Development of a decision support system for managing Heterodera schahtii in sugar beet production

The database consisted of two units, (i) a sugar beets unit with a list of “Standard,” “Semi-tolerant” and “Tolerant” varieties, and (ii) an intercrop unit. The varietal unit included values of parameters for calculating sugar yield and SBN population dynamics. These values were obtained from Swedish and Danish commercial field trials (Olsson, 2011). The intercrop unit included Rf values of resistant class 1 and 2 of mustard and oil seed radish cultivars (Olsson, 2009).

To start using the SBN-Watch, certain data are required to be entered by the user. The most important input is the average initial population density in the soil Pi (eggs g⁻¹ soil). Field soil has to be sampled according to, guidelines from the Nematode Laboratory (NL) laboratory, which provides services for growers for analyzing plant parasitic nematodes in soil. Other data inputs are financial information regarding the expected sugar beets price (€ tonne⁻¹) and exchange rate (€ to SEK). Using a drop-down list, sugar beet varieties and crop sequence can be selected. Additionally, another option concerning tolerance (T) can be made by the user through entering a specific (T) value for the selected sugar beet variety.

Changes in the population of SBN in the soil under presence of the host (sugar beet) or non-host crop plants have been described by different equations in the literature. A modified equation by Burt and Ferris (1996) was used to describe SBN population dynamics when a non-host crop was selected to be cultivated in the crop rotation: (1) P ( m ) = S m × P i , $${{\rm{P}}_{({\rm{m}})}} = {{\rm{S}}^{\rm{m}}} \times {{\rm{P}}_{\rm{i}}},$$ where s = proportion of the population survived (s = 0.35); Pi (eggs g⁻¹ soil) = initial population of SBN at sowing expressed as number of eggs g⁻¹ soil; and m = number of years the non-crop was grown in succession. If WOSR was selected to be cultivated in the rotation, a multiplication factor of 1.5 was used instead (Olsson and Olsson, 2006).

When a sugar beets variety was selected for cultivation, the final population Pf was calculated by the equation of Seinhorst (1967): (2) Pf = a ( − ln q ) − 1 ( 1 − qPi ) + sPi , $${\rm{Pf}} = a( - \ln q) - 1(1 - q{\rm{Pi}}) + {\rm{sPi}},$$ where a = Reproduction factor at very low Pi values; s = proportion of the population survived (s = 0.35); q = constant < 1.

The reproduction factor (Rf) of a sugar beet variety was then calculated as Rf = Pf/Pi. To estimate the number of years required for (Pi) to reach (T) for specific sugar beets variety, the equation of Burt and Ferris (1996), Equation (1), was used modification: (3) m = log ( T sort ) − log ( P i × RF ) log ( S ) $${\rm{m}} = {{\log ({{\rm{T}}_{{\rm{sort}}}}) - \log ({{\rm{P}}_{\rm{i}}} \times {\rm{RF}})} \over {\log ({\rm{S}})}}$$ where Tv = tolerance limit for specific sugar beets variety was used instead of population P; Rfv = reproduction factor (Rf = Pf/Pi) of a given sugar beets variety; s = survived proportion of SBN = 0.35.

To estimate yield loss of sugar beets due to infestation by SBN, the equation of Seinhorst (1965), which relates sugar beet yield to SBN population density at sowing was used: (4) Y = m + ( 1 − m ) Z ( Pi − T ) , $${\rm{Y}} = {\rm{m}} + (1 - {\rm{m}}){{\rm{Z}}^{({\rm{Pi}} - {\rm{T}})}},$$ where Y = relative yield; m = minimum yield at the highest SBN population; Z = constant; Pi = initial SBN population at sowing (eggs g⁻¹ soil); and T = tolerance limit. When Y × Z^(Pi−T) is < m × Y the value of m × Y is used to express yield. When Y × Z^(Pi−T) > m × Y the yield is estimated by the equation Y × Z^(Pi−T).

Different crop rotation scenarios were chosen to illustrate the influence of sugar beet varieties, non-host crops, WOSR and intercrops on the SBN population and sugar yield as shown in Table 2. Pi value of 2 eggs g⁻¹ soil was assumed in the initial year of sugar beet cropping.

In this demonstration example, the sugar beet variety “Rosalinda” was selected to be grown in year zero followed by non-host crops. Pi value of 4 eggs g⁻¹ soil was assumed in the initial year of sugar beet cropping.

A demonstration example was used to illustrate the relationship between the initial population of SBN and the reproduction factor. Three sugar beet varieties “Mixer,” “Rosalinda” and “Julietta” were selected to be tested at six different Pi (2; 4; 7; 9; 13; 17 eggs g⁻¹ soil) similar to the levels used in microplot experiments described below. The final SBN populations (Pf) were estimated by SBN-Watch and the corresponding reproduction factors were calculated as Rf = Pf/Pi.

Two experiments were conducted according to Fatemy et al. (2007), but with the following modification. The field soil was not sterilized in order not to affect the biological and physical soil characteristics, instead different inoculum levels of SBN were generated by mixing different proportions of a naturally infested soil with a nematode free soil collected from the same commercial field in 2013 and 2014. The selection criteria of the field soils were a conducive light soil with high sand content (73%). Soil chemical characteristics were analyzed at Eurofins Food and Agro Testing Sweden AB, Kristianstad, Sweden (Table 2). The infestation levels were analyzed in the inoculum soil and the nematode free soil before soil collection and transportation to experimental site. Soil inocula and the nematode free soil were mixed by a cement blender to generate different Pi (eggs g⁻¹ soil) levels confirmed by analysis of subsamples from each level (Table 3). The experiments were set up as microplots consisting of 12 l plastic buckets (28 cm diameter; height 24.5 cm) filled with 12 kg soil each (n = 4). The Each microplot was fertilized with 3.6 g NPK 15-4-8 mineral fertilizer, which corresponds to a standard level of 130 N ha⁻¹ used in the commercial sugar beet fields, then sown with three seeds. Three sugar beets cultivars (Table 3) were grown each year in the microplots, arranged randomly in four blocks in a net house under natural conditions. The buckets were buried into the soil to give similar soil temperature as in the field. Watering was done directly after sowing and then when it is needed during the growing season. After germination the plants were thinned to one sugar beets plant per microplot. All experiments were conducted at the Swedish University of Agricultural Sciences (SLU), Alnarp. The initial densities and final SBN populations in the microplots were analyzed at the HS Nematology laboratory, Alnarp using Seinhorst elutriation method (Seinhorst, 1964).

The SBN-Watch starts with information about the SBN biology and description of models used for calculation of population dynamic and sugar yield followed by three user interfaces: in the first interface soil analysis of the initial SBN population Pi (eggs g⁻¹ soil), sugar beets price and exchange rate (€ to SEK) must be entered. The second interface comprises the choice of crop rotation (Fig. 1) and the third interface shows population dynamic of SBN upon choice of certain variety at a given Pi and T value, which can optionally be changed by the user. SBN-Watch delivers results in a form of report, including information on the selected crop rotation, sugar yield, economical value, SBN population dynamics within the selected crop rotation and finally the number of years to wait for the next sugar beet crop to be grown when specific variety was selected followed by non-host crops.

Figure 1:

The first two crop rotations illustrates the guidelines that give recommendations to growers for cultivating sugar beet every three years together with non-host crops or every four years upon cultivation of WOSR in the same crop rotation. In the first crop rotation, the standard variety “Mixer” was selected to be cultivated in a three years rotation in sequence with cereals. In this case, the calculated Pf values were as follows: 19, 6.6, 2.3 and 21.1 eggs g soil⁻¹ (Fig. 2). The Rf values were 9.5 and 1.9. The calculated sugar yields were 10.0 and 9.9 tonnes hectare⁻¹, respectively. In crop rotation 2 “Mixer” was selected followed by a cereal crop, WOSR and a second cereal crop before cultivating “Mixer” again in the crop rotation. The calculated Pf value increased to 10.0 after WOSR then declined to 3.5 after the second cereal crop and finally increased to 27.4 egg g soil⁻¹ after “Mixer” (Fig. 3). The Rf values were 9.5 and 1.2 and the calculated sugar yields were 10.0 and 9.3 tonnes hectare⁻¹, respectively (Fig. 3). Selecting a resistant oil radish intercrop before the second sugar beets crop in crop rotation 3, decreased the Pf value to 2.0 and the sugar yield 10.0 tonnes hectare⁻¹, was 700 kg higher than the corresponding yield in crop rotation 2 (Fig. 4). The Rf value of the second sugar beet crop was 2.0.

Figure 2:

Figure 3:

Figure 4:

Choosing a tolerant sugar beet variety together with cereals in crop rotation 4, the Pf values were as follows: 4.6, 1.6, 0.6, and 1.4 (Fig. 5). Sugar yields were 9.6 and 10.7 tonnes hectare⁻¹ and Rf values were 2.3 and 3.7 (Fig. 5). When Julietta and WOSR were grown together according as in crop rotation 5, the estimated Pf values were 4.6, 1.6, 2.4, 0.8, and 2.1 (Fig. 6). The sugar yield slightly increased from 9.6 to 10.7 tonnes hectare⁻¹ as Pf after the second cereal crop decreased to 0.8 (Fig. 6). The Rf values of the first and second sugar beet crops were 2.3 and 3.6, respectively. Growing a resistant oil radish intercrop before “Julietta” in crop rotation 6, further decreased Pf value to 0.5 and the calculated sugar yield was 10.7 tonnes hectare⁻¹ (Fig. 7).

Figure 5:

Figure 6:

Figure 7:

To answer the question of how many years it takes for a SBN population to decrease to the tolerance limit when a specific sugar beet variety was selected to be grown at a specific Pi (eggs g⁻¹ soil) value, a demonstration example with the Semi-tolerant variety “Rosalinda” at Pi (eggs g⁻¹ soil) = 4.0 is shown in Fig. 8. In this example, SBN-Watch gives a recommendation for the user to wait four years before cultivating “Rosalinda” again in the crop rotation.

Figure 8:

The predicted Rf values of three sugar beet varieties at different Pi values corresponded well with the varietal characteristics where the multiplication rates of the “Standard” variety “Mixer” and the “Semi-tolerant” variety “Rosalinda” were higher than in the “Tolerant” variety “Julietta” (Fig. 9A). Furthermore, the highest Rf values were estimated at the lowest Pi (eggs g⁻¹ soil) values (Fig. 9A).

Figure 9:

The means of reproduction factors from 2013 and 2014 were pooled together giving initial population densities from 2.24 to 17 (Fig. 9B). In general, the mean Rf values were higher at the lowest Pi (eggs g⁻¹ soil) level (2.24) and the “Standard” variety “Mixer” had the highest Rf values. The Rf values of the “Tolerant” variety “Elora” and “Mixer” at the lowest Pi (eggs g⁻¹ soil) were similar, but the variation was high at this particular level. The reproduction factors of all tested varieties were then decreased to Rf < 4 at Pi (eggs g⁻¹ soil) > 4 (Fig. 9B). A similar trend was obtained by SBN-Watch, even two different “Tolerant” varieties were used, “Elora” in microplots and “Julietta” in SBN-Watch (Fig. 9A).