Crop Science Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via ISI Web of Science (1)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Brown, P. W.
Right arrow Articles by Kopec, D. M.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Brown, P. W.
Right arrow Articles by Kopec, D. M.
Agricola
Right arrow Articles by Brown, P. W.
Right arrow Articles by Kopec, D. M.
Related Collections
Right arrow Turfgrass Management
Right arrow Turfgrass
Crop Science 41:1197-1206 (2001)
© 2001 Crop Science Society of America

TURFGRASS SCIENCE

Penman Monteith Crop Coefficients for Use with Desert Turf Systems

Paul W. Brown*,a, Charles F. Mancinob, Michael H. Youngc, Thomas L. Thompsona, Peter J. Wierengaa and David M. Kopecd

a Dep. of Soil, Water, and Environmental Sci., Univ. of Arizona, Tucson, AZ 85721
b The Scotts Company, 14111 Scottslawn Rd., Marysville, OH 43401
c Division of Hydrological Sci., Desert Research Institute, 755 Flamingo Rd., Las Vegas, NV 89119
d Dep. of Plant Sciences, University of Arizona, Tucson, AZ 85721

* Corresponding author (pbrown{at}ag.arizona.edu)


   ABSTRACT
TOP
 NOTES
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Irrigation scheduling systems which estimate actual evapotranspiration (ETa) by adjusting reference evapotranspiration (ETo) with crop coefficients (Kcs) have been suggested as a means of improving irrigation management of turfgrass in the desert southwest. The objective of this study was to develop turfgrass Kcs for use with ETo computed by the Penman Monteith Equation recommended by the United Nations Food and Agricultural Organization (FAO). Crop coefficients were developed for fairway quality ‘Tifway’ bermudagrass (Cynodon dactylon L. x C. transvaalinsis Davy) in summer and overseeded ‘Froghair’ intermediate ryegrass (Lolium perenne x L. multiflorum) in winter by relating daily measurements of ETa obtained from weighing lysimeters to ETo computed with meteorological data. Monthly and seasonal Kcs were developed by (i) computing the mean of individual daily Kcs, (ii) dividing cumulative ETa by cumulative ETo for the period, and (iii) computing the slope of least squares regression lines relating ETa to ETo. The three computation procedures did not greatly affect the resulting Kc value. Bermudagrass Kcs ranged from 0.78 in June to 0.83 in September, with monthly variation related to turf growth rate. Use of a constant Kc of 0.80 would suffice for estimating ETa in summer. Monthly Kcs for intermediate ryegrass ranged from 0.78 in January to 0.90 in April and varied in relation to mean air temperature. Increased bulk surface resistance resulting from chill-induced reductions in stomatal conductance and/or a reduction in turf growth rate and leaf area index may lower ETa and Kcs during the colder winter months, making use of a constant seasonal Kc less suitable in winter. An inverse linear relationship was obtained between the coefficient of variation of mean monthly Kc and the ratio of measured to theoretical clear sky solar radiation, indicating Kcs are less reliable during periods of cloudy weather.

Abbreviations: agl, above ground level • DSW, desert southwest • ETa, actual evapotranspiration • ETo, reference evapotranspiration • FAO, United Nations Food and Agricultural Organization • gr, growth rate • Kcs, crop coefficients • KDTRF, Karsten Desert Turf Research Facility • LAI, leaf area index • P, precipitation • VER, vertical extension rate


   INTRODUCTION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
POPULATION GROWTH and an expanding tourism industry have led to a rapid increase in the amount of land dedicated to irrigated turfgrass in the desert southwest (DSW). The water requirement of DSW turfgrass is quite high because of the arid climate which generates high rates of evaporative demand, limited precipitation, and mild winter temperatures that allow for year round culture of turfgrass. Local, state, and federal agencies charged with monitoring and/or securing water supplies are particularly concerned about the rapid expansion of urban irrigation in the DSW, and there is growing public pressure to regulate irrigation of turfgrass strictly. Scientific irrigation scheduling regimes which compute irrigation water requirements on the basis of estimates of actual evapotranspiration (ETa) have been suggested as one means of improving irrigation management of turfgrass. The required values of ETa are usually obtained by multiplying estimates of reference ET (ETo) computed from meteorological data by a correction factor known as a crop coefficient (Kc).

Daily values of ETo are readily available from public weather networks in the region (Snyder et al., 1985; Brown et al., 1988; Mott and Sammis, 1990; Devitt et al., 1992), and many golf courses and parks operate on-site weather stations to provide local ETo data. With this widespread availability of ETo data, access to reliable Kcs becomes a limiting factor when implementing scientific irrigation scheduling systems for turfgrass. A number of studies have addressed the water requirements of turfgrass grown in the DSW (Erie et al., 1982; Kneebone and Pepper, 1982; Devitt et al., 1992; Garrott and Mancino, 1994), but only a few studies have presented Kcs. Tovey et al. (1969) used drainage lysimeters to examine the water use of Tifway and ‘Tifgreen’ bermudagrass during the summer at Reno, NV. Bermudagrass Kcs based on the Penman Equation (Penman, 1963) averaged 0.89 when irrigation was applied weekly. Kopec et al. (1991) used ETa obtained from small bucket lysimeters to compute Kcs for ‘Midiron’ bermudagrass in summer and perennial ryegrass (Lolium perenne L.) in winter. Bermudagrass Kcs based on the modified Penman Equation (Snyder and Pruitt, 1985; Brown, 1998) ranged from 0.83 during mid-season to 0.72 during transition to winter dormancy. Crop coefficients for perennial ryegrass averaged 0.87, but decreased to about 0.50 during cold weather.

Devitt et al. (1992) installed drainage lysimeters on two golf courses and one park at Las Vegas, NV, to examine the water requirements and Kcs of common bermudagrass [C. dactylon (L.) Pers.] overseeded in the fall with perennial ryegrass. They found golf turf used 41% more water than park turf and attributed the difference in water use to differences in fertilizer management. Monthly Kcs based on the modified Penman procedure described by American Society of Civil Engineers (ASCE, 1973) ranged from 0.43 in February to 0.89 in June and July for fairway turf. Crop coefficients for the lower maintenance park turf ranged from 0.33 in February to 0.60 in August.

Garrot and Mancino (1994) examined the water requirements of ‘Texturf-10’, Tifgreen, and Midiron bermudagrass subjected to a deep and infrequent irrigation regime which returned the soil to field capacity only after visible wilt was observed during the afternoon hours. Crop coefficients based on a modified Penman Equation (Snyder and Pruitt, 1985; Brown, 1998) were computed for the entire drying cycle and ranged from 0.57 for Midiron to 0.64 for Texturf-10. Crop coefficients ranged from 0.70 to 0.80 during the first few days after irrigation.

Meyer and Gibeault (1987) summarized the results of a series of applied water studies at Santa Ana, CA, and concluded that monthly Kcs based on the modified Penman Equation, used by the California Irrigation Management Information System [(CIMIS); Snyder and Pruitt, 1985], ranged from 0. 55 to 0.79 for warm season grasses and 0.60 to 1.04 for cool season grasses. They reported that the quality of bermudagrass remained acceptable when irrigated using a constant annual Kc of 0.6 while a higher annual Kc of 0.8 was required to maintain acceptable quality in cool season grasses.

Allen et al. (1998) recommend using Kc of 0.80 to 0.85 for warm season grasses and 0.90 to 0.95 for cool season grasses when using the FAO Penman Monteith Equation to estimate ETo. However, they do not provide direct references for research supporting their recommendations.

The Kcs reported from previous DSW studies of turf water use exhibit considerable variation which is due in part to differences in cultural factors such as turf height, turf quality and irrigation regime. Another factor contributing to the variation in Kcs is the differing computation procedures used by the various researchers to estimate ETo. Brown et al. (1998) compared the procedures used by public weather networks in the DSW to estimate ETo and found ETo could differ by as much as 25% under identical meteorological conditions. Such extreme variation in ETo makes transfer of Kcs from one state or region to another difficult because Kcs should be matched to a specific ETo procedure to ensure good estimates of ETa. The FAO and ASCE have identified this disparity in ETo computation procedures as a confusing and potentially limiting factor in the implementation of effective scientific irrigation scheduling systems, and have recently recommended using a standardized computation procedure for ETo based on the Penman Monteith Equation (Allen et al., 1998, 2000). Movement toward a benchmark or standardized procedure for computing ETo will simplify the use and/or transfer of Kcs, provided Kcs are available for the new procedure.

The objectives of this study were to develop monthly and seasonal Kcs for use with Tifway bermudagrass in summer and overseeded Froghair intermediate ryegrass in winter when ETo is computed using the FAO Penman Monteith Equation.


   MATERIALS AND METHODS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
The study was conducted between November 1994 and September 1997 at the University of Arizona Karsten Desert Turf Research Facility (KDTRF) located at Tucson, AZ. The KDTRF resides in an alluvial valley at an elevation of 713 m above mean sea level. Soil at the site is classified as an Agua sandy loam (coarse-loamy over sandy or sandy skeletal, mixed, calcareous, thermic Typic Torrifluvent).

Two large weighing lysimeters, centrally located within the 2.2-ha field research area, were used to measure turf ETa. The lysimeters were constructed and installed by Precision Lysimeters Inc. (Red Bluff, CA) and are cylindrical in shape with a diameter of 2.5 m and depth of 4 m. A Vinton fine sand (sandy, mixed, thermic Typic Torrifluvent) serves as the soil for both lysimeters. The soil was sieved to remove coarse fragments in excess of 3 mm prior to placement in the lysimeters. Bulk density of the lysimeter soil is uniform with depth and averages 1.50 ± 0.01 Mg m-3 (Young et al., 1996).

Each lysimeter rests on a modified truck scale (Model FS-8, Cardinal Scale Mfg., Webb City, MO) which is connected to load cell with a 45-kg capacity (Model Z-100, Cardinal Scale Mfg.). Outputs from scale load cells are monitored using a CR7 data logger (Campbell Scientific Inc., Logan, UT) programmed to sample load cell output every 2 s and output 10-min averages of lysimeter mass. Scale calibration was checked each October by placing calibration weights on covered lysimeters during nighttime hours. Scale accuracy resulting from calibration was found to be ±300 g (equivalent to water depth of 0.06 mm).

Water draining to the bottom of the lysimeters is removed using a vacuum pump that is attached to a series of suction candles located 25 cm above the bottom of the lysimeters. A suction of 15 kPa is applied to the suction candles for a period of 4 h each day to remove drainage water. Drainage water is stored in tanks on the lysimeters; drainage therefore does not affect lysimeter mass readings until water is removed and quantified by lysimeter technicians (typically once/week). Additional details on lysimeter design and operation are provided in Young et al. (1996).

Tifway bermudagrass, established on the lysimeters and the surrounding 0.1-ha area by sprigging during the summer of 1994, served as the turf surface during the summers of 1995, 1996, and 1997. Froghair intermediate ryegrass was overseeded into the bermudagrass at a rate of 670 kg ha-2 on a pure live seed basis in October of each year and served as the turf surface during the winters of 1994–1995, 1995–1996, and 1996–1997. The turf received N at a rate of approximately 35 kg ha-1 mo-1 from irrigation water and chemical fertilizer (NH4SO4 in liquid form). Monthly applications of fertilizer N were adjusted based on the irrigation rate and N concentration in the irrigation water. Potassium and phosphorus were applied every 6 wk at rates of 24 and 16 kg ha-1, respectively. Granular K2SO4 (0-0-52) and Ca(H2PO4)2 (0-20-0) served as fertilizer sources for K and P, respectively. The turf was mowed two to three times per week during the summer and one to two times per week during the winter with a reel mower. Mowing height was set at 2.2 cm in summer and 2.5 cm in winter. Clippings were captured during each mowing event between April 1995 and September 1997 to facilitate measurement of turf growth rate in g m-2 d-1. Clippings were dried to a constant mass in an oven set at 65°C and the resulting dry mass divided by the area of the lysimeter and the number of days since the last mowing event.

A dual irrigation system serves the lysimeter site, allowing use of either tertiary effluent or groundwater for irrigation of the lysimeters. To accommodate a companion research study, one lysimeter was irrigated with effluent while the other lysimeter was irrigated with groundwater. The two waters differed in quality in two areas of potential importance: electrical conductivity (0.4 dS-1 for groundwater and 1.0 dS-1 for effluent) and total nitrogen (3 mg N L-1 for groundwater and 13 mg N L-1 for effluent). Irrigation was applied in excess of evaporative demand to ensure necessary leaching of soluble salts, and N fertilization was adjusted to account for N differences in the two water sources. Irrigation was supplied to each lysimeter with four low trajectory Rainbird 1804 Series pop-up sprinkler heads (Rainbird Intl., Glendora, CA) installed in a square pattern with head spacing set at 3.65 m. Each lysimeter was centered within the square sprinkler array. The precipitation rate and distribution uniformity of the irrigation system were monitored on a regular basis by placing catch cans on each lysimeter prior to an irrigation event. Over the course of the study, precipitation rate averaged 48 mm h-1 ± 3 mm h-1 and the low quarter mean distribution uniformity was 0.75 ± 0.05.

Irrigation was applied daily in the predawn hours with irrigation amount set equal to 100% of ETo as computed with the Rainbird Maxi-5 irrigation control system and its attendant weather station. The Maxi-5 system computes ETo using the modified Penman Equation described by ASCE (1973). During periods of rainfall, irrigation amount was determined by subtracting rainfall from ETo during the previous 24-h period. Irrigations were eliminated on days when rainfall exceeded ETo. Rainfall amounts in excess of ETo were assumed stored in the soil and used to offset future evaporative demand (ETo) with the proviso that this total stored rainwater could never exceed 12.5 mm. Irrigations resumed once this stored supply of rainwater was depleted.

Fetch upwind of the lysimeters was not completely uniform because of the presence of various research turf blocks. Wind flow at the site is dominated by a mountain valley flow regime which causes wind to blow from the west during the day and from the east at night (Brown et al., 1995). The 0.1-ha turf block containing the lysimeters was maintained in the same manner as the lysimeter turf and thus provided a uniform fetch of approximately 15 to 20 m to northwest and 20 to 25 m to the east. An additional 60 m of mixed fetch existed east and west of the lysimeter turf block. Irrigated turfgrass occupied approximately 70% and 85% of the area to the east and west of the lysimeter area, respectively. Access roads and plot alleyways not planted to grass occupied the remaining fraction of the fetch lengths. Abrupt changes in turf height were not deemed a major problem in the mixed fetch areas; turf located upwind of the lysimeters was generally maintained at heights within 2 cm of the lysimeter turf.

Evapotranspiration was determined daily in units of mm for the 24-h period ending at 0000 h by the following soil water balance equation:

(1)
where I is the amount of irrigation, P is precipitation, {Delta}S is the daily change in soil moisture storage and D is the amount of drainage. Irrigation was applied on most days during a 15-min period just prior to sunrise to facilitate measurement of I from the resulting gain in lysimeter mass. Evaporation was assumed negligible during the brief irrigation period. Precipitation was measured with a tipping bucket rain gauge located 20 m south of the lysimeters. The change in lysimeter mass was assumed equal to {Delta}S, and D in units of kg was obtained by multiplying the volume of drainage water in liters by a specific gravity of 1.0 kg L-1. Changes in lysimeter mass and the mass of water removed in D were converted to an equivalent depth of water in mm by dividing by the lysimeter conversion factor of 4.91 kg mm-1. Daily ETa values obtained from the two lysimeters were typically within 3% of one another and were averaged to provide the ETa value used in this study.

Meteorological data were acquired 20 m south of the lysimeters over well watered tall fescue maintained at a height of approximately 8 cm. Air temperature and relative humidity were measured at 1.5 m above ground level (agl) in a naturally aspirated radiation shield (12-plate Gill shield) using an HMP35C temperature and humidity sensor (Campbell Scientific Inc., Logan, UT). Solar radiation was measured at a height of 2 m agl with a LI-COR LI 190 silicon cell pyranometer (LI-COR, Inc., Lincoln, NE) while wind speed and direction were measured at 3 m agl with a Wind Sentry cup anemometer and wind vane (model 03101-5, R.M. Young Co., Travers City, MI). Precipitation was measured at 2.5 m agl with a TE525 tipping bucket rain gauge (Texas Electronics, Dallas, TX). The meteorological sensors were connected to a 21X micrologger (Campbell Scientific, Inc.) programmed to monitor sensor outputs at 0.1 Hz and generate relevant means, totals, and extremes of the meteorological parameters each hour and for the 24-h period ending at midnight.

Reference evapotranspiration (ETo) was computed by means of the simplified form of the FAO Penman Monteith Equation for use with 24-h time steps (Allen et al. 1994, 1998):

(2)
where Rn is net radiation in MJ m-2 d-1, G is soil heat flux in MJ m-2 d-1, {gamma} is the psychrometer constant in kPa °C-1, T is mean daily air temperature in °C, U2 is the mean wind speed measured at 2 m agl in m s-1 (ea - ed), is the vapor pressure deficit in kPa, and {Delta} is the slope of the saturation vapor pressure curve in kPa °C-1. The coefficients 900 and 0.34 vary slightly with measurement height of the temperature and wind speed, but are recommended for the purposes of standardization of ETo computation. The coefficient 0.408 converts units of MJ m-2 d-1 into mm d-1 with a constant latent heat flux of 2.45 MJ kg-1.

Net radiation, Rn, is computed by means of the following:

(3)
where Rs is measured incoming solar radiation in MJ m-2 d-1; Rso is computed clear sky solar radiation in MJ m-2 d-1; ed is the measured average vapor pressure in kPa; {sigma} is the Stefan-Boltzmann constant equal to 4.90 x 10-9 MJ m-2 K-1 d-1; Tkx is the daily maximum air temperature in K; Tkn is the daily minimum air temperature in K; and ac, bc, a1, and b1 are constants equal to 1.35, -0.35, 0.34, and -0.14, respectively. Clear sky solar radiation was computed from:

(4)
where z is station elevation in m and Ra is the computed extraterrestrial radiation in MJ m-2 d-1. Extraterrestrial radiation is obtained from:

(5)
where dr is relative distance between the earth and sun, {omega}s is the sunset hour angle in rad, {phi} is the latitude in rad, and {delta} is the solar declination in rad. The relative distance between the earth and sun is computed as follows:

(6)
where J is the day number in the year. Sunset hour angle is obtained from:

(7)

Solar declination is computed by:

(8)

Soil heat flux, G, was assumed equal to zero for the 24-h period.

The psychrometer constant is computed by:

(9)
where P is atmospheric pressure in kPa and {lambda} is assumed to be a constant value of 2.45 MJ kg-1. Atmospheric pressure was estimated from elevation in m above sea level, z, by:

(10)

Slope of the saturation vapor pressure curve is computed by:

(11)
where T is the mean air temperature for the day in °C.

Wind speed at 2 m (U2) was obtained by adjusting wind speed measured at 3 m above ground level (U3) from the following:

(12)
where z2 is the adjusted height of 2 m, d is the displacement, zom is the roughness length for momentum, and z3 is the measurement height of 3 m. Displacement length was estimated from the crop height, h, in m by the following:

(13)
where h was set equal to the height of the standardized reference of 0.12 m. Roughness length for momentum was also computed from crop height by the following:

(14)
with h again set equal to 0.12 m.

Vapor pressure deficit was computed using estimates of ea and measured values of mean daily ed. Saturation vapor pressure was computed from:

(15)
where T is air temperature in °C. Mean ea for the day was computed from the average of the ea values computed at maximum and minimum temperature (Tx and Tn):

(16)

The period of study encompassed three turf years which were defined as the period 1 November through 30 September. No measurements were made during the period of overseed establishment in October. The three turf years will subsequently be referred to as turf years 1994–1995, 1995–1996, and 1996–1997 and abbreviated TY94/95, TY95/96, and TY96/97, respectively. Within each turf year, the winter turf season ran from 1 November through 31 May while the summer turf season covered the period 1 June through 30 September.

Crop coefficients were determined on a seasonal and monthly basis for winter and summer turf seasons using three different procedures identified as (i) period bulk summation (PBS), (ii) daily mean computation (DMC), and (iii) least squares regression (LSR). The PBS involved summing the total ETa for the monthly or seasonal period and dividing by the summation of ETo for the same period. The DMC procedure involved computing a Kc value for each day in the period, then computing the mean value of the daily Kcs values for the period in question as well as the coefficient of variation (CV) for the mean Kc value. The LSR procedure generated crop coefficients from the slopes of least squares regression lines computed with ETo and ETa as the independent and dependent variables, respectively. Two forms of the regression line were evaluated (i) with finite computed intercept and (ii) with intercept set equal to zero. Lines with computed intercepts were used to test the validity of using regression lines with forced zero intercepts by testing whether the intercept was significantly different from zero. The slope of the resulting regression lines, dETa/dETo was assumed equal to Kc.


   RESULTS AND DISCUSSION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Seasonal ETo and Precipitation
Total ETo and P for each season and turf year are presented in Table 1. Winter totals represent the summation of ETo or P for the months of November through May. Summer totals represent similar summations for the months of June through September. Turf year totals cover the 12-mo period ending 31 October.


View this table:
[in this window]
[in a new window]
  		Table 1. Seasonal and annual totals of Penman Monteith reference evapotranspiration (ETo) and precipitation (P), and mean values of ETo and P for the three turf years. Winter and summer totals include the months of November through May and June through September, respectively. Annual ETo includes ETo for the month of October when overseeded turf was established.

 
Turf year and summer totals of ETo were reasonably consistent over the three years of the study. Reference ET ranged from 1803 mm in TY94/95 to 1908 mm in TY95/96, and averaged 1865 mm for the three years of study. Cumulative ETo during the four summer months accounted for slightly less than 50% of annual ETo and ranged from 860 mm in TY94/95 to 891 mm in TY95/96, with a 3-yr average of 876 mm. Summer P was quite variable in terms of both total accumulation and the number of P days, ranging from 87 mm on 14 d in TY94/95 to 183 mm on 25 d in TY95/96. The nearly two-fold difference in summer P between TY94/95 and TY95/96 did not greatly affect ETo because summer precipitation is short-lived, convective in nature and skewed toward the late afternoon hours—characteristics that do not greatly affect incoming solar radiation, the dominant energy source for evaporation.

Winter ETo totals exhibited considerably more year-to-year variation about the 3-yr mean of 851 mm, ranging from 774 mm in TY94/95 to 916 mm in TY95/96. In contrast to the summer situation, winter ETo was heavily influenced by the amount and frequency of P. The amount and frequency of winter P ranged from 168 mm on 35 d in TY94/95 to 58 mm on 14 d in TY95/96. The inverse relationship between P and ETo occurs in winter because P results from passage of low pressure centers which generate extended periods of cloudiness and P.

Computation Procedure for Crop Coefficients
The three computation procedures have the potential to generate different Kc values since they effectively weight the daily ETa and ETo data differently. The PBS approach computes Kcs from monthly or seasonal sums of ETa and ETo and provides greater weight to days when evaporative demand (ETo) is high, since such days affect monthly sums more than days with low evaporative demand. In contrast, the DMC procedure provides equal weights to Kcs computed for each day, regardless of the level of evaporative demand. For the LSR technique, Kc values are assumed equal to the slope of the least squares regression line which is computed by dividing the sum of the cross products of ETo and ETa by the sum of squares of ETo. This computation procedure places greater emphasis on extreme values of ETa and ETo in a given data set since the cross products and sum of squares for such data points are large. Biases or errors in data collected at the extremes of ETa and ETo can therefore be expected to generate greater impact on Kcs than similar biases or errors associated with the remainder of the data.

The procedure used to compute Kcs did not significantly affect the resulting value, suggesting no serious bias resulted from any of the three computation procedures. Table 2 presents the monthly and seasonal Kcs for TY96/97 computed by means of the PBS, DMC, and LSR procedures. The Kcs derived from LSR represent the slopes of the regression lines when the intercept is forced through the zero. Crop coefficients obtained from the three computation procedures differed by 0.01 or less in nine of the 11 months and for both turf seasons as a whole. The range in Kcs resulting from computation increased to 0.02 to 0.03 in colder months (e.g., December 1996) and months when turf growth rate (gr) was changing rapidly (e.g., May 1997). During the three summer turf seasons studied (12 mo total), the range in monthly Kc resulting from computation exceeded 0.02 in only one month. In winter, the range in Kc exceeded 0.02 in just four of the 21 months examined. In months where the Kc range exceeded 0.02, the DMC and PBS procedures were in general agreement while the LSR procedure tended to produce the most extreme Kcs. One factor affecting Kcs derived from the LSR approach is the assumption that the resulting regression line runs through the origin. Such an assumption, while seemingly valid, should be checked by testing whether the intercept of the regression line is significantly different from zero. Completion of this statistical test on the 33 months of study data revealed the intercept was significantly different from zero about 40% of the time in both winter and summer, indicating a need to include the intercept in the regression equations. Use of intercepts in the LSR approach lessens the practical usefulness of the resulting Kcs since the intercept would cause the Kcs to vary with ETo.


View this table:
[in this window]
[in a new window]
  		Table 2. Monthly and seasonal crop coefficients (Kc) for turf year 1996–1997 derived using the period bulk summation (PBS), daily mean (DMC) and least squares regression (LSR) computation procedures. Range designates the largest difference in Kc resulting from computation procedure.

 
Given the identified limitations of the LSR computation procedure and the similarity of Kcs derived from the PBS and DMC procedures, we chose to use the DMC approach to present the results of this study. The DMC approach provides the additional advantage of allowing the computation of various statistical parameters such as standard deviations and coefficients of variation (CVs) for mean Kcs derived for the monthly and seasonal periods.

Summer Crop Coefficients
Bermudagrass Kcs are presented by month in Fig. 1 and Table 3. The Kcs presented in Fig. 1 represent the average Kc for each summer month over the three years of study; error bars represent the root mean square deviation (RMSD) about the mean. Crop coefficients obtained for individual months in each year are presented in Table 3 along with CVs for each monthly Kc. Crop coefficients were relatively constant when averaged over the three years of the study and ranged from 0.78 during June and July to 0.83 in September. Year-to-year variation in monthly mean Kc, as indicated by the RMSD, was greatest in June and declined with each subsequent summer month. The slight increase in Kc and the reduced year-to-year Kc variation during the later summer months is related to gr and density of the bermudagrass. Growth rate and visual density of bermudagrass are lowest in June following the death of the overseeded ryegrass, then increase with each subsequent summer month. Higher year-to-year variation in June and July reflect the normal year-to-year variation in turf performance in early summer. Bermudagrass Kcs increased linearly with turf gr derived from clipping weights (Fig. 2). The least squares regression line presented in Fig. 2 is Kc = 0.75 + 0.021 x gr (r2 = 0.70) where gr is in units of g m-2 d-1. The slope of the regression line is significantly different from 0 at P < 0.05.



View larger version (42K):
[in this window]
[in a new window]
  		Fig. 1. Average crop coefficient (Kc) by month for Tifway bermudagrass for the period June through September of 1995, 1996 and 1997. Error bars represent the RMSD obtained for the three years of study.

 

View this table:
[in this window]
[in a new window]
  		Table 3. Mean monthly and seasonal crop coefficients (Kc) and coefficients of variation (CV) for Kc for bermudagrass in turf years 1994–1995, 1995–1996 and 1996–1997. Mean Kc values obtained over the three years of the study as well as the root mean square deviation (RMSD) of the mean Kc values are presented under All Years.

 


View larger version (14K):
[in this window]
[in a new window]
  		Fig. 2. Monthly crop coefficient (Kc) of Tifway bermudagrass plotted as a function of the monthly bermudagrass growth rate (gr) in g m-2 d-1. The resulting least squares regression line is Kc = 0.75 + 0.21*gr with r2 = 0.70.

 
The relationship between Kc and growth rate likely reflects seasonal changes in leaf area index (LAI) which affect both absorption of Rs and bulk surface resistance (rs) to ET. Monteith and Unworth (1990) indicate LAI >4 is required to absorb ~90% of available Rs in most plant canopies. Measurements of LAI were not obtained in this study; however, research conducted by others indicates LAI of turf is typically <4 (Brede and Duich, 1984; Kopec et al., 1987; Johns et al., 1981). Changes in gr observed in this study were due to changes in both turf density and vertical extension rate (VER) and likely were associated with changes in LAI which would alter radiation absorption by the canopy.

Changes in gr could also affect rs through changes in canopy structure (Kim and Beard, 1988; Shearman, 1989) and VER which alters LAI (Kim and Beard, 1988; Allen et al., 1998). Johns et al. (1981) examined the resistance to ET of St. Augustinegrass [Stenotaphrum secundatum (Walter) Kuntze], which they divided into three components consisting of aggregate diffusive resistance of foliage, resistance to air mass exchange within the canopy which they termed canopy resistance, and resistance to turbulent exchange between the canopy and the bulk air (aerodynamic resistance). They found diffusive resistance of foliage, computed from measurements of stomatal resistance and LAI, and total resistance declined with increasing canopy height. Kim and Beard (1988), using the Johns et al. (1981) definition of canopy resistance, indicated canopy resistance increased with increasing shoot and leaf density and more horizontal leaf orientation; however, they also found ET increased with VER and leaf area, suggesting an increase in LAI reduces the diffusive resistance of the foliage. In practice, the diffusive foliage and canopy resistance terms described by Johns et al. (1981) are commonly combined into one resistance term referred to as rs (Allen et al., 1998) which typically declines with increasing LAI (Allen et al., 1994). It is therefore possible that the relationship between Kc and gr observed in this study reflects variations in rs caused by growth-induced changes in LAI.

The results of this study suggest bermudagrass Kcs values decline by ~0.10 during periods of slow growth when the turf is irrigated daily. Crop coefficients during similar slow growth periods may decline more than 0.10 with less frequent irrigations which would allow the soil surface to dry more thoroughly between irrigations. Our results are in agreement with previous work by Kopec et al. (1991) who found Kcs of Midiron bermudagrass decreased from ~0.83 in mid-summer to 0.73 during the fall when growth declined as the grass progressed into dormancy. Even larger decreases in water use and/or Kcs have been reported for bermudagrass when nitrogen fertility limits growth (Kneebone and Pepper, 1982, Devitt et al., 1992).

The rather small seasonal variation in bermudagrass Kcs suggests a constant Kc would suffice for estimating bemudagrass water use during the summer season. The mean Kc for the entire summer period averaged 0.80 with a RMSD of just 0.02. Use of a constant Kc of 0.80 as compared with monthly Kcs (Fig. 1 or Table 3) would generate differences in estimated monthly ETa at Tucson, AZ, of approximately 3%, or <8 mm mo-1—an amount unlikely to cause water stress in months with higher Kcs, nor serious over watering in months with lower Kcs.

The summer Kc values for bermudagrass agree with the findings of Qian et al. (1996). They did not compute Kcs in their work, but did report that the slope of the regression line relating ETa of Midiron to Penman Monteith ETo was equal to 0.80. Our bermudagrass Kcs also agree quite well with those of Devitt et al. (1992) who found Kcs ranged from 0.82 to 0.89. However, the Nevada Kcs are based on the modified Penman Equation described by ASCE (1973) which generates summer ETo values at Tucson that are 10 to 15% higher than ETo computed by the Penman Monteith Equation. The Nevada coefficients would therefore need to be increased by 10 to 15% for use with the Penman Monteith Equation. Following this adjustment, the Nevada Kc values range from 0.92 to 0.99 or ~20% higher than the Kcs reported here. Slightly higher Kcs should be expected from the Nevada study since their summer turf consisted of 40% ryegrass and 60% bermudagrass. Cool season grasses typically use 15 to 20% more water than warm season grasses when grown side by side under similar conditions (Feldhake et al., 1983; Kneebone and Pepper, 1982; Meyer and Gibeault, 1987). The mixed stand of summer turf should therefore use 5 to 10% more water than a pure bermudagrass stand. However, the adjusted Nevada Kc values approach 1.0 in June and July which would be considered unlikely for a short turf surface containing 60% bermudagrass. We suspect our adjustment of the Nevada data, which was based on the differences in ETo procedures, may be in error. It is possible the two ETo procedures agree more closely when compared in Las Vegas where wind flow is typically higher than in Tucson (Ruffner and Bair, 1977).

The Kc values obtained are larger than those reported by reported by Meyer and Gibeault (1987) for bermudagrass (0.62–0.71) growing during the summer months near Santa Ana, CA. In later work, Gibeault et al. (1988) suggested a Kc of 0.60 was optimum for bermudagrass in California. It is important to note that the California Kc values were developed and validated from applied water studies in which total water applied was increased ~35% above ETa (computed from Kc x ETo) to accommodate irrigation nonuniformity. Meyer and Gibeault (1987) reported minimal subsurface drainage in their studies; therefore, it is likely some of the excess water applied to accommodate irrigation nonuniformity was in fact used by the turf. A 35% increase in the California Kc value of 0.60 produces a seasonal Kc of 0.81, which agrees quite well with the seasonal Kc reported here.

The CVs for monthly Kcs (Table 3) provide insight into the day-to-day variation in Kcs. Coefficients of variation ranged from 0.057 to 0.185 over the three summers studied, with the lower CVs occurring in summer months that produced relatively cloud free skies and consistent temperatures (e.g., June and September). Higher CVs resulted in the mid-summer monsoon months of July and August when cloudiness, higher humidity and precipitation are more common. Higher CVs suggest the Penman Monteith Equation is less accurate at estimating ETo during periods of unsettled weather. Qian et al. (1996) reported a similar finding when relating ET of Midiron bermudagrass to ETo estimated with the Penman Monteith Equation.

Winter Crop Coefficients
Monthly Kcs for overseeded ryegrass are presented for each winter month in Fig. 3 and in Table 4. The ryegrass Kc values in Fig. 3 represent the average monthly value obtained over the three winter turf seasons; error bars in Fig. 3 depict the RMSD of the monthly Kcs. Monthly Kc values changed in a systematic pattern during the winter turf season and ranged from 0.78 in January to 0.89 in April. Lower Kcs were evident during the colder, winter months while the peak Kcs were observed during the warmer winter months of November, March, April and May. The seasonal trend in winter Kcs resembles the trend in mean air temperature in Tucson with the exception of May when Kcs declined because of the negative impact of high daytime temperatures on ryegrass.



View larger version (56K):
[in this window]
[in a new window]
  		Fig. 3. Average crop coefficient (Kc) by month for Froghair intermediate ryegrass overseeded into bermudagrass during the winters of 1994–1995, 1995–1996 and 1996–1997. Error bars represent the RMSD obtained for the three years of study.

 

View this table:
[in this window]
[in a new window]
  		Table 4. Mean monthly and seasonal crop coefficients (Kc) and coefficients of variation (CV) for Kc for overseeded ryegrass in turf years 1994–1995, 1995–1996 and 1996–1997. Mean Kc values obtained over the three years of the study as well as the root mean square deviation (RMSD) of the mean Kc values are presented under All Years.

 
Lower Kcs during the colder winter months could result if low temperatures delay and/or lessen stomatal opening during the daylight hours and thereby increase rs. Temperatures at or below freezing occurred with some regularity at the research site during the months of December through February. Slack and Kopec (1988) reported canopy temperatures of well-watered perennial ryegrass were well in excess of air temperature following cold winter nights and suggested chill was reducing stomatal conductance during the daylight hours. Moon et al. (1990) observed that photosynthetic efficiency and stomatal conductance of perennial ryegrass were reduced for several days following just 1 d of chill (8°C day and 5°C night). They concluded that reduced stomatal conductance resulting from chill stress was mainly due to reduced photosynthetic efficiency which increased intracellular CO2 and led to reduced stomatal aperture. Both Kc and gr (from clippings) were found to increase in a linear fashion with mean air temperature (T) during the winter period (Fig. 4; gr data not shown) which would support the hypothesis that chill was impacting photosynthetic efficiency and stomatal conductance. The resulting least squares regression line for the relationship between Kc and T was Kc = 0.66 + 0.01 x T with r2 = 0.50. The slope of 0.01 is significantly different from zero at P < 0.05.



View larger version (15K):
[in this window]
[in a new window]
  		Fig. 4. Monthly crop coefficients (Kc) of overseeded Froghair intermediate ryegrass plotted as a function of mean monthly air temperature (T) for the three years of the study. The resulting least squares regression line is Kc = 0.66 + 0.01*T with r2 = 0.50.

 
Temperature-induced changes in gr could also account for the apparent temperature effect on Kc. Turf gr declined from initially high values in November to minimum levels during the cold months of December and January, then increased rapidly in late winter and early spring in response to warming temperatures which enhanced VER and tillering. Kim and Beard (1988) reported that turf species exhibiting the highest ET rates typically have higher VER and greater leaf area than species with lower ET rates. Shearman (1989) evaluated the water use of 12 varieties of perennial ryegrass and found ET to be positively correlated with VER, but negatively correlated with verdure, and suggested both factors were affecting rs. Higher VER could enhance ET by impacting both the turbulent transport of air above the canopy (aerodynamic resistance) and rs. Higher VER effectively increases the average height of the turf at each mowing event and results in a slightly higher mean turf height which should lower aerodynamic resistance (Allen et al., 1994). A higher VER may also lead to a higher LAI (Kim and Beard, 1988) which affects rs. Allen et al. (1998) indicate the LAI of clipped ryegrass increases with canopy height, and that rs is inversely proportional to LAI. While LAI was not directly measured in this study, gr data derived from clipping weights should be related to leaf area production and LAI. We believe the higher Kcs obtained during periods with higher gr (and temperatures) are in part due to reductions in aerodynamic resistance and rs resulting from a taller effective turf canopy and greater LAI. A number of other studies have reported that water use and/or Kc is impacted by turf density (e.g., Kneebone and Pepper, 1982, Feldhake et al., 1983, Devitt et al., 1992) and turf height (e.g., Feldhake et al., 1983, Biran et al., 1981; Shearman, 1989).

Devitt et al. (1992) reported a similar yet more pronounced seasonal pattern in Kcs for perennial ryegrass grown during the winter months at Las Vegas, NV. We increased the NV Kcs by ~ 20% to reflect computational differences in winter between the ASCE Penman and Penman Monteith ETo procedures. The adjusted NV Kcs declined from a value of 0.96 in November to 0.51 in February, then increased to 0.90 in April (Fig. 5). The Kcs derived from this study agree quite closely with the adjusted NV Kcs during the warmer spring months of March through May, but differ markedly during the colder months. Two possible reasons for the lower NV Kcs in the colder months are (i) lower mean temperatures at Las Vegas than at Tucson and (ii) infrequent winter irrigation in the Nevada study. Las Vegas winter temperatures run 3 to 4°C cooler than Tucson which should generate slower turf growth and greater potential for chilling conditions. The relationship between Kc and mean air temperature observed in this study would lower Kcs in Las Vegas (relative to Tucson) by only 0.03 to 0.04, suggesting this relationship may not hold up in cooler climates. It is possible the greater level of chill in Las Vegas leads to much greater reductions in stomatal conductance which would lead to lower Kcs.



View larger version (25K):
[in this window]
[in a new window]
  		Fig. 5. Comparison of monthly turfgrass crop coefficients (Kc) obtained in this study (AZ) with Kcs recommended for California (Meyer and Gibeault, 1987) and NV (Devitt et al., 1992). California Kcs for cool season turf are used for the period November through May. Both California and Nevada Kcs were adjusted to reflect computational differences in ETo between the Penman Monteith Equation and the modified Penman Equations used by California and Nevada.

 
Infrequent irrigation may also explain the lower Kcs observed in Nevada. While irrigations in this current study were applied daily, except during period with rainfall, irrigations in the Nevada study were applied as infrequently as one time per week during the colder months (Devitt et al., 1992). Infrequent irrigations would presumably allow the soil surface to dry between wetting events and thus reduce the soil evaporation component of ET leading to a lower overall Kc. The Nevada Kcs come into agreement with the Kcs obtained in this study later in the spring when irrigations return to a more frequent and perhaps daily schedule.

Meyer and Gibeault (1987) also reported a significant month-to-month variation in Kcs during the winter months for cool season grasses. Winter Kcs, after adjustment for computational differences between the CIMIS Penman Equation (Snyder and Pruitt, 1985) and the Penman Monteith Equation, ranged from 0.53 in December to 1.09 in April, and averaged 0.83 for the period November through May (Fig. 5). The range in monthly Kcs obtained in California is much larger than the range observed in this study (0.53–1.09 vs. 0.78–0.90); however, the average winter Kcs are the same (0.83).

The monthly variation in Kcs and the relationship between Kc and mean air temperature call into question the use of a constant Kc for the entire winter season. The peak monthly Kc in April runs 7% higher and the lowest monthly Kc in January runs about 6% below the overall mean Kc for the winter of 0.83, respectively. Use of a constant Kc in winter would lead to over watering from December through February and under watering from March through May.

Day-to-day variation in Kcs was generally greater in winter as compared to summer. Coefficients of variation for monthly Kcs ranged from 0.047 to 0.267 (Table 4) and tended to be inversely related to ETo. The highest CVs were obtained in the months of November through February, while CVs approached summer levels (Table 3) in the warmer spring months of March through May. Higher CVs were also obtained in months exhibiting more changeable or unsettled weather. To further assess the impact of weather on Kc variation, we pooled the CVs for both summer and winter monthly Kcs and plotted the CVs against a measurement of cloudiness defined as the ratio of actual solar radiation to theoretical clear sky solar radiation (Rs/Rso; see Eq. [4] and [5]). Monthly CVs declined as the level of cloudiness declined (Fig. 6), indicating Kcs are less variable during clear weather conditions. Coefficients of variation are generally 0.10 or less when skies are clear (Rs/Rso > 0.90), but nearly double during months with significant cloudiness (Rs/Rso ~ 0.75). The cause of higher Kc variation during cloudy weather was not examined in this study; however, we feel estimation of Rn during cloudy weather could be one factor contributing to less reliable estimates of ETo and thus higher Kc variation.



View larger version (16K):
[in this window]
[in a new window]
  		Fig. 6. Coefficient of variation (CV) of monthly turfgrass crop coefficient plotted as a function of the ratio of actual solar radiation (Rs) to theoretical clear sky solar radiation (Rso) for both winter and summer months. The resulting least squares regression line is CV = 0.53 - 0.46 (Rs/Rso) with r2 = 0.31.

 

   CONCLUSIONS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
Crop coefficients appropriate for use with the FAO Penman Monteith Equation were developed for Tifway bermudagrass and overseeded Froghair intermediate ryegrass subjected to a high maintenance regime which included adequate plant nutrition and daily irrigation. Monthly Kcs for bermudagrass varied with turf growth rate, but the seasonal range in Kc was only ±3% of the overall summer mean Kc value of 0.80. A constant Kc of 0.80 would therefore be effective for estimating ETa for overseeded bermudagrass during the summer months. Use of a constant Kc of 0.80 is not recommended for non-overseeded bermudagrass which has extended periods of slow growth and lower ETa during the spring and fall.

Monthly Kcs for overseeded Froghair intermediate ryegrass varied from 0.78 in January to 0.90 in April, suggesting winter Kcs are dependent upon temperature. Increased rs resulting from reduced ryegrass growth and/or chill-induced reductions in stomatal conductance may be the cause of lower Kcs during the colder winter months. This apparent interaction of Kc with chill/temperature makes use of a constant Kc less practical in winter in the DSW. While the authors believe the winter Kcs will transfer to many other DSW locations, appropriate caution should be exercised when the winter Kcs are used in regions with significantly more or less winter chill.

Crop coefficients were more variable during periods of unsettled or cloudy weather, primarily during the winter months, suggesting irrigation scheduling based on weather data may be less reliable when these conditions prevail. Because ETa and irrigation requirements are often low during periods of unsettled weather due to increased cloudiness and precipitation, the overall impact of less reliable Kc values on estimated ETa may prove minimal.

Finally, it is clear from previous studies that a number of factors affect turf water use and thus Kcs, including turf species (e.g., Biran et al., 1981; Kneebone and Pepper, 1982; Kim and Beard, 1988) and/or variety (Shearman, 1986, 1989), canopy characteristics (e.g., Johns et al., 1981; Shearman, 1986; Kim and Beard, 1988; Shearman, 1989), mowing height (e.g., Madison and Hagan, 1962; Biran et al., 1981), nutrition (e.g., Kneebone and Pepper, 1982; Feldhake et al., 1983; Devitt et al., 1992), irrigation frequency (e.g., Biran et al., 1981; Garrot and Mancino, 1994), and the procedure used to estimate ETo (e.g., Tovey et al., 1969; Brown et al., 1998) The computation procedure for ETo is one factor that adds greatly to the confusion associated with selection and use of turfgrass Kcs. Use of the FAO Penman Monteith Equation for estimation of ETo in future studies related to water use and irrigation of turfgrass would lessen this confusion and is therefore recommended.


   NOTES
TOP
 NOTES
ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
This research supported by grants from the United States Golf Assn.

Received for publication July 28, 2000.


   REFERENCES
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 





This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via ISI Web of Science (1)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow