Agronomy Journal 92:423-427 (2000)
© 2000 American Society of Agronomy
TURFGRASS MANAGEMENT
Using Near Infrared Reflectance Spectroscopy to Schedule Nitrogen Applications on Dwarf-Type Bermudagrasses
Ian R. Rodrigueza and
Grady L. Millerb
a Horticulture Dep., Clemson Univ., Clemson, SC 29634 USA
b Environmental Horticulture Dep., Univ. of Florida, Gainesville, FL 32611 USA
glmi{at}gnv.ifas.ufl.edu
 |
ABSTRACT
|
|---|
Due to the high rates of N fertility necessary for producing high-quality turfgrasses, quick, reliable methods of determining the N status of turfgrasses would be valuable management tools. The first objective of this study was to evaluate the use of near infrared reflectance spectroscopy (NIRS) to schedule N fertilization on two dwarf-type bermudagrasses [Cynodon dactylon (L.) Pers. x C. transvaalensis Burtt Davy]. The second objective was to test the accuracy of NIRS-predicted mineral tissue concentrations. The third objective was to study the effect of N fertility on thatch development. `Tifdwarf' and `FloraDwarf' bermudagrasses grown on sand–peat (9:1 by volume) were subjected to five treatments using time, NIRS-predicted N thresholds, and a visual quality rating threshold to schedule applications of (NH4)2SO4 for 20 wk per growing season in 1997 and 1998. There were positive linear relationships between total Kjeldahl nitrogen (TKN) and NIRS-predicted N in 1997
and 1998
. NIRS-scheduled fertility resulted in similar quality with less fertilizer than time or visual quality-based fertility. The NIRS mineral concentration predictions for K, Ca, Mg, Fe, Zn, Mn, and Cu were positively correlated with traditional laboratory methods, but there was not sufficient precision in measurements to use NIRS for determination of these nutrients. Thatch development and yields were greater in treatments receiving higher rates of fertilizers, suggesting that excessive growth rates due to high rates of applied fertilizer may have contributed to thatch development.
Abbreviations: NIRHi, high NIRS N-threshold • NIRLow, low NIRS N-threshold • NIRS, near infrared reflectance spectroscopy • SCHHi, scheduled high N • SCHLow, scheduled low N • TKN, total Kjeldahl nitrogen • TNC, total nonstructural carbohydrate(s) • VIS, fertilized based on visual rating
|
INTRODUCTION
|
|---|
NEAR INFRARED REFLECTANCE SPECTROSCOPY analysis is a nondestructive method for measuring the chemical composition of materials with simple sample preparation. NIRS technology is based on near infrared absorption properties, which can be measured and used to differentiate one compound from another in a tissue sample (Marten et al., 1985). NIRS has been in development for 30 yr, and new uses for this technology are being developed as it is refined. NIRS techniques have been used to measure such forage quality criteria as digestibility, energy intake, and botanical composition as accurately as conventional laboratory analyses (Eckman et al., 1983; Moore et al., 1990) and to evaluate moisture, oil, and starch concentrations in many food commodities (Halgerson et al., 1995; Hattey et al., 1994; Roy et al., 1993). Research has also shown the potential of using NIRS to accurately predict the response of corn (Zea mays L.) to fertilizer and to the N-supplying capability of a soil (Fox et al., 1993). Mineral analysis of Ca, P, and K in forages using NIRS has been proven to be reliable (r2 > 0.74) (Clark et al., 1987). NIRS has been shown to significantly decrease time and labor involved in measuring total nonstructural carbohydrates (TNC) in several turfgrasses (Shepard et al., 1990), with a correlation of
between laboratory TNC and NIRS predictions in Tifdwarf and `Tifway' bermudagrasses (Miller and Dickens, 1996). NIRS techniques have been used to predict creeping bentgrass (Agrostis palustris Huds.) thatch composition (Couillard et al., 1994), and to assess fungal infection levels in tall fescue (Festuca arundinacea Schreb.) (Hill et al., 1987; Roberts et al., 1988). NIRS-predicted N concentrations have been correlated with TKN for perennial ryegrass (Lolium perenne L.)
and creeping bentgrass
(Murphy, 1993). The main advantages of NIRS analysis for many constituent determinations are time savings, simple sample preparation, and nondestructiveness of the sample. Scanning a sample typically takes less than 3 min. Sample preparation usually only involves drying and pulverizing, and the samples can be tested repeatedly by NIRS or by another procedure, as they are not consumed (Marten et al., 1985).
Although N fertilizers are required to maintain quality turfgrasses, excessive N applications can lead to turf problems. Excessive N can result in decreased root growth (Oswalt et al., 1959), hindering water and nutrient uptake and recuperative ability. Thatch accumulation is commonly associated with excessive N application (Cisar et al., 1991; Turgeon, 1980). Excessive thatch can cause scalping, hydrophobic properties (Taylor and Blake, 1982), and can provide habitat for insects and disease organisms (Duble, 1996). Frequent applications of soluble forms of N have been shown to cause greater thatch accumulation than infrequent applications and slow-release N fertilizers (Sartain, 1985). Excessive application of N fertilizers can lead to environmental problems. Groundwater contamination has become a serious problem in North America (Hubbard and Sheridan, 1989; Weil et al., 1990). High annual rainfall, sandy soils, and shallow water tables place Florida at high risk for groundwater contamination (Calvert and Phung, 1972; Mansell et al., 1986). It has been shown that in Florida, doubling the amount of a single N application during periods of excessive rains can result in nearly eight times as much N leaching as under normal amounts of precipitation (Snyder et al., 1980). Turfgrass tissue analysis, coupled with proper interpretation of soil testing, may play a significant role in maximizing efficiency in fertility programs. Our first objective was to evaluate the use of near infrared reflectance spectroscopy to schedule N fertilization on two dwarf-type bermudagrasses. Our second objective was to test the accuracy of NIRS-predicted mineral tissue concentrations. Our third objective was to study the effect of N fertility on thatch development.
|
Materials and methods
|
|---|
A 2-yr field study was conducted from 19 May 1997 to 15 Oct. 1997 and from 31 Mar. 1998 through 3 Aug. 1998 to evaluate the utility of using NIRS for developing N fertilization schedules. Studies were conducted on 10-month-old stands of FloraDwarf and Tifdwarf bermudagrasses [Cynodon dactylon (L.) Pers. x C. transvaalensis Burtt Davy] grown on a U.S. Golf Association (USGA) green (9:1 sand:peat by volume) at the University of Florida Turfgrass Envirotron in Gainesville, FL. Test design was a randomized complete block design with five treatments replicated four times as blocks on each cultivar. Plot size was 3.0 by 3.0 m. Irrigation was applied as needed to maintain high-quality turf. Grasses were maintained at a height of 3 mm by mowing three to four times per week, with clippings removed at each mowing.
Nitrogen fertilizer treatments were applied in the form of (NH4)2SO4 (21-0-0) according to one of five treatments. Treatments were designed to compare the effectiveness of using visual turf quality, time schedules, and NIRS predictions to schedule N fertilization. NIRS-scheduled treatments were based on preliminary observations of visual turf quality and NIRS-predicted N. High NIRS N-threshold (NIRHi) treatments were fertilized at 24 kg N ha-1 when NIRS analysis showed a leaf N concentration <45 g kg-1. Low NIRS N-threshold (NIRLow) treatments were fertilized at 24 kg N ha-1 when NIRS analysis showed a leaf N concentration <40 g kg-1. Scheduled high N (SCHHi) treatments were fertilized at 24 kg N ha-1 wk-1. Scheduled low N (SCHLow) treatments were fertilized at 12 kg N ha-1 wk-1. One treatment was fertilized on the basis of a visual rating (VIS). All plots were rated on a scale of 1 to 10, with 10 representing an optimum rating and 1 representing a dead stand. When a rating fell below 7 on a VIS plot, N fertilizer was applied at 24 kg N ha-1.
Predicted levels of N tissue concentrations were determined weekly for all plots using NIRS methods. Grass clippings were collected 6 d after treatment applications using a walking greens mower with a catch basket. Clippings were rinsed with tap water to remove dust and soil particles. Approximately 250 mL of rinsed clippings were placed on paper plates and dried in a 1000-W microwave oven for 6 to 8 min. Dried samples were ground in a cyclone mill (Sample Mill, Udy Corp., Fort Collins, CO) to pass through a 1.0-mm screen, placed into sample cups, and loaded into a NIRS scanning instrument (Model 5000, Foss NIRSystems, Silver Spring, MD). A software package installed on an accompanying computer predicted N tissue concentrations for each plot based on spectral data (Toro Diagnostic Software Version 2.4, The Toro Co., Bloomington, MN).
Due to ongoing developments in NIRS, a generalized turfgrass NIRS prediction equation was used in 1997 and an updated, bermudagrass-specific equation was obtained and used in 1998. Five samples from each grass cultivar were randomly selected without regard to treatment from weekly harvests for NIRS-predicted N concentration verification using TKN methods (Jones, 1991). To analyze NIRS accuracy for mineral concentration predictions, 32 randomly selected samples of Tifdwarf were analyzed for K, Ca, Mg, Fe, Zn, Mn, and Cu using dry ash procedures (Hue and Evans, 1986). Results from NIRS were compared to lab results using the SAS General Linear Model procedure (SAS Inst., 1987).
Thatch accumulation expressed as percent organic matter loss to ignition (Callahan et al., 1997) was determined once in each plot per growing season. On 10 Oct. 1997 and 14 Oct. 1998, three 2-cm-diam. plugs were randomly collected to a depth of 6 cm in each plot and dried in a forced-air oven at 75°C for 48 h. Samples were weighed, ashed in a muffle furnace for 4.5 h at 450°C, and weighed again to determine sample weight loss on ignition. Clipping yields were determined twice on each plot during the 1998 growing season. Clippings were collected after a 2-d growth period and dried in a forced-air oven at 75°C for 48 h. Dried clippings were weighed and yields were recorded as kg ha-1 d-1.
|
Results and discussion
|
|---|
There were significant year x N treatment and cultivar x N treatment interactions for all data collected (Table 1)
, so means were analyzed separately for years and grass cultivars. Significant differences between years were expected since an updated, bermudagrass-specific NIRS-prediction equation was used in 1998. Since NIRS treatments were fertilized based on NIRS results, application frequencies were dictated by the prediction equation for these treatments. This led to higher fertilizer rates in 1998 due to lower NIRS-predicted N concentrations (Table 2)
. FloraDwarf was higher in NIRS-predicted N concentrations for both years, even though Tifdwarf received significantly more N in 1998 (Table 2). This suggests that there are differences between cultivars in terms of N response, as previously reported in bermudagrasses (Dudeck et al., 1985).
View this table:
[in this window]
[in a new window]
|
Table 1 Mean square from analysis of variance for NIRS-predicted N (NIRS–N), N applied, and visual ratings for dwarf-type bermudagrasses grown in 1997 and 1998
|
|
View this table:
[in this window]
[in a new window]
|
Table 2 Mean NIRS-predicted N concentrations (NIRS-N), N applied, and visual ratings for FloraDwarf and Tifdwarf, as influenced by fertilization treatment during 1997 and 1998
|
|
In 1997 the SCHHi treatments were highest in NIRS-predicted N, visual ratings, and applied fertilizer for both bermudagrass cultivars (Table 2). Higher NIRS-predicted N and visual ratings for SCHHi treatments can be attributed to this treatment receiving two to three times more fertilizer than the other treatments (Table 2). FloraDwarf had similar NIRS-predicted N levels for SCHLow and NIRHi treatments, even though NIRHi treatments received significantly less fertilizer (Table 2). NIRLow treatments applied to Tifdwarf resulted in higher visual ratings than VIS treatments, even though NIRS-predicted N concentrations and applied N were similar (Table 2). These data suggest that using NIRS techniques for scheduling N-fertility may result in similar or greater visual quality and tissue N concentrations while using less fertilizer compared with time-based or visual-based fertilization scheduling. Cultivar did not affect NIRS-predicted N or TKN results for the 200 randomly selected samples used in wet chemistry verification. There was a positive linear relationship
between total Kjeldahl N and NIRS-predicted N using the 1997 prediction equation (Fig. 1)
.
View larger version (20K):
[in this window]
[in a new window]
|
Fig. 1 Relationship between NIRS-predicted N (NIRS-N) and total Kjeldahl nitrogen (TKN) in two dwarf-type bermudagrasses in (A) 1997 and (B) 1998
|
|
In 1998, SCHHi treatments were highest in NIRS-predicted N, visual ratings, and applied fertilizer for both bermudagrass cultivars (Table 2). This can be attributed to the SCHHi treatments receiving from one to two times more fertilizer than other treatments. Nitrogen applications based on NIRS on FloraDwarf resulted in less fertilizer being applied to NIRLow than to SCHLow treatments (Table 2). Despite less total fertilizer being applied, NIRSLow and SCHLow treatments had similar visual ratings when averaged across the growing season. This suggests that NIRS-scheduled fertilization may result in similar quality with less fertilizer than time-scheduled fertilization. There was a positive linear relationship
between total Kjeldahl N and NIRS-predicted N using the 1998 prediction equation (Fig. 1). These results are similar to the findings of Murphy (1993), which correlated total Kjeldahl N with NIRS-predicted N for creeping bentgrass and perennial ryegrass.
NIRS was not accurate in predicting essential mineral concentrations. The bermudagrass-specific equation used in this study did not produce correlation r2 values as high as those found by Clark et al. (1987) for minerals in tall fescue and wheatgrass.
Traditional laboratory mineral analysis showed positive relationships between K
, Ca
, Mg
, Fe
, Zn
, Mn
, and Cu
and NIRS-predicted mineral concentrations. Although these were positive correlations, r2 values were too low (<0.50) to acceptably account for errors in the models. There was no correlation between laboratory-determined P and NIRS-predicted P.
Effects of thatch accumulation became apparent in 1998. Observable scalping in SCHHi treatments was easily noticed after mowing late in the 1998 growing season. Since plots were rated visually just prior to mowing, scalping-damaged plots had recovered before being rated for visual quality and had little influence on visual ratings. Thatch development was higher in FloraDwarf than in Tifdwarf for both growing seasons (Table 3)
. Differences among treatments were not apparent until 1998, with thatch development being greatest (Table 3) in treatments receiving higher rates of N (Table 2). These results suggest that high rates of N and frequent applications are more conducive to the development of thatch. These responses are in agreement with data reported by Sartain (1985). Mean separations by treatment for clipping yields (Table 4)
were similar to results for N applied (Table 2), which suggests that excessive growth rates due to high rates of applied N may have contributed to thatch development. Although Tifdwarf had higher yields (Table 4) and received more N during both growing seasons than FloraDwarf (Table 2), Tifdwarf developed less thatch (Table 3). This suggests that cultivar growth habit may have a greater influence than growth rate and fertilizer rate on thatch development. This conclusion is in agreement with the findings of Dunn et al. (1980).
View this table:
[in this window]
[in a new window]
|
Table 3 Thatch development over two years for FloraDwarf and Tifdwarf, as influenced by cultivar and fertilization treatment
|
|
Results suggest that NIRS can be used to increase the efficiency of a N fertilization program. Although a more accurate equation was used in 1998 (Fig. 1), NIRS-scheduled treatments outperformed other treatments receiving similar amounts of fertilizer more often when the 1997 equation was used (Table 2). This can be attributed to lower N concentration predictions in 1998 (Table 3). With careful establishment of N thresholds based on data from a specific site, results should improve. It was apparent that optimum N thresholds used to schedule N fertilization may not be consistent across cultivars, as responses differed between FloraDwarf and Tifdwarf in this study. While high-fertility, time-scheduled treatments were highest in visual quality and N concentration, they were kept at those levels at the cost of twice as much fertilizer as NIRS-scheduled treatments. For example, N rate increases of 100% and 128% only resulted in visual quality increases of 20% and 24% in SCHHi treatments over SCHLow and NIRHi treatments, respectively, in FloraDwarf in 1997. Higher rates of N led to higher yields and greater thatch development, both translating into increased disposal and labor costs from a management standpoint. At some point N fertilization becomes excessive and NIRS may be a useful tool to determine that point, allowing a turfgrass manager to avoid N losses due to leaching and excessive clipping removal due to surge growth. Using NIRS for scheduling N applications may allow a turfgrass manager to maintain high-quality turfgrass while avoiding budgetary and environmental costs that accompany excessive N fertilization.
In summary, NIRS is a rapid means of accurately determining N concentrations of bermudagrasses. Scheduling N fertility using NIRS can result in similar quality with less fertilizer than time or visual quality-based scheduling. NIRS did not accurately predict mineral concentrations in bermudagrass tissue. Inherent differences in N response, thatch development, and yield are apparent between cultivars of dwarf-type bermudagrasses.SAS Institute 1987
|
NOTES
|
|---|
Journal Paper no. R-06689, Univ. of Florida Agricultural Experiment Station.
Received for publication January 19, 1999.
|
REFERENCES
|
|---|
- Callahan L.M., Sanders W.L., Parham J.M., Harper C.A., Lester L.D., McDonald E.R. Comparative methods of measuring thatch on a creeping bentgrass green. Crop Sci. 1997;37:230-234.[Abstract/Free Full Text]
- Calvert D.V., Phung H.T. Nitrate-nitrogen movement into drainage lines under different soil management systems. Soil Crop Sci. Fla. Proc. 1972;31:229-232.
- Cisar J.L., Snyder G.H., Nked-Kizza P. Maintaining quality turfgrass with minimal nitrogen leaching. Bull. 273. Gainesville: Univ. of Florida, Inst. of Food and Agric. Sci, 1991.
- Clark D.H., Maryland H.F., Lamb R.C. Mineral analysis of forages with near infrared reflectance spectroscopy. Agron. J. 1987;79:485-490.[Abstract/Free Full Text]
- Couillard A.A., Turgeon A.J., Shenk J.S., Westerhaus M.O. Comparison of near-infrared reflectance spectroscopy and wet chemical analysis on thatch composition. HortScience 1994;29:249.[Abstract/Free Full Text]
- Duble R.L. Turfgrasses: Their management and use in the southern zone, 2nd ed College Station, TX: Texas A&M Univ. Press, 1996.
- Dudeck, A.E., C.H. Peacock, and T.E. Freeman. 1985. Response of selected bermudagrasses to nitrogen fertilization. p. 495–504. In Proc. Int. Turfgrass Res., 5th, Avignon, France. 1–5 July 1985. INRA, Paris.
- Dunn J.H., Nelson C.J., Sebaugh J.L. Characterization of thatch, rhizomes, carbohydrates, and spring deadspot in twenty cultivars of bermudagrass. J. Am. Soc. Hort. Sci. 1980;105:653-657.
- Eckman D.D., Shenk J.S., Wangsness P.J., Westerhaus M.O. Prediction of sheep responses by near infrared reflectance spectroscopy. J. Dairy Sci. 1983;66:1983-1987.
- Fox R.H., Shenk J.S., Piekielek W.P., Westerhaus M.O., Toth J.D., Macneal K.E. Comparison of near-infrared spectroscopy and other soil nitrogen availability quick tests for corn. Agron. J. 1993;85:1049-1053.[Abstract/Free Full Text]
- Halgerson J.M., Sheaffer C.C., Hesterman O.B., Griffin T.S., Stern M.D., Randall G.W. Prediction of ruminal protein degradability of forages using near infrared reflectance spectroscopy. Agron. J. 1995;87:1227-1231.[Abstract/Free Full Text]
- Hattey J.A., Sabbe W.E., Basten G.D., Blakeney A.B. Nitrogen and starch analyses of cotton leaves using near infrared reflectance spectroscopy (NIRS). Commun. Soil Sci. Plant Anal. 1994;25:1855-1863.
- Hill N.S., Petersen J.C., Shelby R.A., Dalrymple L.W., Barton F.E., II. Quantification of Acremonium coenophialum in tall fescue seed using near infrared reflectance spectroscopy. Crop Sci. 1987;27:1291-1295.[Abstract/Free Full Text]
- Hue N.V., Evans C.E. Procedures used for soil and plant analysis by the Auburn University Soil Testing Laboratory. Auburn, AL: Dep. of Agron. and Soils. Dep. Ser. 106. Auburn Univ, 1986.
- Hubbard R.K., Sheridan J.M. Nitrate movement to groundwater in the southeastern coastal plain. J. Soil Water Conserv. 1989;44:20-27.
- Jones J.B. Kjeldahl method for nitrogen determination. Athens, GA: Micro-Macro Publishing, 1991.
- Mansell R.S., Selim H.M., Calvert D.V., Stewart E.H., Allen L.H., Graetz D.A., Fiskell J.G.A., Rogers J.S. Nitrogen and water distributions in fertilized sandy soil during irrigation and drainage. Soil Sci. Soc. Am. J. 1986;44:95-102.
- Marten, G.C., J.S. Shenk, and F.E. Barton II (ed.). 1985. Near infrared reflectance spectroscopy (NIRS): Analysis of forage quality. Agric. Handb. No. 643. USDA, Washington, DC.
- Miller G.L., Dickens R. Bermudagrass carbohydrate levels as influenced by potassium fertilization and cultivar. Crop Sci. 1996;36:1283-1289.[Abstract/Free Full Text]
- Moore K.J., Roberts C.A., Fritz J.O. Indirect estimation of botanical composition of alfalfa–smooth bromegrass mixtures. Agron. J. 1990;82:287-290.[Abstract/Free Full Text]
- Murphy, J.A. 1993. Near-infrared reflectance spectroscopy to quantify leaf nitrogen concentration in turfgrass. p.162. In 1993 Agronomy abstracts. ASA, Madison, WI.
- Oswalt D.L., Bertrand A.R., Teel M.R. Influence of nitrogen fertilization and clipping on grass roots. Soil Sci. Soc. Am. Proc. 1959;23:228-230.
- Roberts C.A., Barton F.E., II, Moore K.J. Estimation of Acremonium coenophialum mycelium in infected tall fescue. Agron. J. 1988;80:737-740.[Abstract/Free Full Text]
- Roy S., Anatheswaran R.C., Shenk J.S., Westerhaus M.O., Beelman R.B. Determination of moisture content of mushrooms by vis-NIR spectroscopy. J. Sci. Food Agric. 1993;63:335-360.
- Sartain J.B. Effect of acidity and N source on the growth and thatch accumulation of Tifgreen bermudagrass and on soil nutrient retention. Agron. J. 1985;77:33-36.[Abstract/Free Full Text]
- SAS Institute. SAS user's guide: Statistics, 6th ed Cary, NC: SAS Inst, 1987.
- Shepard, D.P., J.M. Dipaola, J.C. Burns, and C.H. Peacock. 1990. Near infrared reflectance spectroscopy and its role in turf research. p. 182. In 1990 Agronomy abstracts. ASA, Madison, WI.
- Snyder, G.H., E.O. Burt, and J.M. Davidson. 1980. Nitrogen leaching in bermudagrass turf: Effect of nitrogen sources and rates. p. 313–324 In R.W. Sheard (ed.) Proc. Int. Turfgrass Res. Conf., 4th, Guelph, ON, Canada. 19–23 July. Int. Turfgrass Soc. and Ontario Agric. Coll., Univ. of Guelph, Guelph, ON.
- Taylor D.H., Blake G.R. The effect of turfgrass thatch on water infiltration rates. Soil Sci. Soc. Am. Proc. 1982;46:616-619.
- Turgeon A.J. Turfgrass management. Reston, VA: Reston Publishing, 1980.
- Weil R.R., Weismiller R.A., Turner R.S. Nitrate contamination of groundwater under irrigated Coastal Plain soils. J. Environ. Qual. 1990;19:441-448.
This article has been cited by other articles:
|
|
|
|
 
S. Narra, T. W. Fermanian, and J. M. Swiader
Analysis of Mono- and Polysaccharides in Creeping Bentgrass Turf Using Near Infrared Reflectance Spectroscopy
Crop Sci.,
January 1, 2005;
45(1):
266 - 273.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
