Effects of different amendments (organic matter and hydrogel) on the actual evapotranspiration and crop coefficient of turf grass under field conditions *

The irrigation schedule in arid areas has to be efficient in order to reduce losses due to evaporation and deep infiltration. Irrigation optimization poses the need to establish with precision the value of actual evapotranspiration (ETa), and the crop coefficient (Kc). The water soil availability can be increased using hydrogel and organic matter amendments, and their effects could vary ETa and Kc. The aim of this study was to determine the ETa, and Kc of an experimental site with lysimeters on the Spanish Mediterranean coast cropped with a turf grass variety, Agrostis stolonifera ‐L‐93, under field conditions, and amended with hydrogel and organic matter.


M O T S C L É S
lysimètre, coefficient d'agriculture (K c ), évapotranspiration (ET a ), matière organique, climat méditerranéen 1 | INTRODUCTION Evapotranspiration is the combination of two separate processes whereby water is lost from the soil: evaporation and transpiration. Evaporation consists of the vaporization of water due to solar radiation, temperature, wind and other meteorological factors, and transpiration consists of the vaporization of liquid water contained in plant tissues and its removal to the atmosphere. Since both processes occur simultaneously and there is no easy way of distinguishing between them, they are compiled in a single term: evapotranspiration. Evapotranspiration can be measured with experimental lysimeters (actual evapotranspiration, ET a ) or estimated from meteorological data (reference evapotranspiration, ET 0 ).
The ET a can be determined from the water balance and depends on the type of crop, the characteristics of the substrate, soil moisture, agronomic activities and climatic conditions (intensity and frequency of rainfall, temperature, solar radiation, wind speed and relative humidity) (Shearman and Beard, 1973;Xinmin et al., 2007;Wherley et al., 2015;Amgain et al., 2018). In addition, as pointed out by Biran et al. (1981) and Kneebone and Pepper (1984), we must account for the fact that the ET a increases when water is available. Aronson et al. (1987) and Blankenship (2011) noted that evapotranspiration was governed mainly by meteorological factors when there was enough moisture in the soil, but that it declined after a critical level of moisture was reached.
On the other hand, ET 0 is estimated from meteorological data (precipitation, solar radiation, maximum and minimum temperature, wind speed and relative humidity) using the FAO-Penman-Monteith equation (Smith et al., 1992;).
Under standard conditions (well-watered conditions) the ET a of a crop can be related to the ET 0 through the crop coefficient, K c (ASCE, 1990;Zhang et al., 2010;Marin et al., 2016). The K c refers to the characteristics that distinguish the studied crop from a reference crop under standard (well-watered) conditions. It varies with the nature of the crop, its height and stage of development, the supporting substrate and the climatic characteristics of the area. The K c shows daily variation and, to minimize complexity, is expressed as the average over a period, either monthly, yearly, by stage of crop development or season.
The installation and maintenance of golf courses constitute a demanding agricultural activity involving the intensive cultivation of large areas of grass that require significant quantities of water for irrigation (Rodriguez Diaz et al., 2007). The use of different grass according to weather conditions seeks to increase the efficiency of irrigation, and ET a varies according to the variety of grass. ET a from cool-and warm-season grasses ranges from 3 to 8 mm daȳ¹ and from 2 to 6 mm daȳ¹, respectively (Augustin, 2000;Huang, 2006;Xinmin et al., 2007;Wherley et al., 2015;Colmer and Barton, 2017). When water availability drops, the grass responds to the shortage by activating biological mechanisms that result in lower water consumption. As reported by McCann and Huang (2008), the Agrostis stolonifera-L-93 variety generally suffers a sharp decline in the rate of ET a in low water stress conditions and, as indicated by Xu and Huang (2000), Liu and Huang (2001) and DaCosta and Huang (2006a, b), also suffers from biological changes that are triggered to reduce water consumption. Numerous studies have reported different values of K c for the same grass variety, reflecting the influence of the growing area. For example, the K c value of the Bermuda grass (Cynodon dactylon) variety ranges is between 0.17 and 0.99 in south-eastern USA (Wherley et al., 2015). Kentucky bluegrass has K c values between 0.80 and 1.40 in Beijing, and K c of tall fescue is 0.5 and 0.8 in Colorado, and ranges from 0.84 to 1.49 in Beijing (Ervin and Koski, 1998;Fu et al., 2004;Xinmin et al., 2007).
For this study, carried out under Mediterranean climatic conditions, an experimental golf green comprised of four sand-based lysimeters was built. The actual water requirement (ET a ) of a maintained Agrostis stolonifera-L-93 creeping turf grass was determined under both total water availability and water stress conditions. Since the sand-based lysimeters were amended with organic matter (OM) and hydrogel, an evaluation of the effect of these amendments on the ET a and K c could be made.
The addition of OM and hydrogel amendments is a common practice since they increase efficiency in water and agrochemicals use (Aamlid et al., 2009;Ullah et al., 2015;Martin del Campo et al., 2019). Hydrogels are hydrophilic polymers that absorb water, improve soil porosity, aeration, infiltration, nutrient transport and release, and water absorption that promote plant growth (Akhter et al., 2004;Abedi-Koupai et al., 2008;Ullah et al., 2015).

| Description of the experimental green
Four lysimeters were built, each with a surface of approximately 40 m 2 and a volume of 11 m 3 . The substrate is composed of a 26-40 cm sandy base (substrate categorized by the United States Golf Association (USGA) as siliceous sand), overlaying a 10-cm gravel layer containing drainage pipes (7.5 cm diameter) that collect water and drain them toward the exit. At the exit, recipients collect drainage water for control purposes. Water drainage samples were collected daily.
Each lysimeter is coated on the bottom and sides with a geomembrane that independently collects and channels all infiltrated water toward the drainage exit.
The addition of the OM and hydrogel in the lysimeters was carried out on the already deposited sand, and was mixed with the first 10 cm of the sandy substrate. The lysimeters were amended as follows: P-1 amended with both: 20% OM (peat) and 145 g m 2 hydrogel, P-2 amended with 20% OM (constructed according to USGA requirements) and P-3 amended with 145 g m 2 hydrogel (TerraCottem®). P-4 is sand only.
Each lysimeter has an independent irrigation system. Each irrigation system comprises eight diffusers (Model 6,406-ADV Nelson Turf®) equipped with 15 cm body type nozzles (7,370 Multiarc). Each system is controlled by an electric pump and a counter. Although irrigation is programmed, the flow is not always the same and depends on different factors, such as water pressure in the main pipes and water availability. Flow rates in lysimeters vary between 23.4 and 39.0 mm h¹. The determination of the water that falls within each lysimeter was made assuming that the irrigation is uniform. Irrigation during the investigation was scheduled according to rainfall and the objectives pursued: (i) total water availability: the condition of total water availability was maintained through most of 2010; (ii) tracer tests: tests that involved high water inputs were carried out from December 2010 to May 2011; and (iii) water stress: a slight water stress was imposed in the period from June to December 2011 to determine whether irrigation water could be saved in comparison to 2010.
The lysimeters were equipped with three moisture sensors installed vertically (DECAGON). Two sensors were the 10HS type that measures the volumetric moisture at depths of 12 and 24 cm, respectively, while the other one is the 5TE type, installed at a depth of about 18 cm, which also measures electrical conductivity and temperature. They were all calibrated for the substrate in which they were installed and were set up to record data every 2 min.
A meteorological station (Weather Rain Bird Smart), installed next to the green, provided hourly precipitation, solar radiation, maximum and minimum temperature, wind speed and relative humidity data. We used the data from this station to calculate the ET 0 from the FAO-Penman-Monteith equation (Smith et al., 1992. Apart from the irrigation rates, which were modified to meet the requirements of each lysimeter, the experimental site was treated in the same way (watering, mowing, fertilizing, phytosanitary treatment, pricked and verticutting) as the other greens on the golf course.

| Climatic characteristics of the area
The experimental green is located a few kilometres from the Mediterranean coast in Spain ( Figure 1). The area is characterized by a mild and humid Mediterranean climate. According to the meteorological data obtained from the meteorological station of the experimental green, the average temperatures in the warmer months during the study period were about 23 C, with peak point temperatures close to 30 C. On the other hand, the average temperatures for the winter months were between 8 and 10 C, with minimum temperatures of 2-3 C. During the period from 2009 to 2011, the months with the lowest rainfall were July 2010 and August 2011, with no rainfall. In contrast, the rainiest month was September 2009 with a rainfall of 360 mm, followed by November 2011 with 182 mm (Figure 2).

| Reference evapotranspiration (ET 0 )
ET 0 is usually estimated from meteorological data, which were obtained from the installed meteorological station. The FAO-Penman-Monteith equation is the most widely accepted method for calculating ET 0 (Smith et al., 1992): F I G U R E 1 Location of the field study site From Equation (1), ET 0 is calculated for an area planted with a hypothetical reference crop that has an assumed height of 12 cm, a fixed surface resistance of 70 s m −1 and an albedo of 0.23. ET 0 depends on the net radiation (R n ), the heat flux on the ground (G), the air temperature measured 2 m from the ground (T), the average wind speed (U 2 ), the saturation vapour pressure (e s ), the actual vapour pressure (e a ), the slope of the vapour pressure curve versus temperature (Δ) and the psychrometric constant (γ).

| Determination of actual evapotranspiration (ET a ) using the water balance
ET a can be calculated using the water balance (Equation (2)) between two dates on which substrate moisture values were approximately the same; thus, the variation in moisture storage was zero (ΔV = 0). Under this premise ET a is the difference between the input water (rainfall and irrigation) and the output water (drainage). This condition was used for determining ET a in 2009, since no moisture sensors were installed that year.
In 2010 and 2011 data from the moisture sensors were used to determine ΔV and ET a could be calculated on a daily and monthly basis.
The condition of total water availability was maintained through most of 2010. Tracer tests that involved high water inputs were carried out from December 2010 to May 2011. When tracer tests were performed, a restriction on irrigation was set in the second half of 2011 (June to December) in order to maintain the soil moisture at lower levels than those from June to December 2010.

| RESULTS AND DISCUSSION
3.1 | Effect of amendments under total water availability

| Effect of the OM amendment
To test the effect of OM on the water balance, the values of ET a for P-2 (amended with OM) and P-4 (100% sand) are compared. Figure 3 shows that the ET a values for the two lysimeters are similar; in fact, for a few months (March, April, May and July 2010), ET a in P-2 is lower than in the not amended lysimeter, while in other months (June, August, September and October 2010) it is up to 23% higher. The highest values of ET a were reached in June-August 2010: 2.76-12.2 mm daȳ¹ in P-2 and 3.06-10.3 mm daȳ¹ in P-4. These values are similar to those obtained by Green et al. (1990) and Bowman and Macaulay (1991): 7.7-12.7 and 4.57-13.0 mm daȳ¹, respectively. Research carried out in Norway by Aamlid et al. (2016) showed that, under daily irrigation conditions, they obtained ET a values of 5-10 mm daȳ¹, lower F I G U R E 2 Temperatures and rainfall (2009)(2010). Data from the meteorological station of the experimental site than P-2, probably due to climatic conditions. To achieve these results, they installed mini lysimeters on a green with Agrostis stolonifera -L-93.
Under this condition of total availability of water, the edaphic factor (in this case, the presence of OM) is barely relevant and the presence of OM does not show its water retention capacity, as Bigelow et al. (2000), Waltz et al. (2003) and McCoy et al. (2007) already showed in previous research.

| Effect of the hydrogel amendment
Over the same period, ET a values of the lysimeters amended with hydrogel, P-1 and P-3, are greater than those for the not amended lysimeters, as shown in Figure 4(A) (P-1 compared with P-2, amended with OM) and 4(B) (P-3 compared with P-4, which is 100% sand). The moderate increase of 23% in the ET a of P-1 (OM and hydrogel), and the large increase of 61% for P-3 (hydrogel) may be explained by the ability of the hydrogel to retain water, which facilitated evaporation, and/or a high K c value generated by transpiration. Mohawesh and Durner (2019) suggested that soil amendments such as hydrogel improved soil water retentivity across the whole moisture saturation range (from total water availability to water stress conditions) and, also, improved the water availability of the sandy soils for a longer period (almost 22 days). The ET a values achieved in P-1 and P-3 (P-1: 3.71-13.18 mm daȳ¹; P-3: 3.34-15.16 mm daȳ¹) exceed F I G U R E 3 ET a values for P-2 (OM) and P-4 (100% sand), to show the effect of OM for 2010. ET a variation: difference between the ET a value of P-2 and the ET a value of P-4 F I G U R E 4 (A) ET a values for P-1 (hydrogel and OM) and P-2 (OM). ET a increase: percentage by which the ET a value of P-4 increased with respect to P-2. (B) ET a values for P-3 (hydrogel) and P-4 (100% sand), to show the effect of the hydrogel for 2010. ET a increase: percentage by which the ET a value of P-3 increased with respect to P-4 the maximum values of Green et al. (1990) (12.7 mm daȳ¹) and Bowman and Macaulay (1991) (13.0 mm daȳ¹).
It is noteworthy that the increase in water storage under existing conditions of water availability may become more damaging to the grass than a lack of water, especially in summer. Surface water absorbs heat from the sun and transfers it to the root zone, such that the temperatures may be several degrees above the ambient temperature, causing damage to the roots (Dernoeden, 2006).
Monthly variations of ET a and ET 0 between 1 March 2010 and 31 December 2011 for all lysimeters are presented in Figure 5. All the curves follow the same trend: the highest values are reached in the months from June to August and the lowest in the months from November to February.
It is important to point out that ET a values were greater than ET 0 between July and August 2010 (total availability water) in all lysimeters, and especially noticeable in the hydrogel treated P-1 and P-3. Detailed analysis indicated that, in these months, the water requirements (ET a ) of P-1 and P-3 are greater than ET 0 ( Figure 5), because of the extra water needed when air temperatures approach 30 C. There is a a clear influence of agronomic activities and the FAO-Penman-Monteith equation underestimates the water requirement. Qian et al. (1996) and Lecina and Martínez-Cob (2000) reached the same conclusion from studies of other grass varieties that had high values of evapotranspiration.
From March to June and September to November 2010 ( Figure 5), when the temperature dropped, ET 0 provided a reasonable reflection of the water requirement in all the lysimeters. Irrigation was increased when tracer tests were done in December 2010 and moisture in the substrates was high. Excess moisture resulted in an increase in ET a in December in P-2 and P-3 (no data for P-4 and P-1), which shows that the level of moisture in the substrate also influenced the value of ET a , as mentioned by Biran et al. (1981) and Kneebone and Pepper (1984).

| Effect of amendments under slight water stress condition
A slight water stress was imposed in the period from June to December 2011 to determine whether irrigation water could be saved in comparison to 2010. The result was a low-quality turf and a decline in the ET a in P-2 (with OM) and P-4 (sand) (there are no data for P-1); however, each lysimeter reacted differently to water deficit, depending on the amendment. Gómez-Armayones et al. (2018) showed that adverse effects of deficit irrigation on turfgrass quality are more evident when turf is subject to environmental and/or management stresses such as long intervals between irrigation, short mowing heights or high temperatures.
F I G U R E 5 Monthly ET a and ET 0 values for the period from 1 March 2010 until 31 December 2011 for P-1 (hydrogel and OM), P-2 (OM), P-3 (hydrogel) and P-4 (100% sand)

| Effect of the OM amendment
Evapotranspiration in P-2 (OM) and P-4 (100% sand) are compared in Figure 6. Given the water restriction from July to December 2011, the effect of the OM was manifested in a lower decrease in ET a of P-2 than in P-4. For example, in July 2011, a decrease of 13% in storage in P-2 ( Figure 6(A)) caused the ET a to decrease 16% (Figure 6(B)), while in P-4, a 5% decrease in storage in P-4 ( Figure 6(C)) caused a decrease of 38% in the ET a (Figure 6(D)). In August 2011, the ratios were lower, but the OM still prevented a decrease in ET a and plant heat stress. In September, when the decrease in soil moisture was similar in both lysimeters, the ET a for P-2 was still greater than that for P-4. The effect of OM was very low in October as the moisture was too low in P-2 (40% less than in 2010); the ET a was therefore lower in P-2 than in P-4, for which the humidity was only 8%, less than in the previous year. The values obtained by Aamlid et al. (2016) under non-irrigation conditions (one single irrigation at the beginning of the period) varied between 3 and 5 mm daȳ¹, and in the P-2, for these stress conditions, ET a presented values between 1.1 and 5.4 mm daȳ¹. In 2011, the difference in the behaviour of the substrates (edaphic factor) was not very evident as the standard conditions of the FAO (total water availability) were kept in all lysimeters (2010). When the sand was amended with OM, the decline of ET a was minimized only under conditions of water deficit, and grass stress caused by high temperatures was reduced.
When water availability decreased, as occurred in the second half of 2011, the evapotranspiration rate was higher from the OM-amended lysimeter than from the 100% sandy lysimeter, which indicates that the retention capacity of OM was significantly lower than that of the hydrogel.

| Effect of the hydrogel amendment
Hydrogel-amended P-3 is compared with not amended P-4 (100% sand) to determine the influence of hydrogel on ET a in Figure 7. It is shown that P-3, despite having a greater decrease in storage in the summer of 2011 ( Figure 8) than P-4 ( Figure 6(C)), has an ET a about 40% higher than P-4 from July to September, indicating that the hydrogel provided water to the roots that could be used by the plant.
It appears that the distribution of water within the substrate, which in turn determines the availability to the F I G U R E 6 (A) Values of the average monthly water storage for P-2 (OM), (B) values of ET a for P-2 (OM), (C) values of the average monthly water storage for P-4 (100% sand), (D) values of ET a for P-4 (100% sand) for 2010 and 2011. ET a variation: difference between the 2010 ET a value and the 2011 ET a value grass, is more important than the amount of water stored; this is especially important in the summer months when the grass faces heat stress.
Monthly variations of ET a and ET 0 in the summer of 2011 (Figure 5), when the levels of humidity in the lysimeters were lower than in 2010, show values of ET a still deviated from ET 0 in P-2 and P-3 but with a smaller gap; there was no difference, however, between the values for P-4. Also, when water availability decreased, as occurred in the second half of 2011, the ET a was lower for P-2 (with OM) than for P-3 (hydrogel), which indicates that the retention capacity of OM was significantly lower than that of the hydrogel. Values in P-1 could not be calculated because sensors were broken.
3.3 | ET 0 , ET a and K c from the annual water balance To calculate ET a when ΔV = 0 (variation in moisture storage was zero), we identified periods of time when the F I G U R E 7 ET a values for P-3 (hydrogel) and P-4 (100% sand) in 2011 to highlight the effect of the hydrogel amendment. ET a variation: difference between the ET a value of P-3 and the ET a value of P-4 F I G U R E 8 Comparison of the average monthly storage in P-3 (hydrogel) during 2010 and 2011. Storage variation: difference between the 2010 storage value and the 2011 storage value moisture profiles showed that the condition of total water availability was met (well-watered condition). These events corresponded to a first interval from 30 March to 22 September 2009, a second interval between 4 March and 12 October 2010, and a third interval from 12 March to 21 November 2011. For these three intervals, daily values of ET 0 were calculated from weather station data using the FAO-Penman-Monteith equation. ET a was calculated from water balance from the irrigation, drainage and daily precipitation data. The ET a and crop coefficient, K c = ET a /ET 0 , are presented in Table I. Table I  Under the standard conditions maintained in 2010 and 2011, ET a is 10-50% bigger than ET 0 depending on the amendment: the hydrogel-amended P-1 and P-3 had the highest values of K c every year. The high values of K c of P-1 and P-3 do not indicate higher water requirements; rather, evapotranspiration of the water retained by the hydrogel was increased since drainage was minimized, and storage showed less variation. Evaporation increases when the water is maintained in the first few centimetres of the profile; further, when there is water available for the roots, transpiration is facilitated, and is responsible for maintaining the temperature in the leaves and reducing heat stress (Throssell et al., 1987;Carrow, 1996;Liu and Huang, 2001;McCann and Huang, 2008).
The results show that, in 2010, the amount of water needed to maintain an acceptable quality of grass in P-2 (OM) and P-4 (100% sand) was between 16 and 17% higher than that determined by the weather station (ET 0 ), resulting in a K c value of 1.17-1.16; the values of K c for P-2 and P-4 in 2009 and 2011-2 years of low-quality grass-were 0.99 and 0.92, and 1.04 and 1.11, respectively. Aamlid et al. (2016) determined a K c of 2.39 on the first day after irrigation and 0.79 a subsequent day (mean following day) on an experimental green constructed similar to P-2. Labranche (2005) for mime grass and substrate obtained a K c of 0.85. The range of values of K c between 0.8 and 1.09 obtained by Aronson et al. (1987) is low compared to those obtained in P-2 and P-4. The reason for this difference may be the lower water consumption of Poa, Festuca and Lolium varieties studied by Aronson et al. (1987) against the variety Agrostis stolonifera-L-93, that presents a greater capacity of transpiration as a resource to protect its photosynthetic metabolism from stress due to high temperatures (Liu and Huang, 2001). Maximum and minimum K c values for the hydrogel-amended lysimeters (P-1 and P-3) were higher (P-1: 1.09-1.26); P-3: 1.04-1.52) than those from P-2 and P-4 because of the ability of this compound to retain water, which facilitated evapotranspiration.
T A B L E I ET 0 , ET a and K c values for the water conditions ΔV = 0 (P-1: hydrogel + OM, P-2: OM, P-3: hydrogel, P-4:100% sand; I: irrigation, R: rainfall, D: drainage) 3.4 | ET 0 , ET a and K c from the daily water balance If moisture sensors are available, the variations in water storage between any two given times can be calculated, and the daily ET a can be obtained from the water balance equation. Monthly ET a data are the result of the sum of the daily ET a values when enough storage data are available. For months with incomplete data, monthly ET a was extrapolated. Figure 9 shows that ET a and ET 0 (daily data) follow the same trend but with a slight lag because of the interval chosen for the calculation. It also shows that, although the monthly K c value was greater than 1 (K c = 1.08), the values of ET a did not always outperform ET 0 , but were sometimes above or below this value, indicating that the FAO-Penman-Monteith equation may over-or underestimate ET a in certain circumstances.

| CONCLUSIONS
The results obtained in this research allow verification of the effect of the amendments with hydrogel and OM on the values of ET a and K c .
The maximum and minimum K c values for lysimeters with hydrogel were higher than other lysimeters due to the ability of this compound to retain water. The water retention effect of the hydrogel generates a greater availability of water for the root system. Its effect is particularly noticeable in conditions of low humidity and high evapotranspiration (summer).
Thus it is important to note that the addition of hydrogel can be a good measure for optimizing the use of water without impairing the quality of the grass.
It is possible to observe, in all lysimeters, that when there is total availability of water (well-watered conditions) ET a is greater than or equal to ET 0 and therefore K c is higher than 1.
The monthly variation of K c shows that the ET 0 calculated from meteorological parameters seems to be a reliable measure of the annual water requirement. However, while adequate in rainy periods, it is inadequate for months with dry situations because of the high water requirement of the L-93 turf grass variety during summer, especially when the temperature exceeds 29 C, above which the grass suffers heat stress. In such situations, the water requirement is 37% more than that calculated by replacing the ET 0.
To maintain the L-93 variety in optimal condition, the K c must be higher than or equal to 1.2, because the grass is of poor quality when the K c values are kept between 0.9 and 1.1.

ACKNOWLEDGEMENT
We wish to acknowledge the staff and direction from the Club de Campo del Mediterraneo golf course for their kind support.