Preliminary Evaluation of Autonomous Mowing for Sustainable Turfgrass Management in Mediterranean Climates

Abstract

Turfgrass provides significant functional, environmental, recreational and aesthetic benefits; however, its high management inputs raise sustainability concerns due to intensive irrigation, fertilization and mowing. The aim of this study is to evaluate whether adopting a new mowing technology can support or enhance current low-input strategies in turfgrass management, such as reducing synthetic fertilization and deficit irrigation. This study was conducted from September 2023 to October 2024 at the Centre for Research on Turfgrass for Environment and Sports (CeRTES) in Pisa, Italy. Two turf compositions, pure tall fescue and tall fescue–microclover mixture, were managed using an autonomous mower operating daily at three mowing heights, 20, 40 and 60 mm. Turf quality, color, the NDVI, weed cover, leaf morphology, and clover presence were assessed throughout the growing season, including a drought and recovery period. The experimental design consisted of a two-factor split-plot randomized complete block design with four replications, and the statistical approach used was two-way and one-way ANOVAs with Fisher’s LSD at p = 0.05. The results of the study indicated that, under conditions where an autonomous mower was set to operate on a daily basis, the selected mowing height had minimal influence on drought response or recovery when water availability was a limiting factor. Furthermore, when subjected to the lowest mowing heights, the legume species included in the turfgrass mix demonstrated strong resilience, maintaining its presence and performance. In addition, when mowing with a high mowing frequency and at low mowing heights, the overall quality of the turfgrass appeared enhanced. These results serve as an important starting point for considering autonomous mowing technology as an innovative strategy in advancing toward turf management systems that prioritize sustainability and efficient use of resources.

1. Introduction

Turfgrass is a unique and vital resource for our developing world and, when designed and managed in an appropriate fashion, affords a myriad of benefits [1]. Turf is an important component of the urban and rural landscape, has a unique role in aesthetics and provides an irreplaceable surface for recreational sports/activities [2]. Despite the great and various advantages that turfgrass areas can provide, the high management inputs required have led to controversial issues, which include high water consumption and incorrect use of fertilizers, herbicides and pesticides [3]. In addition, mowing is considered the most energy-consuming practice in turfgrass management [4]. Specifically, as stated by Gu et al. (2015) [5], emissions when maintaining the lawn with a gas-powered mower lead to an increase in the anthropogenic Hidden Carbon Costs (HCCs).

A low-input turfgrass management approach focuses on reducing resource consumption, such as water, fertilizers and all inputs associated with mowing. Within this framework, water conservation strategies have been explored extensively. A study by Blankenship et al. (2020) [6] proved that tall fescue maintained at high mowing heights can be an effective strategy to help conserve water. Similarly, Fry and Huang (2004) [7] demonstrated that a higher mowing height translates to a deeper root system in turfgrass; thus, increasing the mowing height is linked to enhanced drought tolerance.

Nutrient management has also been identified as a critical area for input reduction. One possible approach for reducing fertilizer applications may be the incorporation of nitrogen-fixing legumes species into the turf stand. In their study, Bigelow et al. (2022) [8] state that one possible solution to excess nutrient applications is cultivating lawns containing a mixture of grasses with legumes. With their ability to biologically fix atmospheric N, legumes can improve soil N levels, replace N lost by crop removal and reduce dependency on supplemental N fertilization [9]. This practice is not new; historically, legumes have been used to improve soil fertility and benefit subsequent crops [10]. Moreover, the adoption of a mowing management approach with no clipping collection may further enhance nitrogen availability for plants. Indeed, grass cycling facilitates an increase in carbon (C) and nitrogen (N) content in the soil [11,12,13,14,15,16].

Beyond water and nutrient management, mowing practices themselves are increasingly being reconsidered for sustainability. One strategy for reducing gasoline emissions from mowing is to carefully select an appropriate mower type. While gasoline-powered rotary mowers are the most common technology used currently, there is a growing demand for the use of battery-powered equipment [17]. These machines, which rely on rechargeable batteries rather than fossil fuels, are increasingly valued for their lower noise, reduced local air pollution and improved energy efficiency compared to conventional gasoline-powered mowers. Specifically, among battery-powered equipment, of particular interest are the recently introduced battery-powered Lightweight Autonomous Mowers (LAMs). A study from Grossi et al. (2016) [18] showed that battery-powered mowers save energy compared to conventional gasoline-powered mowers.

Minimizing management inputs associated with irrigation, fertilization and mowing can significantly influence societal perceptions of conventional turfgrass maintenance, promoting a transition toward more ecologically sustainable practices. The aim of this study is to evaluate whether adopting a new mowing technology can support or enhance the current low-input strategies in turfgrass management, such as reducing synthetic fertilization and deficit irrigation. In particular, this study investigates how different mowing heights, attained with the same mowing frequency, influence the drought tolerance, recovery capacity, long-term persistence and stand quality of a tall fescue–clover mixture.

2. Material and Methods

A study was conducted at the experimental station of the Centre for Research on Turfgrass for Environment and Sports (CeRTES), Department of Agriculture, Food and Environment, University of Pisa, located at S. Piero a Grado, Pisa (43°40′ N 10°19′ E, 6 m asl), Italy, from September 2023 to October 2024. The experimental area is characterized by a silty loam soil (Calcaric Fluvisol, 28% sand, 55% silt, and 17% clay) with a pH of 7.8 and 18 g kg⁻¹ organic matter. Soil preparation consisted of two weeding sessions on August 3rd and September 6th to achieve complete removal of all weeds. Glyphosate was applied at 3.42 kg a.i./ha as Roundup PRO^® Concentrate Herbicide (600 g a.i./L; Monsanto Company, St. Louis, MO, USA). On 19 September 2023, soil rototilling was performed for seed bed preparation. For turf establishment, one turfgrass and one turf-type legume were used, namely Schedonorus arundinaceus (Schreb.) Dumort cultivar ‘Raptor III’ (tall fescue) and Trifolium repens variety ‘Pipolina’ (microclover). In order to avoid seed segregation, seeding was performed separately for tall fescue and the legume species. On 20 September 2023, both tall fescue and the legume were hand-seeded. The seeding rate was 50 g m⁻² Pure Live Seeds (PLS) for tall fescue and 2.5 g m⁻² for the legume. Fertilization at seeding was 72 kg ha^–1 of N and 184 kg ha^–1 of P₂O₅ from diammonium phosphate ((NH₄)₂HPO₄) to facilitate the establishment. No further fertilizer was applied during the trial, relying on grass cycling and legume nitrogen fixation, both of which are widely adopted in low-maintenance turfgrass strategies. To improve seed germination, the experimental area was entirely covered with geotextile (30 g m⁻² specific weight) and light frequent irrigations were applied to encourage establishment. Irrigation was applied until full maturity, replacing 100% evapotranspiration.

In order to induce water stress starting from 10 May, irrigation was withheld until some of the plots reached unacceptable visual quality. On 24 June 2024, unacceptable visual quality loss was detected, and irrigation was applied to induce a drought recovery period. No chemical treatments were applied during the trial period. Monthly mean maximum and minimum air temperatures and monthly precipitation during the period (September 2023–October 2024) are reported in

Table 1. Monthly mean maximum and minimum air temperatures and monthly precipitation during study period (September 2023–October 2024).

	2023			2024
	Air Temperature (°C)		Precipitation (mm)	Air Temperature (°C)		Precipitation (mm)
	Min.	Max.		Min.	Max.
Jan	-	-	-	4.8	12.7	41.4
Feb	-	-	-	6.5	15.8	125.4
Mar	-	-	-	7.5	17.1	99.1
Apr	-	-	-	9.2	21.2	89.2
May	-	-	-	13.4	24.3	96.7
Jun	-	-	-	16.0	28.1	41.9
Jul	-	-	-	20.0	34.4	0.7
Aug	-	-	-	21.5	35.4	80.4
Sep	17.7	27.8	0.7	16.2	26.2	106.9
Oct	16.1	23.6	5.7	14.3	21.9	168.6
Nov	8.0	16.2	4.7	-	-	-
Dec	6.0	13.0	45.3	-	-	-

.

On 9 October 2023, mowing was performed with a lightweight Husqvarna EPOS 550 autonomous mower (Husqvarna, Stockholm, Sweden). The lightweight autonomous mower was set to operate at a daily mowing frequency (every day, seven days per week) and adopting systematic trajectories (Figure 1). This operating mode is based on Real-Time Kinematic Global Navigation Satellite Systems (RTK-GNSS) and allows the autonomous mower to follow parallel contiguous lines within a working area that is defined by the user through the same navigation system [19]. Plots were oriented north–south, with no buffer alleys, and mower passes included a 10 cm overlap to reduce edge effects.

The experimental design was a two-factor split-plot randomized complete block design with four replications. The main treatment consisted of three mowing heights, 20, 40 and 60 mm, with plots measuring 4.0 × 2.0 m (8 m²) for a total of twelve plots. Secondary treatment consisted of two turf mixes: pure tall fescue and tall fescue in a mixture with turf-type clover. Main plots were divided into two subplots, each measuring 2.0 × 2.0 m, (4 m²) for a total of twenty-four subplots. A daily mowing frequency was adopted with clipping returned.

When using light autonomous mowers, mowing frequency can be increased without causing soil compaction. We chose to use such a high mowing frequency for this trial to fully exploit autonomous mowers’ potential in terms of technical efficiency and stress reduction for plants. From October 2023 to the end of March 2024, daily mowing management was carried out to create stable turf mixes of turf and legume species. Data collection was carried out from April to October 2024 and the assessed parameters were as follows:

(a) 

Turf quality rated on a scale of 1 to 9, with 1 equaling the poorest and 9 equaling the best. A rating of 6 is considered acceptable [20].

(b) 

Turf color rated on a scale of 1 to 9, with 1 equaling light green and 9 equaling dark green. A rating of 6 is considered acceptable [20].

(c) 

The NDVI, using a GreenSeeker handheld crop sensor (model 505, Trimble Corp., Sunnyvale, CA, USA) held 70 cm above the turf canopy. The NDVI ranges from 0.00 to 0.99, which correlates positively with turf quality [21].

(d) 

Weed cover, visually assessed and reported as the percentage of the plot surface covered by weeds.

(e) 

Turf composition, determined with the line-intersect grid method. The number of intersects where the presence of the clover is recorded as positive or negative is divided over the total number of intersects to calculate the percentage of clover cover [8]. Only clover percentage is quantified and reported, while the complement is assumed to be tall fescue and weeds.

Furthermore, to evaluate variations in leaf size across different mowing heights, leaf texture and size measurements were conducted on two separate dates: 15 April and 16 October 2024. Tall fescue leaf texture and clover leaf size were measured by randomly collecting ten fully expanded mature grass and legume leaves per plot.

Collected leaves were subsequently scanned and subjected to image analysis for size determination using Imagej software (version 1.54, Java 1.8.0_345 64-bit, National Institute of Health, Bethesda, MA, USA). Legume leaf size was determined by measuring the midrib of the central leaflet [22]. Grass leaf texture was determined by measuring leaf width (mm). Starting from 5 June 2024, a loss in visual color among plots was detected; thus, NDVI assessments were carried in order to monitor differences. Moreover, volumetric water content measurements (FieldScout TDR 350 sensor, Spectrum Technologies, Inc., Aurora, IL, USA) were carried out as ancillary measurements, but the data are not reported here. The purpose was to assess soil water consistency across the entire experimental area to ensure that the NDVI values were interpreted under consistent soil moisture conditions at the moment of measurement. Assessments were carried out every other day for a total of eight assessments (from 5 June to 21 June 2024) until visual quality loss fell below acceptable values. Similarly, to monitor quality recovery, a total of eight NDVI assessments were conducted during the drought recovery period (from 24 June to 10 July 2024). Statistical analysis was performed with R, Version 2024.12.0+467 (Copyright (C) 2024 by Posit Software, Version 2024.12.0+467, PBC, Boston, MA, USA) and additional packages including lm for fitting linear models and predictmeans for Fisher’s protected least significant difference test at p = 0.05. Two-way ANOVA was performed separately for each date using linear models to test the effect of mowing height, turf mix and their interaction on the measured parameters: turf quality and color, the NDVI, weed cover and leaf texture. Leaf texture data analysis refers to tall fescue values in the pure stand and the grass–legume mixture (

Table 2. The results of the two-way ANOVA testing the effect of mowing height, turf mix and their interaction on the measured parameters throughout seven months (from April to October 2024) and leaf texture in April and October 2024.

	Source of Variation
Measured Parameters		Mowing Height (M)	Turf Mix (T)	M × T
Turf quality	Apr	ns	ns	ns
	May	ns	ns	ns
	Jun	ns	**	ns
	Jul	ns	**	ns
	Aug	ns	ns	ns
	Sep	ns	ns	ns
	Oct	ns	*	ns
Turf color	Apr	*	ns	ns
	May	ns	ns	ns
	Jun	*	***	ns
	Jul	ns	ns	ns
	Aug	ns	*	ns
	Sep	ns	ns	ns
	Oct	ns	**	ns
NDVI	Apr	**	ns	ns
	May	***	**	ns
	Jun	ns	ns	ns
	Jul	ns	**	ns
	Aug	ns	ns	ns
	Sep	*	**	ns
	Oct	*	ns	ns
Weed cover	Apr	***	ns	ns
	May	*	*	ns
	Jun	ns	ns	ns
	Jul	ns	ns	ns
	Aug	ns	ns	ns
	Sep	ns	ns	ns
	Oct	ns	ns	ns
Leaf texture (tall fescue)	Apr	**	ns	ns
	Oct	**	ns	ns

ns, not significant at the 0.05 probability level. * significant at the 0.05 probability level. ** significant at the 0.01 probability level. *** significant at the 0.001 probability level.

).

Subsequently, one-way ANOVA was performed separately for each date using linear models to test the effect of mowing height on the turf composition (%clover) and leaf size. Turf composition and leaf size data analysis refers to clover values within the grass–legume mixture (

Table 3. The results of the one-way ANOVA testing the effect of mowing height on turf composition (%clover) throughout seven months (from April to October, 2024) and leaf texture in April and October 2024.

	Source of Variation
Measured Parameters		Mowing Height (M)
Turf composition (% clover)	Apr	ns
	May	ns
	Jun	ns
	Jul	ns
	Aug	ns
	Sep	ns
	Oct	*
Leaf size (clover)	Apr	*
	Oct	***

ns, not significant at the 0.05 probability level. * significant at the 0.05 probability level. *** significant at the 0.001 probability level.

).

In addition, data referring to the dry down and drought recovery period have separately been analyzed with two-way ANOVA for each assessment date using linear models to test the effect of mowing height and turf mix and their interaction on the NDVI (

Table 4. The results of the two-way ANOVA testing the effect of mowing height and turf mix and their interaction on the NDVI during the dry down and drought recovery period from 5 June to 10 July 2024.

	Source of Variation
Measured Parameter		Mowing Height (M)	Turf Mix (T)	M × T
NDVI-DD	5 June	***	ns	ns
	7 June	ns	ns	ns
	10 June	*	*	ns
	12 June	ns	*	ns
	14 June	ns	ns	ns
	17 June	ns	ns	ns
	19 June	ns	ns	ns
	21 June	ns	ns	ns
NDVI-DR	24 June	ns	*	ns
	26 June	ns	***	ns
	28 June	ns	**	ns
	1 July	ns	***	ns
	3 July	ns	***	ns
	5 July	ns	***	ns
	8 July	ns	***	ns
	10 July	*	***	ns

ns, not significant at the 0.05 probability level. * significant at the 0.05 probability level. ** significant at the 0.01 probability level. *** significant at the 0.001 probability level.

). To distinguish data referring to the two different assessment periods, the measured parameters are defined in the table as NDVI-DD (dry down) and NDVI-DR (drought recovery). Separate ANOVAs were adopted, as they best serve the aim of capturing the patterns of treatment effects at each assessment date, rather than producing generalized statistical inferences across time. In the Results section, p-values are reported while the tables present treatment means with the LSD.

3. Results

The results of the two-way ANOVA (Table 2) showed that turf quality was only affected by turf mix, specifically in June, July and October, while turf color was affected by mowing height in April and June and turf mix in June, August and October. The NDVI was affected by mowing height on all assessment dates except for June and July and it was affected by turf mix on all assessment dates excluding April, June and October. Weed cover was affected by mowing height in April and May, while it was affected by turf mix only in May. Tall fescue leaf texture was affected by mowing height on both dates (April and October 2024).

The results of the one-way ANOVA (Table 3) revealed that turf composition was affected by mowing height only in October. Clover leaf size was affected by mowing height on both dates (April and October 2024).

The results of the two-way ANOVA (Table 4) revealed that during the dry down period, NDVI-DD was affected by mowing height only on the assessment dates of 5 and 10 June, while turf mix affected NDVI-DD on 10 and 12 June. Concerning the drought recovery period, NDVI-DR was affected by mowing height on 10 July only, while the ANOVA revealed that turf mix affected NDVI-DR across the entire drought recovery period.

Mowing height did not significantly affect turf quality during the entire trial period. Average data are reported in

Table 5. The effect of mowing height on turf quality rated on a scale from 1 = poorest to 9 = best, with a rating of 6 = acceptable, throughout seven months at the Centre for Research on Turfgrass for Environment and Sports (CeRTES). Data were analyzed independently for each month.

Mowing Height (mm)	Turf Quality (1–9 Score)
	15 April	15 May	14 June	15 July	19 August	16 September	16 October
20	6.6	6.9	6.1	6.1	6.8	7.0	7.1
40	6.6	6.8	5.6	5.7	6.4	6.9	7.1
60	6.4	6.8	5.7	5.8	6.6	6.8	7.0
LSD (p = 0.05)	ns	ns	ns	ns	ns	ns	ns

Means are significantly different at the 0.05 probability level as determined by Fisher’s protected LSD. ns, not significant at the 0.05 level of probability.

. On the contrary, turf mix affected turf quality in June (p = 0.0026), with the grass–legume mixture showing a higher quality compared to the pure stand (6.1 and 5.4, respectively). The same trend was observed in July (6.2 and 5.4) (p= 0.0033). A significant difference was also observed in October (p = 0.020), with a slight difference between the pure stand and the grass–legume mixture (7.2 and 7.0, respectively) (

Table 6. The effect of turf mix on turf quality rated on a scale from 1 = poorest to 9 = best, with a rating of 6 = acceptable, throughout seven months at the Centre for Research on Turfgrass for Environment and Sports (CeRTES). Data were analyzed independently for each month.

Turf Mix	Turf Quality (1–9 Score)
	15 April	15 May	14 June	15 July	19 August	16 September	16 October
Fa	6.5	6.8	5.4	5.4	6.5	6.9	7.2
Fa + Tr	6.5	6.9	6.1	6.2	6.6	6.9	7.0
LSD (p = 0.05)	ns	ns	0.4	0.5	ns	ns	0.2

Means are significantly different at the 0.05 probability level as determined by Fisher’s protected LSD. ns, not significant at the 0.05 level of probability.

); the pure stand recorded the best quality.

Color was affected by mowing height only in April and June. Specifically, in April, significantly higher values were observed when mowing at 40 and 60 mm compared to 20 mm (6.9, 6.9 and 6.4, respectively) (p = 0.048) (

Table 7. The effect of mowing height on turf color visually assessed and rated on a 1–9 scale (1 = light green to 9 = dark green, with a rating of 6 = acceptable) throughout seven months at the Centre for Research on Turfgrass for Environment and Sports (CeRTES). Data were analyzed independently for each month.

Mowing Height (mm)	Turf Color (1–9 Score)
	15 April	15 May	14 June	15 July	19 August	16 September	16 October
20	6.4	6.8	6.3	6.0	6.8	7.0	7.1
40	6.9	6.9	5.7	5.3	6.3	7.3	7.0
60	6.9	6.9	6.1	5.8	6.7	7.1	7.1
LSD (p = 0.05)	0.4	ns	0.5	ns	ns	ns	ns

Means are significantly different at the 0.05 probability level as determined by Fisher’s protected LSD. ns, not significant at the 0.05 level of probability.

). In June, significantly higher values were observed when mowing at 20 mm compared to 40 mm (6.3 and 5.7, respectively) (p = 0.04). Color was also affected by turf mix in June (p = 0.00054), August (p = 0.036) and October (p = 0.0019).

The same trend was observed in June and August, with the mixture stand showing significantly higher values compared to the pure stand (6.4 and 5.6 in June and 6.8 and 6.4 in August). On the contrary, the pure stand showed significantly higher values in October (7.2 compared to 6.9) (

Table 8. The effect of turf mix on turf color visually assessed and rated on a 1–9 scale (1 = light green to 9 = dark green, with a rating of 6 = acceptable) throughout seven months at the Centre for Research on Turfgrass for Environment and Sports (CeRTES). Data were analyzed independently for each month.

Turf Mix	Turf Color (1–9 Score)
	15 April	15 May	14 June	15 July	19 August	16 September	16 October
Fa	6.8	6.8	5.6	5.4	6.4	7.1	7.2
Fa + Tr	6.6	6.9	6.4	6.0	6.8	7.1	6.9
LSD (p = 0.05)	ns	ns	0.4	ns	0.4	ns	0.2

Means are significantly different at the 0.05 probability level as determined by Fisher’s protected LSD. ns, not significant at the 0.05 level of probability.

).

Regarding the NDVI, significant differences were observed during the entire trial period except for June, July and August. In April (p = 0.001), a significant difference was observed, showing higher NDVI values when mowing at 40 and 60 mm (0.86 and 0.87, respectively) compared to 20 mm (0.81) (

Table 9. The effect of mowing height on the NDVI (0.00–0.99) (GreenSeeker handheld crop sensor, model 505, Trimble Corp., Sunnyvale, CA, USA) throughout seven months at the Centre for Research on Turfgrass for Environment and Sports (CeRTES). Data were analyzed independently for each month.

Mowing Height (mm)	NDVI (0.00–0.99)
	15 April	15 May	14 June	15 July	19 August	16 September	16 October
20	0.81	0.87	0.71	0.77	0.82	0.86	0.81
40	0.86	0.88	0.74	0.77	0.83	0.87	0.85
60	0.87	0.90	0.76	0.81	0.85	0.87	0.85
LSD (p = 0.05)	0.03	0.01	ns	ns	ns	0.01	0.03

Means are significantly different at the 0.05 probability level as determined by Fisher’s protected LSD. ns, not significant at the 0.05 level of probability.

). In May (p = 0.00089), significantly higher NDVI values were observed when mowing at 60 mm, followed by 40 and 20 mm (0.90, 0.88 and 0.87, respectively). Similarly to April, the same trend was observed in September (p = 0.032) and October (p = 0.037), where NDVI values were significantly lower when mowing at 20 mm. As for the turf mix, significant differences were observed in May (p = 0.0013), July (p = 0.002) and September (p = 0.0026), with the grass–legume mixture showing significantly higher values compared to the pure stand (

Table 10. The effect of turf mix on the NDVI (0.00–0.99) (GreenSeeker handheld crop sensor, model 505, Trimble Corp., Sunnyvale, CA, USA) throughout seven months at the Centre for Research on Turfgrass for Environment and Sports (CeRTES). Data were analyzed independently for each month.

Turf Mix	NDVI (0.00–0.99)
	15 April	15 May	14 June	15 July	19 August	16 September	16 October
Fa	0.85	0.87	0.72	0.75	0.83	0.86	0.84
Fa + Tr	0.85	0.89	0.75	0.81	0.85	0.87	0.83
LSD (p = 0.05)	ns	0.01	ns	0.03	ns	0.01	ns

Means are significantly different at the 0.05 probability level as determined by Fisher’s protected LSD. ns, not significant at the 0.05 level of probability.

).

Weed cover (%) was affected by mowing height in April (p = 0.00088) and May (p = 0.017) and by turf mix only in May (p = 0.05). Significantly lower weed cover percentages were observed when mowing at 40 and 60 compared to 20 mm (2.4, 2.8 and 8.0% in April, and 2.4, 2.0 and 4.8% in May) (

Table 11. Effect of mowing height on weed cover (%) throughout seven months at Centre for Research on Turfgrass for Environment and Sports (CeRTES). Data were analyzed independently for each month.

Mowing Height (mm)	Weed Cover (%)
	15 April	15 May	14 June	15 July	19 August	16 September	16 October
20	8.0	4.8	2.0	2.8	4.5	4.5	3.8
40	2.4	2.4	2.8	2.0	2.8	2.4	2.8
60	2.8	2.0	2.4	3.8	5.4	5.0	3.8
LSD (p = 0.05)	2.8	1.9	ns	ns	ns	ns	ns

Means are significantly different at the 0.05 probability level as determined by Fisher’s protected LSD. ns, not significant at the 0.05 level of probability.

). Significantly lower weed cover percentages were also observed in May within the grass–legume mixture compared to the pure stand (2.3 and 3.8%, respectively) (

Table 12. Effect of turf mix on weed cover (%) throughout seven months at Centre for Research on Turfgrass for Environment and Sports (CeRTES). Data were analyzed independently for each month.

Turf Mix	Weed Cover (%)
	15 April	15 May	14 June	15 July	19 August	16 September	16 October
Fa	5.2	3.8	2.9	2.5	4.3	3.8	2.8
Fa + Tr	3.6	2.3	2.9	3.2	4.1	4.1	4.1
LSD (p = 0.05)	ns	1.6	ns	ns	ns	ns	ns

Means are significantly different at the 0.05 probability level as determined by Fisher’s protected LSD. ns, not significant at the 0.05 level of probability.

).

Turf composition was affected by mowing height only in October (p = 0.03). In particular, a significantly higher percentage of clover was observed within the grass–legume mixture when mowing at 60 mm (50.0%) compared to 20 and 40 mm, which did not statistically differ from each other (17.5 and 25.0%, respectively) (SE 6.50854) (

Table 13. Effect of mowing height on turf composition determined with line-intersect grid method (%) throughout seven months at Centre for Research on Turfgrass for Environment and Sports (CeRTES). Data were analyzed independently for each month.

Mowing Height (mm)	Turf Composition (% Clover)
	15 April	15 May	14 June	15 July	19 August	16 September	16 October
20	35.0	32.5	57.5	67.5	45.0	42.5	17.5
40	30.0	32.5	47.5	67.5	35.0	47.5	25.0
60	45.0	40.0	52.5	70.0	50.0	50.0	50.0
LSD (p = 0.05)	ns	ns	ns	ns	ns	ns	22.5

Means are significantly different at the 0.05 probability level as determined by Fisher’s protected LSD. ns, not significant at the 0.05 level of probability.

).

Tall fescue leaf texture was affected by mowing height both in April (p = 0.0015) and October (p = 0.0014). In April, a significantly finer texture was observed when mowing at 20 and 40 mm (2.3 and 2.7 mm, respectively) compared to 60 mm (3.3 mm) (

Table 14. Effect of mowing height on leaf texture (mm) in April and October 2024 at Centre for Research on Turfgrass for Environment and Sports (CeRTES). Data were analyzed independently for each month.

Mowing Height (mm)	Leaf Texture (mm) Tall Fescue
	15 April	16 October
20	2.3	1.7
40	2.7	2.1
60	3.3	2.5
LSD (p = 0.05)	0.5	0.4

Means are significantly different at the 0.05 probability level as determined by Fisher’s protected LSD.

). During October, tall fescue leaf texture appeared to be the finest when mowing at 20 mm (1.7), which statistically differs from a 40 mm mowing height (2.1 mm) and a 60 mm mowing height (2.5 mm). Concerning clover leaf size, a significant difference was observed in April (p = 0.021) between 20 mm and 60 mm mowing heights (7.0 and 15.8 mm, respectively).

The same trend in October (p = 3.41 × 10⁻⁵) was observed for clover, showing a significantly progressive reduced leaf size when mown at 20 mm (6.3 mm), 40 mm (10.3 mm) and 60 mm (13.7 mm) (

Table 15. Effect of mowing height on leaf size (mm) in April and October 2024 at Centre for Research on Turfgrass for Environment and Sports (CeRTES). Data were analyzed independently for each month.

Mowing Height (mm)	Leaf Size a (mm) Clover
	15 April	16 October
20	7.0	6.3
40	12.3	10.3
60	15.8	13.7
LSD (p = 0.05)	5.4	1.4

Means are significantly different at the 0.05 probability level as determined by Fisher’s protected LSD. ^a Clover leaf size was determined as the length of the midrib of the central leaflet.

).

During the dry down period, mowing height affected NDVI-DD only on 5 June (p = 0.00037) and 10 June (p = 0.039) (

Table 16. Effect of mowing height on NDVI-DD during dry down period (from 5 June to 21 June 2024) at Centre for Research on Turfgrass for Environment and Sports (CeRTES). Data were analyzed independently for each assessment date.

Mowing Height	NDVI-DD
	June
	5th	7th	10th	12th	14th	17th	19th	21st
20	0.76	0.76	0.75	0.74	0.71	0.67	0.66	0.67
40	0.81	0.79	0.79	0.77	0.74	0.70	0.69	0.70
60	0.81	0.79	0.80	0.78	0.76	0.70	0.71	0.71
LSD (p = 0.05)	0.02	ns	0.04	ns	ns	ns	ns	ns

Means are significantly different at the 0.05 probability level as determined by Fisher’s protected LSD. ns, not significant at the 0.05 level of probability.

). In particular, the same trend was observed for the two assessment dates, recording significantly higher values when mowing at 40 and 60 mm compared to 20 mm. Turf mix affected NDVI-DD on 10 June (p = 0.012) and 12 June (p = 0.011) (

Table 17. Effect of turf mix on NDVI-DD during dry down period (from 5 June to 21 June 2024) at Centre for Research on Turfgrass for Environment and Sports (CeRTES). Data were analyzed independently for each assessment date.

Turf Mix	NDVI-DD
	June
	5th	7th	10th	12th	14th	17th	19th	21st
Fa	0.78	0.77	0.76	0.74	0.72	0.68	0.67	0.67
Fa + Tr	0.80	0.79	0.80	0.78	0.75	0.69	0.70	0.71
LSD (p = 0.05)	ns	ns	0.03	0.03	ns	ns	ns	ns

Means are significantly different at the 0.05 probability level as determined by Fisher’s protected LSD. ns, not significant at the 0.05 level of probability.

), where significantly higher values were recorded for the grass–legume mixture.

During the drought recovery period, the ANOVA detected a significant effect of mowing height only on July 10th (p= 0.04) (

Table 18. Effect of mowing height on NDVI-DR during drought recovery period (from 24 June to 10 July 2024) at Centre for Research on Turfgrass for Environment and Sports (CeRTES). Data were analyzed independently for each assessment date.

	NDVI-DR
Mowing Height (mm)	24 June	26 June	28 June	1 July	3 July	5 July	8 July	10 July
20	0.70	0.73	0.72	0.73	0.74	0.73	0.74	0.76
40	0.72	0.74	0.75	0.72	0.72	0.73	0.74	0.76
60	0.74	0.74	0.74	0.75	0.75	0.76	0.76	0.79
LSD (p = 0.05)	ns	ns	ns	ns	ns	ns	ns	0.03

Means are significantly different at the 0.05 probability level as determined by Fisher’s protected LSD. ns, not significant at the 0.05 level of probability.

). Specifically, significantly higher values were recorded when mowing at 60 mm compared to 40 and 20 mm. On the contrary, the ANOVA detected a significant effect of turf mix on all assessment dates (24 June, p = 0.01; 26 June, p = 0.00051; 28 June, p = 0.0025; 1 July, p = 0.0002; 3 July, p = 0.00020; 5 July, p = 0.00053; 8 July, p= 0.00035; and 10 July, p= 1.78 × 10⁻⁵) (

Table 19. Effect of turf mix on NDVI-DR during drought recovery period (from 24 June to 10 July 2024) at Centre for Research on Turfgrass for Environment and Sports (CeRTES). Data were analyzed independently for each assessment date.

	NDVI-DR
Turf Mix	24 June	26 June	28 June	1 July	3 July	5 July	8 July	10 July
Fa	0.68	0.69	0.71	0.69	0.70	0.70	0.72	0.74
Fa + Tr	0.75	0.78	0.77	0.77	0.76	0.78	0.78	0.80
LSD (p = 0.05)	0.05	0.04	0.04	0.03	ns	0.04	0.03	0.02

Means are significantly different at the 0.05 probability level as determined by Fisher’s protected LSD. ns, not significant at the 0.05 level of probability.

). Within the entire drought recovery period, the same trend was observed, with significantly higher values for the grass–legume mixture compared to the pure stand.

4. Discussion

Concerning turf quality, turf mix exhibited a significant influence on turf quality, particularly in the warmer months of June and July. During this period, the grass–legume mixture demonstrated superior turf quality compared to the pure stand (6.1 and 5.4 in June; 6.2 and 5.4 in July). This result is in accordance with what was previously observed by Bigelow et al. in 2022 [8], confirming that this difference could be attributed to the presence of legumes, which may enhance soil nutrient availability through nitrogen fixation, thereby improving overall plant vigor and aesthetics. Additionally, mixed stands, as observed by Boyle in 2023 [23] and Sparks in 2014 [24], often provide better resilience against environmental stresses such as drought and temperature fluctuations, which could further explain the observed improvements in quality.

On the contrary, the trend reversed in October, with the pure stand recording the highest turf quality (7.2 vs. 7.0). The slight but significant difference observed during this period may be related to seasonal growth dynamics. The same study conducted by Boyle in 2023 [23] showed that while legumes typically contribute to improved turf quality during peak growing conditions, they may become less competitive in cooler months, allowing the pure stand to maintain a more uniform and better appearance.

Regarding turf color, a discrepancy between two different periods was observed. In June, higher mowing height (40 and 60 mm) resulted in significantly higher color values compared to a mowing height of 20 mm. This suggests that taller mowing heights may contribute to improved turf color, possibly due to enhanced leaf area and chlorophyll content, leading to better light absorption and photosynthesis. However, an opposite trend was observed at the 20 mm and 40 mm mowing heights, where the lower mowing height (20 mm) exhibited significantly higher color values than the 40 mm height (6.3 and 5.7, respectively). This discrepancy may be influenced by specific environmental conditions or turf stress responses during that period. Concerning turf mix, a similar pattern can be outlined, as already observed for turf quality. In particular, in the warmer months (June and August), the grass–legume mixture exhibited significantly higher color values than the pure stand (6.4 and. 5.6 in June and 6.8 and 6.4 in August). In contrast, the trend reversed in October, with the pure stand demonstrating superior color compared to the grass–legume mixture (7.2 and 6.9, respectively). This shift could be attributed to seasonal variations in species adaptation; the dominant species in the pure stand might have performed better under cooler autumn conditions.

Mowing height significantly affected the NDVI in April, May, September and October, following a consistent trend where higher mowing heights (40 mm and 60 mm) resulted in higher NDVI values compared to the lowest mowing height (20 mm) when water was not a limiting factor. This trend suggests that the increased leaf area at higher mowing heights may enhance light capture, which may translate to a better health status of the plants, which is in accordance with what was reported by Bremer et al. (2011) [25]. Regarding turf mix, the grass–legume mixture consistently exhibited significantly higher NDVI values than the pure grass stand in May, July and September.

Mowing height had a significant effect on weed cover in April and May, with higher mowing heights (40 mm and 60 mm) consistently associated with significantly lower weed infestation compared to the shorter 20 mm height. Specifically, weed cover was reduced by approximately 70% in April and by over 50% in May when mowing height was increased. This finding suggests that higher mowing heights enhance turfgrass competitiveness by promoting denser canopy closure and greater shading at the soil surface, which inhibits weed germination and establishment.

Turf mix influenced weed cover only in May, with the grass–legume mixture exhibiting significantly lower weed cover compared to the pure grass stand. This result may reflect the combined competitive effects of both grass and legume species on weed suppression. Legumes, in particular, can contribute to ground cover that might otherwise be exploited by weeds. Furthermore, the improved nutrient status of the turf through nitrogen fixation by legumes may enhance overall sward density, indirectly reducing opportunities for weed invasion.

The data indicate that mowing height significantly influenced turf composition, but this effect was only evident in October. Specifically, clover cover within the grass–legume mixture was higher at the 60 mm mowing height (50.0%) compared to 20 mm (17.5%) and 40 mm (25.0%), which did not differ significantly from each other. This result is in accordance with what was reported by Boyle (2023) [23]. However, no significant differences were observed for all other assessment dates, which indicate that during most of the observation period, the persistence of the clover in the mix stand can be achieved regardless of the height of the turf. On the contrary, Sparks (2014) [24] reported that low cutting heights promote clover stolon production, which can improve overall stand density.

The results demonstrate a clear influence of mowing height on both tall fescue leaf texture and clover leaf size, with effects observed in both early (April) and late (October) growing season assessments. For tall fescue, lower mowing heights consistently resulted in finer leaf texture. In April, leaves were significantly narrower under 20 mm and 40 mm mowing (2.3 mm and 2.7 mm, respectively) compared to 60 mm (3.3 mm). This pattern persisted in October, with the finest leaf texture again observed under the lowest mowing height (1.7 mm), progressively increasing with taller mowing regimes.

In addition to the effect of mowing height, our results clearly show that the daily autonomous mowing regime adopted in this trial contributed to a significant reduction in leaf width and overall texture. This outcome is consistent with findings from Pirchio et al. (2018) [26], who also reported reduced leaf width under high-frequency autonomous mowing regimes.

Clover leaf size also responded significantly to mowing height, with smaller leaves observed under the lowest mowing regimes. In April, clover leaves at a 20 mm height averaged 7.0 mm, in contrast to 15.8 mm at 60 mm. A similar gradient was observed in October, with leaf size increasing progressively from 6.3 mm at 20 mm to 13.7 mm at 60 mm. These results suggest that frequent mowing associated with low mowing heights inhibits leaf expansion in clover, likely due to limited energy reserves and reduced photosynthetic surface area.

During the dry down period, mowing height had a significant effect on the NDVI, limited to only two dates, June 5th and 10th. On these dates, turf mowed at 40 mm and 60 mm exhibited higher NDVI values compared to turf mowed at 20 mm. No significant differences were observed for all other assessments dates, which suggests that mitigation of early drought stress can be achieved equally regardless of the height of the turf. In contrast, turf mix influenced the NDVI later in the dry down period, with significant differences observed on 10 and 12 June. Specifically, the grass–legume mixture recorded higher NDVI values than the pure fescue plots, indicating improved drought tolerance in mixed stands. This may be attributed to the legumes’ deeper rooting systems or their contribution to nitrogen availability, both of which could enhance resilience under water-limited conditions.

During the drought recovery period, the results show that mowing height had a limited effect, with a significant difference detected only on 10 July. This suggests that mowing height did not have a major influence on better recovery from drought for most of the recovery period. On the other hand, turf mix consistently influenced recovery across all assessment dates, with the grass–legume mixture outperforming the pure grass stand throughout the period. This consistent trend underscores the resilience advantage provided by legumes, likely due to their nitrogen-fixing ability and deeper root systems.

5. Conclusions

In conclusion, within a low-input management framework, these site-specific and preliminary results highlight the dual role of autonomous mowers in promoting both turf quality and environmental sustainability. While previous studies have consistently reported that higher mowing heights improve drought tolerance by promoting deeper root systems, our results suggest that under daily autonomous mowing, even low mowing heights (20 mm) do not show major disadvantages during drought stress or recovery compared to higher mowing regimes. This indicates that the frequent, non-stressful mowing enabled by lightweight autonomous mowers may help mitigate the expected negative effects of low mowing heights on drought tolerance and recovery. Consequently, the potential to minimize irrigation inputs may not depend exclusively on maintaining taller canopies, but may also rely on the mowing frequency and technology adopted. Furthermore, within the limited context of a single-site and a single-year trial, the use of the autonomous mower has not affected the persistence or stability of the legume component within the turf stand across various mowing heights, thereby preserving its role in reducing fertilizer inputs. Notably, this management approach also contributes to enhanced turf quality, as evidenced by the development of finer, narrower leaves in both tall fescue and microclover under lower mowing regimes. Collectively, these findings provide a valuable preliminary basis for adopting autonomous mowing technology as a key tool in the transition toward more sustainable, resource-efficient turf management systems.