Crop Phenology for Irrigated
Spring Cantaloupes (Cucumis melo L.)
Roberto Soto-Ortiz and Jeffrey C. Silvertooth
Department of Soil, Water and Environmental Science
University of Arizona
Abstract
Field experiments were conducted in 2007 to evaluate a cantaloupe (Cucumis melo L.)
plant development model as a function of heat units accumulated after planting (HUAP).
Field experiments were conducted in 2007 in the Yuma Valley, Arizona (32o 42’ N, 114o 42’
W), about 150 feet (~ 32 m) elevation in four commercial cantaloupe fields managed by a
cooperator-grower using four varieties. Plant measurements were made on regular 14-day
intervals and the following growth stages were identified in relation to plant measurement
data collection: pre-bloom, early fruit set, early netting, and physiological maturity
(harvest). The model was evaluated by comparing the observed HUAP versus the predicted
HUAP values using a repeated measures design. Mean differences within each sampling
stage were separated using the Fishers’ protected least significance difference (LSD) test at
P≤ 0.01. In addition, regression models were performed for all in-season data collected
and the accuracy of the model was evaluated on the basis of the R2 values with a specified
level significance (α = 0.01). No statistical differences were found between the observed
phenological data and the predicted values from the model throughout the study period.
Also, the model presented an overall accuracy of 54 ± 37 HUAP (2 ± 1 day) in predicting
cantaloupe-harvesting time. It can be concluded that the model can be used as a useful tool
to assist cantaloupe growers in predicting and identifying critical stages of growth for
irrigated spring cantaloupe crops in Arizona and the desert Southwest.
Keywords
cantaloupe, Cucumis melo L., crop phenology, crop growth and development, heat units.
Introduction
Arizona cantaloupe (Cucumis melo L.) production ranks second to California in the United States. In
2006, approximately 23,000 acres of spring cantaloupes were harvested in Arizona, and yields
averaged approximately 250 Cwt/acre (Arizona Agricultural Statistics, 2007). In 2006, total farm value
of spring cantaloupes for Arizona amounted to more than $100 million dollars.
Vegetable Report (P-152), January 2008
113
Harvest timing or seasonality of many vegetable crops, including cantaloupes, plays a major role in
determining produce prices in the marketplace (Tronstad, 1995). We can often monitor and predict
development based on measuring the thermal conditions in the plants environment. Various forms of
temperature measurements and units commonly referred to as heat units (HU) or growing degree units
(GDU), have been utilized in numerous studies to predict phenological events for both agronomic and
horticultural crops (Baker and Reddy, 2001).
Wurr et al. (2002) stated that to describe crop growth and development there is first the need to
determine rate functions for various processes. These include the identification of distinct stages and
phases of growth and development, as well as the duration of developmental phases for given
temperature regimes. Marcelis et al. (1998) stated that the most straightforward models in this field are
based on heat sums. The uses of HU accumulation methods are efficient techniques for modeling and
predicting growth stages in crops (such as cantaloupes) as compared with the traditional days after
planting (DAP) method since variations among seasons and locations can be better normalized by the
use of HU calculations rather than days after planting (DAP).
Soto-Ortiz et al. (2006) developed a cantaloupe plant development model as a function of HUs
accumulated after planting (HUAP) that could be extended as a crop management tool for growers.
However, this model is lacking sufficient validation. Similarly, there are no reports as to the accuracy
of any other similar models to predict developmental stages and the harvest time of cantaloupes.
Therefore, the objective of this study was to evaluate the accuracy of the Soto-Ortiz et al. (2006) model
for cantaloupe plant development and its ability to predict crop developmental stages and harvest dates
of irrigated spring cantaloupes in Arizona.
Materials and Methods
The time line of the Arizona cantaloupe phenological model (Soto-Ortiz et al., 2006), as well as the
regression equations to predict some growth and development parameters are shown in Figure 1 and
Table 1. To provide in field-testing and validation of this model and to provide additional data for the
model, a set of field experiments were conducted in the Yuma Valley, Arizona (32o 42’ N, 114o 42’ W,
about 150 feet elevation) with four commercial cantaloupe fields managed by a cooperator-grower.
These four sites employed four different cantaloupe varieties. The general description of the sites and
some experimental methods are described in Tables 2 and 3. In all fields, melon seeds were dry planted
Vegetable Report (P-152), January 2008
114
on 80-inch beds and furrow irrigated. Each field was thinned to approximately 12-inch plant spacings
(6,450 plants per acre). General fertilization at all sites included 145 lb. N/acre, 30 lb. P/acre and 15 lb.
S/acre. Twenty five percent of the total N and all of the P fertilizers were applied pre-plant using liquid
ammonium phosphate (10-34-0). Fifty percent of the total N and all of the S fertilizer was side-dress
injected before the first irrigation using a mixture of UAN-32 and sulphuric acid (20-0-0-5). The
remaining 25% of the fertilizer N was water-run applied on the second irrigation, using UAN-32 (32-00). Water and pest control procedures were managed throughout the season on an as-needed basis. In
addition, all crop management decisions were made in conjunction with the grower-cooperator.
An automated Arizona Meteorological Network (AZMET) station located close to the experimental sites
monitored weather conditions on a daily basis throughout the growing season. The AZMET station is
used to determine a complete series of meteorological measurements including the hourly maximum and
minimum temperature values. Consequently, the HU accumulations (86/55 ºF thresholds) are calculated
by a method presented in Baskerville and Emin (1969) and modified by Brown (1989). The daily HU
accumulations are summed from the time of planting and reported as HUAP.
In-season data collection for each variety was taken at five randomly selected locations in each field on
regular 14 day intervals and they include the following basic plant growth and development
measurements: number of mainstem nodes, number of fresh flowers on each vine, and length (cm) of
each fruiting vine. Also, the number of melons larger than “golf ball” size in two meter row segments
were counted. Also, the following growth stages were identified: pre-bloom, early fruit set, early netting,
and physiological maturity (time of first harvest). The model was evaluated by comparing the predicted
HUAP values versus the observed HUAP values using a repeated measures design. Mean differences
within each stage of sampling were separated using the Fishers’ protected least significance difference
(LSD) test at P≤ 0.01. In addition, regression models were performed for all in-season data collected and
the accuracy of the model evaluated on the basis of the R2 values using a designated significance level
for testing (α = 0.01). Statistical procedures were consistent with those outlined by Steele and Torrie
(1980) and SAS (SAS Institute, 1999a and 1999b).
Results and Discussion
Figure 2 presents the observed phenological stages for all varieties as expressed in terms of HUAP and
calendar days, versus the projected by the model. Despite the natural variation among varieties,
planting dates, and sites the observed phenological stages matched the projected from the model
Vegetable Report (P-152), January 2008
115
relatively well. The Tuscan and Ocotillo varieties were better described by the model than the Olympic
Gold and Esteem varieties (or sites). The HUAP and calendar day deviation from the model averaged
21 HUAP and 1.5 days for Tuscan, and 26 HUAP and 1.5 days for Ocotillo. The greatest deviation
from the model was observed in Olympic Gold, with an overall deviation of 71 HUAP and 4 calendar
days. The model presented an overall accuracy of 54 ± 37 HUAP in predicting the time of cantaloupe
harvesting. Expressed in terms of calendar days, the predicted time of harvest had an accuracy of 1 to 3
days.
Figure 3 presents the overall statistical comparisons among the observed and predicted phenological
stages. No statistical differences were found between the observed and the predicted data throughout
the study period.
The model was also able to accurately predict some cantaloupe growth parameters (Figure 4). The
corresponding R2 values for the number of nodes, main vine length, and number of melons per 2m2
were highly significant (α = 0.01), and corresponded to 0.91, 0.88 and 0.88 respectively. The R2 value
for the regression between the observed numbers of blooms versus the model predictions was
significant (α = 0.05) with a value of 0.58.
The Arizona model demonstrated a reasonable level of accuracy and precision as a management tool
for commercial spring cantaloupe growers under irrigated conditions common to Arizona and the
desert Southwest.
Acknowledgements
The financial support and valuable cooperation provided by Mr. T. T. Havins, (Grower/cooperator) is
greatly appreciated. Also the hard work and technical assistance provided by the research assistants from
the UA Agronomy program are greatly appreciated.
Vegetable Report (P-152), January 2008
116
References
Arizona Agricultural Statistics Service. 2007. United States Department of Agriculture. National
Agricultural Statistics Service. Annual Statistics Bulletin.
Baker, J.T., and V.R. Reddy. 2001. Temperature effects on phenological development and yield of
muskmelon. Annals of Botany. 87:605-613.
Baskerville, G.L. and P. Emin. 1969. Rapid estimation of heat accumulation from maximum and
minimum temperatures. Ecology 50:514-517.
Brown, P. W. 1989. Heat units. Ariz. Coop. Ext. Bull. 8915. Univ. of Arizona, Tucson, AZ.
Marcelis, L.F.M., E. Heuvenlink, and J. Goudriaan. 1998. Modelling biomass production and yield of
horticultural crops: a review. Scientia Horticulturae. 74:83-111.
SAS Institute. 1999a. The SAS system for Windows. Version 8.0. SAS Inst., Cary, NC.
SAS Institute. 1999b. SAS/STAT user’s guide. Version 8.0. SAS Inst., Cary, NC.
Soto-Ortiz, R., J. C. Silvertooth, and A. Galadima. Crop Phenology for Irrigated Cantaloupes (Cucumis
melo L.) in Arizona. 2006 Vegetable Report. A College of Agriculture and Life Sciences
Report. University of Arizona. Series P-144. http://cals.arizona.edu/pubs/crops/az1419/
Steel, R.G.D. and J.H. Torrie. 1980. Principles and procedures of statistics. McGraw-Hill, New York.
Tronstad, R. 1995. Importance of melon type, size, grade, container and season in determining melon
prices. Journal of Agricultural and Resource Economics. 20:785-790.
Wurr, D.C.E., J.R. Fellows, and K. Phelps. 2002. Crop Scheduling and prediction – Principles and
opportunities with field vegetables. In: Advances in Agronomy. D.L. Sparks (Editor). Volume
76. Academic Press. p.p. 201-234.
Vegetable Report (P-152), January 2008
117
Table 1. Cantaloupe growth models, 2007.
Parameters
Equation
R2
SE
Vine length
y = -44.9 + 0.3x
0.8
31.9
Number of nodes
y = -2.9 + 0.05x
0.8
5.1
Blooms
y = -0.2 + 0.01x – 1.3 x 10-5x2
0.2
1.5
Melons per 2 meter
y = -6.1 + 0.03x
0.6
3.8
R2.- coefficient of determination, SE.- standard error.
Table 2. Cantaloupe experimental sites and basic agronomic information, 2007.
Field
Wet date
Variety
Soil type
Barkley Ranch
8 March
Tuscan Gasden Clay (Fine, montmorillonitic
[calcareous], hyperthermic Typic
Torrifluvents
Amigo Ranch
30 January
Ocotillo Gasden Clay (Fine, montmorillonitic
Block 2
[calcareous], hyperthermic Typic
Torrifluvents
Amigo Ranch
18 March
Esteem Gasden Clay (Fine, montmorillonitic
Block 25
[calcareous], hyperthermic Typic
Torrifluvents
Richardson Ranch 1 March
Olympic Kofa clay (Clayey over sandy or sandy
Gold
skeletal, montmorillonitic [calcareous],
hyperthermic Vertic Torrifluvents
Vegetable Report (P-152), January 2008
118
Table 3. Weather conditions at the Yuma Valley throughout the cantaloupe
phenology study period (January-June), 2007.
Average air
Maximum and
Maximum and
Cumulative
temperature at minimum mean air
minimum mean
precipitation
planting dates
temperature
relative humidity
-------------------oF---------------%
inches
64 ± 11
85 – 52
43 - 29
Planting Date First Bloom Golf-ball size Early Netting
0 HUAP
500 HUAP
700 HUAP
900 HUAP
0.02
Physiological Maturity
1350 HUAP
Figure 1. Spring cantaloupe developmental stages predicted by the model, 2007.
Vegetable Report (P-152), January 2008
119
Tuscan
Observed
Planting Date
0 HUAP
8 March
First Bloom
508 HUAP
13 April
Golf-ball size
730 HUAP
30 April
Early Netting
918 HUAP
10 May
Physiological Maturity
1379 HUAP
31 May
Projected
Planting Date
0 HUAP
8 March
First Bloom
500 HUAP
12 April
Golf-ball size
700 HUAP
28 April
Early Netting
900 HUAP
8 May
Physiological Maturity
1350 HUAP
30 May
Ocotillo
Observed
Planting Date
0 HUAP
30 January
First Bloom
498 HUAP
23 March
Golf-ball size
722 HUAP
9 April
Early Netting
886 HUAP
23 April
Physiological Maturity
1283 HUAP
14 May
Projected
Planting Date
0 HUAP
30 January
First Bloom
500 HUAP
23 March
Golf-ball size
700 HUAP
7 April
Early Netting
900 HUAP
23 April
Physiological Maturity
1350 HUAP
16 May
Olympic Gold
Observed
Planting Date
0 HUAP
1 March
First Bloom
517 HUAP
9 April
Golf-ball size
802 HUAP
30 April
Early Netting
990 HUAP
10 May
Physiological Maturity
1451 HUAP
31 May
Projected
Planting Date
0 HUAP
1 March
First Bloom
500 HUAP
7 April
Golf-ball size
700 HUAP
24 April
Early Netting
900 HUAP
4 May
Physiological Maturity
1350 HUAP
27 May
Esteem
Observed
Planting Date
0 HUAP
18 March
First Bloom
566 HUAP
30 April
Golf-ball size
754 HUAP
10 May
Early Netting
1042 HUAP
24 May
Physiological Maturity
1371 HUAP
8 June
Projected
Planting Date
0 HUAP
18 March
First Bloom
500 HUAP
26 April
Golf-ball size
700 HUAP
7 May
Early Netting
900 HUAP
16 May
Physiological Maturity
1350 HUAP
6 June
Figure 2. Observed phenological development (HUAP and calendar days) versus projected
(model) development, 2007.
Vegetable Report (P-152), January 2008
120
Heat Units Accumulated After Planting (86/55oF)
NS
1400
1200
Observed
Predicted
NS
1000
NS
800
600
NS
400
Early bloom Early fruit set Early netting P. maturity
Growth Stages
Figure 3. Observed versus predicted phenological stages (HUAP), 2007.
(NS.- not significant)
Vegetable Report (P-152), January 2008
121
200
y = -12.0 + 1.19x, R2 = 0.88**
2
y = -2.75 + 1.07x, R = 0.91**
n = 85
n = 85
Observed main vine length
Observed nodes
30
20
10
150
100
50
a
b
0
0
0
10
20
30
0
50
Predicted nodes
150
20
y = -3.22 + 1.43x, R2 = 0.88**
-5 2 2
y = -0.2 + 0.01x - 1.3x10 x , R = 0.58*
n = 85
Observed number of melons per 2 m
2
n = 85
5
4
3
2
15
10
5
c
1
d
0
0.0
0.5
1.0
1.5
2.0
Predicted number of blooms
200
Predicted main vine length
6
Observed number of blooms
100
2.5
3.0
0
5
10
15
20
Predicted number of melons per 2 m2
Figure 4. Observed number of nodes (a), main vine length (b), number of blooms in main vine
(c), and number of melons per 2m2 (d) versus predicted values from the Arizona
model, 2007. * 0.05 and ** 0.01 levels of significance for R2 values.
Vegetable Report (P-152), January 2008
122