Perennial grasses as circular strips improve rainfall conservation and crop growth in center pivot agriculture

The dwindling groundwater reserve of the Ogallala Aquifer (OA) is threatening agricultural production, rural livelihoods, and the ecosystem of the semi-arid southern high plains (SHP) of the USA. Nearly 90% of the water withdrawal from OA is used for irrigation in the region to partially meet the crop’s evapotranspiration (ET) demand for water¹. The water needs of crops are further exacerbated during drought years, which are becoming increasingly frequent and prolonged due to climate change. Over time, excessive water withdrawal surpassing natural recharge rates has depleted more than 50% of the OA’s groundwater reserves across the region². This continued reduction in water availability threatens the long-term viability of irrigated agriculture in the SHP, potentially leading to food insecurity, given the critical role of irrigated farming in food production³. Ensuring the sustainability of irrigated agriculture by improving its water productivity is crucial for maintaining food security⁴.

Rainfall is an important component of the water budget of annual crop production in SHP. With the declining nonrenewable groundwater, efficient utilization of renewable rainwater is becoming increasingly important⁵. The SHP receives an average of 330 to 500 mm of rainfall annually⁶, with most precipitation occurring during the summer growing season⁷. However, a substantial portion of this rainfall is received as intense, scattered thunderstorms, leading to significant runoff due to low surface residue, soil organic matter content, and infiltration rate. Rainfall loss to runoff increases as the intensity and amount increase more than 25 mm⁸. Climate change projections indicate an increase in the fraction of rainfall delivered through intense storms in this region⁹. To mitigate these challenges, strategies are needed to improve rainwater conservation of the irrigated production systems and reduce their dependence on dwindling groundwater reserves.

Center pivot is the major irrigation system used in the SHP to irrigate crops. This system is known for its efficiency, reduced labor and energy requirements, adaptability to various crops, and lower costs compared to other irrigation methods. Since its invention in the 1940s, the fraction of center pivot irrigated area has increased in the US¹⁰. In the SHP, over 85% of the irrigated area was watered with center pivot systems¹¹. While irrigation has been critical for boosting agricultural productivity and stability in the region, excessive extraction of groundwater has led to significant OA depletion and reduced outputs. This has forced many farmers to transition portions of their pivot circles to minimally irrigated or dryland areas¹². These partial pivots are less efficient in conserving and utilizing intense rainfall.

The increasing prevalence of partial pivots in the SHP presents an opportunity to explore a unique concept of Circular Perennial Grass Buffer Strips (CBS). This approach involves rearranging the minimally irrigated or dryland portion of the partial pivots into concentric circles of native perennial grass buffer strips alternating with crop strips to offer many benefits to the production system¹³ (Fig. 1). Native perennial grasses occurring naturally in the region are well-suited to the region’s semi-arid climate, have deep root systems, high drought tolerance, extended growing seasons, and low input and management requirements. Conservation practices like perennial grass buffer strips or wind breaks, typically associated with rainfed agriculture, can be integrated into center pivot irrigated systems through CBS. This could provide agronomic and ecological benefits, such as enhanced rainwater conservation.

Perennial grass buffers positioned across slopes have been shown to increase rainfall infiltration by slowing down water runoff and increasing ponding in rainfed agriculture^14,15. The dense growth and flexible stems of grasses provide resistance to water movement, slowing runoff and creating a ponding effect that enhances water infiltration into the soil profile¹⁶. Even narrow grass strips, measuring 0.5 to 1.2 m in width, effectively reduce runoff from croplands¹⁵. Further, the wind moderation benefit of buffer strips would reduce evaporation loss of water from the soil surface^17,18. The strategic design of CBS, incorporating native perennial grasses, buffer placement across wind and water flow paths, multiple strips, and a circular layout, enhances these benefits. The circular design is expected to act as barriers for water runoff and incoming wind flow in all directions, while the use of multiple strips increases the system’s overall efficiency, extending benefits across the entire pivot area. The underlying hypothesis is that CBS can better retain the resources, such as rainwater and block off external stressors such as wind within the center pivot irrigated production system, thereby enhancing water availability and water productivity (WP).

Maize (Zea mays L.) is one of the major irrigated crops in the region¹⁹ characterized by its extensive cultivated area, high economic value, and substantial water requirements. To our best knowledge, no prior study has explored the storage and utilization of natural rainfall in maize production under center pivot irrigated systems. This study focused on two large and intense rainfall events ( ≥ 25 mm) in 2019 and 2020. The two-year field experiment was conducted to test whether the presence of perennial circular grass buffer strips would influence storage (conservation) and use of rainwater by irrigated maize (Zea mays L.) during intensive storm events. A secondary objective was to evaluate the impact of conserved rainwater on crop microclimate, water stress levels, growth, and WP in maize cultivated with and without CBS.

Weather Conditions

Average maximum and minimum air temperature during maize summer growing period (May to September) was 31 and 14 °C, 29 and 12 °C in 2019 and 2020, respectively (Fig. 2). Although 2019 season was slightly warmer, there was a heat wave around stress-sensitive tasseling stage (just after the rainfall observation period) in 2020 with maximum temperature reaching 40 °C (Fig. 2). The site received a total of 304 and 158 mm of rainfall during two respective growing seasons. Due to drier conditions, a total of 356 mm of irrigation was applied to both CBS and CT in 2020 compared to 267 mm in 2019.

Characterization of rainfall events

Rainfall received on 6th and 5th July in 2019 and 2020, respectively, were selected for this study because both events were heavy and intense, and happened around the same maize growth period. The two rain events were 50 and 42 mm, respectively, in 2019 and 2020. Both the events occurred when maize was at V14 growth stage and headed towards the highly water-sensitive tasseling growth stage. In 2019, the rain event started at 3:00 am on July 6 and lasted for 7 h and 15 min (Fig. 3). The peak rainfall rate of 40 mm h⁻¹ occurred during the 4:45 to 5:00 AM time slot (Fig. 3). There were 8 other 15-minute intervals when the rainfall rate was ≥7.6 mm h⁻¹. In 2020, the rain event was relatively more intense and 90% of it was received within 2 h and 15 min. During the first 30 min, 29 mm of rain was received with a peak rainfall rate of 68 mm h⁻¹. The overall rainfall rate of the two events was 7.7 and 18 mm h⁻¹. The American Meteorological Society classifies rain with an intensity ≥7.6 mm h⁻¹ as “heavy rainfall”²⁰. Further dissection of the rain data at 5-min interval revealed that the peak rain intensity of the two events was 58 and 107 mm h⁻¹ in 2019 and 2020, respectively (Fig. S1). Therefore, the two heavy and intense rainfall events were accompanied by near torrential rainfall periods (≥60 mm h⁻¹)²¹. The intensity and amount of rainfall received during both events were more than enough to cause runoff from the fine-textured Olton clay loam soil that is characterized by a low infiltration rate and organic matter²².

Rainfall conservation

Significant interactions between treatment and sampling distance were observed for rainwater storage during both rainfall events. Notable variations were detected at different soil depths and throughout the soil profile up to 1.4 m. In the CBS treatment, a similar amount of water was stored after the rainfall at all measured depths in both D1 and D2 (Figs. 4 and 5). In contrast, the CT treatment showed significantly less rainwater stored at the edge (D1) compared to further inside the control pivot (D2) (Figs. 4 and 5). Averaged over two distances, CBS increased soil profile water by 43 and 21 mm up to 1.4 m of depth after the rain events in 2019 and 2020, respectively, while CT only managed increases of 18 and 11 mm (Table 1). Of the total rainwater captured, 49 and 30% were stored at deeper soil depths (below 0.8 m) in CBS during 2019 and 2020, respectively. In CT, this proportion was 42% in 2019, and negligible in 2020. Profile recharge has been used as a primary metric for rainfall retention when surface runoff cannot be measured at the plot scale^14,23. Anderson et al¹⁴. monitoring a no-till corn-soybean watershed on a Putnam silt loam (a fine-textured claypan soil), reported 23% higher profile recharge in the top 0.5 m under agroforestry buffers (112 mm) than under row crop (91 mm) over a three-event recharge week with no runoff measured from the watershed. Sahin et al²³., on the same watersheds, reported that profile recharge under the agroforestry buffer accounted for as much as 101% of event rainfall, compared with only 41% under row crop, again with no runoff recorded over the recharge week. The interpretation rests on a mass-balance argument that when antecedent moisture is known and the profile is monitored to sufficient depth, rainfall not accounted for as profile recharge is assumed to have been lost from the monitored profile through surface runoff, deep drainage, or shallow lateral redistribution. In our study, CBS and CT shared the same Olton clay loam soil texture, conventional tillage, maize hybrid, and irrigation history up to the rainfall events, and antecedent soil water was measured before each event. Differences in profile recharge between CBS crop strips and CT during a given rain event are therefore likely influenced by the presence of adjacent perennial grass buffer strips, given that soil type and agronomic management history were comparable between treatments.

High-intensity rainfall can lead to rapid ponding by overwhelming the soil’s infiltration capacity, resulting in surface runoff²⁴. However, a barrier to the runoff that stabilizes the ponding results in greater and deeper infiltration by providing more time, and additional pressure exerted on the soil by ponding depth^25,26. Previous research comparing different types of buffer strips has highlighted grass buffer strips as the most effective in reducing field runoff¹⁶. Flexible stems and robust root growth provide resistance to above and below surface water movement in grass buffer strips²⁷. Native warm-season grasses such as switchgrass (Panicum virgatum), effectively slow down runoff and create water ponding in the adjacent areas²⁸. The present dataset (Figs. 4 and 5) is consistent with grass buffer strips acting as a physical barrier to surface runoff, thereby slowing overland flow and providing more time for rainwater to infiltrate the adjacent crop strip. Vegetative filter strips function primarily by increasing surface roughness and hydraulic resistance through the frictional drag of aboveground biomass (stems, leaves, and accumulated thatch), which dissipates the kinetic energy of overland flow and reduces its velocity^29,30,31. As runoff encounters the dense network of stiff grass stems, a hydraulic backwater effect is induced, creating water ponds upstream of the strip, concentrated flow paths are disrupted, and the water is forced to spread laterally into uniform sheet flow^32,33,34. This reduction in velocity and induction of ponding increases the contact time between the runoff and the soil surface, providing a greater opportunity for water to percolate downward into the profile³⁵. Infiltration enhancement through increased contact time is, in fact, consistently cited as the dominant mechanism by which grass buffers reduce total runoff volume^36,37.

Our data were consistent with this barrier-ponding-infiltration chain at three levels. First, profile recharge in CT recovered only 36% (18 of 50 mm) and 26% (11 of 42 mm) of event rainfall in 2019 and 2020, while the CBS crop strip recovered 86% (43 mm) and 50% (21 mm), even though both pivots received the same rainfall and had identical soil, crop, and management conditions. Given that deep drainage below 1.4 m on a fine-textured Olton clay loam over the 24–36 h between N₀ and N₁ is expected to be small, the most parsimonious interpretation consistent with the data is that the grass strips likely impeded surface runoff, allowing more water to enter and remain within the monitored soil profile. Second, the spatial pattern of recharge inside the crop strip provides insight into the underlying processes. Within CT, recharge at D1 (the outer pivot edge) was substantially lower than at D2, consistent with unimpeded runoff carrying water off the outer edge. Within CBS, recharge at D1 (located in the crop strip 1.5 m of the outer grass barrier) was statistically similar to D2, which is 9.1 m away from two surrounding grass strips. Because soil, tillage, hybrid, and irrigation history were identical at both CBS positions, this within-crop-strip equalization is most plausibly explained by the influence of the grass buffer barrier on D1. This spatial pattern is consistent with the hydraulic backwater and ponding mechanisms described in previous studies, where the barrier may have slowed runoff leaving the crop strip, promoted localized ponding in the immediate vicinity of D1, and increased the opportunity for infiltration of rainwater rather than being lost from the monitored crop strip^30,37,38. We did not measure ponding depth directly in this study. We infer that ponding contributed because the spatial moisture signature in the CBS crop strip matches the signature reported in studies that did measure runoff and ponding directly.

Third, a larger fraction of stored rainwater reached deeper layers (below 0.8 m) under CBS than under CT (49 versus 42% in 2019, and 30% versus negligible in 2020). Deeper redistribution is consistent with the prolonged surface contact time predicted when a grass barrier slows runoff velocity and induces ponding: the longer water remains on or near the soil surface, the greater the opportunity for vertical percolation into deeper horizons^23,35. However, because we did not measure infiltration rate or saturated hydraulic conductivity at the sampling locations, we cannot quantitatively partition the additional retention in the CBS crop strip into a runoff-reduction component (less water leaving the field) and an infiltration-enhancement component (more water entering the soil per unit time). Both mechanisms are well documented for grass buffer strips^23,36,38, and both are plausible contributors here.

Beyond their direct physical role in reducing runoff by acting as a barrier, grass buffer strips also potentially influence rainwater conservation through several indirect mechanisms. Before rainfall (N₀), the maize Leaf Area Index (LAI) in CBS was 0.44 and 0.56 in 2019, and 0.78 and 1.25 in 2020 at D1 and D2, respectively, compared to 0.29 and 0.54 in CT for 2019, and 0.42 and 0.99 in 2020. The semi-arid SHP is a windy region with high-speed winds that negatively impacts water, soil, and plant growth^13,39. CBS treatment reduced near-surface wind speeds by up to 62% as compared to CT, thereby decreasing wind stress and conserving surface soil moisture, particularly at D1⁴⁰. Notably, both rainfall events occurred during the tasseling stage of maize, when plants had developed substantial biomass and leaf area. Raindrop impact is known to create a “crust” or “soil seal” that impedes infiltration of rainwater, whereas standing and flat crop residue improves infiltration by reducing kinetic energy of raindrops^41,42,43. Improved maize growth conditions in CBS compared to CT might have indirectly supported rainwater conservation. A denser maize canopy may have intercepted a greater fraction of raindrops and reduced their kinetic energy at the soil surface, potentially mitigating the surface crusting that suppresses infiltration on fine-textured soils.

Given the current and future scenario, where irrigation demand and water availability are increasingly at odds in regions like the SHP, enhancing rainfall capture and soil storage is vital for the sustainability of irrigated cropping systems. Even during the drier 2020 season, when the site received only 141 mm of rainfall during the March-July period and aboveground growth of perennial grass buffer strips was 29% lower as compared to 2019 (data not shown), CBS reported a 91% increase in the storage of rainfall in the soil profile. These findings offer promising evidence that adopting CBS can help farmers extend irrigation intervals after rainfall events, potentially conserving groundwater and supporting economically viable, long-term irrigated production. The approach may also be relevant to other center-pivot systems in the northern and central Great Plains, although additional site-specific testing would be needed. The approach could be adapted to address region-specific challenges such as declining groundwater, runoff contamination of adjacent water bodies, and erosion of nutrient-rich topsoil.

Crop micro-climate dynamics

Crop micro-climate includes atmospheric conditions within and/or above the crop canopy, and soil profile conditions such as moisture and temperature. In this study, atmospheric conditions like air temperature and relative humidity at 10 cm height from the surface and soil moisture and temperature at 5 cm depth were measured to evaluate and compare the micro-climate around the two rainfall events in CBS and CT. Results showed that the mean air temperature (AT) within the corn canopy was up to 0.4 and 0.3 °C cooler in CBS as compared to CT during the two observation periods in 2019 and 2020, respectively (Fig. 6a). This cooling effect was more pronounced at D1 in 2019, with AT in CBS being 0.6 °C lower than CT. Furthermore, CBS exhibited narrower ATs ranges at both measured distances, indicating reduced temperature variability relative to CT. The mean relative humidity (RH) within the canopy was slightly higher in CBS as compared to CT (Fig. 6b). The minimum values for RH during the observation period were lower in CT as compared to CBS. Looking at the surface soil conditions, the mean soil moisture (SM) at 5 cm depth during the observation period was considerably higher, with D1 showing as much as 35% higher moisture in CBS as compared to CT (Fig. 6c). Whereas the soil temperature (ST) was lower in CBS as compared to CT at both the distances, with the difference reaching a maximum of 1.2 °C at D1 in 2019 and at D2 in 2020 (Fig. 6d). This trend for ST was mirrored in both the mean and median values (Fig. 6d).

Data suggests that the microclimate within the crop strip of CBS exhibited a lower evaporative demand and created an environment more favorable to maize growth, as characterized by its reduced variability, cooler temperatures, and higher humidity compared to CT. Greater rainwater conservation in CBS facilitated increased water uptake and transpiration by crop plants, regulating the atmospheric conditions within the crop canopy. Soil moisture plays an important role in surface energy balance through its impact on the evapotranspiration⁴⁴. Wetter soils, due to their higher thermal conductivity and heat capacity, contribute to a more balanced micro-climate by facilitating moisture evaporation during warmer times, whereas storing energy during cooler times. Further, the presence of grass buffers substantially attenuated the velocity of incoming winds at the boundary of the crop area (Fig. 6e), thereby reducing kinetic energy and diminishing soil moisture loss via evaporation. Consequently, these improved hydrological dynamics in CBS led to more efficient thermal regulation, manifested in lower air and soil temperature (AT and ST) and diminished fluctuation, thereby creating conditions more conducive to enhanced crop growth and development.

Crop response

Soil water extraction following the two large and intense rainfall events were more in treatments where more rainwater storage was observed. For instance, in the CBS treatment, maize extracted 20 and 19 mm of rainwater from deeper depths (below 0.70 m), where 23 and 20 mm of rainwater storage was recorded in 2019. In contrast, in CT, maize extracted only 5 and 15 mm of water at D1 and D2, respectively, where storage was limited to 6 and 10 mm. In 2020, negligible soil water extraction occurred from deeper depths in CT due to the absence of rainwater storage. In total, CBS maize extracted 48 and 19 mm of soil water, while only 29 and 10 mm of water was extracted in CT over the subsequent 7 and 4 days following two rainfall events in 2019 and 2020, respectively. The amount of water available in the soil profile is a key factor influencing water uptake by plants. Almost all the rainwater stored was extracted, and the water status of the soil profile returned to pre-rainfall (N₀) levels (Figs. 4 and 5).

A significant interaction between treatment and distance was observed for leaf water potential (LWP) in 2019. In CBS, there was no significant difference was observed between the D1 and D2, with LWP values of −1.9 and −1.8 MPa, respectively. However, in CT, maize at D1 had an LWP of −2.3 MPa, significantly lower than −2.0 MPa recorded at D2 (Fig. 7). In 2020, the main effect of treatment and distance was significant (Table 1). Compared to CT, LWP was 14% higher in CBS. Between the two distances, D2 had 19% higher LWP than D1. The LWP has been rigorously used to evaluate plant water status and stress index⁴⁵. Maize LWP measured on the 9 and 7 days after the two rain events (at N₂) was 10 and 13% higher in CBS than CT in 2019 and 2020, respectively (Fig. 7 and Table 1). This indicates that CBS maize was less water stressed compared to that in CT, few days after the two rainfall events. Both CBS and CT received the same amount of irrigation until the rainfall events. During the short study period (N₀ to N₂), rainfall was the primary water source, and the presence of a perennial circular grass buffer strip contributed to its conservation. Greater rainwater storage and subsequent extraction by maize were pivotal in reducing water stress in CBS. Previous studies on maize and safflower also reported an increase in LWP and relieve of water stress with extraction of stored soil profile water in semi-arid regions^46,47,48.

Water stress impedes crop physiological processes such as photosynthesis, which reduces the ability of the plants to produce additional biomass. Although crop plants can be tolerant to water deficit at certain growth stages, elevated water stress levels at critical stages like tasseling in maize can drastically impact plant growth and development. Since maize in CBS was relatively less water stressed, maize leaf photosynthetic rate measured on the same day was 56 and 11% higher in CBS than in CT in 2019 and 2020, respectively (Table 1). A decline in photosynthetic rate due to water stress is well documented in various crops^47,49,50,51 including maize^52,53. Consequently, relatively higher photosynthesis activity due to reduced water stress enabled maize in CBS to accumulate 1400 and 1230 kg ha⁻¹ of aboveground biomass, 41% higher than CT over the short measurement period of 16 and 10-day measurement period around the two rain events in 2019 and 2020, respectively. Additionally, between the two distances, D2 accumulated 21 and 35% more biomass than D1 during these periods in two respective years (Table 1).

Evapotranspiration (ET) and water productivity (WP)

To assess the efficiency of CBS and CT maize crop in utilizing heavy rainfall, short-term ET and WP were calculated for 16 and 10 days around the two rain events in 2019 and 202, respectively. Rainfall was the main source of water, contributing most to ET during these short periods. In 2020, the ET amount was similar to the rainfall received (42 mm) and a negligible contribution of soil water extraction was measured (Table 2). In 2019, the observation period was 6 days longer as compared to 2020. Due to this longer period, soil water extracted by maize exceeded the amount of rainwater (50 mm) received in 2019.

The biomass WP of maize was significantly different between the two systems in both years (Table 2). In CBS, maize had a WP of 27 and 29 kg ha⁻¹ mm⁻¹, which was 50 and 38% higher than the CT during the 2019 and 2020 seasons, respectively (Table 2). The presence of a grass buffer strip increased storage of intense rainfall in the soil profile. In contrast, only 27 and 38% of rainfall was retained in CT during the two respective events. Comparatively more amount of water available during high water demanding growth phase in CBS mitigated water stress levels and enhanced photosynthetic activity (Table 1). As a result, more biomass was produced in CBS compared to CT, which drastically improved WP.

The effect of distance on WP was significant in 2020, with D2 reporting higher WP than D1. Such a significant difference was not observed in 2019. In 2020, the air temperature reached 39 °C, the average wind speed was 3.6 m s⁻¹, and southwest was the predominant wind direction during the following 4 days after the rainfall. On the other hand, in 2019, the maximum air temperature recorded was 36 °C and the average wind speed of 2.5 m s⁻¹ with southeast as predominant wind direction over the next 7 days after the rain event. This resulted in a more stressful microclimate at the pivot’s edge in 2020, as reflected by a 23% lower LWP at D1 compared to D2, versus only 9% in 2019. Consequently, 2019 reported less biomass accumulation per unit of water at D1 than at D2.

Improving WP of irrigated production system is essential for ecologically achieving water and food security⁴. Irrigated agriculture consumes large amounts of blue water and is the primary reason for declining groundwater reserves⁵⁴. However, it contributes 40% to the global food production while using only 20% of total agricultural land³. A shift to dryland agriculture due to reduced irrigation capacity can decrease irrigated production by up to 47%⁵⁵. Therefore, a sustainable future of irrigated agriculture is crucial for achieving and sustaining food security. In intensive irrigated agriculture, very little importance has been given to the conservation of rainfall. However, with the declining nonrenewable groundwater, efficient use of renewable rainwater is becoming increasingly important⁵. Further, an improvement in rainwater conservation would provide ecosystem services, such as a reduction in rainwater runoff from irrigated farms would help mitigate non-point source pollution by reducing loss of applied chemicals and fertilizers⁵⁶ as well as soil erosion. Enhanced rainwater conservation in CBS increased water availability, reduced water stress levels, and improved growth of the irrigated maize plants especially during critical growth stage without relying on groundwater.

This study demonstrated that integrating native perennial grasses as circular buffer strips (CBS) within partial pivot systems significantly improved the capture and utilization of intense rainfall, increasing rainwater storage by up to 121%. Enhanced rainfall retention improved water availability, reduced crop water stress, and supported higher photosynthetic rates, resulting in greater biomass accumulation during critical growth stages. These findings highlight the potential of CBS to promote sustainable agriculture in semi-arid regions, particularly amid declining Ogallala Aquifer (OA) resources and increasing climate-induced water demands. By optimizing rainwater use, CBS may reduce reliance on groundwater and help extend the viability of OA-dependent systems. While this study focused on soil moisture responses around rainfall events, future research should evaluate additional factors—such as field slope, ponding depth, and runoff through larger-scale, full-pivot studies to better assess CBS performance across varied landscapes.

Location and experimental design

The study was conducted at the Agricultural Science Center (ASC) of New Mexico State University (NMSU) at Clovis, New Mexico (34° 36’ N, 103° 12’ W, elevation 1331 m) during maize growing seasons of 2019 and 2020, respectively. The soil type of the experimental field was an Olton clay loam (fine, mixed, super-active, thermic Aridic Paleustolls). The location has a semi-arid climate with an annual average precipitation of 445 mm. The trial was a large landscape study using southwest-facing wedges of two adjacent center pivots at the research center (Fig. 8a). The experimental field consisted of 1/4th of the pivot circle with circular grass buffer strips, hereafter called the CBS pivot and the 1/6th of the adjacent pivot without any buffer strip, called the control pivot (CT). In the CBS pivot, five perennial grass buffer strips of 9.1 m width are arranged alternately with 18.2 m wide irrigated annual crop strips. This large landscape-level study covered an area of 6 ha. Grass strips were planted on August 8th and 9th, 2016 and were fully established during the time of study. Both aerodynamic principles and the width of major farm implements were taken into consideration in deciding the widths of grass and crop strips. For example, a 9.1 m wide grass strip is two passes with a 4.5 m swather. Typically, aerodynamic literature reports wind moderation to a distance of up to 15 times the height of the barrier⁵⁷. Therefore, if grass strip grows 1.2 m tall, it should protect 18.2 m wide annual crop strip completely. The width of the crop strip was equivalent to three passes of a grain combine or of an eight-row planter. A mixture of tall-growing native grass species that included five warm and two cool-season species were used to establish grass buffer strips. Switchgrass (Panicum virgatum), Big Bluestem (Andropogon gerardi), Sand Bluestem (Andropogon halli), Sideoats Grama (Bouteloua curtipendula), and Indiangrass (Sorghastrum nutans) were 5 warm-seasons species. Tall Wheatgrass (Thinopyrum ponticum) and Western Wheatgrass (Pascopyrum smithii) were two cool-season grass species. A seed rate of 8 kg ha⁻¹ was used to establish grass buffer strips. They were given establishment irrigation during the first year to get good grass strips establishment. In 2019, 117 mm of irrigation was applied to grass strips to enhance regrowth, which was suffering due to drought conditions during spring and early summer. No irrigation was applied to grass buffer strips in 2020. Grass buffer strips were treated with a mixture of Atrazine @ 4.1 L ha⁻¹, Prowl H₂O @ 4.1 L ha⁻¹, and Express @ 0.02 L ha⁻¹) on 26 March 2020 to control broadleaf weeds. Another round of herbicide mixture of Roundup Power @ 4.7 L ha⁻¹, and Acuvart @ 0.4 L ha⁻¹) were applied to grass buffer strips on 29 April 2020. Once the warm-season grasses reached maturity, grass strips were swathed and bailed in August 2019 for animal feeding. However, drought like condition in 2020 impacted grass growth. Thus, instead of bailing, grass strips were shredded, and the residue was scattered over the strips in August 2020.

The area intended for maize planting in both control pivot and CBS pivot were conventionally tilled with a disk, followed by ploughing using DMI Ripper and land finisher just before planting. Maize (Zea mays L) was planted in the crop strips of the CBS pivot and the entire conventional control pivot on 7 and 1 May in 2019 and 2020, respectively. Maize Hybrids ‘Pioneer 1151’ and “Pioneer P1138 AMS” were planted during 2019 and 2020, respectively. Both hybrids belonged to a medium maturity group and were suitable for limited water availability. A John Deere 8-row planter was used to plant maize at a row spacing of 0.76 m. A population of 54,362 plants ha⁻¹ was used during both years. At planting, maize was fertilized with 207 and 179 kg ha⁻¹ nitrogen (N), 90 and 67 kg ha⁻¹ phosphorus, 29 and 33 kg ha⁻¹ sulfur, and 1.7 and 1.2 L ha⁻¹ of zinc in 2019 and 2020, respectively. A total of 267 and 356 mm of irrigation was applied to maize over the entire growing season in 2019 and 2020, respectively. Both CBS and control pivots were irrigated at the same time. Due to limited well output, the outer half of both pivots were irrigated first and then the inside half was irrigated. This ensured all sampling points in both pivots were similarly treated. A mixture of Balance Flex @ 0.1 L ha⁻¹, Atrazine @ 2.8 L ha⁻¹, and Brawl II @ 2.1 L ha⁻¹ in 2019 and Corvus @ 0.4 L ha⁻¹, Atrazine @ 2.8 L ha⁻¹, and Warrant @ 4.7 L ha⁻¹ in 2020 were applied as preemergent herbicide to maize in both pivots. Maize was sprayed with a mixture of Steward EC @ 5.9 L ha⁻¹, Oberon 45 C @ 3.7 L ha⁻¹, and Stratego @ 0.8 L ha⁻¹ to control spider mites and western corn root borers (adults) on 29 July 2019. In 2020, Roundup Power Max @ 4.7 L ha⁻¹, Onager @ 1.2 L ha⁻¹, and Warrant @ 4.7 L ha⁻¹ were applied to check Johnson grass and spider mites infestation in maize. Mention of trade names or commercial products, such as chemicals and cultivars, in this publication is solely for the purpose of providing specific information and does not imply recommendations or endorsement by the US Department of Agriculture. The USDA is an equal opportunity provider and employer.

Weather, soil water measurements, and evapotranspiration

The observation period for 2020 was slightly shorter (10 vs 16 days) compared to 2019 (10 vs 16 days). With a predicted heat wave coming over the next few days, relatively lower rainwater storage and depletion of stored rainwater, irrigation was applied relatively earlier in 2020 after the rainfall, reducing the observation period around the rainfall in 2020.

Weather information, including rainfall distribution, was collected from a ZiaMet Weather Station located at NMSU ASC, Clovis, NM. The weather station was located 0.6 km away from the experimental area. All the weather parameters were recorded at 5-min intervals.

Neutron access tubes of 1.5 m length were installed at two distances in CT, and crop strips of CBS to measure rainwater storage and use by maize (Fig. 8b). The first set of 3 tubes (D1) were installed 1.5 m from the inside edge of the grass strips in CBS and 1.5 m from the outer edge of the pivot in CT pivots. This allowed us to assess the grass barrier effect on slowing down rainfall runoff and increase infiltration due to the ponding effect. A second set of 3 tubes (D2) were installed 18.2 m from the outer edge of the maize in both the CBS and CT pivot. Within CBS, D2 is in the middle of the crop strip and at the farthest distance away from both outer and inner grass buffer strips which allowed us to test spatial aspect of CBS benefit in rainfall conservation. Similarly, D2 in CT is at the distance where the border effect on the crop growth diminishes and can differ in rainwater storage as compared to D1. Since both pivots are fairly leveled and only the slope is in the southwest direction, any difference between D1 and D2 in the two systems will be due to the grass barrier increasing ponding time in CBS by slowing down water runoff. While water use differences will be due to microclimate changes, due to CBS blocking desiccating wind entering the system. Thus, the placement of D1 and D2 in both systems allowed us to have a better depiction of rainwater storage and use in the outermost 18.2 m (24 maize rows) of the maize crop with or without buffer strip effect.

Prior to rainfall events, volumetric soil water content (SWC [m⁻³ m⁻³]) of the soil profile from 0 to 1.4 m depth at 0.2 m increment was measured using a calibrated soil moisture neutron gauge (Model 503 DR, Campbell Pacific Nuclear Inc., CA, US), referred to as N₀ in the text. Ideally, N₀ reading should be taken immediately before a rainfall event. However, it is not possible to predict the exact date and time of extreme rainfall events. Alternatively, seasonal neutron probe readings taken for another objective of the project were used for this study. Therefore, 1- and 7-day gap between N₀ and initiation of rain event was there in 2020 and 2019, respectively. A second set of measurements (N₁) were taken as soon as the field was dry enough to reenter after the rainfall event (24–36 h after the rainfall event stopped). Subtracting the second from the first soil water content measurements provided the amount of rainwater stored in the profile due to the extreme rainfall event. The third set of measurements (N₂) were taken seven and four days after the second measurement. The N₂ measurement was used to calculate the amount of stored rainwater extracted by the maize at two distances under the two systems (CBS and CT) during the short post-storm period. To calculate evapotranspiration during the short period around the two rain events in 2019 and 2020, respectively, we used the following equation:

where ET is evapotranspiration, I is irrigation, which was 0, RF is rainfall amount, and ∆SW is the change in soil water content (N₀ – N₂ measurements). All terms in Eq. 1 have units of mm.

Crop micro-climate, agronomic and physiological response, and rainwater use efficiency

Three micro-climate stations representing three replications were set up in the first crop strip of CBS and the outer edge of CT just after maize plants emerged. At each micro-climate station, a set of 6 sensors were installed with three each at two distances (D1 and D2). At each distance, a Gill 2D WindSonic anemometer, Steven’s Hydraprobe II, and Vaisala HMP60 sensors were installed to measure wind speed at the soil-air boundary layer (5 cm height), surface soil moisture at 5 cm depth, air temperature and relative humidity at 10 cm height. All the sensors were connected to a CR1000 datalogger powered externally using a car battery. The datalogger was programmed to collect micro-climate data at 15-min intervals using a scan rate of 9 seconds.

Along with the N₀ and N₂ SWC measurements, four randomly selected maize plants were hand harvested at two distances similar to locations of neutron tubes for soil moisture measurement in CBS and CT (replicated three times). Biomass sampling locations were far from neutron tubes and did not interfere with moisture readings. The increase in biomass (BMInc), measured as the difference between two readings, was the biomass accumulated by maize during the period. At the time of N2 measurement, leaf water potential (LWP) of freshly cut fully opened maize leaves were measured using a pressure bomb (Model 615, PMS Instrument Company, Albany, OR, USA) to assess crop water stress levels. Mid-day leaf photosynthetic rate (P_r, μmolm⁻²s⁻¹) was measured using a portable photosynthesis system (Model LI-COR 6400, Lincoln, NE, USA) to evaluate any change in the photosynthetic response of maize plants. Leaf area index (LAI) was measured using a SunScan canopy analyzer (Type SS1, Delta-T Device). Water productivity (WP; kg ha^-1 mm^-1) for the short measurement period after heavy rainfall was calculated as follows:

Statistical analyses

A linear mixed-effect model was used to determine the effects of treatment and distance on rainwater storage, maize water extraction, leaf water potential, photosynthetic rate, biomass increase, and WP. Treatment (CBS and CT) and distance (D1 and D2) were considered as fixed factors, while replication was used as a random factor nested within the treatment. The dataset of the two rain events were analyzed separately and tested for normality of residuals and homogeneity of variance. Non-normal parameters were transformed using an ultra-fine transformation procedure⁵⁸. For pairwise separation of least square means at p ≤ 0.05 and p ≤ 0.10, the pdmix800 procedure was used⁵⁹. All statistical analysis was performed using SAS 9.4⁶⁰.