Agronomy Journal Grow Your Career With ASA
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published in Agron J 98:451-461 (2006)
DOI: 10.2134/agronj2005.0083
© 2006 American Society of Agronomy
677 S. Segoe Rd., Madison, WI 53711 USA
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 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 HighWire
Right arrow Citing Articles via Web of Science (4)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Izquierdo, N. G.
Right arrow Articles by Cantarero, M. G.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Izquierdo, N. G.
Right arrow Articles by Cantarero, M. G.
Agricola
Right arrow Articles by Izquierdo, N. G.
Right arrow Articles by Cantarero, M. G.
Related Collections
Right arrow Agroclimatology
Right arrow Crop Physiology & Metabolism
Right arrow Seed Quality
Right arrow Sunflower
Right arrow Other Models

Modeling

Modeling the Response of Fatty Acid Composition to Temperature in a Traditional Sunflower Hybrid

Natalia G. Izquierdoa, Luis A.N. Aguirrezábala,*, Fernando H. Andradea and Marcelo G. Cantarerob

a Unidad Integrada Facultad de Ciencias Agrarias (UNMdP), Estación Experimental Agropecuaria INTA Balcarce, CC.C. 276, 7620 Balcarce, Argentina, and Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET, Argentina)
b Facultad de Ciencias Agropecuarias, Universidad Nacional de Cordoba, Av. Valparaiso s/n. Ciudad Universitaria, 5000 Córdoba, Argentina

* Corresponding author (laguirre{at}mdp.edu.ar)

Received for publication March 21, 2005.

   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Oil quality highly depends on its fatty acid composition. In traditional sunflower (Helianthus annuus L.) hybrids, fatty acid composition is affected by night temperature during grain filling. It is unknown if the increase of oleic acid concentration when night temperature increases saturates at a temperature threshold. Modeling the response of fatty acid composition to temperature could lead to oil quality prediction. The objectives of this work were: (i) to develop precise models to estimate fatty acid composition, and (ii) to use the established models to assess differences in final fatty acid composition among regions, sowing dates, and years. The traditional hybrid Dekasol 3881 was exposed to different day–night temperature regimes during grain filling (28 and 20°C, 25 and 23°C, and 20 and 28°C). To model the response of fatty acid composition to temperature data from two field experiments, eight field crops and five growth chamber experiments were analyzed. Night minimum temperature during the 100 to 300 ddaf (degree-days after flowering) period (base temperature = 6°C) accounted for most of the variability in oleic acid concentration (r2 = 0.84). The relationship was linear up to 22.6°C, the temperature at which the maximum value of oleic acid was reached. The model also estimated other fatty acid contents. The relationships accurately predicted independent data from Dekasol 3881 and other hybrids. Our model explained most of the variation in oleic acid concentration observed in a large region were sunflower is cultivated (27°–37° S) and it is feasible for a wide range of environmental conditions.

Abbreviations: ddaf, degree-days after flowering • CYT, comparative yield trials • LCS, lack of correlation weighted by the standard deviations • MSD, mean square deviation • SB, square bias • SDSD, squared difference between standard deviations


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
THE ENVIRONMENT affects fatty acid composition in the oil of traditional sunflower hybrids. The increase in the oleic/linoleic acid ratio with increasing temperature during grain filling has been widely reported (Canvin, 1965; Nagao and Yamazaki, 1984; Sobrino et al., 2003). Prediction of fatty acid composition using temperature at different periods of grain filling has been attempted in a few studies. Harris et al. (1978) established a model to predict linoleic acid concentration based on temperature from flowering to physiological maturity. Nagao and Yamazaki (1984) reported that linoleic acid concentration was best estimated by temperature from 12 d after flowering to physiological maturity. Moreover, Izquierdo et al. (2002), under field conditions, observed in four sunflower hybrids a stronger effect of temperature during the period from flowering to 400 ddaf than during other periods. Temperature affects total activity of oleate desaturase, the enzyme that catalyzes the conversion of oleic to linoleic acid, the two major fatty acids of sunflower oil. Total activity of this enzyme is greatest at early grain filling (Garcés and Mancha, 1991; Garcés et al., 1992; Kabbaj et al., 1996), so temperature at early grain filling would have a stronger effect on fatty acid composition than during other grain-filling stages. More precision about the period during which fatty acid composition is most sensitive to temperature would improve the performance of oil quality prediction models.

It is also unclear which is the best temperature predictor for fatty acid composition. Variations in oleic acid concentration in the field were better explained by maximum (Séller, 1983), minimum (Harris et al., 1978), or daily mean temperature (Nagao and Yamazaki, 1983). Rochester and Silver (1983) observed that low night temperatures reduce oleic acid concentration. Izquierdo et al. (2002), using field and growth chambers experiments, concluded that the best temperature predictor for oleic acid concentration was night temperature and not daily minimum temperature; however, it is still not known which night temperature variable (mean, maximum, or minimum) is the best oil quality predictor for modeling purposes.

The literature is unclear about the type of the response (linear or curvilinear) of oleic acid concentration to temperature. Linear relationships between oleic (or linoleic) acid concentration and temperature were established (Harris et al., 1978; Goyne et al., 1979; Silver et al., 1984) for ranges of daily mean temperature between 15 and 27°C. Trémolières et al. (1982) reported a curvilinear relationship between oleic acid concentration and mean temperature, with a maximum value at approximately 27°C. It appears then that there is an optimum temperature for maximum oleic acid concentration. Moreover, the fact that traditional hybrids never reach oleic acid concentrations >60% would also indicate the existence of an optimum temperature. It is necessary to study this response under a wide range of night temperatures since sunflower is cultivated at several latitudes including subtropical environments.

An oil quality model would also need to predict concentrations of other fatty acids. These concentrations could be estimated through their relationships with oleic acid concentration (Izquierdo et al., 2002).

Improving the relationship between fatty acid concentration and night temperature would allow more accurate prediction of sunflower oil composition across locations, sowing dates, and seasons (Hall, 2004). The objectives of this study were: (i) to develop precise models to estimate fatty acid composition as a function of temperature during a specific period during grain filling and (ii) to use the established models to assess differences in final fatty acid composition among regions, sowing dates, and years.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Description of Crops and Experiments
The traditional hybrid Dekasol 3881 was cultivated in 10-L pots filled with soil (Exp. GC1). Five seeds per pot were sown on 7 January (Replication 1) and on 25 January (Replication 2). Seven days after seedling emergence, plants were thinned to one plant per pot. Soil was fertilized with N, P, S, and B (Izquierdo et al., 2002) and irrigated every 12 h to avoid water stress. Plants were kept under field conditions until treatment initiation. Capitula were covered with self-pollination bags. Flowering of a plant was registered when all florets from the capitulum's outer ring showed their stamens (R5.1, Schneiter and Miller, 1981). Flowering (95% of plants of the plot at R5.1) occurred on 12 March (Replication 1) and 19 April (Replication 2). Plants were exposed to three day–night temperature regimes during grain filling: 28 and 20°C, 25 and 23°C, and 20 and 28°C. Each temperature regime was achieved using growth chambers (Refrimax S.R.L., Mar del Plata, Argentina) with 12-h photoperiod and incident photosynthetically active radiation at the top of the plants of 690 ± 75 µmol m–2 s–1. Shorter plants were raised to receive the same incident radiation at the top. Plants were harvested at physiological maturity.

For modeling purposes, data from two field experiments, eight field crops, and five growth chamber experiments (including GC1) were analyzed. Field crops and experiments were conducted at the Instituto Nacional de Tecnología Agropecuaria (INTA) Balcarce Experimental Station (37° S), Universidad Nacional de Córdoba (31° S), and Tandil (37° S), Argentina (Table 1). The cultivar was the same as in GC1. Nutrient availability was sufficient to support crop yields ≥5000 kg ha–1 (Sosa et al., 1999; Andrade et al., 2000). In field experiments, rainfall was complemented with irrigation to avoid water stress. Pest and disease control was not necessary. Flowering of a plant was registered as in GC1 and flowering of a plot was recorded when 95% of the plants had flowered. Capitula were covered with bags to prevent cross-pollination except for official CYTs (comparative yield trials), where plants were free pollinated. A base temperature of 6°C was used to calculate thermal time.


View this table:
[in this window]
[in a new window]
  		Table 1. Flowering dates, duration of the period flowering to physiological maturity (PM), latitude, and mean night temperature during grain filling for the field experiments, crops, and growth chamber experiments (FB = field at Balcarce; FC = field at Cordoba; FP = field at Paraná; CYT = comparative yield trial).

 
The two field experiments, in which night temperature was increased at different periods during grain filling, were conducted under field conditions at Balcarce (FB1 and FB2). Treatments were heating from 0°C (flowering) to 200°C daf, from 200 to 400°C daf, from 400 to 600°C daf, and an untreated control. Heating during the night (1900–0600 h) was achieved within enclosures and with electrical fan heaters.

The eight field crops were conducted at Balcarce (FB3, FB4, FB5, and FB6), at Cordoba (FC1 and FC2), and in official CYTs at Balcarce and Tandil (CYT Balcarce and CYT Tandil). No treatments were applied to these crops.

In the growth chamber experiments (GC2, GC3, GC4, and GC5), plants were sown in pots filled with soil as described for GC1 and grown under field conditions or in greenhouses until treatment application. During grain filling, plants were exposed to different temperatures using the growth chambers described in GC1. Day–night temperature regimes were: 26 and 16°C, 22 and 20°C, and 16 and 26°C (GC2); 16 and 16°C, 21 and 21°C, 26 and 26°C, and 26 and 16°C (GC3); and 26 and 16°C, 22 and 20°C, and 26 and 26°C (GC4). In GC5, plants were exposed to a regime of 20 and 28°C for 10 d at the beginning, the middle, or the end of grain filling, the temperature regime being 25 and 23°C during the rest of the period. The control was exposed to a regime of 25 and 23°C from flowering to physiological maturity. Further details about FB1, FB2, and GC2 are given in Izquierdo et al. (2002).

In all the experiments, physiological maturity was estimated visually from the hard yellow color of the capitulum's back face and from the brown color of its bracts (Farizo et al., 1982). Plants were harvested at physiological maturity.

Air temperature was measured with thermistors (Stow Away XTI, ONSET, Computer Corp., Pocasset, MA) or Cu constantan thermocouples (Termoquar, Buenos Aires, Argentina). Data were recorded with data loggers. All temperature sensors were previously cross-checked and the maximum difference between sensor measurements was <0.1°C. In CYTs, temperatures were obtained from meteorological stations placed ≤8000 m away from the crops. Mean, maximum, and minimum temperatures were calculated for the entire day (0700–0600 h of the next day) and the night period. The night corresponded to the dark period in the growth chamber experiments or to the time elapsed between sunset and sunrise, corrected for the duration of twilight, in the field crops and field experiments.

Sample and Data Analysis
Fatty acid composition was determined by gas chromatography and expressed as a percentage of total fatty acid content as described by Izquierdo et al. (2002). Data of fatty acid composition from GC1 was processed by analysis of variance. Residuals of fatty acid concentrations were homogeneously distributed around zero so data were not transformed. Treatment means were compared by Duncan test (p < 0.10). The concentration of oleic acid of all the experiments was related to night mean, night minimum, night maximum, daily mean, and daily minimum temperature. In a previous study (Izquierdo, 2000) and in data reported in the literature (Harris et al., 1978; Rochester and Silver, 1983), no effect of daily maximum temperature was observed on oleic acid concentration, so daily maximum temperature was not considered. The relationships were adjusted considering these temperatures at different periods of 100 or 200 degree-days during grain filling as described by Aguirrezábal et al. (2003) for intercepted radiation. Longer periods such as flowering to 400 ddaf or flowering to physiological maturity were also analyzed.

Linear or linear-plus-plateau equations were fitted to the data set. As some functions may be driven by a few data points, a bootstrap analysis was applied to select the most critical period and the best temperature predictor for fatty acid composition. This method is currently used in different fields of biology (Antinuchi and Luna, 2002; Reymond et al., 2003). Adjustments were done with Pop-Tools (a macro-complement for Microsoft Excel downloaded from www.cse.csiro.au/poptools/). One hundred random samplings with replacement were performed and an adjustment was obtained for each sample. Adjustments were evaluated by comparing the distribution of the 100 r2 values. The deviation of data estimated by the model with respect to measured data was quantified by regressions with the Integrated Resources for Evaluating Numerical Estimates (IRENE, software beta version 1.00, Fila et al., 2003) and analyzing the components of the mean square deviation (MSD) as proposed by Kobayashi and Salam (2000). Calculated components were: square bias (SB), squared difference between standard deviations (SDSD), and the lack of correlation weighted by the standard deviations (LCS). A larger SDSD means that the model fails to simulate the magnitude of fluctuation among measurements, and a larger LCS indicates that it fails to simulate the pattern of the fluctuation across measurements. The concentrations of other fatty acids were calculated through their relationship with oleic acid concentration.

Independent data to validate the selected models were obtained from experiments conducted with hybrid DK3881 in growth chambers (GC6) and under field conditions in two locations: Paraná (31° S, FP1) and Balcarce (37° S, FB7). In Exp. GC6, plants were grown under conditions similar to those of the other growth chamber experiments. During grain filling, plants were exposed to the following night minimum temperatures: 19, 21, and 24°C (first sowing date) and 10, 12, and 18°C (second sowing date). For FB7 and FP1, sowing dates were 25 October and 9 December, respectively. Further details about these experiments are shown in Table 1. Prediction quality of the relationships was also evaluated using data from other cultivars. The traditional hybrids ACA 885 (Asoc. Cooperativas Argentinas), DK 3915 (Monsanto S.A.), and VDH 488 (Advanta Semillas S.A.I.C.) were sown at Balcarce (37° S), Manfredi (31° S), Reconquista (29° S), and P.R.S. Peña (27° S) in official CYTs conducted by the Instituo Nacional de Tecnología Agropecuaria of Argentina. Fatty acid composition was measured as described above and estimated using the established relationships. For all validations (DK3881 and other cultivars), the hypothesis of intercept = 0 and slope = 1 for the regressions between observed and estimated data was tested with IRENE (p < 0.05).

The selected relationship between oleic acid concentration and temperature was used to assess differences in oleic acid concentration among four locations in Argentina where the beginning of the grain-filling period typically occurs on different dates. Some of these locations have contrasting mean temperatures during the sunflower growing season. The locations were: Balcarce, Rio Cuarto, Venado Tuerto, and Presidente Roque Saenz Peña (P.R.S. Peña; Table 2). Hourly temperatures were available for Balcarce and P.R.S. Peña. Hourly temperatures for Rio Cuarto and Venado Tuerto were estimated by the model of Parton and Logan (1981) using daily maximum and minimum temperatures as input data. Temperatures from Balcarce estimated by this method were in close agreement with those measured at this location (r2 = 0.85).


View this table:
[in this window]
[in a new window]
  		Table 2. Latitude, longitude, analyzed period, and mean temperature during grain filling at four locations where sunflower is cultivated in Argentina.

 
The relationship between oleic acid concentration and temperature was also used to assess the variability in this fatty acid concentration among different years within a location. Data from Balcarce (1986 and 1987) and P.R.S. Peña (1990 and 1992) were analyzed. Selected years were those in which mean temperature during the typical grain filling period was 1°C higher or lower than the 1980 to1999 average at each location.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Experiment at High Night Temperature
In GC1, oleic acid ranged from 39 to 52% among treatments (Table 3). Increasing night temperature from 20 to 23°C resulted in a higher oleic acid concentration (p < 0.08). Further increases in night temperature did not affect fatty acid composition. An inverse trend was observed for linoleic acid. The maximum value for this fatty acid corresponded to the minimum night temperature, whereas similar values were observed for the other treatments. Finally, palmitic acid concentration was highest at the 28 and 20°C termperature regime (p < 0.01) and stearic acid concentration did not differ among treatments (p > 0.59).


View this table:
[in this window]
[in a new window]
  		Table 3. Fatty acid composition in the oil of plants grown under different day–night temperature regimes during grain filling. Experiment conducted in growth chamber (GC1). The photoperiod was 12 h.

 
Relationship between Oleic Acid Concentration and Temperature
Oleic acid concentration in the oil ranged from 17 to 41% in the field and from 38 to 59% in the growth chamber experiments. When data from all the experiments were pooled and analyzed, the r2 values of the linear functions between oleic acid concentration and temperature ranged from 0.05 to 0.80, depending on the temperature used (mean, minimum, etc.) and the moment during grain filling (Table 4). The r2 values of the relationships were low at flowering and maximum at around 200 ddaf. In general, regardless of the period, the highest r2 values were obtained with night mean or night minimum temperatures as the independent variable and not with daily temperatures or night maximum temperatures (data not shown). The maximum r2 value corresponded with the relationships between oleic acid concentration and night minimum temperature during the 100 to 300 ddaf period.


View this table:
[in this window]
[in a new window]
  		Table 4. Values of r2 for linear relationships between oleic acid concentration and temperature at different periods during grain filling (0 degree-days is flowering, PM = physiological maturity).

 
At high temperature, some of these linear equations overestimated oleic acid concentration. Considering this overestimation and the results from GC1, linear-plus-plateau functions were adjusted. The relationships between oleic acid concentration and temperature were improved with these adjustments. As observed with linear functions, better adjustments (selected by a high mean r2 and a short range of r2 distribution in the bootstrap analysis) corresponded with night temperatures and not with daily temperatures (Table 5). The wide range of r2 values distribution obtained with daily temperatures indicates that these variables are not good oil quality predictors. The closest relationships were those between oleic acid concentration and night minimum temperature during 100 to 300 ddaf, and longer periods such as 0 to 400 ddaf (the period reported by Izquierdo et al., 2002) and flowering to physiological maturity; however, night minimum temperature from 100 to 300 ddaf presented the maximum r2 and the shortest range of r2 distribution (Fig. 1 ). The linear -plus-plateau relationships between oleic acid concentration and other temperature descriptors or periods were weaker and presented wider ranges of r2 than that obtained with night minimum temperature during the period from 100 to 300 ddaf.


View this table:
[in this window]
[in a new window]
  		Table 5. Mean, median, and range of 100 r2 values obtained with a bootstrap analysis for the linear-plus-plateau relationships between oleic acid concentration and night minimum temperature at different periods during grain filling (0 degree-days is flowering, PM = physiological maturity).

 

Figure 1
View larger version (15K):
[in this window]
[in a new window]
  		Fig. 1. Distribution of r2 for linear-plus-plateau relationships between oleic acid concentration and night minimum temperature during (a) the 100 to 300 degree-days after flowering (ddaf) period, (b) flowering to 400 ddaf, and (c) flowering to physiological maturity. Adjustments were done for 100 random samplings with replacement from the original data pool.

 
Data estimated with night minimum temperature during the period from 100 to 300 ddaf were linearly related to observed data for Dekasol 3881, with an intercept and slope not different from 0 and 1, respectively (p > 0.19). The relationship between oleic acid concentration and night minimum temperature for the 100 to 300 ddaf period showed lower MSD, SB, SDSD, and LCS than the other relationships (data not shown). Night minimum temperature from 100 to 300 ddaf was considered for further analysis (Fig. 2 ). Increasing night minimum temperature up to 22.6°C during such period linearly increased oleic acid concentration (maximum value of 54.2%). Higher night minimum temperatures did not increase oleic acid concentrations. The variation coefficient for night minimum temperature during the selected period was similar to the mean variation coefficient for all the analyzed periods and temperatures (24.5 vs. 23.4 ± 1.3%, respectively). Thus, the higher regression coefficient for the selected period and temperature was not explained by a wider range of data for this variable.


Figure 2
View larger version (12K):
[in this window]
[in a new window]
  		Fig. 2. Oleic acid concentration as a function of night minimum temperature during the 100 to 300 degree-days after flowering period. Adjusted function is: oleic acid (%) = –23.1 + 3.4x for x < 22.6 and 54.2% for x ≥ 22.6, n = 34, r2 = 0.84.

 
Relationship between Fatty Acids
Oleic and linoleic acid concentrations were linearly and negatively related (Fig. 3a ). The slope of this relationship was –0.938, indicating that an increase in oleic acid concentration corresponded with a decrease of similar magnitude in linoleic acid concentration. The relationship between oleic and palmitic acid concentrations was also inverse (Fig. 3b) but not as strong as the previous one. No significant association was found between oleic acid and stearic acid concentrations (p > 0.089, Fig. 3c).


Figure 3
View larger version (14K):
[in this window]
[in a new window]
  		Fig. 3. Relationship between oleic acid and (a) linoleic acid, (b) palmitic acid, and (c) stearic acid concentrations in sunflower grain oil. The dotted line represents the function y = 100 – x.

 
For modeling purposes, total saturated fatty acids could be estimated through their relationship with the sum of oleic and linoleic acid percentages (y = –0.4156x + 47.726, r2 = 0.40, p < 0.001). This approach predicts 98 to 99% of the fatty acids of the sunflower oil. Figure 4 shows these predictions at different night minimum temperatures. Saturated fatty acids could also be estimated as 100 – (oleic + linoleic acid percentages); however, as other minor fatty acids are present in the oil, this relationship would present a higher prediction error than the previous one.


Figure 4
View larger version (19K):
[in this window]
[in a new window]
  		Fig. 4. Fatty acid composition estimated by the model for plants grown at different night temperatures.

 
Validation
Night minimum temperature during the period 100 to 300 ddaf varied from 10 to 24°C among treatments and experiments (Exp. GC6, FP1, and FB7). For these conditions, oleic acid concentration of the traditional hybrid Dekasol 3881 varied from 13 to 50%. The relationship established between oleic acid concentration and night minimum temperature (Fig. 2) adequately estimated the concentration of this fatty acid (Fig. 5 ). Linoleic acid concentration ranged from 36 to 74% and was well estimated by the relationship with oleic acid concentration (Fig. 3a). For both fatty acids, observed and estimated values were linearly related, with intercepts and slopes not different from 0 and 1, respectively (p > 0.12; Fig. 5). Saturated fatty acid values estimated by the model were similar to observed values (11.4 vs. 10.9%).


Figure 5
View larger version (17K):
[in this window]
[in a new window]
  		Fig. 5. Observed vs. estimated data for (a) oleic acid concentration and (b) linoleic acid concentration for the traditional hybrid DK 3881. Oleic and linoleic acid concentrations were calculated based on the function presented in Fig. 2 and 3a, respectively. Observed values correspond to Exp. GC6 (circles) and FB7 and FP1 (triangles). Dotted lines represent y = x and solid lines show the linear regression between observed and estimated values.

 
Night minimum temperature during 100 to 300 ddaf for the traditional hybrids ACA 885, DK 3915, and VDH 488 grown at Balcarce, Manfredi, Reconquista, and P.R.S. Peña varied from 14 to 21°C. For these conditions, oleic acid concentration ranged from 18.4 and 50.3%. Regressions between observed and estimated values showed intercepts and slopes not different from 0 or 1, respectively (p > 0.28; Fig. 6 ). The relationship established for Dekasol 3881 adequately estimated oleic acid concentration for hybrids ACA 885 and DK 3915. For hybrid ACA 885, the model slightly overestimated this variable. The greatest prediction error was observed in VDH 488.


Figure 6
View larger version (38K):
[in this window]
[in a new window]
  		Fig. 6. Observed vs. estimated data for oleic acid and linoleic acid concentrations for the traditional hybrids (a and b) ACA 885, (c and d) DK 3915, and (e and f) VDH 488. Estimated values for a, c, and e were calculated based on the function presented in Fig. 2 and for b, d, and f based on functions shown in Fig. 3a. Dotted lines represent y = x and solid lines show the linear regression between observed and estimated values.

 
Linoleic acid concentration ranged from 40 to 71% among hybrids and locations. Linoleic acid concentrations for these hybrids were well estimated by the relationship between this fatty acid and oleic acid concentration in Dekasol 3881 (Fig. 6). Regressions between observed and estimated values gave intercepts and slopes not different from 0 or 1, respectively (p > 0.53).The error estimates were similar to the errors described for oleic acid estimates. Total saturated fatty acids varied between 8 and 11% among hybrids and locations. Estimated values were slightly higher than observed ones (9 vs. 12%).

Simulating Sowing Date, Location, and Year Effects on Oleic Acid Concentration
Location strongly affected oleic acid concentration. Maximum values were observed at P.R.S. Peña and minimum values at Balcarce. Estimated oleic acid concentration increased when the beginning of the selected period (100–300 ddaf) was delayed in the growing season (Fig. 7 ). When the delay of the selected period was extreme, oleic acid concentration tended to decrease because of a decrease in night temperature. At each location, estimated oleic acid values were greatest when the selected period started in January, when night temperatures are at their maximum. The sunflower crop at P.R.S. Peña is harvested in December or early January, so the beginning of the selected period usually occurs with lower temperatures than those registered in January. Nevertheless, concentrations >54% are not expected at this location since night minimum temperatures >22.6°C do not result in an increase in the oleic acid concentration (Fig. 2).


Figure 7
View larger version (23K):
[in this window]
[in a new window]
  		Fig. 7. Oleic acid concentration as a function of the date when the 100 to 300 degree-days after flowering period began at four locations where sunflower is cultivated in Argentina (see Table 2 for further information). Data from P.R.S. Peña were not continued after 22 January because sunflower is not cultivated beyond that date. Horizontal bars indicate the period during which grain filling typically occurs at each location. Oleic acid concentration was estimated using the function from Fig. 2.

 
The highest oleic acid concentration estimated at Balcarce was lower than the lowest value estimated at P.R.S. Peña. When the typical grain-filling period was considered, location had a greater effect on oleic acid concentration than sowing date. Combining both effects, the estimated oleic acid concentration varied from approximately 17% at Balcarce to 52% at P.R.S. Peña. The highest value was estimated for crops that flowered in late December at the warmest location. Delaying the sowing date increased oleic acid concentration at locations where grain filling occurred between December and January, and decreased this concentration at locations where grain filling occurred between January and February.

Linoleic and palmitic acid concentrations showed inverse trends to those described for oleic acid. They were higher in Balcarce (61 and 7%, respectively) and lower in P.R.S Peña (43 and 5%, respectively) for crops sown at recommended dates. Saturated acid concentration (palmitic plus stearic) was estimated, on average, nearly two units lower in the warmest location than in the coldest one, with extreme differences of 3.3 units.

Important differences in oleic acid concentration among years were estimated. Oleic acid concentration varied up to 31 and 33% points between the 2 yr at Balcarce and P.R.S. Peña, respectively (Fig. 8 ). Although the mean oleic acid concentration estimated for P.R.S. Peña was between 36 and 43% (Fig. 7), a cool year at this location could produce an oleic acid concentration as low as the average value for Balcarce (Fig. 8). Finally, an oleic acid concentration as high as 44% could be expected in a warm year at Balcarce.


Figure 8
View larger version (21K):
[in this window]
[in a new window]
  		Fig. 8. Oleic acid concentration as a function of the date when the 100 to 300 degree-days after flowering period began at Balcarce and P.R.S. Peña for a cold year, a warm year, and the average for the 1980 to 1999 period. Oleic acid concentration was estimated using the function from Fig. 2.

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
CONCLUSIONS
 REFERENCES
 
The model established in this study for a traditional sunflower hybrid is based on experimental results recently reported (Izquierdo et al., 2002) and other novel data presented in this study. The range of temperature explored in the experiments used to establish the model covers most of the variation in temperature under which sunflower is usually cultivated around the world. The model considers the most critical period and the best temperature predictor for fatty acid composition. We have identified that oleic acid concentration saturates at high temperatures, which represents an improvement in the modeling of fatty acid composition relative to previous linear relationships.

The response of oleic acid concentration to temperature was bilinear. Increases in night minimum temperature from 10.7 to 22.6°C resulted in a strong increment of oleic acid concentration in the traditional hybrid Dekasol 3881. Higher night temperatures did not increase the concentration of this fatty acid (Exp. GC1). Preliminary studies with various traditional hybrids confirm this bilinear response between oleic acid concentration and night temperature. This saturation response contributes to an explanation of why traditional hybrids in high-temperature environments do not attain oleic acid concentrations >60%. The maximum oleic acid concentration was obtained at a lower temperature (night or daily average) than those responsible for a reduction in the embryo growth rate and grain weight (>35°C, Rondanini et al., 2003). Rondanini et al. (2003) also observed an increase in oleic acid concentration under such conditions, suggesting that other mechanisms could operate when grains are exposed to extremely high temperatures.

The largest effect of temperature on fatty acid composition occurs between 100 and 300 ddaf. In previous research, the final fatty acid composition in the oil of the hybrid Dekasol 3881 was explained by night temperature during the 0 and 400 ddaf period. (Izquierdo et al., 2002); however, the period during which this effect was more pronounced was different between two experiments. The 100 to 300 ddaf period identified in this study as the most critical for fatty acid determination was found by combinig data from several field crops, field experiments, and growth chamber experiments, covering a wider range of temperatures and environmental conditions than in our previous work. This period is included within the period reported by Izquierdo et al. (2002). Night minimum temperature during other periods of grain filling, even periods longer than 200 ddaf, did not improve the relationship between oleic acid concentration and temperature. By knowing the temperature during this short period, it is feasible to predict fatty acid composition before crop harvest.

Critical periods are those when the sensitivity to an environmental variable is highest. They are widely used in crop ecophysiology and modeling. For example, Aguirrezábal et al. (2003) determined that the period 250 to 450 ddaf was critical for seed weight and oil concentration in sunflower. Moreover, the period 20 d before flowering to 7 d after flowering is critical for wheat (Triticum aestivum L.) yield (Abbate et al., 1998). In this study, we selected the period 100 to 300 ddaf as the most critical for fatty acid composition in sunflower. This does not mean that temperature during other periods does not affect fatty acid composition.

Oleic acid concentration was better related to night minimum temperature than to other temperature descriptors. The underlying mechanism of this correlation is unknown. More biochemical research is needed to understand the effect of low temperatures on fatty acid desaturation (Harris et al., 1978; Izquierdo et al., 2002); However, for modeling purposes, night minimum temperature is the best predictor for oleic acid concentration. As the model only requires this temperature, it can be used in a wide range of environments. When only daily mean, maximum, and minimum temperatures are available, night minimum temperature can be estimated using prediction models like the one used in this study (Parton and Logan, 1981).

Approximately 90% of the total fatty acids in the oil corresponds to oleic and linoleic acids and the other 10% to the saturated stearic and palmitic acids. Thus, predicting oleic, linoleic, and total saturated fatty acid concentrations by using the relationships herein presented would provide a good estimation of oil quality of traditional sunflower hybrids.

The statistical methods used to select the model were complementary and coincident. The selected model presented the lower components of the MSD with respect to the other models. The validation with DK 3881 for field and growth chamber experiments and with other traditional hybrids indicates that the model is feasible for use under a wide range of conditions. The slopes and intercepts of the relationships between observed and estimated data did not differ statistically from 1 and 0, respectively; however, it seems that variations in the response of oleic acid concentration to temperature could exist among hybrids. In this case, a few data would be enough to find the parameters of the relationships for each hybrid. Mean palmitic acid concentration was well estimated by the model; however, mean stearic acid concentration was not well predicted. To improve modeling of this saturated fatty acid, more research is needed on environmental effects and its genotypic variability.

The oleic acid concentration of traditional hybrids in Argentina varies from 16.1 to 57.9% among locations, sowing dates, and hybrids (Muratorio et al., 2003). Using the model established in this study, we concluded that most of this variation was explained by night temperature differences among locations and sowing dates. Variations in oleic acid concentration among locations were markedly higher than among sowing dates. Combining both effects, a wide range of oleic acid content could be obtained. Moreover, with a late sowing date in a warm location, a traditional hybrid could produce an oleic acid concentration close to that typical of midoleic hybrids. This oil is often preferred for cooking because of its higher stability, compared with the typical oil of traditional hybrids. The simulations suggest also that oils produced under such warm conditions have a lower concentration of saturated (stearic plus palmitic) acids. Such oil composition is preferred to prevent cardiovascular diseases (Velasco and Fernandez Martinez, 2002).

The oleic acid concentration obtained at extreme locations (Balcarce and P.R.S. Peña) could be similar because of the high variation in temperature among years. A change in night minimum temperature during the selected period of at least 1°C relative to the average value across years produced a change of 15 units of oleic acid concentration. The probability of such change in night minimum temperature was 40% for Balcarce and 55% for P.R.S. Peña for crops sown at recommended dates. This variation in temperature could be used to determine the probability of obtaining a given oil quality among years.

Temperature is the main environmental factor responsible for changes in fatty acid composition; however, other factors, such as photoperiod, intercepted radiation, and N availability, have been mentioned as responsible for such changes (Talha and Osman, 1975; Steer and Seiler, 1990; Santalla et al., 1995). In experiments involving peeled seeds, reducing the O2 concentration decreased the activity of the enzyme, but this decrease was faster as temperature increased (Martínez Rivas et al., 2000). Further research is needed to know if O2 concentration affects fatty acid desaturation at the whole-plant level. Results presented in this study show that, for prediction purposes, night minimum temperature is a good predictor of fatty acid composition.

The established model, coupled with sunflower growth models (Texier, 1992; Chapman et al., 1993; Steer et al., 1993; Villalobos et al., 1996), could be useful for evaluating interactions between grain yield and oil quality. For example, the optimum sowing date for grain yield could be different than the sowing date that results in the best oil quality.


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
REFERENCES
 
We have demonstrated that the oleic acid concentration of the traditional sunflower hybrid Dekasol 3881 responds linearly to night minimum temperature up to 22.6°C. Higher temperatures (up to 28°C) during the 100 to 300 ddaf period did not affect the concentration of this fatty acid. Night minimum temperature during this period adequately predicted independent data. Location had a stronger effect on oleic acid concentration than sowing date. With an adequate location and sowing date, a traditional hybrid could produce an oil quality close to that typical of midoleic hybrids.


   ACKNOWLEDGMENTS
 
This work was supported by Agencia Nacional de Promoción Científica y Tecnológica (PICT 08-09711), Asociación Argentina de Girasol ("Proyectos 2002"), Universidad Nacional de Mar del Plata, Instituto Nacional de Tecnología Agropecuaria, Dow Agroscience S.A., Nidera S.A., and Aceitera General Deheza S.A. Thanks are given to Dow Agroscience S.A (Dr. G. Pozzi Jauregui) for fatty acid analysis, to S. Mascioli, J.C. Berto, R. Vilardo, A. Do Santos, G. Peluso, A. Digilio, C. Camurri, and G. Pereyra Irujo for their technical assistance, and to Eileen Din for the revision of English style. Seeds obtained in CYTs were provided by Dr. V. Pereyra and Dr. F. Quiroz, and at Paraná by Dr. O. Valentinuz. This work is part of a thesis submitted by N.G. Izquierdo in partial fulfillment for the requirements for the Doctor's degree, Universidad Nacional de Mar del Plata, Argentina.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 




This article has been cited by other articles:


Home page
 Agron. J.Home page
V. D. Zheljazkov, B. A. Vick, B. S. Baldwin, N. Buehring, T. Astatkie, and B. Johnson
Oil Content and Saturated Fatty Acids in Sunflower as a Function of Planting Date, Nitrogen Rate, and Hybrid
Agron. J., July 7, 2009; 101(4): 1003 - 1011.
[Abstract] [Full Text] [PDF]

Home page
 Agron. J.Home page
V. D. Zheljazkov, B. A. Vick, M. W. Ebelhar, N. Buehring, B. S. Baldwin, T. Astatkie, and J. F. Miller
Yield, Oil Content, and Composition of Sunflower Grown at Multiple Locations in Mississippi
Agron. J., May 7, 2008; 100(3): 635 - 642.
[Abstract] [Full Text] [PDF]

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 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 HighWire
Right arrow Citing Articles via Web of Science (4)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Izquierdo, N. G.
Right arrow Articles by Cantarero, M. G.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Izquierdo, N. G.
Right arrow Articles by Cantarero, M. G.
Agricola
Right arrow Articles by Izquierdo, N. G.
Right arrow Articles by Cantarero, M. G.
Related Collections
Right arrow Agroclimatology
Right arrow Crop Physiology & Metabolism
Right arrow Seed Quality
Right arrow Sunflower
Right arrow Other Models

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Crop Science Vadose Zone Journal
Journal of Natural Resources
and Life Sciences Education		 Soil Science Society of America Journal
Journal of Plant Registrations Journal of
Environmental Quality		 The Plant Genome