Annual Irrigation Water Use for Arkansas Rice Production

Throughout Arkansas, Louisiana, and Mississippi, the three predominant land forms used for rice production are contour levees, precision graded or straight levees, and zero grade. Levees are typically approximately 30–45 cm (12–18 in.) in height and are pushed up after dry seeding the crop on grade every 3–9 cm (0.1–0.3 ft) to allow for a flood to be maintained on the paddy area between levees. Contour-levee fields have minimal land improvements, and the levees follow the natural contour of the land. Irrigation water is applied at the top of the field and cascades through the paddies through levee gates or spillways. Upper paddies must be filled before water is distributed to lower paddies. This water management system is used to maintain a flood without the additional cost of land leveling. Straight levees, as a result of precision grading, involve adjusting the grade to a desired slope before levees are pushed up. This approach facilitates the use of furrow irrigation on crops rotated with rice and is usually less costly to construct than zero-grade systems.

Traditional flooded rice production consists of a well or riser located in the highest-elevation portion of the field. Water spills into lower paddies as the upper paddies fill. In an alternative method, known as multiple-inlet irrigation, rather than the water being discharged directly into the highest paddy, a disposable, low pressure pipe (polypipe) is connected to the water source and laid either next to or across paddies extending down the field. Gates or holes are placed in the polypipe within each paddy so that all paddies are concurrently watered instead of receiving overflow from a higher paddy. Multiple inlet can be used on either contour-levee or straight-levee fields.

Zero-grade fields are precision graded in all directions to have little to no grade to minimize the flood depth variation. A small canal is built on three or four sides of the field. Water is pumped into the canal at a single or multiple sites, and the field is flooded from multiple sides. Most growers using this land form provide small ditches (drain furrows) across the field for faster water distribution. An advantage of this system is that the entire field can be quickly flooded and the flood depth is constant across the field. A disadvantage of zero grade is that these fields do not drain as well as sloping fields, which is a concern for crops rotated with rice. Therefore, many producers opt for continuous rice production in zero-grade fields. In Arkansas, Wilson et al. (2008) reported that between 2006 and 2008, approximately 51% of the total rice acres in Arkansas were in contour levees, 43% in straight levees, and 6% in zero grade fields.

The University of Arkansas Cooperative Extension Service established an interdisciplinary rice educational program in 1983 focused on maximizing returns through management intensity and integrated pest management. The Rice Research Verification Program (RRVP) was established to verify the profitability of University of Arkansas recommendations in fields with less than optimum yields or net financial returns. The program goals include the following: (1) to educate producers on benefits of utilizing University of Arkansas recommendations for improving yields and/or net returns; (2) to conduct on-farm field trials to validate research-based recommendations; (3) to help researchers identify production areas that require further study; (4) to improve or refine existing recommendations to increase profitability; and (5) to utilize data from the RRVP in Extension educational programs at both the county and state levels. Producer check-off funds through the Arkansas Rice Research and Promotion Board support the RRVP (Vories et al. 2006).

Since 1983, over 401 commercial rice fields in more than 30 rice-producing counties in Arkansas have been included in the RRVP. The RRVP cooperators and fields are selected before the growing season. Cooperators agree to pay production expenses and provide expense data, as well as to implement the current University of Arkansas management and timing recommendations from planting to harvest. The RRVP coordinator and a designated county agent collect data, scout the field, and maintain regular contact with the producer (Vories et al. 2006). The importance of this effort is underscored by the persistent reduction in groundwater in the region and reports of unsustainable pumping from the alluvial aquifer for agricultural use (Fugitt et al. 2011).

During the first 8 years of the program (1983–1990), the water supplies of 42 program fields were equipped with flow meters that recorded water use during the permanent flood period (Vories et al. 2006). Vories et al. (2006) reported that from 1983 to 1990 overall irrigation water use ranged from 381 to 1,714 mm (15.0 to 67.5 in.), with an average of 879 mm (34.6 in.). Water use did not differ between predominant soil types—nine clay fields used on average 902 mm (35.5 in.), whereas 25 silt loam fields used on average 907 mm (35.7 in.). After 8 years, water use data from the rice fields were collected less frequently; sparse data were collected from 1991 to 2002.

Beginning in 2003, a new effort was initiated to evaluate whether irrigation water use had changed over the years. In comparison to when the RRVP was initiated 20 years earlier, many of the newer cultivars were shorter maturing. Furthermore, following a multiyear educational program, many producers had adopted multiple-inlet irrigation. From on-farm comparisons during the 1999–2002 growing seasons, Vories et al. (2005) found a 24% reduction in irrigation water use, without a yield penalty, on fields utilizing multiple-inlet irrigation compared to fields practicing conventional flooding.

Vories et al. (2006) examined water use from 10 RRVP fields in 2003, 10 in 2004, and 13 in 2005, finding that irrigation water use ranged from 460 to 1435 mm (18.1 to 56.5 in.), with the 2003–2005 average of 780 mm (30.7 in.); annual averages were 724 mm (28.5 in.) in 2003, 621 mm (24.4 in.) in 2004, and 946 mm (37.2 in.) in 2005. Water use was similar among predominant soil types—17 clay fields used on average 781 mm (30.8 in.), whereas 13 silt loam fields used on average 763 mm (30.1 in.). In addition, Vories et al. (2006) indicated that long-term water use may be lower than the 3-year (2003–2005) averages suggested. The 2005 rice growing season was unusually dry, and most producers had to flush their fields at least once before applying the flood, after not needing to flush during the 2003 and 2004 seasons. In addition, several producers reported requiring more pumping to establish a flood in 2005 than they could ever remember, and if they ever lost the flood owing to pump failure or draining for straighthead management, they had difficulty reestablishing it. The Vories et al. (2006) study projected future growing seasons should indicate whether the values for 2005 were unusually high.

Smith et al. (2007) examined irrigation water use for rice production systems in Arkansas and Mississippi during the 2003 and 2004 growing seasons. In Mississippi, on average, rice required 895 mm (35.2 in.) of irrigation water. Contour-levee systems used 1,034 mm (40.7 in.), straight-levee systems used 856 mm (33.7 in.), and zero-grade systems used 382 mm (15.0 in.) of irrigation water. In Arkansas, contour-levee and straight-levee systems used 789 and 653 mm (31.1 and 25.7 in.) of irrigation water, respectively. Fields utilizing multiple-inlet irrigation systems reduced water use by 28% in Mississippi and 11% in Arkansas compared to systems utilizing a single-point irrigation system.

This paper presents on-farm water use information from 10 years of RRVP in Arkansas. This information will supplement previous shorter-term summaries, assist in structuring future water management programs, and is timely owing to the updated Arkansas Water Plan report released in 2014 (Arkansas Natural Resource Commission 2014). The release of the water plan has created interest in improving water management for rice production in the mid-south. Rather than detailing the precise water needs of the rice crop, these data incorporate all of the factors that affect the amount of irrigation water applied on Arkansas rice farms and as such will be useful to those that plan, design, manage, or construct irrigation systems for rice production.

Irrigation water use in the RRVP was measured using portable McCrometer propeller-style flow meters that are installed on grower fields during the crop season. The coordinators of the RRVP read the totalizers at the beginning, during, and end of the season to determine irrigation water use. Rainfall amounts at each location were collected using manually read rain gauges or tipping bucket rain gauges. Generally, coordinators visited sites each week during the growing season to make agronomic observations about the fields in the program. Annually, there are between 15 and 25 fields tracked in the program. The main purpose of the RRVP is to verify and promote university Extension recommendations to farmers. The water use data is not a primary focus of the RRVP, but the collected data from this program and their reports were used to improve the understanding of water use trends and measured application depths in rice. The data set from the RRVP for 2003 to 2012 included year, soil texture (e.g., clay, silt loam), multiple inlet use information, water source (e.g., well, surface), irrigation water use, total water use, and rice yield. In addition, the data set from 2005 to 2012 included irrigation system type (contour levee, straight levee, multiple inlet, zero grade), and the data set from 2008 to 2012 included pump type (e.g., electric, diesel). All data used in this analysis were derived from the published annual RRVP reports (UACE 2016), were checked for completeness and reasonableness, and were cross-checked where possible between the published reports and manuscripts, spreadsheets, and other internal documents and files to ensure the data were reasonable and accurate. Inconsistent data were either excluded or corrected.

SAS software, version 9.3, was used to analyze the data using the general linear model ANOVA. Data collection was not consistent over years, which resulted in a need to group data parameters such as irrigation system, irrigation management, soil type, and pump type variables that could be compared across a number of years. There was no replication at each site; thus, the site was used as a replication within each year. Least-squares means (LSM) were determined, and Tukey’s honestly significant difference test was used to determine treatment differences at the significance level of $P<0.05$.

Three separate ANOVA tests were computed to appropriately analyze each of the irrigation system types for irrigation water use and total water use. First, the data from 2005–2012 were analyzed using soil texture and irrigation system type ($N=84$). Second, the data from 2005 to 2012 were analyzed using a soil texture, two of the three irrigation system types, and whether or not multiple inlet irrigation management was used ($N=72$). Fields using the zero-grade irrigation system were not included in this analysis because the multiple inlet variable does not apply to the zero-grade irrigation system; zero-grade irrigation systems inherently use a single water-entry point, and there is no expected water savings from using multiple inlet with a zero-grade system. Third, the data from 2003 to 2012 were analyzed according to soil texture and whether or not they used multiple inlet irrigation ($N=94$). The irrigation system type variable was not included in this analysis because it was not reported in the RRVP data from 2003 to 2004. Again, fields using the zero-grade irrigation system were not included in this analysis because the multiple inlet variable does not apply. The 2003–2012 data set was analyzed to determine effects of multiple inlet use on yield. In addition, the 2003–2012 data set was analyzed to determine the effects of water source on irrigation water use and irrigation pumping cost per acre ($N=105$). The 2008–2012 data were analyzed to determine effects of the pump type on irrigation water use and irrigation cost per acre ($N=53$). Yield by irrigation and total water use was analyzed ($\mathrm{yield}/{\mathrm{m}}^3$ irrigation or total water use) to determine if total water use was related to grain yield. The same trends and conclusions were drawn from this analysis as is presented subsequently, so these results were not included.

Irrigation water use by year, from 2003 to 2012, is presented in Table 1. From 2003 to 2012, irrigation water use by fields ranged from 254 to 1,880 mm (10.0–74.0 in.), with an average of 763 mm (30.0 in.). Irrigation water use was significantly higher in 2005 [985 mm (38.8 in.)] than in 2004 [621 mm (24.4 in.)] and in 2005 than in 2008 [620 mm (24.4 in.)]. There was significantly more rain (growing season only) in 2009 than in every other year from 2003–2012. In addition, 2005 had significantly less rain than 2004 and 2011. Total water applied (irrigation plus rain) of fields ranged from 559 to 2,283 mm (22.0–90.0 in.), with an average of 1,136 mm (44.7 in.). Total water applied was not significantly different between any of the years from 2003 to 2012. Rice yield was not related to irrigation water use or total water use (Fig. 1).

Irrigation water use from 2003 to 2012 was 17 mm (0.7 in.) less than the 2003–2005 average of 780 mm (30.7 in.) for Arkansas reported by Vories et al. (2006), and 132 mm (5.2 in.) less than the 2003–2004 Mississippi average of 895 mm (35.2 in.) reported by Smith et al. (2007). Growers may have improved their efficiency in overall water management or a short-term cycle of wetter climate may explain this trend. In addition, the Vories et al. (2006) data set may simply include more relatively dry years than this 2003–2012 data set. Additionally, the Vories et al. (2006) data set used preliminary data for 2005, and there was a small discrepancy between some of the data points compared to the published verification report. One concern is the potential for human error and lack of quality control of the data. For example, sometimes coordinators would estimate flushes because the meters were not installed in time or may not have known that the field was flushed. However, it is the best and only available historical data of its kind available for Arkansas and the mid-south. Furthermore, this information will help the researchers determine which factors to concentrate on in future studies.

This study investigated irrigation water use in rice from the University of Arkansas Rice Research Verification Program between the years of 2003 and 2012. For this time period, rice producers applied an average of 763 mm (30.0 in.) of irrigation water. This is 17 mm (0.7 in.) less irrigation water than reported by Vories et al. (2006) for 2003–2005. Water use ranged between 254 and 1,880 mm (10.0 and 74.0 in.). From 2003 to 2012, silt loam soils annually applied, on average, 145 mm (5.7 in.) more irrigation water than clay soils (845 versus 700 mm). A significant 40% water savings was reported for rice grown under a zero-grade irrigation system (486 mm) compared to contour-levee and straight-levee systems, although fewer zero-grade fields were included in the data set. No difference in irrigation water use was found between contour-levee fields (814 mm) and straight-levee fields (822 mm). This data set indicates that producers implementing the multiple-inlet water conservation practice on contour-levee or straight-levee systems, irrespective of soil type, did not realize a water savings but did experience a yield increase [$0.5\mathrm{t}/\mathrm{ha}$ ($10\mathrm{bu}/\mathrm{acre}$)]. Previous studies indicated a water savings potential from the adoption of multiple inlet (Smith et al. 2007; Vories et al. 2005, 2006). Therefore, the results from this study suggest that an educational effort may be helpful to improve the water and energy savings performance for those utilizing multiple inlet in Arkansas.

The authors would like to thank Phil Tacker and all the past verification coordinators and faculty project leaders throughout the history of the RRVP program for their efforts in collecting these data. These data would not have been available without the long-term financial support of the Arkansas Rice Research and Promotion Board.