Lawn irrigation with domestic wastewater by subsurface drip or sprinkling

To maintain the quality of a lawn and allow it to perform its functions (esthetics, enhancement of property, reduction of air temperature, etc.) efficient use of irrigation and adequate fertilization are necessary. The use of wastewater from sewage treatment, which has a nutrient load, in irrigation, can be an alternative especially if applied via subsurface, to reduce contamination risks. This study was conducted to examine the development and quality of an ornamental lawn of Zoysia japonica Steud (zoysiagrass) irrigated with wastewater depths, by two methods. The experiment was carried out in Botucatu-SP, Brazil, during winter and spring (June to December 2016). The treatments consisted of two irrigation methods (sprinkling and subsurface drip) and four domestic-wastewater irrigation depths (50%, 100%, 150% and 200% of the average daily evapotranspiration) plus control, where treated water was supplied by the Basic Sanitation Company of São Paulo State (Sabesp) of Botucatu, SP, Brazil, replacing 100% of average daily evapotranspiration. The following characteristics were studied: clipping dry matter weight, relative chlorophyll content, lawn height, total coliforms and Escherichia coli in the plant tissue and nutrient concentration in the leaves. Subsurface drip irrigation with wastewater, using an irrigation depth of 100% of the average evapotranspiration, allowed an adequate development of the zoysiagrass lawn by maintaining its quality, without contamination by total coliforms or Escherichia coli .


Introduction
Among various existing grass species and cultivars, zoysiagrass (Zoysia japonica Steud) is the most widely sold and recommended for residential areas in Brazil. However, to maintain lawn quality, efficient irrigation as well as adequate fertilization are necessary. Irrigated grassy areas are mostly serviced by sprinklers, especially the retractable type, which does not compromise the esthetics and allows free transit over the lawn; as well as by mechanized mowers. Nonetheless, significant losses may occur due to runoff, drift and evaporation (Abed El-Bast et al., 2016). When using a subsurface drip system, water is applied directly to the root zone, avoiding these problems (Leinauer and Devitt, 2013). There is, however, a risk of emitters getting clogged (Silva et al., 2017), mainly when wastewater is used.
The reuse of domestic wastewater has been practiced in agriculture and considered an excellent measure to mitigate the problem of water scarcity, especially in areas surrounding cities. Use of wastewater in agriculture has grown in the last decades, as it is a natural nutrient source proven to improve the yield of irrigated crops. Owing to its significant nitrogen levels (20 to 70 mg L -1 ) (Ferreira et al., 2019), domestic wastewater can be a fertigation option for lawns, as N is an important nutrient to maintain good lawn quality (green color, density and recovery from disease, drought and stress caused by wear) (Caturegli et al., 2016). In Brazil, few studies have investigated lawn irrigation with wastewater, mainly by subsurface dripping.
The aim of this study was to evaluate the development of a Zoysia japonica lawn fertigated with domesticwastewater irrigation depths by two methods, as an alternative for the proper use of domestic wastewater.

Material and Methods
The experiment was carried out at the Department of Soils and Environmental Resources of the Faculty of Agronomic Sciences of Botucatu -Unesp (São Paulo State, University "Júlio de Mesquita Filho") (22º50'48"S-48º26'07"W, 817 m above sea level). The climate of the region is a Cfa type (subtropical, mesothermal, humid), according to the Köppen classification, characterized by cold, dry winters and hot, humid summers.
The experiment was laid out in a split-plot design with four replicates, consisting of two irrigation systems (sprinkling and subsurface drip) as the main plot and four domestic wastewater (DW) irrigation depths (50%, 100%, 150%, and 200% of the average daily evapotranspiration [ADEv]) as subplots. Another treatment was installed (control), where treated water (TW) was supplied by the Sanitation Company of São Paulo State (Sabesp) of Botucatu, SP, Brazil, replacing 100% of the ADEv. Each main plot measured 80 m² and was divided into 20 subplots of 1.5 m × 1.5 m, with 0.5-m spacing between the subplots.
For the conventional sprinkler system, each subplot was equipped with four Rain Bird® 1804 retractable sprinklers (15 cm high) containing a nozzle adjusted to release the water at a 90º angle, flow rate of 0.07 m³ h -1 and pressure of 2 x 10 5 N m -2 . For subsurface drip, the XFS-09-12-500 hose was installed with in-line drippers, with a flow rate of 3.48 L h -1 , and a 17-mm diameter tube, with emitters protected against root intrusion by the Copper Shield™ technology. Each subplot contained four 1.5-m lines of the hose, which were spaced 0.375 m apart. Each row contained five drippers that were installed in a closed system, buried at a depth of 12 cm. For the automation of the irrigation system, two Rain Bird ® ESP -Modular controllers were used together with 100-DVF electric pulse (24 v) solenoid valves. Coupled to each controller, two rain sensors were installed to interrupt the irrigation whenever rainfall occurred exceeding 5 mm. After the irrigation systems were installed, the area was limed with 335 g m -2 dolomitic limestone (PRNT 95%) and fertilized one month later with 28.5 g m -2 of the 4-14-8 formulation. The Zoysia japonica cv 'Esmeralda TM ' was planted as sods immediately after fertilization.
Daily data on average daily evapotranspiration, precipitation and minimum, mean and maximum temperature were obtained from the meteorological station of Unesp during the experimental period, from May to December 2016 ( Figures 1A and 1B). The average daily evapotranspiration data were calculated using the Penman-Monteith equation, which was used to define the daily irrigation depth to be applied, referring to the average weekly evapotranspiration.
Lawn irrigation with domestic wastewater by subsurface drip or sprinkLing The domestic wastewater (DW) used in the experiment was provided by the Sewage Treatment Station of Botucatu, SP (Estação de Tratamento de Esgoto de Botucatu -ETE). The water was captured using a tank trailer coupled to a tractor and transported to the reservoir of the experiment. To pump the wastewater, a 0.5-hp motor-pump set with a working pressure of 2.458 x 10 5 Pa and a flow rate of 2.7 m³ h -1 was used. The easel at the irrigation outlet had a 38,1 mm 100 mesh disk filter and a pressure gauge. Irrigation with DW started in May 2016, according to the treatments. From May to December 2016, DW samples were collected monthly and sent to the laboratory for nutrient content determination, using the adapted methodology of Malavolta et al. (1997), as well as pH and electrical conductivity ( Table 1). Samples of treated water were also analyzed in May and December, following the same procedure.
The lawn was mowed fortnightly, to a height of 25 mm, using the GreenMaster 1000 TM (Toro) helical mower with a cutting collector box. Cuttings were collected from each subplot at 70, 135, 160 and 175 days after the start of treatments (DAS), packed in a paper bag and oven-dried for 72 h at 65 ºC until the quantification of clipping dry matter weight using a precision scale.
Relative chlorophyll content (RCC) was determined on the grass leaves using the FieldScout CM-1000 portable chlorophyll meter, which was positioned 1 m above the lawn, allowing a sampled area of 82.8 cm 2 . The RCC was determined at 60, 125, 145, and 160 DAS and recorded as the average of five readings, which were obtained at random in each subplot. Lawn height was measured using a Grass Height Prism Gauge TM (Turf-Tec). This measurement was carried out before some of the mowings (at 60,100,160,and 190 DAS).
In December, at the end of the experiment, total coliform and Escherichia coli counts were determined in the grass leaves. The area was prepared 15 days before the sample was removed and a 1-m tarp was installed, separating the irrigation systems so that there was no drift. Using gloves, masks and sterile scissors, 10 g of leaves were removed from the lawn of each subplot, starting with the control until the highest dose. These samples were packed in sterile plastic bags and sent to the Department of Microbiology and Immunology of the Institute of Biosciences at Unesp, Botucatu, SP, where a solution with 0.86% salt in deionized water was prepared. Ninety milliliters of this solution were added to the plastic bag, containing 10 g of plant tissue, which was then placed in a Stomacher TM 400 Circulator machine for 30 s, at 230 RPM. Subsequently, samples were sent to the water quality laboratory of the Department of Rural Engineering at FCA at Unesp, Botucatu, SP, for the count of total coliforms and Escherichia coli present in the samples, using the Colilert method (Lewis and Mak, 1989). Results were expressed as the most probable number (MPN) of organisms for 100 mL of sample. The cuttings obtained from the September mowings (135 DAS) were used to determine the macronutrient concentration, in accordance with the modified methodology of Malavolta et al. (1997).
Results were subjected to analysis of variance and, subsequently, to regression analysis, Tukey's test and contrast analysis at the 5% probability level using Sisvar TM software.

Clippings dry matter weight
The irrigation methods influenced dry matter weight on the first three evaluations, with higher values obtained with subsurface drip irrigation compared with sprinkling (Table 2). At 175 DAS, results were very similar. It is noteworthy that at 70 and 135 DAS, the treatment with DW by dripping provided a 3.2 and 3.9 times heavier clipping dry matter weight, respectively than sprinkling. Even when we compare only the systems with TW, drip irrigation provided greater grass growth in relation to the sprinkler, at all evaluation dates.  Lawn irrigation with domestic wastewater by subsurface drip or sprinkLing Table 2. Summary of analysis of variance of the results of clipping dry matter weight according to the treatment.  Although there was no significant interaction between irrigation systems and DW depths, regression analysis revealed a linear response of average clipping weight to the DW depths when applied via drip irrigation. Based on the angular coefficients of the equations, the plant biomass was estimated to increase by 0.16, 0.07, 1.84, and 0.64 g at 70, 135, 160, and 175 DAS, respectively, with every 1% of irrigation depth applied (Table 3). The higher dry matter accumulation rates at the end of the experiment were due to the higher air temperatures. The increasing grass growth observed with the increasing DW depths is due, in large part, to its nutrient load, considering that at depth of 100% ADEv, applied via drip irrigation, the grasses accumulated up to 10 times more mass when compared with treatment with TW at the same irrigation depth. Domestic wastewater contributed to the greater nutrient accumulation in millet plants, representing an alternative for reducing the demand for water and the use of chemical fertilizers (Reis et al., 2019). Thus, the largest accumulation of dry matter in zoysia grass was provided by the application of the highest DW depth via subsurface dripping. However, for ornamental lawns, increased growth is not a highly desired feature, as it will require a higher mowing frequency -in other words, greater maintenance. L -linear; Q -quadratic; ns -not significant; *, ** and *** -significant at the 5, 1 and 0.1% probability levels. Means followed by common letters do not differ from each other according to Tukey's test at the 5% level.

Lawn height
Lawn height differed between the irrigation depths from 100 days after the start of irrigation (DAS). For the irrigation method × depth interaction (M × D), differences were only seen at 60 DAS, with the highest values obtained at the irrigation depths greater than or equal to 150% ADEv in the drip method. Lawn height increased linearly as the levels were elevated in the subsurface drip system (Tables 4 and 5).  Lawn irrigation with domestic wastewater by subsurface drip or sprinkLing The plant height values indicate two very distinct periods: the first at 60 and 100 DAS, when the grass height ranged between 16 and 22 mm; and the second, which was characterized by the highest temperature, when plant height varied from 32 to 44 mm. It is important to stress, however, that the mowing period was different between the measurements, which were performed at intervals of 60, 40, 60, and 30 days after each mowing.
For the irrigation depth of 100% ADEv, a comparison can be made between the treatments where DW and TW were applied. The lawn was taller in the plots irrigated with DW, both by drip irrigation and sprinkling, when compared with the grass irrigated with TW. This greater plant growth can be attributed to the presence of ions, mainly N, found in larger amounts in DW (average 56 mg L -1 ), as N is considered the main nutrient responsible for shoot growth, in lawns (Gazola et al., 2016, Lima et al., 2018, Mota et al., 2019. Lawn height rose linearly as the DW depths were increased, regardless of irrigation method, i.e., the grass was taller where a greater water volume and a larger amount of nutrients were applied. As was discussed for the accumulation of dry matter, the increased height provided by higher irrigation depths is not a desirable characteristic for the grass, since this increased shoot growth will result in greater expenditure on mowing (Gazola et al., 2016). The heights reached by the lawn in the treatments were within the range deemed adequate by Patton et al. (2017) for grasses of the genus Zoysia (13 to 63 mm). According to Christians et al. (2016), it is recommended to maintain Zoysia japonica grass at a height of 15 to 30 mm. Using the 100% ADEv depth, grass height was maintained between 17 and 34 mm, which is very close to ideal, whereas higher depths would require a higher mowing frequency to maintain the ideal height ( Table 5). The 200% ADEv depth provided a grass height of 63 mm, at 160 DAS.
A caveat should be emphasized, however, that the plots with drip irrigation were not homogeneous in grass height. Therefore, the grass nearest the dripper was taller ( Figure  2). This effect can be minimized by reducing the distance between the drip lines, which, in this case, was 38 cm.

Relative chlorophyll content (RCC)
Higher RCCs were observed in the treatments using the drip irrigation method compared with the sprinkler, only at the beginning of the experiment (60 DAS). On all evaluation dates, the RCC of the lawn was higher when irrigation was performed using DW than with TW. The domestic wastewater had a N content of 40 to 67 mg L -1 , whereas the treated water contained only 2 to 3 mg L -1 of N (Table 1). Nitrogen is one of the main nutrients influencing the chlorophyll content of plants, as it is part of the chlorophyll molecule as well as of various compounds involved in the synthesis of the pigment (Malavolta et al., 1997).
In the sprinkling system, RCC increased during the evaluation period, with the highest value observed at 190 DAS. When the lawn was drip-irrigated, the highest RCC was attained earlier, at 145 DAS, indicating superior N utilization efficiency with this irrigation system. It must be considered, however, that there might have been a higher chlorophyll content in the grasses near the drip line (Figure 2), which generated these higher rates (Table 6).  Lawn irrigation with domestic wastewater by subsurface drip or sprinkLing There was a linear increase in RCC, which is attributed to the contribution from water. However, this result is mainly due to the nutrients in the water, especially N, present in a greater concentration-, besides Mg. Considering the angular coefficient of the generated equations, each 1% of the applied irrigation depth would provide increases of 1.2, 1.9, 1.2 in RCC at 125, 145 and 190 DAS, respectively (Table 7). Increasing RCC values in response to application of N-containing products was also observed by Gazola et al. (2016), Lima et al. (2018) and Mota et al. (2019). Lima et al. (2012) indicated, for bermudagrass cv. 'Celebration', a species with a darker green color than Zoysia cv. 'Esmeralda', adequate RCC values as 374 to 471 chlorophyll units. In the present experiment, from 145 DAS onwards, RCC values higher than those were obtained using irrigation depths equal to or greater than 100% ADEv. In other words, there is no need for irrigation depths greater than 100% to obtain an adequate green color (Figure 2). L -linear; Q -quadratic; ns -not significant; *, ** and *** -significant at the 5, 1 and 0.1% probability levels. Spk -sprinkling; Ssd -subsurface drip.
Based on the irrigation depth applied and the average N concentration present in DW, it was possible to calculate the amount of N applied to the soil in each treatment. In the eight months of experiment, the irrigation depths of 50%, 100%, 150% and 200% ADEv corresponded to 703; 1,364; 1,924; and 2,440 L of water per treatment in each subplot and the respective average N doses of 170, 340, 480, and 610 kg ha -1 . Thus, the irrigation depth of 100% ADEv, with DW, which was sufficient to maintain adequate green color as well as adequate growth (dry matter accumulation and height) corresponded to the N dose of 340 kg ha -1 . Schwartz et al. (2018) found the average N dose of 171 kg ha -1 as adequate for grasses of the genus Zoysia, for the state of Florida, USA, using mineral nitrogen fertilizer. In Brazil, Gazola et al. (2016) considered the N dose of 100 to 200 kg ha -1 as adequate for Zoysia cv. Esmeralda. The N dose applied with the irrigation depth of 100% of ADEv can meet the annual requirement of zoysiagrass in less than a year, even in sandier soil.

Total coliforms and Escherichia coli
In the application of DW by sprinkling, the MPN of total coliforms and Escherichia coli were at the upper limit of the reading table, that is, 214.3 × 10 -4 MPN/100 mL. The opposite occurred for the drip irrigation method, where the results were less than 100 MPN/100 mL, which was also found in the control treatment, without DW. Application by subsurface drip can make it difficult for pathogens to access the soil surface, keeping them close to the emitter, in the soil, and not in the plant tissue. Similar results were obtained by Cuba et al. (2015), who cultivated lettuce in a hydroponic system using domestic sewage effluent, where the shoots had no contact with the solution. Based on the obtained values, we may infer that application of water via drip irrigation, even with a high coliform content, does not reach the shoots, serving as a protection for the plant and those that benefit from it.

Leaf nutrient content
The P, B, Fe and Mn contents in the leaf varied with the irrigation methods, and the DW depths influenced the levels of N, P, K, Fe and Mn. The Ca, Mg, S, Cu and Zn levels in the zoysia grass leaves were not affected by the treatments (Tables 8 and 9).      Lawn irrigation with domestic wastewater by subsurface drip or sprinkLing Phosphorus was significantly affected by the M × D interaction, with higher contents obtained by sprinkling with DW compared with drip irrigation, at the highest depths (Table 10). The leaf Fe and Mn contents were also higher in the sprinkling treatments than with drip irrigation, both with DW and TW. This result may be related to the low mobility of P, Fe and Mn in the soil, which limits their absorption by plants closer to the drip line, and the foliar absorption of these nutrients, which is more effective than via soil, since water would always remain on the leaves after irrigation. Additionally, these nutrients have low mobility in the plant. As for B, which is a very mobile element in the soil, because of the lack of charge of its main absorbed form (B(OH) 3 ) (Dhassi et al., 2019), the highest leaf content of this nutrient was observed in the drip irrigation systems, both with DW and TW. No difference was detected between the irrigation systems for the leaf N content. However, considering only TW, a lower N content was obtained with the sprinkling system, despite the low N levels in this water (Table 1). Comparing only the treatments at the irrigation depth of 100%, between TW and DW, a higher N content was obtained with DW using both irrigation methods, and a higher P content was obtained with drip irrigation. Although DW has higher K contents than TW, the leaf K content was lower in the treatment with DW in the drip system. This effect may have been due to the higher Ca content in DW, which can occupy the colloid charges, making K more susceptible to leaching and, therefore, less available to the plant, besides competitive inhibition during absorption (Fernandes et al., 2019).
There was a linear increase in the leaf P, K and Mg contents as the reused-water irrigation depths were increased when applied via sprinkling; and in the leaf N content regardless of the irrigation system, which shows that DW can be an efficient means to provide these nutrients for grasses. In terms of micronutrients, the leaf Fe and Mn contents decreased with the increasing depths applied via sprinkling, probably due to the competitive inhibition with sodium, which is the element present at the highest concentration in DW, together with N ( Table 1).
The treatments with DW irrigation depths greater than 100% ADEv provided adequate leaf nutrient contents, except for S and Zn, in which the leaves were deficient (Shadoxx, 2017). Regarding Zn, the presence of this nutrient in DW was not detected in the analysis, whereas S, despite being present in DW at concentrations higher than P and Mg, was not sufficient to meet the plant's requirements, likely because it is easily leachable and grasses have their roots more concentrated in the surface layer. Very high Fe, Mn, and Cu contents were also found in the leaf, which was not due to irrigation with DW, since the treatments with TW also showed high levels of those nutrients, but rather due to contamination with soil particles.

Conclusions
Subsurface drip irrigation with wastewater at 100% of the average daily evapotranspiration depth allowed proper development of lawn with Zoysia japonica (zoysiagrass), maintaining its quality, with no contamination by total coliforms or Escherichia coli.

Author contribution
RBM: preparation of the introduction, methodology and tables of results. Interpretation and discussion of results; RLVB: elaboration of the methodology and critical review; LJGG: contribution to the introduction, statistical analysis, interpretation and discussion of results, and critical review; RLS: formatting of the article and contribution to the discussion of results.