A simplified method to improve water distribution and application uniformity for sprinkler irrigation on sloping land: adjustment of riser orientation

To improve the water application uniformity for sprinkler irrigation on sloping land, indoor tests were conducted on an artificial slope (slopes of 0, 0.05, 0.10 and 0.15) to evaluate the effects of two riser orientations, vertical (VO) and perpendicular (PO) to the slope, on the uniformity of sprinkler rotation, radius of throw, water distribution of an individual sprinkler and the overlapped water application uniformity (WAU). Compared with the VO, the PO could effectively improve the water distribution on sloping land and minimize the risk of soil erosion. Additionally, the PO was superior in the WAU, and a rectangular arrangement could dramatically enhance the WAU at smaller sprinkler spacing, while larger acceptable sprinkler spacing was accepted in a triangular arrangement. The riser orientation and sprinkler spacing had the most significant effect on the WAU, followed by the slope and sprinkler arrangement, suggesting that the adjustment of riser orientation or sprinkler spacing was helpful in improving the WAU. However, from the aspects of investment cost and installation convenience for irrigation projects, the method of PO was recommended. Therefore, when designing the sprinkler irrigation systems on the slope, choosing PO is the simplest and most effective way to achieve good irrigation uniformity.


GRAPHICAL ABSTRACT INTRODUCTION
China has 135 million ha of cultivated land, of which slope cultivated land accounts for 33.33 million ha, approximately 1/4 of the total cultivated land area. Sloped farmland is the food production base for the survival and development of people in China's hilly or mountainous areas. Due to the long-term influence of terrain slope, the soil on sloping farmland is sensitive to drought, which drastically reduces crop yield and quality (Zhang et al. ). Thus, it is of great importance to determine a reasonable irrigation method for crops on sloping land. Compared with traditional irrigation methods, sprinkler irrigation on sloping land has been shown to have both higher efficiency and better adaptability to the terrain (Elwadie et al. ).
Sprinkler distribution is an important parameter in the design of irrigation engineering, and a key indicator for evaluating the water application uniformity of irrigation systems (Zhu et al. ). There are many factors that affect water distribution, including the sprinkler head, distribution system, and climatic and management factors (Mateos ). Because of the terrain slope, the water distribution of a single sprinkler on sloping land is quite different from that on flat land. When no wind is present, the curves of water distribution on flat ground approximate a set of concentric circles centred on the location of the sprinkler, and the water application rates at the same distance from the sprinkler are almost equal. However, when the sprinkler is used on sloping land, the water distribution patterns approximate a set of eccentric circles, and at the same distance from the sprinkler, the water application rate for uphill is greater than that for downhill, thereby resulting in poor water application uniformity (Zhang et al. ).
Extensive research has been conducted on sprinkler water distribution patterns and their application uniformity on sloping land. Montazar & Moridnejad () revealed that soil moisture uniformity was more sensitive to the water application uniformity than to terrain slope. uniformity of slope irrigation, adjusting the orientation of the sprinkler riser has been gradually applied in recent years, due to its simple operation and low investment cost.
A preliminary analysis on the effects of sprinkler riser orientation on rotation uniformity and throw radius was investigated by Li (), but the effects on water distribution and application uniformity were not analysed. Soares et al. () established a model to evaluate the sprinkler water distribution under different slopes and riser orientations and found that the riser perpendicular to the soil surface could maximize the uniformity of water application. However, this model was developed using water data collected on flat ground combined with ballistic theory, and there were differences from the actual water distribution of the slope.
Accordingly, the effects of the riser orientation on the sprinkler water distribution and application uniformity on sloping land still requires further investigation.
In this study, an experiment on water distribution patterns was conducted under two commonly used riser orientations (vertical (VO) and perpendicular (PO) to the slope) and four slopes (0, 0.05, 0.10, and 0.15). The objectives were (1) to evaluate the effects of the two riser orientations on the sprinkler rotation uniformity, radius of throw and the water distribution of individual sprinklers on sloping land; (2) to investigate the overlapped water application uniformity (WAU) of multiple sprinklers with various riser orientations, sprinkler arrangements, spacings and slopes; and (3) to further propose a new method to improve the water application uniformity for sprinkler irrigation on sloping land. The flowchart of research methodology for this study is shown in Figure 1.

Experimental setup
The experiment of sprinkler irrigation on sloping land was conducted indoors under no wind conditions. The experimental slope surface was artificially constructed using height-adjustable brackets and steel channels. There were a total of 11 rows of steel channels of 12 m in length. The horizontal distance between the two adjacent rows was 1 m. The brackets were installed under the steel channels, and their heights were adjustable according to the test slope. To obtain the sprinkler water distribution on sloping land, catch cans with an opening diameter of 10.6 cm and a height of 15 cm were placed in the steel channels and arranged in a grid pattern. The grid size on the ground was 1 × 1 m. In the experiment, 11 catch cans were arranged on each row of steel channels, so there were 121 catch cans in 11 rows of steel channels ( Figure 2). An LF1200 rotating sprinkler (Rain Bird Corp., Azusa, California, USA), commonly used in agricultural irrigation, was selected for this study. The sprinkler has a 2.18 mm nozzle diameter and 17 jet angle, with a recommended operating pressure ranging from 210 to 410 kPa. In addition, a pressure transducer (Xi'an Xinmin model CYB, accuracy of ±0.1%) with a range from 0 to 500 kPa was installed at the sprinkler inlet and connected to a data logger. The pressure was recorded at 5 s intervals during each 1 h sprinkler test, and the average value was calculated for each test.

Experimental design
The sprinkler was tested with two riser orientations of VO and PO and four terrain slopes of 0, 0.05, 0.10, and 0.15. This setting was selected mainly because when the riser was PO, the riser angles (inclination of the riser from the vertical) corresponding to the various slopes (0, 0.05, 0.10 and 0.15) were 0 , 2. 86 , 5.71 , and 8.53 , respectively, in accordance with the requirements of ISO 7749-1 (ISO standards ) that the riser angle should not exceed 10 .
In total, there were seven trials in the overall experiment, and each of them were performed with three replicates in order to obtain reliable experimental data. During the sprinkler test, the working pressure of the sprinkler was stabilized at the designed pressure of 300 kPa, indoor air and water temperature were approximately 30 C and 26 C, respectively, and relative humidity was 60%. The test indicators included sprinkler rotation speed, radius of throw, as well as water distribution on sloping land.

Test of water distribution and throw radius of an individual sprinkler
A full grid collector array method was used to test the sprinkler water distribution on sloping land. For convenience of the test, the sprinkler was installed at the bottom and the top of the slope surface, respectively. The height of the sprinkler riser was 30 cm, according to the manufacturer's recommendation. At the given operating pressure of 300 kPa, the sprinkler was first installed at the top of the slope, and the water distribution on the downhill was recorded. After a 1 h test, the sprinkler was installed at the bottom of the slope, and the water distribution on the uphill was recorded under exactly the same experimental conditions. The combination of the water distribution on the downhill and uphill was the water distribution on the whole slope surface.
Additionally, according to the ISO 15886-3 (ISO standards ), the radii of throw on the slope was determined by measuring the distances to the farther points with the minimum effective water application rate of 0.26 mm h À1 .

Data analysis
The maximum relative sprinkler rotation deviation (MRD) (Li & Kawano )  The MRD can be calculated by Equation (1): where T j is the sprinkler rotation times of jth quadrant, s, and T is the average sprinkler rotation time of four quadrants, s.
The CU can be calculated by Equation (2): where h i is the measured water depth from an individual catch can, mm; h is the average measured water depth of all catch cans, mm; and n is the number of catch cans.
The DU can be calculated by Equation (3): where h m is the average of the lowest 1/4 of ranked catch can measurements, mm.

Uniformity of individual sprinkler rotation
Sprinkler rotation uniformity is a key factor affecting the quality of sprinkler irrigation systems (Li & Kawano ).
Generally, when the riser was VO, the sprinkler had uniform rotation speed, whereas the actual rotation speed varied when the riser was inclined from the vertical (Soares et al. ). The sprinkler rotation time per quadrant and MRD value versus terrain slope when the riser was PO are presented in Table 1. Overall, sprinkler rotation periods among the four slopes were basically the same, whereas a large variation was found in the MRD values. For instance, the rotation periods under the slopes of 0, 0.05, 0.10 and 0.15 all remained at approximately 18.08 s, but the corresponding MRD values were 0.96%, 3.19%, 6.81% and 9.21%, respectively (Table 1).
This finding indicated that when the riser was PO, the increase in terrain slope caused the sprinkler rotation uniformity to decrease, regardless of the minimal effect on the sprinkler rotation period, which was consistent with the find-

ings of Nderitu & Hills () and Li & Kawano ().
Further analysis showed that the rotation times of T 1 and T 4 were always less than that of T 2 and T 3 at the slopes of 0.05, 0.10 and 0.15, which meant that when the One possible explanation for this performance was that the riser inclined from the vertical led the gravity centre of the sprinkler to shift, so that the sprinkler rotating torque changed inevitably during the full circle spraying process.
The sprinkler slowed down due to the reduction in the rotational torque when the nozzle turned towards the uphill direction, while the nozzle turned towards the downhill direction, a reverse result was observed.
In addition, from the analysis, the maximum MRD value appeared at the slope of 0.15, reaching 9.21%. This value was lower than the maximum allowed MRD value of 12% stipulated in Agricultural Irrigation Equipment-Rotating Sprinklers (ISO - ), suggesting that in this study, as the riser was perpendicular to a slope not exceeding 0.15, the uniformity of the sprinkler rotation could be guaranteed.

Throw radius of an individual sprinkler
The throw radius of a sprinkler plays an important role in the optimal selection of sprinkler spacing and lateral spa- This finding was partly because this kind of riser orientation changed the initial jet angle of the nozzle, making the uphill and downhill jet trajectories similar to that of flat ground. As a result, the PO could effectively minimize the influence of terrain slope on the sprinkler throw radius and reduce the difference between the uphill and downhill throw radii, which was conducive to the design of these sprinkler irrigation projects.

Water distribution of an individual sprinkler
The riser orientation affects not only the rotation uniformity and throw radius of a sprinkler but also the trajectory of spray jet towards different directions, thereby affecting the sprinkler water distribution (Seginer et al. ). Figure 3 presents the water distribution patterns of individual sprinklers under the two riser orientations and four terrain slopes.
The coordinate (0, 0) marks the location of the sprinkler.
When the riser was VO, the water distribution on sloping land differed drastically from that on flat ground. From the perspective of water distribution shape, it was similar to a 'heart-shape' for the slopes of 0.05, 0.10 and 0.15, whereas it resembled a 'circle-shape' for the slope of 0 (Figure 3 This result was primarily caused by the change in water jet landing time due to the influence of terrain slope, as in the aforementioned analysis. In contrast, when the riser was PO, the influence of the terrain slope on the water jet trajectory was greatly reduced. It was clear that as the slope increased, the water distribution curve on sloping land was gradually transformed from a 'heart-shape' to a more circular shape, and it finally approximated the water distribution curve on flat ground (Figure 3(b)).
Additionally, from the perspective of water application rate (WAR) on sloping land, the higher WAR was closer to the sprinkler and the lower WAR was further away from the sprinkler, regardless of whether the riser was VO or PO In an attempt to quantitatively study the water distribution of individual sprinklers on sloping land, the distribution proportions of water application rates (P WAR ) under the two riser orientations and four terrain slopes are presented in Figure 4. It was notable that most WAR values were below 1.5 mm h À1 for the different combinations of riser orientation and terrain slope, covering more than 83%. This finding demonstrated that the overall WAR values were relatively uniform, although the water distribution shape was affected by the riser orientation and slope, which might be attributed to the spray pattern of a single jet rotating sprinkler. From the aforementioned analysis, the WAR values on the slope were mainly distributed within 1.5 mm h À1 , but the applied water in the range of 1.5-4.0 mm h À1 must also be considered, because high WAR values are prone to cause soil erosion (Levy et al. Figure 4 revealed that, in terms of the WAR values within 1.5-4.0 mm h À1 , the P WAR of the PO was lower than that of the VO. Using the three slopes (0, 0.05 and 0.15) selected in this study as an example, the P WAR within 1.5-4.0 mm h À1 was 13.0%, 13.9% and 16.7% for the VO, respectively, whereas they were 12.0%, 13.5% and 13.2% for the PO, respectively. In summary, the PO could not only improve the water distribution of an individual sprinkler on sloping land but also minimize the risk of soil erosion,   respectively. It was not difficult to find that, with the rectangular arrangement, the CU and DU values for the PO were on average 4.0 and 5.5 percentage points higher than those for the VO, while with the triangular arrangement, the values were 5.0 and 6.5 percentage points higher. One possible explanation for this performance was that when the riser was VO, the sprinkler produced a more irregular profile on the slope than that of the PO (Figure 3), and this consequently influenced the overlapped distribution pattern, as previously reported by Nderitu & Hills ().    Furthermore, the influence of the slope on the WAU also cannot be ignored. When the riser was VO, the WAU increased with the higher slope at the same sprinkler spacing,   Despite the water distribution shape of individual sprinklers was affected by the riser orientation and slope, the overall WAR on sloping land was relatively uniform for both the VO and PO, and their WAR values below 1.5 mm h À1 all covered more than 83% at different combinations of riser orientation and terrain slope. Additionally, the distribution frequency of the high WAR values within 1.5-4.0 mm h À1 was observed to be lower for the PO than for the VO, suggesting that the PO was conducive to minimizing the risk of soil erosion.

Further analysis of
The PO had an absolute superiority in overlapped WAU compared to the VO. Its mean CU and DU coefficient with rectangular arrangement were 4.0 and 5.5 percentage points higher than those for the VO, while with the triangular arrangement, the above values were changed to 5.0 and 6.5 percentage points. Moreover, for the PO, the rectangular arrangement could dramatically enhance the WAU at smaller sprinkler spacing, while the triangular arrangement had larger acceptable sprinkler spacing (CU ! 75%, DU ! 60%); thus, both arrangements should be used in combination for the design of sprinkler irrigation projects on sloping land.
From the ANOVA results of the effect of various factors on the WAU, the riser orientation and sprinkler spacing were found to have the most significant effect on the CU and DU values, followed by the slope and sprinkler arrangement, revealing that the adjustment of the riser orientation or sprinkler spacing was important in improving the WAU values, regardless of the sprinkler arrangement and slope.
However, in view of investment cost and installation convenience for irrigation projects, the method of PO is a worthy recommendation. In summary, in order to obtain a good irrigation uniformity when designing the sprinkler irrigation system on slopes, choosing PO is undoubtedly the simplest and most effective way.
This paper presented the comparison results of the water distribution and application uniformity for the two riser orientations using a Rainbird LF1200 rotating sprinkler on an artificial slope. These results provided a basis for the design of sprinkler irrigation systems on sloping land, but there were still some limitations. In this paper, a rotating sprinkler with a unique structure was selected for experimental research. The water collected in catch cans were not weighed timely during the tests; this might lead to the measurement errors due to water evaporation. This study was conducted under indoor conditions, ignoring the influence of external factors such as wind speed and temperature, which might affect the design of sprinkler irrigation systems in the field. In addition, this study did not involve the soil water movement on sloping land, and further research is therefore required. ACKNOWLEDGEMENT