- Top of page
- Materials and methods
This study was conducted in 58 producer-field locations in Manitoba from 2003 to 2006 to understand how cropping practices influence Fusarium graminearum inoculum levels on stubble of various crops, including wheat, collected from the soil surface. Colonies per m2 (CN) were determined and converted to base-10 logarithm values (log10CN). Mean log10CN of the sampled field for various crops and groups of crops grown in the 3 years prior to sampling were tested to find significant differences. Average log10CN values were also used to determine significant differences between tillage systems and the effect of number of years. Average log10CN values for zero and minimum tillage systems were not different but were significantly higher than values for conventional tillage. A series of three crop rotation scenarios were tested using weighted log10CN values for crop, tillage, their interaction and their squared terms in step-wise regression models to identify which model was the best predictor of log10CN. This was selected as the cropping practice index (CPI) model and was expressed as: CPI = 1·98423 + 0·55975 (C2 × C1 × T)2 + 0·4390 (C2 × T)2, where C1, C2 and T represent the weighted log10CN values for crops grown 1 and 2 years previously and tillage system, respectively. R2 value for this model was 0·933 (P < 0·0001). The reliability of the CPI model was tested using jack-knife full cross-validation regression. The resulting R2 was 0·899. The CPI model was tested using data collected from seven wheat fields in 2006 (R2 = 0·567). The relationship between CPI and FHB index (R2 = 0·715) was significant.
- Top of page
- Materials and methods
Wheat is one of the most important crops in the world, and fusarium head blight (FHB) disease has become one of its most significant diseases. This destructive disease is caused by the fungus Gibberella zeae (anamorph: Fusarium graminearum) in North America. Since 1993, FHB disease has become a major problem in Canada (Gilbert & Tekauz, 2000). FHB causes reduced wheat yield due to shrunken kernels, lower grades and physical quality due to reduction of test weight, and reduced bread loaf volume. FHB also threatens the safety of human food and animal feed because of mycotoxins produced by the pathogen (Dexter & Nowicki, 2003).
In Canada, economic losses to FHB for wheat were estimated at US $220 million in Central Canada, and at US $300 million in Manitoba from 1993 to 1998 (Windels, 2000). In the United States, wheat and barley losses to this disease were estimated at US $3 billion in the 1990s. Epidemics of FHB are favoured by high-moisture weather conditions, planting of susceptible wheat cultivars, and increasing levels of zero or minimum tillage, which contribute to a buildup of fusarium inoculum (Windels, 2000).
The prevalence of FHB has resulted in considerable research aimed at controlling it through agronomic practices, most notably via crop rotation and tillage. FHB was significantly reduced by wheat rotation with non-wheat and corn crops in US based studies (McMullen et al., 1997). Elsewhere, crop rotation was found to significantly control FHB (Sutton, 1982; Parry et al., 1995; Dill-Macky & Jones, 2000). The effect of wheat rotation with non-host crops on reducing FHB is most likely due to reduction of the amount of infested wheat stubble on the soil surface (Teich & Nelson, 1984; Windels & Kommedahl, 1984; Parry et al., 1995; Dill-Macky & Jones, 2000). Recently, canola, field pea, bean and flax stubble were found to harbour F. graminearum (Chongo et al., 2001). The stubble of these crops would likely become a potential inoculum reservoir for a wheat crop grown in the following year. The importance of controlling FHB disease by tillage is also significant (Miller et al., 1998). Intense mouldboard ploughing significantly reduces FHB disease compared to chisel and zero tillage (Dill-Macky & Jones, 2000). In the soil, wheat stubble is partly decomposed in 24 months (Parry et al., 1995; Pereyra et al., 2004). Cereal stubble burning in the previous year was found to lower FHB disease levels in the current year (Dill-Macky & Salas, 2002). Therefore, epidemics of FHB disease are obviously influenced by crop species, sequence of rotated crops, frequency of host crops in the rotation, and tillage practices employed.
Fusarium graminearum is the predominant Fusarium species in North America. Its presence and its level of inoculum in crop stubble is undoubtedly a reflection of the effect of cropping practices. It is important to further understand how cropping practices affect FHB in order to provide wheat growers with more reliable and accurate strategies for disease management. Accordingly, the objectives of this study were to quantify the effects of cropping practices, i.e. crop rotation and tillage, on F. graminearum inoculum levels in various crop stubble including wheat, and to develop a model to express this relationship.
- Top of page
- Materials and methods
When individual crops grown 1 year before the sampling year were tested, wheat was significantly different from soybean-pea, and fallow. There were neither significant differences in log10CN level between wheat, canola and flax, nor between canola, soybean-pea, flax and fallow (Table 2). When all crops except for wheat were grouped together, there was no significant difference in log10CN levels between wheat and the grouped crops. However, when wheat and canola were assigned to one group, and other crops to a second group, log10CN values were significantly different (Table 2). Therefore, the average log10CN values for wheat-canola and for all other crops were used for weighting the effects of crops grown 1 year before on log10CN.
When CN levels between individual crops grown 2 years before the sampling year were similarly tested, there were significant differences between wheat, soybean-pea, flax and fallow, and between canola, flax and fallow. There were no differences between wheat and canola, and between flax and fallow (Table 2). When the log10CN values for each crop were grouped and averaged, the wheat value was significantly higher than the average for canola-soybean-pea, which was significantly higher than the average of flax-fallow. When all crops except for wheat were grouped together, the log10CN level for wheat was significantly higher than the average for the other crops. When wheat and canola were grouped together, the average log10CN level was significantly higher than the average value for all the other crops. Therefore, three different crop combinations of log10CN values were considered for weighting the effects on crops grown 2 years before on log10CN.
When log10CN levels between individual crops grown 3 years before the sampling year were tested, the wheat value was significantly less than the canola value, but there were neither significant differences between wheat and other crops nor between canola and other crops (Table 2). When all crops were grouped except for wheat, the wheat log10CN level was significantly less than the other crops. There was no significant difference between the average log10CN for wheat-canola and the average log10CN for the other crops (Table 2). The effects of crops grown 3 years before did not exhibit significant effects on log10CN and, therefore, were not considered in the crop effect weightings.
In regards to tillage effects, there was no significant difference in the average log10CN values between zero and minimum tillage. However, the average zero-minimum tillage log10CN value was significantly higher than that for conventional tillage (Table 3). The average log10CN of the zero-minimum tillage fields and the conventional tillage fields were considered for weighting the effects of tillage on the log10CN value.
Table 3. Logarithm of the numbers of Fusarium graminearum colonies on various crop stubbles per square metre of soil for different tillage systems
|Tillage||Mean log10CNa||Tillage||Mean log10CN|
|Zero tillage||3·62 a||Zero/minimum tillage||3·60 a|
|Minimum tillage||3·59 a||Conventional tillage||3·34 b|
|Conventional tillage||3·34 b|| || |
The average log10CN value for crops grown 1 year before was 1·1 times greater than the average log10CN value for crops grown 2 years before. Thus, the year weightings for crops grown 1 and 2 years previous were assigned values of 1·10 and 1·00, respectively.
Three different rotation scenarios were evaluated for development of a CPI model (Table 4). Weighted log10CN values (log10CN value for a crop or group of crops in a rotation scenario divided by the smallest log10CN value in the same scenario and multiplied by the corresponding year score) for each crop or group of crops is shown in Table 4. The weighted value for wheat-canola in rotation scenario 1, for example, was calculated as 1·27 = 3·62 × 1·10/3·12 (Table 2). The score for zero-minimum tillage systems was expressed as their average log10CN value divided by the mean log10CN value for conventional tillage fields (1·08 = 3·60/3·34) (Table 4).
Table 4. Mean base-10 logarithm of the numbers of Fusarium graminearum colonies per square metre (log10CN) of soil for different crops and tillage systems
| || ||Rotation||Scoreb|
|Scenario 1a||1 year before||Wheat/Canola||1·27|
|2 years before||Wheat||1·15|
|Scenario 2||1 year before||Wheat/Canola||1·20|
|2 years before||Wheat||1·09|
|Scenario 3||1 year before||Wheat/Canola||1·24|
|2 years before||Wheat/Canola||1·12|
|Tillage|| ||Zero/Minimum tillage||1·08|
| ||Conventional tillage||1·00|
Three CPI models were developed based on the weighted log10CN values from Table 4 for various crops in each rotation scenario, tillage, their interactions and their squared values (Table 5). The step-wise regression models for rotation scenarios 1 and 2 selected very similar independent factors. The interaction of crops grown 1 and 2 years before and tillage was squared in scenario 2 but not in scenario 1. The interaction of crop grown 2 years before and tillage was squared in both scenarios. The stepwise regression model for scenario 3, selected the squared value for crop grown 1 year before and tillage, as well as tillage. The CPI model with the highest R2 value (0·933) was that for rotation scenario 2 (Table 5). The VIF value for the two factors, (C2 × C1 × T)2 and (C2 × T)2, was <10, indicating that multicollinearity was not significant and, therefore, the CPI model for rotation scenario 2 was statistically valid.
Table 5. Regression models developed for base-10 logarithm of CN values (Fusarium graminearum colonies per square metre) in the three scenarios for crop combinations in the previous 2 years
| ||Modela||R 2||VIFb|
|Scenario 1||Log10CN = 0·92413 + 1·29901C2 × C1 × T + 0·47880 (C2 × T)2||0·828|| |
|Scenario 2||Log10CN = 1·98423 + 0·55975 (C2 × C1 × T)2 + 0·4390 (C2 × T)2||0·933||5·0|
|Scenario 3||Log10CN = 0·98539 + 0·94807 (C1 × T)2 + 0·90830T||0·923|| |
The CPI model was tested using the actual data that were collected in seven fields in 2006 (Table 6). The root mean square error for the predicted CPI was 0·14 and the mean bias error was slightly positive at 0·07.
Table 6. The actual and predicted cropping practice index (CPI) values for the model from data collected from seven fields in 2006
| ||Field 52a||Field 53||Field 54||Field 55||Field 56||Field 57||Field 58|
The CPI model was also evaluated using the jack-knife full cross-validation method (Walker, 2007). The R2 value for prediction of log10CN logarithm was 0·899 (Fig. 1). The slope value for the regression was close to 1. The CPI model, using crop rotation and tillage practices in the 2 years prior to the current crop, accounted for almost 90% of the variation in F. graminearum inoculum levels in the crop stubble of the crop monitored in 2006.
Figure 1. Jack-knife full cross-validation of the cropping practice index (CPI) model for rotation scenario 2 (Table 4), in which wheat or canola was assigned score 1·20, and other crops were 1·12 for the first year; wheat was assigned 1·09 and others were 1·00 for the second year. In the regression model, CN represents Fusarium graminearum colonies per m2, variables x and y indicate actual and predicted log10CN, respectively.
Download figure to PowerPoint
There was a significant relationship between CPI and FHB index (R2 = 0·715, P < 0·0001) as shown in Fig. 2. There were overlapped data points in the figure; therefore some single data points could include more than one set of data.
Figure 2. Relationship between logarithm of fusarium head blight (FHB) index and cropping practice index (CPI) calculated with the prediction model, log10CN = 1·98423 + 0·55975 (C2 × C1 × T)2 + 0·4390 (C2 × T)2 in the scenario 2 (Table 4). In the regression model, variables x and y indicate CPI and FHB index, respectively.
Download figure to PowerPoint
- Top of page
- Materials and methods
This is the first study to quantify previous year cropping practices according to cumulative F. graminearum inoculum levels on wheat stubble from the soil surface. A regression model was developed for the quantified cropping practices using a cropping practice index derived from the various rotation scenarios. Prediction accuracy for the CPI model was approximately 90% based on the jack-knife full cross-validation method (Walker, 2007). The model was also used successfully to predict CPI values for an independent year and was strongly correlated to FHB index.
Cropping practices were quantified based on the significant effects of the preceding crops on fusarium inoculum level on various crop stubbles on the soil surface in this study. Wheat and canola crops played more important roles in accumulation of inoculum than other crops (Table 2). Rotation of wheat with non-host crops, such as clover, alfalfa, potato and soybean, reduces the amount of infested wheat stubble and FHB (Sutton, 1982; Teich & Nelson, 1984; Windels & Kommedahl, 1984; Parry et al., 1995). This study demonstrated that canola grown in the previous 2 years showed no difference from wheat in its effect on fusarium inoculum level, which could be associated with canola’s capability of harbouring F. graminearum (Chongo et al., 2001).
Tillage was another important factor in quantifying cropping practices. Zero or minimum tillage contributed significantly more to fusarium inoculum level than conventional tillage (Table 3). The effect of tillage on wheat stubble and pathogen survival has been well documented. Fusarium pathogens can survive on stubble and wheat kernels on the soil surface (zero or minimum tillage) for more than 2 years (Inch & Gilbert, 2003; Pereyra et al., 2004). This study found that conventional tillage made a significant reduction in fusarium inoculum levels (Table 3). Wheat stubble partly decomposes and approximately 30% of dry matter stubble is left after 24 months in the soil (Parry et al., 1995; Pereyra et al., 2004). There was no significant difference between zero and minimum tillage systems in this study. Both zero tillage and minimum tillage leave large amounts of crop stubble on the soil surface. Because G. zeae requires light to form perithecia (Inch & Gilbert, 2003), this condition could be satisfied only for stubble remaining on the soil surface. Therefore, zero and minimum tillage can build up a larger reservoir of G. zeae perithecia compared to conventional tillage.
The CPI model, developed using cropping practice factors, indicated that fusarium inoculum level increased when wheat and canola crops were grown in the previous 2 years and when more zero or minimum tillage was employed in the previous 2 years.
A significant relationship between CPI and FHB index was found in this study; however, there are other factors affecting FHB index. The CPI value reflected the level of F. graminearum colonies on crop stubble from the soil surface only and not the G. zeae perithecial level or the airborne-inoculum level. Gibberella zeae perithecial formation occurs over a long time period from the early spring to the late summer in Canada (Inch & Gilbert, 2003), and needs appropriate temperature, moisture and light conditions (Sutton, 1982; Inch & Gilbert, 2003). This is affected by more than just the cropping practices. Gibberella zeae ascospore dispersal occurs throughout the same period as the perithecial formation (Fernando et al., 2000; Inch et al., 2005), and only spores discharged at the wheat flowering stage are real inoculum (Wilcoxson et al., 1992; Groth et al., 1999). Weather conditions during this period control spore release (Sutton, 1982). The CPI model was developed to quantify the effects of cropping practices employed before the current year, which cause potential changes in fusarium inoculum levels. CPI can be coupled with the number of ascospores dispersed at the flowering stage to improve the reliability of the CPI as a predictor of FHB index.
The CPI model excluded the factor of cultivars used in the current years, as they had no effects on the fusarium inoculum levels on the crop stubble. There are some limitations for the CPI model. For example, it did not take into account fusarium resistance or susceptibility of crop cultivars because there were a limited number of fields in the study. However, the CPI model could be improved with additional sampling of fields and by including a fusarium susceptibility factor.
Importantly, the CPI score developed in this study can be integrated into a broader disease management strategy. It can be used with weather conditions for prediction of F. graminearum airborne inoculum level, fusarium head blight disease and mycotoxins produced by the pathogen. This application will be reported in a separate paper.