Agronomy Journal 93:1136-1141 (2001)
© 2001 American Society of Agronomy
TROPICAL SOIL AND CROP MANAGEMENT
Seasonal Variation in Linear Increase of Taro Harvest Index Explained by Growing Degree Days
Hsiu-Ying Lu*,a,
Chun-Tang Lua,
Lit-Fu Chanb and
Meng-Li Weia
a Dep. of Agron., Taiwan Agric. Res. Inst. (TARI), 189 Chung-Cheng Rd., Wufeng, Taichung Hsien, Taiwan 41301, Republic of China
b Office of Farm Manage., Taiwan Agric. Res. Inst. (TARI), 189 Chung-Cheng Rd., Wufeng, Taichung Hsien, Taiwan 41301, Republic of China
* Corresponding author (iying{at}wufeng.tari.gov.tw)
Received for publication January 4, 2001.
 |
ABSTRACT
|
|---|
Selecting plants with high harvest index (HI) can increase corm yield in wetland taro [Colocasia esculenta (L.) Schott]. Planting time, however, affects the response of HI during the linear increase phase, with temperature being the most important factor affecting taro growth. The objectives of this research were to quantify the relationship of calendar days after planting (DAP) and growing degree days after planting (GDD) to HI in taro and to compare their ability to explain seasonal variation in the linear increase of HI. Data from six planting months of field-grown taro across a 3-yr period were collated to investigate the linear increase in HI. The DAP and GDD (with base temperature of 17°C) were included as independent variables to analyze the three-phase piecewise linear function with HI. Piecewise linear functions based on either DAP or GDD fitted well. The responses during the linear increase phase of HI were, however, more stable across years for the GDD model. Moreover, the model based on GDD was superior to DAP for explaining the seasonal variation of HI in taro. These results indicate that the GDD model is a useful approach for determining weather–crop growth relations in taro. Information gained in this study should help relate phenological responses to seasonal variation in taro.
Abbreviations: DAP, calendar days after planting • GDD, growing degree days after planting • HI, harvest index
|
INTRODUCTION
|
|---|
TARO IS A WETLAND CROP cultivated in many tropical and subtropical areas of the world. Prediction of biomass production is important for scheduling harvest and milling operations (Shih and Snyder, 1984; Mohankumar and Sadanandan, 1990; Goenaga, 1995). Corm yield might be increased by selecting plants with high harvest index (HI) in taro (Chan, 1996). Harvest index is the ratio of economic yield to biological yield and is used to describe the accumulation and redistribution of assimilates to achieve final yield (Donald and Hamblin, 1976; Bange et al., 1998). Both biological yield and HI in taro are sensitive to environmental conditions during the vigorous top-growth and rapid corm-bulking stages (Goenaga, 1995; Chan, 1996). Chan (1996) found that corm dry weight was associated with an increase in HI up to harvest stage or 9 to 10 mo after planting. When total dry weight was held constant, the partial correlation coefficient between HI and corm dry weight was significantly positive. Lu et al. (1999) also reported that the response of HI with time in taro decreased following planting and increased during the vigorous top-growth and rapid corm-bulking stages. The responses of linear increase in HI appeared to be most important for final HI. The linear increase in HI was, however, sensitive to climatic factors and varied with planting times (Lu et al., 1999).
Chan et al. (1998) found that the primary factor governing taro growth rate was temperature. Although moisture stress strongly interacts with temperature in plant growth processes, it had little effect on the growth of wetland taro. The effect of temperature on development rate has been described using a thermal-time concept, which assumes that phenological development is constant per degree of temperature between a base temperature and an upper threshold temperature, above and below which the development rate is zero. Many different thermal indices have been used to predict dates of flowering and maturity in crops (Cross and Zuber, 1972; Coelho and Dale, 1980; Mitchell et al., 1997; Stewart et al., 1998). The most commonly used index is growing degree days after planting (GDD). The GDD unit for a given day is defined as the difference between the daily mean temperature and a growth threshold temperature (Gilmore and Rogers, 1958; Cross and Zuber, 1972). A GDD index is obtained by summing the daily GDD from planting to the phase of plant ontogeny desired, usually flowering or maturity in many cases.
With the GDD method, there is a linear relation between GDD and rate of plant development (Wang, 1960). It is assumed that photoperiod does not influence the rate of crop development (Wang, 1960). For domesticated crops grown in areas where they are adapted, development may seem to be independent of photoperiod (Daughtry et al., 1984). Although temperatures and photoperiod interact to influence the development of corn (Zea mays L.), thermal models are generally accepted as adequate at predicting plant growth and development (Mederski et al., 1973; Kiniry et al., 1983; Daughtry et al., 1984; Russelle et al., 1984; Stewart et al., 1998). The practical impact of using GDD is considerable (Ritchie and NeSmith, 1991). For example, GDD can be used to classify plants for their flowering dates or length of cycle, to estimate harvest maturity, and to predict the duration between two developmental stages (Bonhomme, 2000).
The objectives of this research were to quantify the relationship of calendar days after planting (DAP) and GDD to HI in taro and to compare their ability to explain seasonal variation in the linear increase of HI. To find better ways of estimating linear HI increase in taro, calendar days and thermal-unit formulas were used and compared. Using data from six planting months of taro across a 3-yr period, the DAP and GDD models were tested for their ability to explain seasonal variation in the linear increase of HI.
|
MATERIALS AND METHODS
|
|---|
Three consecutive year-round experiments of wetland taro were conducted at Taiwan Agricultural Research Institute (TARI) in Taichung, Taiwan from 1994 to 1996. Planting months were January, March, May, July, September, and November. Seed corms were planted in a randomized complete block design with four replicates for all of the experiments. Seed corms were spaced at 70 by 30 cm with 200 plants per plot. Routine field management practices were followed. Twelve plant samples were taken from each plot at monthly intervals from planting to harvest. The plants were harvested at 9 mo (in 1994) or 10 mo (1995 and 1996) after planting. Temperature data at TARI were also collected during the experimental periods.
Harvest index was determined from each sample taken after planting in all experiments as the ratio of corm dry weight (economic yield) to dry weight of the aboveground vegetative organs and corm (biological yield) (Chan, 1996). The HI data, calculated from each plot (six planting months by 3 yr), were plotted against DAP and GDD to obtain information on the lag phase, the phase of linear increase, and cessation of linear increase. For each model, a three-phase piecewise linear function (as shown in Fig. 1) was fitted using weighted triple exponential functions in SigmaPlot Version 4 (SPSS, 1997a, 1997b):
 | (1) |
where HI is harvest index as modeled by three separate equations, f1, f2, and f3, representing three line segments joined at their endpoints as shown in Fig. 1. Equations f1, f2, and f3 describe the responses of HI during the three periods: the lag phase, linear increase phase, and maturity stage, respectively. The variable t is the time (using DAP or GDD) after planting. This model has eight parameters: t1 is the start of the lag phase; T1 is the time at the intersection of f1 and f2, i.e., the start of the linear increase phase; T2 is the time at the intersection of f2 and f3, i.e., the cessation of the linear increase phase; t4 is the time at harvest; x1 is the HI value at t1; x2 is the HI value at the intersection of f1 and f2; x3 is the HI value at the intersection of f2 and f3; and x4 is the HI value at t4.
View larger version (22K):
[in this window]
[in a new window]
|
Fig. 1. A three-phase piecewise linear function fitted to harvest index (HI) of wetland taro planted in March 1995. The HI is modeled by three separate equations, representing three line segments joined at their endpoints. The equations, f1, f2, and f3, describe the responses of HI during the three periods: the lag phase, linear increase phase, and maturity stage, respectively. The time parameters, t1, T1, T2, and t4, represent the start of the lag phase, the start of the linear increase phase, the cessation of the linear increase phase, and the time at harvest, respectively. The corresponding HI values at separate times are shown as x1 for the HI value at t1, x2 for the HI value at T1, x3 for the HI value at T2, and x4 for the HI value at t4.
|
|
Because the response of linear increase in HI was most important for final HI (Lu et al., 1999), analysis was focused on this linear increase phase. Rate and duration for the phase of linear increase in HI were calculated for each plot as follows:
 | (2) |
 | (3) |
The GDD index for the period from planting to harvest was defined as:
 | (4) |
where Tmax and Tmin are maximum and minimum air temperatures, respectively, and Tb is the base temperature. For daily mean temperatures, (Tmax + Tmin)/2, less than the base temperature, GDD = 0. In this study, GDD were calculated with a base temperature of 17°C, which for taro, was estimated by Lu et al. (2001).
The DAP and GDD were used to analyze HI data from the three year-round experiments. For each model, an analysis of variance, with years as blocks, was conducted on the rate and duration of linear increase in HI to make comparisons among planting times. Treatment means were separated by Fisher's protected LSD. Treatments were considered significantly different at P < 0.05.
|
RESULTS AND DISCUSSION
|
|---|
There were only slight differences in temperature among years, and thermal time accumulation was nearly identical (Fig. 2). The parameter estimates of the three-phase piecewise linear functions fitted to HI data using DAP and GDD after planting are listed in Table 1. All regressions were significant, and all coefficients of determination (R2) exceeded 0.91. Estimates of the square root of mean square of prediction difference were low for both DAP and GDD models. Piecewise linear functions based on either DAP or GDD fitted well.
View larger version (36K):
[in this window]
[in a new window]
|
Fig. 2. Accumulation of growing degree days (GDD), with base temperature (Tb) of 17°C, during the growth of taro plants, 1994–1996.
|
|
View this table:
[in this window]
[in a new window]
|
Table 1. Parameter estimates of the three-phase piecewise linear functions fitted to harvest index (HI), and responses (rate and duration) of linear HI increase in taro for each planting time (1994–1996) using calendar days after planting (DAP) and growing degree days after planting (GDD).
|
|
The rate and duration of linear increase in HI calculated from each plot are shown in Table 1. Variations in both rate and duration of linear increase were found among different planting times, regardless of whether DAP or GDD was used. The responses during the phase of linear increase were, however, more stable across years for the GDD model (Fig. 3). This indicated that the formula for GDD better explained the seasonal variation in linear increase of HI than the formula for DAP. Three-phase piecewise linear functions fitted to HI data for the GDD model are shown in Fig. 4.
View larger version (27K):
[in this window]
[in a new window]
|
Fig. 3. Rate and duration of linear increase in harvest index (HI) of wetland taro planted in different months of the year as calculated from piecewise linear functions based on (A) calendar days after planting (DAP) and (B) growing degree days after planting (GDD). (• represents mean across 3 yr, and error bars represent 2 x standard error among years.)
|
|
Final HI differed significantly among planting months (Table 2). Taro planted in January and March had the highest final HI, whereas taro planted in July and September had the lowest (Table 2). It was assumed that dry matter production in the aboveground vegetative organs was increased by high temperature and high solar radiation during the vigorous top-growth stage for January and March crops and by a higher amount of N transferred to the corm (Chan et al., 1999; Wei et al., 1999). This resulted in greater source efficiency and rapid corm bulking, which in turn, favored yield at maturity. On the other hand, declining temperature and lower solar radiation during the vigorous top-growth stage for July and September crops reduced production of photosynthate in the aboveground vegetative organs. Less assimilate was transferred to the corm during this stage (Wei et al., 1999), and dry matter and N accumulation in the aboveground vegetative organs still increased during the rapid corm-bulking stage (Chan et al., 1999). This resulted in poor yield at maturity. The May and November crops, with intermediate corm yield, showed similar patterns in the accumulation of dry matter and N. The accumulation rates for May and November crops were lower than those for the January crop (Chan et al., 1999). These results showed that final HI was closely related to plant performance during the linear increase phase in HI.
View this table:
[in this window]
[in a new window]
|
Table 2. Final harvest index (HI) and the responses (rate and duration) of linear HI increase in taro planted in different planting months averaged across 3 yr (1994–1996).
|
|
When using the linear increase phase to predict final HI, the formula based on DAP was less efficient than the one based on GDD (Table 2). In the DAP model, no significant differences were found in the duration of linear increase among planting months, and the effect of seasonal variation on final HI could not be explained satisfactorily by the seasonal differences in the rate of linear increase. In the GDD model, the January and March crops had a longer duration of linear increase that favored final HI. Although the September crop had the highest rate of linear increase, its final HI was the lowest among treatments. This was principally caused by the short duration of linear increase. Both the lower rate and shorter duration of linear increase reduced the final HI for the July crop. There were no significant differences in the rate of linear increase among the May, November, and January crops; but the shorter duration of the linear increase period resulted in lower final HI for taro plants planted in May and November. There was a clear relationship between weather and linear increase in HI for the GDD model. Therefore, the GDD model was superior to the DAP model for explaining the seasonal variation of HI in taro. We anticipate that use of the three-phase piecewise linear function based on GDD may lead to the recognition of phenological responses to seasonal variation in taro.
|
ACKNOWLEDGMENTS
|
|---|
This work was supported by the National Science Council, Grant NSC no. 85-2321-B055-005, 86-2321-B055-002, and 87-2321-B005-003.
|
NOTES
|
|---|
Contribution no. 2051 from TARI.
|
REFERENCES
|
|---|
- Bange, M.P., G.L. Hammer, and K.G. Rickert. 1998. Temperature and sowing date affect the linear increase of sunflower harvest index. Agron. J. 90:324–328.[Abstract/Free Full Text]
- Bohnomme, R. 2000. Bases and limits to using ‘degree days’ units. Eur. J. Agron. 13:1–10.
- Chan, L.F. 1996. Harvest index in relation to dry biomass production and distribution in wetland taro [Colocasia esculenta (L.) Schott]. J. Agric. Res. China 45:174–185.
- Chan, L.F., C.T. Lu, and H.Y. Lu. 1998. Growth analysis of wetland taro [Colocasia esculenta (L.) Schott] under various crop seasons. J. Agric. Res. China 47:220–241.
- Chan, L.F., C.T. Lu, M.L. Wei, and H.Y. Lu. 1999. Effect of planting seasons on accumulation of dry matter and nitrogen in wetland taro [Colocasia esculenta (L.) Schott]. J. Agric. Res. China 48:34–48.
- Coelho, D.T., and R.F. Dale. 1980. An energy-crop growth variable and temperature function for predicting corn growth and development: Planting to silking. Agron. J. 72:503–510.[Abstract/Free Full Text]
- Cross, H.Z., and M.S. Zuber. 1972. Prediction of flowering dates in maize based on different methods of estimating thermal units. Agron. J. 64:351–355.[Abstract/Free Full Text]
- Daughtry, C.S.T., J.C. Cochran, and S.E. Hollinger. 1984. Estimating silking and maturity dates of corn for large areas. Agron. J. 76:415–420.[Abstract/Free Full Text]
- Donald, C.M., and J. Hamblin. 1976. The biological yield and harvest index of cereals as agronomic and plant breeding criteria. Adv. Agron. 28:361–405.
- Gilmore, E.C., Jr., and J.S. Rogers. 1958. Heat units as a method of measuring maturity in corn. Agron. J. 50:611–615.[Abstract/Free Full Text]
- Goenaga, R. 1995. Accumulation and partitioning of dry matter in taro [Colocasia esculenta (L.) Schott]. Ann. Bot. 76:337–341.[Abstract/Free Full Text]
- Kiniry, J.R., J.T. Ritchie, and R.L. Musser. 1983. Dynamic nature of the photoperiod response in maize. Agron. J. 74:115–119.
- Lu, C.T., H.Y. Lu, L.F. Chan, and J.L. Lin. 2001. Prediction of base temperature in wetland taro [Colocasia esculenta (L.) Schott] from dynamic partitioning of current assimilate. J. Agric. Assoc. China 2:1–15.
- Lu, H.Y., C.T. Lu, and L.F. Chan. 1999. Dynamic model of harvest index in wetland taro [Colocasia esculenta (L.) Schott]. J. Agric. Res. China 48:86–99.
- Mederski, H.J., M.E. Miller, and C.R. Weaver. 1973. Accumulated heat units for classifying corn hybrid maturity. Agron. J. 65:743–747.[Abstract/Free Full Text]
- Mitchell, R.B., K.J. Moore, L.E. Moser, J.O. Fritz, and D.D. Redfearn. 1997. Predicting developmental morphology in switchgrass and big bluestem. Agron. J. 89:827–832.[Abstract/Free Full Text]
- Mohankumar, C.R., and N. Sadanandan. 1990. Growth and rooting pattern of taro [Colocasia esculenta (L.) Schott]. J. Root Crops 16:61–63.
- Ritchie, J.T., and D.S. NeSmith. 1991. Temperature and crop development. p. 5–29. In J. Hanks and J.T. Ritchie (ed.) Modeling plant and soil systems. Agron. Monogr. 31. ASA, CSSA, and SSSA, Madison, WI.
- Russelle, M.P., W.W. Wilhelm, R.A. Olson, and J.F. Power. 1984. Growth analysis based on degree days. Crop Sci. 24:28–32.[Abstract/Free Full Text]
- Shih, S.F., and G.H. Snyder. 1984. Leaf area index and dry biomass of taro. Agron. J. 76:750–753.[Abstract/Free Full Text]
- SPSS. 1997a. SigmaPlot user's manual. SigmaPlot 4.0 for Windows. SPSS, Chicago, IL.
- SPSS. 1997b. Transforms & regressions reference manual. SigmaPlot 4.0 for Windows. SPSS, Chicago, IL.
- Stewart, D.W., L.M. Dwyer, and L.L. Carrigan. 1998. Phenological temperature response of maize. Agron. J. 90:73–79.[Abstract/Free Full Text]
- Wang, J. 1960. A critique of the heat unit approach to plant response studies. Ecology 41:785–790.[ISI]
- Wei, M.L., L.F. Chan, C.T. Lu, and H.Y. Lu. 1999. Comparison on the photosynthetic production in wetland taro during plant development between crop seasons. J. Agric. Res. China 48:49–66.
This article has been cited by other articles:
|
|
|
|
 
K. Girma, S. L. Holtz, D. B. Arnall, L. M. Fultz, T. L. Hanks, K. D. Lawles, C. J. Mack, K. W. Owen, S. D. Reed, J. Santillano, et al.
Weather, Fertilizer, Previous Year Yield, and Fertilizer Levels Affect Ensuing Year Fertilizer Response of Wheat
Agron. J.,
November 6, 2007;
99(6):
1607 - 1614.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
H.-Y. Lu, C.-T. Lu, M.-L. Wei, and L.-F. Chan
Comparison of Different Models for Nondestructive Leaf Area Estimation in Taro
Agron. J.,
March 1, 2004;
96(2):
448 - 453.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
, and 
represent observed value, and solid line represents fitted curve using three-phase piecewise linear function.)