-120 cm soil depth profile also encompasses the root zone of Kodiak (Panigrahi and Panda 2003), and a number of other field crops (Wu, Zhang, and Gui 1999;Yadav and Mathur 2008;Dong et al. 2010). The installation of the access tubes was done using a hydraulic soil coring machine (Giddings Machine Company, Windsor, CO). The tubes were installed in mid-December of 2019 and 2020 to monitor soil water in the mustard period and again in mid-May to monitor soil water in the corn period. Each time, 12 tubes were installed, two each in control, four mustard areas and one mixed (White Gold þ Kodiak) area. Although the tubes were removed and re-installed during each crop period, the locations of the tubes within the plots were kept the same, since we used a Keson Electronic Measuring Wheel MP401E (Keson, Aurora, IL) to determine the location of the field plots (located at the center of each plot as seen in Supplemental Figure S2) and for each of the plots, the neutron access tube was placed at its center. This ensured that the soil water measured at each plot was from the same area across 2 years.
Irrigation scheduling was done based on the measured soil water content using a neutron probe. Irrigation was given when the average soil water content at top 40 cm dropped below 50% of available water capacity (in this clay loam soil, it was 0.28 cm 3 / cm 3 ), or when stress symptom of leaf rolling started to show up in corn (if soil water data were not available).

Measurements
From early January to late February in 2019 and 2020, the above-ground biomass and leaf area of the mustard cover crops was measured by sampling using a 0.5 Â 0.5 m 2 metal frame. On each sampling date, three frames were placed randomly within each of the 18 plot areas and the average of the measurement values made from the three frames was considered as the value for the plot at a particular sampling date in subsequent statistical data analysis (n ¼ 3). The sampling was made about once every 2 weeks. The set of samples collected immediately before the cover crop termination was used to analyze nitrogen and phosphorous concentrations using the Kjeldahl digestion method.
In late April 2019, damage by a severe rainstorm led to about 8-10% loss of corn seedlings. In both 2019 and 2020, from mid-May to late July, biomass growth of corn was measured 6-7 times (approximately once every 2 weeks) by harvesting two plants randomly selected from each plot on each sampling date. Once sampled, the plant materials were promptly transported to a laboratory, the leaves, stem and cobs of the samples were separated, and green leaf area of all the sampled plants was measured using a LI-3100 Area Meter (Li-Cor, Inc., Lincoln, NE). Then, the samples were dried at 70 C for 5-6 days to measure dry mass. Corn was harvested in early August using a randomly selected 5-m length of a row in each of the 18 plots. Corn kernels were removed from the cobs using an ear corn sheller. A New Holland Grain Moisture Tester (Holland Scientific Inc. Lincoln, NE) was used to measure grain moisture content for the shelled grain samples and the grain yield was represented based on a common moisture of 15.5%.
Soil water content from the top 120 cm soil depth profile was measured weekly or once every two weeks during both the mustard and corn periods using a CPN 503 Elite Hydroprobe (InstroTek, Inc., Research Triangle Park, NC), which was calibrated using the two-point method on the Uvalde clay soil, based on the manufacturer's recommendation. The measurement was made from seven soil depth intervals, i.e., 0-10, 10-20, 20-40, 40-60, 60-80, 80-100, and 100-120 cm. Available soil water was defined as measured volumetric water content minus water content at the wilting point. In rare cases when soil water content was lower than that at the wilting point, available water was considered as zero. Crop water use (ET) was estimated as where I is irrigation (mm), P precipitation (mm), R runoff (mm; assumed zero), W 1 and W 2 the total water retained in the 0-120 cm soil profile at an earlier time and near end of the season (mm), respectively, and D the deep drainage (mm), estimated according to the Wilcox equation (Ogata and Richards 1957;Miller and Aarstad 1972;Stone et al. 2011) fitted to the measured drainage data at the Uvalde site: in which W represents the total stored water (mm) in the 0-120 cm soil depth profile, and T the drainage time in days (see Supplemental Figure S4 for details of the field method, and Supplemental Computer Code S1 and Dataset S1 for the calculation procedure).

Data analysis
To assess the treatment effects, data were checked for normality and variance homogeneity, and subject to analysis of means (ANOM) or analysis of variance (ANOVA) using the General Linear Model procedure available in Minitab 17 (Minitab, Inc. 2016).
If necessary, appropriate data transformations on variables were performed before using ANOVA/ANOM. With the assurance of statistical significance in ANOVA tests, the Tukey method was used for multiple comparisons (n ¼ 2 for soil water analysis, and n ¼ 3 for other analysis). Unless otherwise indicated, multiple comparisons were made using a ¼ 0.05. The time trends of leaf area indices (LAI) of different mustard cover crops were fitted using a logistic model (Mead, Curnow, and Hasted 2003). The daily growth rates of LAI for different types of mustard cover crops were represented using the slopes of the logistic model.

Results
Prior irrigation treatments, soil water status and seasonal rainfalls The field used for this study was previously planted in 2017 with winter wheat, which was subject to differential irrigation treatment, with the full irrigation (100% ET replacement) applied in the areas occupied by Kodiak, PNW01, and White Gold þ Kodiak mixture, and deficit irrigation (60% ET replacement) applied in the areas occupied by Trifecta and the fallow (no-mustard) stretches of the field used in this study. However, due to the abundance of rainfall, the differential irrigation of 25 mm and 12 mm, respectively for the full-and deficit irrigation sections, was actually applied only once on 21 March 2017, which was followed by a total amount of 199 mm rainfall until 30 May 2017 (Figure 1). Based on neutron probe-measured soil water content on 8 May 2017, the areas under full-and deficit irrigation had similar total water storage in the 0-120 cm soil profile (ANOVA, p > 0.951; see Supplemental Figure S5), with 325 mm water stored on the average in the 0-120 soil profile. Thus, the abundant rainfall received during the wheat season in 2017 had essentially equalized the soil water status across the pivot quarter that was later used for the mustard cover crop study. The rainfall pattern from 2018 to 2020 showed a wet fall in 2018, with 402 mm of rain received from September to November. This was in contrast to the much drier fall in 2019, with only 102 mm of rain received from September to November ( Figure 1). The total amounts of rain received during the mustard period (48 mm from December to February) and the corn period (257 mm from March to July) were almost identical in both years, however. To compensate for the limited water recharge in fall 2019, more water was provided through irrigation in 2020 than in 2019 during the corn period from March to July (203 mm vs. 82 mm; again, see Figure 1).

Leaf area growth of the mustard cover crops
In 2018-2019, the tested four mustard cover crops showed different patterns of growth. As seen in the germination counts measured 11 days after planting (17 December 2018), White Gold had the lowest number of plants per square meter (Supplemental Figure  S6). However, at 48 days after planting and later, its LAI remained highest among the four types of the mustard cover crops, as seen in Figure 2(A). The fitted parameter values of the logistic model for different types of cover crops are shown in Supplemental  Table S2.
In 2019-2020, the growth patterns of the mustard cover crops showed considerable differences when compared with those during the 2018-2019 season as seen in Figure  2(B). First, the maximum value of LAI for White Gold was lower than that in 2018-2019. Also, Kodiak and PNW01 reached peak growth rate sooner than in 2018-2019, but Trifecta reached peak growth rate 11 days later than it did in the 2018-2019 season (Figure 2(C,D)). Yet, in both years, the order in which the four types of mustard cover crops reached their respective peak growth rates remained the same, with Kodiak being the earliest and Trifecta the latest.

Final biomass produced and nitrogen and phosphorus inputs of the mustard cover crops
Over 2 years from 2019 to 2020, the average aboveground biomass of the mustard cover crops prior to incorporation was 4271 kg/ha, with equivalent nitrogen (N) and phosphorus (P) inputs to the soil estimated as 92 kg/ha and 12 kg/ha, respectively. However, the N concentration in leaf/stem tissues reduced from 2.6% in 2019 to 1.8% in 2020, and the P concentration reduced from 0.31% in 2019 to 0.27% in 2020.
From 2019 to 2020 when four different types of mustard cover crops were planted, PNW01 consistently displayed the lowest final shoot biomass, which was significantly lower than that of White Gold or the mixture of White Gold and Kodiak (p ¼ 0.016, Table 2). Yet, tissue N and P concentrations of PNW01 were 55.5% and 68% higher, respectively, than the values of White Gold (p < 0.0005). Also, across 2 years from 2019 to 2020, the tissue P concentration of Kodiak was twice that of White Gold (p < 0.0005). Largely due to the fact that the total biomass and nutrient concentrations of PNW01 and White Gold changed toward the opposite directions, the overall effect of mustard types on the nutrient inputs in the soil upon mustard termination was only marginally significant (p ¼ 0.06); however, the N and P inputs in 2020 were 39% and 26% lower than the corresponding values in 2019 (p < 0.0005 and p ¼ 0.015, respectively).

Biomass growth and yields of corn following different types of mustard crops
In 2019, masses of green leaves, shoots (leaves þ stems), as well as cobs all were significantly lower (p ¼ 0.05) in corn growing in the fallow (no mustard) plots, as compared with corn that followed mustard cover crops (Figure 3(A,C)). In 2020, shoot mass of corn under fallow was marginally lower than that of White Gold (p ¼ 0.058, Figure  3(B)), while no difference was found in cob mass between fallow and mustard treatments (Figure 3(D)). Overall cob mass in 2020 was significantly lower than that in 2019 (p < 0.0005). In either 2019 or 2020, there was no significant difference in corn yield among plots planted with different types of mustard cover crops (Table 3). Overall, corn yields in 2019 were significantly higher than those in 2020 (p < 0.05). Also, as seen in Figure 4, the yields in 2020 (dry year) were positively correlated with water use during the main growth period, while there was no significant correlation in 2019 (a wet year). Table 2. Biomass produced, nitrogen (N) and phosphorous (P) concentrations and inputs for different types of mustard cover crops prior to termination in 2019 and 2020. All measurements were made based on harvested shoot dry mass on February 25, 2019 and February 26, 2020 using a 0.5 Â 0.5 m 2 metal frame (n ¼ 3). †"WG þ Kodiak" ¼ "Mixture of White Gold and Kodiak". A Value significantly higher than the average value in a column according to ANOM (a ¼ 0.05, n ¼ 3). B Value significantly lower than the average value in a column according to ANOM (a ¼ 0.05, n ¼ 3). Soil water content at 0-20 cm depth during the mustard and corn growth periods In the 2018-2019 mustard season, SWC at the top 20 cm depth was near field capacity and similar for areas with different mustard cover crops (p > 0.05; Figure 5(A)). During the 2019-2020 mustard season, SWC under four types of mustard cover crops were significantly lower than that of the fallow area (p < 0.0005; Figure 5(C)), while SWC under the mixture of White Gold and Kodiak had similar SWC as the fallow area.
In the 2018-2019 corn season, SWC generally decreased along with the crop development (p < 0.0005; Figure 5(B)), while there was no significant difference in SWC for plots under different cover crops. In the 2019-2020 corn season, SWC under fallow and White Gold was significantly higher than that under PNW01, the mixture of White Gold and Kodiak, and Trifecta, with the SWC under Trifecta being the lowest (p < 0.0005; Figure 5(D)).

Available soil water and water use by the mustard cover crops
Available soil water measured from 2 January to 20 February 2019 indicated that soil under Kodiak and Trifecta had higher water storage (p ¼ 0.001) than that covered with the mixture of White Gold and Kodiak (averaged 150 mm vs. 137 mm). Soil under White Gold tended also to have low available water, especially when it was deep into the   late growth stage near the end of February (Figure 6(A)). In 50 days, water use by the mustard cover crops averaged at 1.2 mm/day. As a comparison, water use in the area without mustard was 0.87 mm/day and there was no significant difference (Table 4).
Available soil water measured from 4 January to 25 February 2020 indicated that soil under fallow, White Gold and the mixture of White Gold and Kodiak had more available water than the soil under PNW01, Trifecta and Kodiak (averaged 98 mm vs. 78 mm; Figure 6(C)). However, the difference was only marginally significant (p ¼ 0.104, n ¼ 2), due to large variability in soil water content measured from the two tubes within the same areas of the mustard cover crops. In 53 days, water use from the fallow area was marginally lower than the average daily water use from all 12 plots (Table 4).
The estimated amounts of deep drainage during the cover crop period was minimum in 2019, although the estimation for the area under Trifecta was numerically higher (Table 5). In the dry year of 2020, the estimated deep drainage for all plots was negligible for the mustard period. Figure 5. Trends of volumetric soil water content measured at 0-20 cm soil depth for both mustard and corn cropping periods in 2019 and 2020. In 2019, soil water content at 20 cm soil depth in both mustard period and corn period was similar across cover crop treatments (A, B). In 2020, soil water content at 20 cm depth was significantly higher (p < 0.0005) in fallow than in all mustard types except White Gold þ Kodiak mixture (during mustard period) and White Gold (during corn period) (C, D).
Available soil water and water use by corn following the mustard cover crops Available soil water in corn had high values in mid-May 2019. The values for all plots dropped significantly starting 25 May, which was followed by a further significant drop starting 1 July 2019 (Figure 6(B); p < 0.0005). Overall, the plots following different mustard cover crops had similar water availability (p ¼ 0.697). During the 70 days from 15 May to 24 July 2019, water consumption per day by corn was similar according to ANOVA (p ¼ 0.067). However, results of ANOM indicated that the corn plots without mustard cover crop had significantly lower daily water use among the tested plots (Table 4).
Available soil water in corn in 2020 was relatively low as seen in the first measurement made on May 10 ( Figure 6(D)). Following a total of 110 mm rain and 25 mm of irrigation (Figure 1), the amounts of available water from all plots increased significantly as measured on 27 May. Overall, the amounts of available water in plots of fallow (147 mm) and White Gold (142 mm) were significantly higher than those in plots of PNW01 (120 mm) and Trifecta (103 mm) (p < 0.0005). During the 69 days from 10 May to 18 July 2020, water consumption per day by corn was similar (averaged at 4 mm/day), regardless of the types of mustard cover crops growing in the winter-spring season (Table 4).
Following a heavy rain of 56 mm on 6 June 2019, the corn plots preceded with different mustard cover crops showed similar water storage (33 ± 19 mm). Again, following a heavy rain of 63 mm on 24 May 2020, the corn plots preceded with different mustard cover crops showed similar water storage (51 ± 14 mm).
In both 2019 and 2020, significantly higher amounts of deep drainage were estimated for the corn plots that had remained fallow during the previous winter seasons, when compared with the average value of all tested plots (Table 5). The calibrated Wilcox equation also estimated higher amounts of drainage during the corn than the mustard period in both years.

Discussion
To address our first hypothesis, our data showed a complex relationship between the growth patterns of the mustard cover crops and their water consumption. For example, the contrast in final biomass production between White Gold and PNW01 (with a high and low biomass value, respectively) was not matched by a likewise contrast in water use during the peak growth stage. Rather, daily water use for areas planted with different mustard cover crops was largely similar (Table 4). However, the steady leaf growth of Trifecta (also reflected as the significantly delayed time in reaching the peak growth rate in both years; see Figure 2) was compatible with the gradually reduced soil water storage as seen in Figures 5 and 6, implying the effect of water consumption accompanied with the sustained leaf vigor. Meanwhile, the tendency of the low-stature Kodiak Table 4. Average water use (mm/day) during the main growth periods for the mustard (January-February) and corn (May-July) crops in 2019 and 2020. 3.9 4.4 3.9 4.1 †"WG þ K" ¼ "Mixture of White Gold and Kodiak"; ‡No mustard; *Value significantly lower than the average value in a row according to ANOM (a ¼ 0.15, n ¼ 2). **Value significantly lower than the average value in a row according to ANOM (a ¼ 0.05, n ¼ 2). Table 5. Estimated deep drainage (mm) during the main growth periods (days) for the mustard (January-February) and corn (May-July) crops in 2019 and 2020. Mustard 52 4.7 Â 10 À5 6.7 Â 10 À3 1.7 Â 10 À4 1.9 Â 10 À3 6.6 Â 10 À3 1.6 Â 10 À3 Corn 69 1.0 0.04 0.17 1.4 0.36 5.7 ÃÃ †"WG þ K" ¼ "Mixture of White Gold and Kodiak"; ‡No mustard; *Value significantly higher than the average value in a row according to ANOM using the square-root transformed data (a ¼ 0.10, n ¼ 2). **Value significantly higher than the average value in a row according to ANOM using the square-root transformed data (a ¼ 0.05, n ¼ 2).
(with low maximum biomass) reaching maximum leaf growth rate earlier in both years also coincided with overall modest depletion of soil water in the plots planted with Kodiak. These results suggest that the growth patterns of cover crops play a role in influencing their water consumption. Yet, to our knowledge this has not received attention in previous studies, although significant differences in overall water use efficiency among different cover crop species have been demonstrated (DeLaune and Mubvumba 2020), and several studies highlighted the importance of seeding date and density (Reese et al. 2014) and terminating management (Burke et al. 2021;Kasper et al. 2022) for ways to curb excessive water use by cover crops in water limited cropping systems. When water use/loss through evapotranspiration was compared between the mustard cover crops and fallow, data from this study showed that the land under cover crops tended to use more water than that under fallow; however, the differences were only occasionally or marginally significant (Table 4). This suggests that, in our cropping system, water loss through evaporation from the bare soil surface of the fallow area must have been considerably large, and a large portion of it would have been captured by plant transpiration, if the fallow were to be replaced by cover crops. Cover crop's effect in reducing soil evaporation has been demonstrated both in experimental (DeVincentis et al. 2022;Gabriel et al. 2019;Unger and Vigil 1998) and modeling studies (Basche et al. 2016b). In fact, evaporation from the bare soil surface can typically be high for the fine-textured soils (Hillel 1971).
In our experiment, the main reason for incorporating the cover crop tissue into the soil was to maximize the soil biofumigation benefits of the mustard cover crops [Our limited data, however, suggested that the number of total nematodes was increased under mustard cover crops as compared with fallow, and this increase was more in fungus-feeding, than in root-feeding nematodes; the data is not shown as it falls out of the scope of this paper]. If, on the other hand, the termination of the mustard cover crops had been done by leaving the shoot tissues on soil surface and limiting tillage practice to allow the surface residue to stay for an extended time period, we would have expected to see the benefit of further reduction in soil evaporation, as has been observed in previous studies (Unger and Parker 1976;Unger 1978;Unger and Vigil 1998). Future research is needed to develop the best method for cover crop termination depending on which problem is more pressing, e.g., water limitation or disease suppression.
Literature data show mixed results in corn growth/yield in response to the use of winter cover crops, which vary from no response in yield (Basche et al. 2016a) to significantly increased (Segarra, Keeling, and Abernathy 1991;Rankoth et al. 2019) or reduced yield (Lewis et al. 2018) in response to winter cover crops. The reduced cotton yield by cover crop use (Lewis et al. 2018) was accompanied with a significantly increased soil organic carbon in the top 15-cm of the soil, which may likely enhance the long-term productivity of the cropping system. The 2-year data from our study indicate that corn yield was not negatively affected by growing the mustard over crops during the winter season in southwest Texas. This was the case both in 2019 with abundant soil water recharge in the previous fall, as well in 2020 that did not receive significant rainfall in the previous fall. However, the shoot biomass growth of corn following the mustard cover crops increased in 2019 as compared with the fallow control, but the response was neutral in 2020. The ecophysiological mechanisms underlying the different responses are unknown, although they may be associated with other benefits of the mustard cover crops that had not been characterized in this study, such as suppression of certain types of soil-borne pests feeding on corn roots (Waisen, Sipes, and Wang 2019).
In our study, there was a competition for water between the cover crop and the main crop in the dry year of 2019-2020, as evidenced by the patterns of soil water depletion. Yet the cover crops attained a closed canopy same as in the wet year of 2018-2019. This was different from a study in a semi-arid area of South Dakota, where the yield of corn was not affected by the winter cover crop use at a high-water stress site but was at a moderate-water stress site (Reese et al. 2014). The mechanism was that the high-water stress site prevented cover crop from establishing a full canopy, which in turn prevented excessive water use by the cover crop as well as the yield loss of corn. On the other hand, the availability of water on the moderate-water stress site encouraged the growth and water use of the cover crop, which caused soil water depletion and corn yield loss. The competition of water leading to the yield loss of the main crop was also demonstrated in a study in east Texas, in which cowpea was intercropped with sorghum (Neely et al. 2018).
Results from our study indicate the tendency of water depletion by the mustard cover crops in the corn cropping system in southwest Texas, which is a concern especially in a year with limited water recharge from rainfall in the previous fall (Figure 6), when the crop growth can be limited by water stress. This was different from a study in Iowa and Indiana, where the use of rye cover crop in a drought year of 2012 in a corn-soybean rotation system either had no impact or significantly increased soil water conservation (Daigh et al. 2014). But our result was consistent with several other studies that documented the water depletion tendency of winter cover crops when used in semi-arid regions (Mitchell, Shrestha, and Irmak 2015;Unger and Vigil 1998). Because of this, the timing of cover termination becomes critical to ameliorate the development of water stress during the main crop phase. This has been demonstrated in Burke et al. (2021) and Kasper et al. (2022). The key is to allow more time between cover crop termination and main crop planting to increase the chance of soil water being recharged by the spring rainfalls. Given the multitude of ecosystem services provided by cover crops (Schipanski et al. 2014; also see supplemental Table S1 and additional references cited therein), it may also be economically viable to irrigate the field after cover crop termination to recharge soil water before planting the main crop.
Our results also suggested the potential benefit of mustard cover crops in reducing the deep drainage over the crop season, especially under wet soil conditions with frequent irrigation events during the corn growing season. This concurs with the findings from a 10-year crop rotation study with and without winter crop in Aranjuez, Spain (Gabriel et al. 2019). Considering the typically high demand of nitrogen fertilizer in corn production (Stichler and McFarland 2019), this tendency of deep drainage reduction, along with the potential scavenging of soil nitrogen due to the deep-root habit of the mustard cover crops (Bodner, Loiskandl, and Kaul 2007;Dabney, Delgado, and Reeves 2001;Gan et al. 2009;Kumar et al. 2020;Panigrahi and Panda 2003;Thorup-Kristensen 2001; considering also the results of our Table 2), suggests the important role that the winter mustard cover crops can play in improving nitrogen use efficiency of the cropping systems, while also helping to prevent the environmental problems caused by the loss of nitrogen through deep drainage (Strock, Porter, and Russelle 2004;Wortman 2016;Abdalla et al. 2019). The same method for estimating deep drainage as used in this study was applied by Holman et al. (2018) in a no-till wheat system. The modeling studies of Yang et al. (2020) and Gabriel, Muñoz-Carpena, and Quemada (2012) demonstrated the reduction of deep drainage in a corn-soybean and corn cropping system due to the use of the winter cover crops over multiple years with varying rainfall patterns.
Although the mustard cover crops tend to deplete soil water, the multiple benefits that they provide to the soil ecosystems, such as soil nitrogen remobilization, deep drainage reduction, soil evaporation reduction, soil temperature modulation, and soilborne pathogen suppression etc. may help cropping system to become more productive and resilient in the long-run. To reduce excessive cover crop water use through management, factors such as seasonal rainfall pattern, cover crop species selection, soil conditions at planting (wet or dry), proper planting and terminating timing, as well as the growth patterns of cover crops, should all be considered, and the system-level approach (Gabriel, Muñoz-Carpena, and Quemada 2012;Reese et al. 2014;Schipanski et al. 2014) is preferred for integrating multiple factors to achieve actionable solution recommendations for crop management, especially in the semi-arid regions.