Plant materials.

Two consecutive studies were conducted under greenhouse conditions at the University of Florida Turfgrass Envirotron in Gainesville, FL, during 2010 and 2011. For each study, sod pieces of bahiagrass, st. augustinegrass, and zoysiagrass were established in free-standing 0.5-L pots (12.7 cm × 12.7 cm at top by 31 cm deep). The bottoms of the pots were filled with a 2.5-cm depth of rock to facilitate drainage and then field soil was used to fill the pots. The sod and field soil for both studies were collected from the University of Florida Plant Science Research and Education Unit, Citra, FL. The soil type was Candler sand (Hyperthermic, uncoated Lamellic Quartzipsamments; Soil Survey Staff, Natural Resources Conservation Service, United States Department of Agriculture, 2011). Large pieces of organic material and any rocks were removed by hand before placing the soil into the pots.

The mean temperature and relative humidity in the greenhouse during the first study were 29 °C and 68%, respectively. For the second study, the temperature and humidity were 23 °C and 53%, respectively. The photosynthetically active radiation (PAR) for the first study from 1100 to 1500 hr averaged 617 μmol·m⁻²·s⁻¹, and the maximum and minimum PAR for these hours were 1334 and 116 μmol·m⁻²·s⁻¹, respectively. In the second study, PAR averaged 513 μmol·m⁻²·s⁻¹ and ranged from 118 to 949 μmol·m⁻²·s⁻¹ for the 1100 to 1500 hr. The low minimum light readings can be accounted for by the sensor being shaded by the greenhouse structure, clouds, and periods of rainfall. Before beginning drought stress treatments, the turf was maintained without water stress under greenhouse conditions until the roots of each of the grasses had reached the bottom of the pots.

Plant water status.

The night before the start of each drying experiment, each plant was watered until water dripped freely from the bottom of the pot. The next morning, each pot was weighed to obtain the drained pot capacity. The drained pot capacity of each pot mass minus pot weight at the end of the experiment (defined subsequently) was defined as the total transpirable soil water (TTSW) of each pot. Before having final pot weight, TTSW was initially estimated as 20% of the drained pot capacity. Those pots subjected to the well-watered treatment were targeted to a weight of 0.8 × TTSW to allow the soil in the pot to remain moist but not fully saturated.

Pots were weighed at 24-h intervals to determine daily transpiration loss from each pot. Water was added to maintain the control plants at the well-watered (WW) target mass, and drying pots were permitted to lose up to 140 g of water per day (≈0.1 of estimated TTSW). Water loss in excess of 140 g in any pot for a single day was added to that pot to facilitate gradual soil drying.

The water status of the drying plants was monitored daily by calculating the ratio of the daily transpiration loss of each drying plant to the mean daily transpiration loss of the WW plants. Potential variation in daily transpiration rates resulting from environmental conditions in the greenhouse was accounted for in the calculation of this ratio. To account for differences in transpiration rates among replicates within each genotype, the daily transpiration ratio for each pot was normalized so that transpiration values were centered on 1.0 during the initial well-watered phase. This was done by calculating for each pot the average transpiration ratio for the first days of the experiment before the total pot mass had fallen below the WW mass for that pot (≈4 d). A NTR was computed on each day for each pot by dividing the daily transpiration ratio by the average transpiration of that pot under WW conditions.

Turf stress indices.

The turf was evaluated daily for signs of stress throughout both studies. Visual evaluations of canopy wilting and firing were recorded, and changes in canopy reflectance were monitored through measurement of the normalized difference vegetative index (NDVI) using a Field Scout TCM 500 NDVI Color Meter (Spectrum Technologies, Plainfield, IL). Visual rating of firing and wilting was recorded as a percentage of the turf canopy in either condition. Visual turf quality was rated on a scale from 1 to 9, where 1 is a fully fired turf, 9 is a bright green turf with no wilting or firing, and 5.5 is minimally acceptable quality. All observations were collected daily during the dry-down and recovery periods.

2010 study design.

Each pot was planted with the appropriate sod on 5 May 2010 and fertilized with a soluble fertilizer (20N–20P–20K) to supply N at 48.8 kg·ha⁻¹ on 18 June 2010 and on 30 July 2010. The st. augustinegrass replicates succumbed to a chinch bug (Blissus insularis Barber) infestation and were excluded.

Five treatments were established based on daily NTR. The WW treatment was maintained at WW status. Two treatments were defined by allowing the pots to dry until NTR was equal to 0.3 and 0.1 NTR, respectively. The final two treatments were defined by extending the period of drying once NTR had reached 0.1 × 3 d (0.1 + 3D) and 7 d (0.1 + 7D), respectively. As an individual plant reached the end point for the assigned treatment, that pot was watered and maintained at a WW mass for a minimum of 7 d to evaluate recovery. A visual schematic summary of the treatments is illustrated in Figure 1. Five replicate pots were assigned to each of the five treatments, a total of 25 pots per grass species. The experiment began on 2 Aug. 2010.

Visual schematic of water stress treatments. Treatments were determined according to a normalized transpiration ratio (NTR) of drying to well-watered turf. Gray horizontal bars represent the period of drying allowed for each treatment to reach an NTR end point (0.5, 0.3, or 0.1). The black horizontal bars represent additional 3 and 7 d of drying for treatments 0.1 + 3D and 0.1 + 7D, respectively, once an NTR of 0.1 had been reached. The white bars indicate the well-watered, 7-d recovery period when each pot was rewetted and then maintained well-watered after the drought stress period; also note that the well-watered treatment (WW) is represented by a white bar spanning the entire experimental period.

Citation: HortScience horts 46, 11; 10.21273/HORTSCI.46.11.1550

• Download Figure

View Full Size

Visual schematic of water stress treatments. Treatments were determined according to a normalized transpiration ratio (NTR) of drying to well-watered turf. Gray horizontal bars represent the period of drying allowed for each treatment to reach an NTR end point (0.5, 0.3, or 0.1). The black horizontal bars represent additional 3 and 7 d of drying for treatments 0.1 + 3D and 0.1 + 7D, respectively, once an NTR of 0.1 had been reached. The white bars indicate the well-watered, 7-d recovery period when each pot was rewetted and then maintained well-watered after the drought stress period; also note that the well-watered treatment (WW) is represented by a white bar spanning the entire experimental period.

Citation: HortScience horts 46, 11; 10.21273/HORTSCI.46.11.1550

• Download Figure

Visual schematic of water stress treatments. Treatments were determined according to a normalized transpiration ratio (NTR) of drying to well-watered turf. Gray horizontal bars represent the period of drying allowed for each treatment to reach an NTR end point (0.5, 0.3, or 0.1). The black horizontal bars represent additional 3 and 7 d of drying for treatments 0.1 + 3D and 0.1 + 7D, respectively, once an NTR of 0.1 had been reached. The white bars indicate the well-watered, 7-d recovery period when each pot was rewetted and then maintained well-watered after the drought stress period; also note that the well-watered treatment (WW) is represented by a white bar spanning the entire experimental period.

Citation: HortScience horts 46, 11; 10.21273/HORTSCI.46.11.1550

• Download Figure

2011 study design.

Each pot was planted with the appropriate sod on 26 Oct. 2010 and fertilized with a soluble fertilizer (20N–20P–20K) to supply N at 48.8 kg·ha⁻¹ on 10 Dec. 2010, before the start of the drought stress period. Five drought stress treatments were again applied to five replicates each of bahiagrass, st. augustinegrass, and zoysiagrass. The WW, 0.3, 0.1, and 0.1 + 3D treatments were repeated from the previous study without modification. The 0.1 + 7D treatment was replaced with a more modest water deficit by rewatering these pots when an end point of 0.5 NTR was reached as shown in Figure 1. Plants were watered back to WW status and maintained at WW for 7 d after the end point was reached to evaluate recovery.

The pots were watered to excess on 12 Dec. 2010, and drying proceeded from fully saturated status as prescribed for each treatment. Daily NTR and TTSW were calculated to monitor plant water status, and visual evaluations and NDVI measurements were all collected as previously described.

Data analysis.

The data from each study were analyzed using PROC GLM for two-way analysis of variance with grass and treatment as main effects, and differences between means were determined using the LSMEANS/pdiff statement using the Statistical Analysis System (SAS Institute, 2008).

Drying days and total transpired water.

In the first study, the total amounts of water transpired by bahiagrass or zoysiagrass to reach the end points of 0.3, 0.1, and 0.1 + 3D were not different as shown in Figure 2A. Additionally, the accumulated water transpired by either of the grasses to reach an NTR of 0.3 was not different from the accumulated water loss to reach 0.1 NTR, and it remained unchanged for the 3 d after the NTR of 0.1 as shown in Figure 2A. Average WW daily water use for bahiagrass and zoysiagrass was 134 and 122 g·d⁻¹, respectively.

Total transpired water (g) at treatment end points for two consecutive studies. (A–B) Data from the first (2010) and second (2011) studies, respectively. Treatments were determined according to a normalized transpiration ratio (NTR) of drying to well-watered turf. Treatments 0.5, 0.3, and 0.1 represent water stress levels of the corresponding NTR. Treatment 0.1 + 3D continued drying for 3 d once an NTR of 0.1 had been reached. Data shown are means ± 1 se, and different letters indicate different means within a bar grouping (α = 0.05).

Citation: HortScience horts 46, 11; 10.21273/HORTSCI.46.11.1550

• Download Figure

View Full Size

Total transpired water (g) at treatment end points for two consecutive studies. (A–B) Data from the first (2010) and second (2011) studies, respectively. Treatments were determined according to a normalized transpiration ratio (NTR) of drying to well-watered turf. Treatments 0.5, 0.3, and 0.1 represent water stress levels of the corresponding NTR. Treatment 0.1 + 3D continued drying for 3 d once an NTR of 0.1 had been reached. Data shown are means ± 1 se, and different letters indicate different means within a bar grouping (α = 0.05).

Citation: HortScience horts 46, 11; 10.21273/HORTSCI.46.11.1550

• Download Figure

Total transpired water (g) at treatment end points for two consecutive studies. (A–B) Data from the first (2010) and second (2011) studies, respectively. Treatments were determined according to a normalized transpiration ratio (NTR) of drying to well-watered turf. Treatments 0.5, 0.3, and 0.1 represent water stress levels of the corresponding NTR. Treatment 0.1 + 3D continued drying for 3 d once an NTR of 0.1 had been reached. Data shown are means ± 1 se, and different letters indicate different means within a bar grouping (α = 0.05).

Citation: HortScience horts 46, 11; 10.21273/HORTSCI.46.11.1550

• Download Figure

In the second study, average daily water use for WW bahiagrass, st. augustinegrass, and zoysiagrass was less than that of the first study at 102, 95, and 70 g·d⁻¹, respectively. Bahiagrass had used more water than st. augustinegrass by the 0.5 end point, but accumulated transpired water of zoysiagrass was not different from that of the other two grasses as shown in Figure 2B. At the 0.3 NTR end point, zoysiagrass had transpired less water than the other two grasses, but the water transpired for st. augustinegrass and bahiagrass was not different. Accumulated transpired water for bahiagrass was higher than st. augustinegrass at the 0.1 NTR end point, whereas zoysiagrass did not differ from the other two grasses as shown in Figure 2B. There were no differences in total transpired water between the three grasses at the 0.1 + 3D end point.

The number of drying days required for bahiagrass to reach 0.3 NTR was different from the number of days required by zoysiagrass to reach that same point of stress in 2010 as shown in Figure 3A. However, both grasses reached 0.1 NTR after 13 ± 0.8 d of drying. In the second study, there were no differences in drying days required to reach 0.5, 0.3, or 0.1 between bahiagrass and st. augustinegrass, whereas zoysiagrass required more drying days than the other two grasses to reach each of these stress points as shown in Figure 3B.

Days of drying required to reach the 0.5, 0.3, and 0.1 normalized transpiration ratio (NTR) end points. (A–B) Data from the first (2010) and second (2011) studies, respectively. Treatments were determined according to a NTR of drying to well-watered turf. Data shown are means ± 1 se, and different letters indicate different means within a bar grouping (α = 0.05).

Citation: HortScience horts 46, 11; 10.21273/HORTSCI.46.11.1550

• Download Figure

View Full Size

Days of drying required to reach the 0.5, 0.3, and 0.1 normalized transpiration ratio (NTR) end points. (A–B) Data from the first (2010) and second (2011) studies, respectively. Treatments were determined according to a NTR of drying to well-watered turf. Data shown are means ± 1 se, and different letters indicate different means within a bar grouping (α = 0.05).

Citation: HortScience horts 46, 11; 10.21273/HORTSCI.46.11.1550

• Download Figure

Days of drying required to reach the 0.5, 0.3, and 0.1 normalized transpiration ratio (NTR) end points. (A–B) Data from the first (2010) and second (2011) studies, respectively. Treatments were determined according to a NTR of drying to well-watered turf. Data shown are means ± 1 se, and different letters indicate different means within a bar grouping (α = 0.05).

Citation: HortScience horts 46, 11; 10.21273/HORTSCI.46.11.1550

• Download Figure

Miller (2000) reports a similar range of 10.8 to 12.5 d required to reach the 0.1 end point for bermudagrass grown in native soil, sand, and combinations of these soils with various commercial amendments. In contrast, zoysiagrass required 13 d and 16 d in the two experiments as shown in Figure 3, which indicated relatively low daily transpiration rates and thus a longer elapsed time to reach the same level of stress as the other grasses. These results may indicate that zoysiagrass may have the potential for maintenance under a reduced-frequency irrigation scheme.

In a field study using minilysimeters to determine evapotranspiration (ET) loss by mass balance, ‘Argentine’ bahiagrass and ‘Meyer’ zoysiagrass (Z. japonica Steud.) water use was characterized by a medium ET rate (6.3 and 6.0 mm·d⁻¹, respectively) as compared with nine other warm-season grasses, and ‘Emerald’ zoysiagrass (Z. japonica Steud. × Z. tenuifolia Wild. ex Trin.) was characterized as medium low ET rate (5.5 mm·d⁻¹) (Kim and Beard, 1988). In that same study, ‘Texas Common’ st. augustinegrass [S. secundatum (Walt.) Kuntze] had a medium low ET rate of 5.8 mm·d⁻¹ (Kim and Beard, 1988). In a recent field study, also using minilysimeters in field plots, ET under WW conditions was measured from the same three grasses as the current study, and zoysiagrass had approximately the same ET rate as bahiagrass and slightly higher ET rates than st. augustinegrass (Wherley et al., unpublished data). Field ET rates for bahiagrass and zoysiagrass were 4.8 and 4.5 mm·d⁻¹, respectively and st. augustinegrass ET was 3.7 mm·d⁻¹ (Wherley et al., unpublished data). While comparing field and controlled-environment studies for water use, the relative differences between the grasses that we found under WW conditions is consistent with the findings of Kim and Beard (1988). Further comparisons would be difficult to draw as a result of genotypic differences among cultivars of similar grasses and the manner in which those genetic factors interact with environmental conditions such as atmospheric vapor pressure deficit (Green et al., 1991; Wherley and Sinclair, 2009).

Turf response to increasing drought stress.

In the first study, bahiagrass began both wilting and firing before reaching 0.3 NTR, and treatment 0.1 + 3D was fully fired by the end point (Table 1). None of the bahiagrass under water deficit stress had maintained acceptable quality on reaching the respective end points. Rather, the quality rating was reduced to 3.3 for treatment 0.3 NTR and declined with continued drying under the other treatments (Table 1). Conversely, the NDVI for treatment 0.3 NTR was not different from that of WW (Table 1). NDVI for the remaining treatments was different from WW, but 0.1 NTR and 0.1 + 3D were not different (0.41 and 0.44, respectively). Drought stress was not detectable by using NDVI until bahiagrass transpiration was greatly reduced, and further reflectance differences did not occur until drought stress had become quite severe (0.1 + 7D).

Table 1.

Turf quality indices at the drought stress treatment end points for ‘Argentine’ bahiagrass and ‘Empire’ zoysiagrass from the first study (2010).^z

View Table

At least 78% of the zoysiagrass canopy was wilted for all drought stress treatments in the first study, and 100% wilting coincided with 0.1 NTR (Table 1). However, firing did not occur in zoysiagrass until the 0.1 + 3D end point (16%), and at the 0.1 + 7D end point, the canopy was 60% fired (Table 1). Turf quality remained marginally acceptable (5.2 ± 1.1) down to the 0.3 NTR end point, but the increasingly stressed treatments resulted in unacceptable visual quality ratings with 0.1 + 7D resulting in a rating of 1.0 (Table 1). NDVI readings for treatments 0.3 NTR and 0.1 NTR were not different from WW, but NDVI declined by the 0.1 + 3D end point and again declined sharply by the 0.1 + 7D end point (Table 1).

In the second study, bahiagrass had begun to wilt by 0.5 NTR, and full wilting (100%) was concurrent with 0.1 NTR. No firing occurred until the 0.1 + 3D end point (Table 2). Visual quality ratings for bahiagrass remained at an acceptable level until an NTR of 0.3 was reached. NDVI readings differed from that of WW beginning with 0.3 NTR, and there were subsequent differences between 0.1 NTR and 0.1 + 3D as well (Table 2).

Table 2.

Turf quality indices at the drought stress treatment end points for ‘Argentine’ bahiagrass and ‘Empire’ zoysiagrass from the second study (2011).^z

View Table

Zoysiagrass was wilted by ≈50% by 0.3 NTR and was fully wilted by the 0.1 + 3D end point in the second study (Table 2). However, very little firing occurred, even with the driest treatment (10% at 0.1 + 7D; Table 2). Visual quality remained acceptable for WW and 0.5 NTR but declined to marginally acceptable (5.2) with treatment 0.3 NTR (Table 2). NDVI readings were not different from WW until drought stress had reached the 0.1 NTR (Table 2).

In the second study, st. augustinegrass had begun wilting by 0.5 NTR, and nearly the entire canopy was wilted by the time transpiration was reduced to 0.3 NTR (Table 2). Firing occurred concurrent with 0.1 NTR and increased with increasing stress (Table 2). Visual quality ratings declined significantly with each successive increase in drought stress after WW but remained at least marginally acceptable (average 5.2) until NTR was 0.3. Reflectance readings by NDVI were different between treatments 0.1 NTR and 0.1 + 3D, which were both different from WW (Table 2).

Although wilting, firing, and visual rating scores of zoysiagrass generally reflected less stress in the plants as compared with the other two grasses when each end point was reached (Tables 1 and 2), zoysiagrass was generally subject to water deficit stress for a longer period than the other two species, because more days were required to reach each rewatering point. The differences among grass species tended to be especially obvious in the 0.5 NTR and 0.3 NTR treatments. This study was not designed to resolve the basis of such differences, but there were two possibilities. One possibility was that the extra time used by zoysiagrass to reach the stress thresholds might have allowed additional time for acclimation of the grass to the stress conditions. The other possibility was that zoysiagrass has inherent physiological features that allow it to better withstand the various stress levels. White et al. (1992) found that low basal leaf osmotic potential under unstressed conditions as well as osmotic regulation, prolonged positive turgor maintenance, and delayed leaf rolling were all associated with genotypes of tall fescue (Festuca arundinacea Schreb.) with the best survival rate under competition among genotypes for declining soil water. Qian and Fry (1997) found that ‘Meyer’ zoysiagrass had a relatively high level of osmotic adjustment relative to buffalograss [Buchloe dactyloides (Nutt.) Engelm.] and tall fescue and recovered well after a period of gradual drying.

Turf recovery.

In the first study, bahiagrass remained wilted after 24 h recovery after treatments 0.3 NTR, 0.1 NTR, and 0.1 + 3d. Treatments 0.3 NTR, 0.1 NTR, and 0.1 + 3D remained partially fired (36 to 60%; Table 3), although less so than on the previous day (58% to 82% fired; Table 1), and 0.1 + 7D remained 100% fired (Table 3). Treatment 0.3 NTR recovered to a marginally acceptable visual quality (5.3 ± 1) after watering was reinitiated, but all other treatments besides WW still had unacceptable ratings (Table 3). After 7 d under WW conditions, bahiagrass further recovered to where treatment 0.3 NTR was no longer fired, treatment 0.1 NTR had only 25% firing remaining, and even treatment 0.1 + 7D, which had been fully fired, was less than fully fired, indicating that some viable tissue remained (Table 3). Treatments WW, 0.3 NTR, and 0.1 NTR had all recovered to acceptable visual quality after 7 d of watering after the prescribed stresses. NDVI readings from WW and 0.1 + 3D treatments were not significantly different, indicating amelioration of stress, and the 0.3 NTR and 0.1 NTR treatments had higher NDVI readings than WW.

Table 3.

Turf recovery quality indices from the first study (2010).^z

View Table

In the second study, all but the most severe drought stress treatment (0.1 + 3D) for bahiagrass rapidly recovered after being rewatered. After only 24 h, there was no consistent wilting or firing present in the WW, 0.5 NTR, 0.3 NTR, or 0.1 NTR treatments, whereas 0.1 + 3D remained almost entirely wilted (92%) and was 48% fired (Table 4). All of the treatments except for 0.1 + 3D had returned to or maintained an acceptable visual quality (Table 3). NDVI for 0.3 NTR, 0.1 NTR, and 0.1 + 3D was not different but were all lower than WW, and 0.1 + 3D was lower than that of the other three stressed treatments (Table 4). After 7 d of recovery, 0.1 + 3D remained 54% fired, but no other treatments had any firing present (Table 4). The visual quality ratings for all stressed treatments besides 0.1 + 3D were 9.0 (Table 3), and the NDVI values for all treatments but 0.1 + 3D were not different (Table 4). The 0.1 + 3D treatment did not return to an acceptable visual quality after 7 d recovery (Table 4).

Table 4.

Turf recovery quality indices from the second study (2011).^z

View Table

Zoysiagrass stressed to 0.3 and 0.1 NTR, in the first study, fully recovered from wilting after 24 h, and 0.1 + 3D remained only 36% wilted (Table 3). The 0.1 + 3D treatment was 40% fired and 0.1 + 7D was 100% fired, both increases in firing from the previous day (Table 3). The WW, 0.3 NTR, 0.1 NTR, and 0.1 + 3D treatments had all recovered to acceptable visual quality after 24 h (Table 3). The NDVI for 0.3 NTR and 0.1 NTR remained not different from WW, but treatments 0.1 + 3D and 0.1 + 7D were different from WW, indicative of some remaining stress (Table 3). After 7 d at WW status, zoysiagrass under the 0.1 + 7D treatment had recovered from fully fired to 50% fired, and the 0.1 + 3D treatment was only 28% fired (Table 3). Although 0.1 + 7D still did not have acceptable visual quality, treatments 0.3 NTR and 0.1 NTR had visual quality ratings of 9.0, and treatment 0.1 + 3D quality had improved to an average of 6.8 (Table 3).

In the second study, zoysiagrass under WW, 0.5 NTR, 0.3 NTR, and 0.1 + 3D treatments had recovered from wilting, and each was less than 10% fired after 24 h recovery. There was an unexplained delay in wilt recovery for the 0.1 NTR treatment after 24 h, but by 7 d of recovery, all zoysiagrass treatments had recovered visual quality ratings were 9.0, except for treatment 0.1 + 3D, which had a rating of 8.6 (Table 4). The NDVI value for the 0.1 treatment was the only one different from WW after 24 h recovery, and after 7 d recovery, there were no differences among zoysiagrass NDVI readings according to treatment (Table 4).

St. augustinegrass, which was evaluated only in the second study, rapidly recovered after having water reintroduced. After the first 24 h, no treatments had greater than 20% wilt remaining, and firing only remained for the 0.1 NTR and 0.1 + 3D treatments (14% and 50%, respectively; Table 4). The visual quality ratings for all treatments were acceptable, except for 0.1 + 3D (Table 4). The NDVI readings for WW, 0.5 NTR, and 0.3 NTR treatments were not different after 24 h recovery (Table 4). After 7 d under WW conditions, st. augustinegrass had further recovered to acceptable visual quality for all treatments, and 0.1 + 3D was less than 40% fired (Table 4). The NDVI readings for WW were lower than that of 0.3 NTR but equal to that of the remaining treatments (Table 4). These differences may be attributable to waning fertility at the end of the experiment.

The recovery of the grasses from the various stress levels indicated a substantial robustness in all of these grasses after being subjected to water deficit. Within 24 h after watering at the end points of 0.5 NTR and 0.3 NTR, all grasses were showing little sign of wilting or firing (Tables 3 and 4). These results indicate that these grasses could be taken to these levels of water deficit without expressing any permanent evidence of the drought episode. However, for more severe stress, e.g., 0.1 + 3D, there were differences among species in their recovery after 24 h from the stress (Tables 3 and 4). Zoysiagrass expressed an ability to rapidly recover from low levels of wilting and firing even within 24 h after rewatering of the 0.1 + 3D treatment.

Evaluation of plant firing 7 d after rewatering the plants in Study 1 showed that the 0.1 + 7D treatment apparently resulted in permanent damage to the grasses (Table 3). Substantial firing still existed in the bahiagrass and zoyiagrass canopies even after the long recovery period. The severity of this stress on the plants was the basis for dropping this treatment for Study 2.

Although the recovery results indicate the grasses recovered from fairly substantial stresses, poor visual quality during recovery likely indicates difficulty in maintaining acceptability of these stress levels by homeowners. All grasses exhibited reduced visual quality by the time they reached the 0.3 NTR. Although the visual quality recovered for plants subjected to this treatment within 24 h, there was a period in which the plants were visually stressed. The results of this study confirm that the practice of waiting to irrigate turf until visual drought stress symptoms develop may not result in permanent damage to a home lawn. This practice could result in a substantial savings in water, because delaying irrigation allows a greater opportunity for natural rainfall to add water to the soil.

The measurements of NDVI showed differences among treatments but it was only the most severe stress treatments that had decreased NDVI (Tables 3 and 4). There were no significant changes in NDVI at stress levels where wilting, firing, and visual ratings were indicating the initiation of stress. Johnson et al. (2009) measured NDVI on six genotypes of seashore paspalum and found that the readings did not deviate from that of the WW until less than 0.3 NTR on organic soil and ≈0.1 NTR on sand. These results indicated the NDVI measurements would not be useful in guiding water deficit management schemes as a result of the delayed sensitivity of this measurement for detecting stress vs. visual observations.

Overall, these results indicate important options for enhancing the conservation of water used to irrigate turfgrasses. Zoysiagrass may be well suited for a water deficit irrigation scheme because of its lower water use rate than the other grasses and its rapid recovery potential after a period of severe drought. However, these results are based on common root zones among these three grasses, which is not the case in the field. Zoysiagrass tends to have the shallowest root zone than the other two grasses examined, so the delayed visual quality decline resulting from drought stress observed under greenhouse conditions may not translate to the field as a result of the smaller amount of soil water available to this grass. Nonetheless, a key finding is that the grass plants examined can be allowed to be stressed to the onset of visual stress and then rewatered for rapid recovery. The delay in rewatering in this scheme takes advantage of including days of comparatively low water use and increases the opportunity for natural rains to occur to supply the turf with water with minimal risk of permanent damage. Of course, the challenge will be to convince homeowners that it is appropriate to allow grasses to show visual symptoms of stress before irrigating.