Process-based modeling of the control of beet armyworm, Spodoptera exigua, with baculoviruses in greenhouse chrysanthemum

This thesis describes the development of a process-based simulation model for the population dynamics of beet armyworm, Spodoptera exigua , and baculoviruses in greenhouse chrysanthemum. The model (BACSIM) has been validated for two baculoviruses with clear differences in biological characteristics, Autographa californica multicapsid nucleopolyhedro-virus (AcMNPV) and Spodoptera exigua MNPV (SeMNPV). The validated model has been used to generate scenario studies, which are used for the evaluation of the potential of baculoviruses as biological control agents.The first requirement for the construction of BACSIM is quantitative understanding of the processes that determine the chrysanthemum- S. exigua -baculovirus system, such as insect behavior and population dynamics, the baculovirus infection cycle and virus transmission routes. These processes have been studied and described in chapters 2 to 5.S. exigua parameters : The developmental rate and background mortality of populations of S. exigua larvae have been quantified on chrysanthemum in a greenhouse situation (chapter 4). In addition, the preference of feeding sites and leaf visit rate of S. exigua larvae have been studied on chrysanthemum plants under controlled conditions (chapter 4).Baculovirus parameters : The infectivity and speed of action of AcMNPV and SeMNPV has been determined for various stages of S. exigua larvae that were reared on artificial diet (chapter 2). To study the effect of the host plant on the baculovirus infection cycle the infectivity and speed of action of SeMNPV has also been determined for S. exigua larvae reared on chrysanthemum (chapter 4). S. exigua larvae reared on chrysanthemum appeared to be more susceptible to SeMNPV than larvae reared on artificial diet. The polyhedron inactivation rate of AcMNPV and SeMNPV on greenhouse chrysanthemum has been quantified in chapter 3 and 7, respectively. Wild-type AcMNPV and SeMNPV have marked differences in infectivity and speed of action, but the relative inactivation rates of AcMNPV and SeMNPV polyhedra were not significantly different. The deletion of the egt , pp34 or p10 genes from AcMNPV had no or only a marginal impact on the infectivity, speed of action or inactivation rate of this virus (chapter 2). In other virus-host combinations such deletions had a major effect on these parameters. As the AcMNPV recombinants behaved biologically in a similar way as wild-type AcMNPV, wild-type AcMNPV and SeMNPV were chosen as viruses for the model system.Baculovirus transmission : Horizontal transmission of SeMNPV in S. exigua populations with varying densities of primarily SeMNPV-infected larvae has been studied in chrysanthemum under greenhouse conditions (chapter 5). The distribution of polyhedra released from SeMNPV-killed larvae over chrysanthemum plants has been quantified in detail under controlled conditions (chapter 4). The vertical transmission rate of SeMNPV in S. exigua populations has been determined by the quantification of the percentage of first instar larvae originating from egg batches produced by sublethally infected moths that contracted SeMNPV (chapter 5).The process-based simulation model BACSIM is based on a detailed quantitative description of the behavior and population dynamics of insects, plant growth characteristics, spray deposition, the baculovirus infection cycle and transmission routes (chapter 7). These data were obtained from experiments mentioned above, or were retrieved from literature. The evaluation of the biological control of insect pests with BACSIM offers the possibility to include many treatments and situations, to simulate baculoviruses with extreme biological properties and to include many covariables that may modify the efficacy of baculoviruses in practice, such as crop age and temperature. In addition, this simulation approach allows optimization procedures that can assess optimal virus properties and spraying regimes under varying conditions.The BACSIM model has been validated with independent data of ten greenhouse trials in which the efficacy of AcMNPV and SeMNPV against synchronized populations of S. exigua larvae in chrysanthemum was assessed. Dose- and time-mortality relationships of different virus concentrations and S. exigua target stages were determined and crop injury assessed as the total leaf area consumed (chapter 6). The validation of the simulation model BACSIM with the independent data of AcMNPV and SeMNPV applications in greenhouse chrysanthemum was in general satisfying. Mortality levels of AcMNPV and SeMNPV infected larvae were generally predicted within a 25% error margin compared to the observed values. None of the deviations were higher than 40%. All values of simulated foliage consumption, caused by S. exigua populations treated with AcMNPV or SeMNPV applications, fell within the 95% confidence intervals of measurements. However, simulated time-mortality relationships gave generally lower survival times than was measured in experiments. A hypothesis that may explain this discrepancy is the functioning of the underside of chrysanthemum leaves as a refuge area where S. exigua larvae are exposed much less to polyhedra (chapter 8).Scenario studies with BACSIM were carried out to evaluate the effectiveness of different spraying regimes, dosages, UV protection agents and speed of action of viruses in different pest situations (chapter 9). The recommended dosages of SeMNPV applications based on simulations with BACSIM corresponded well with recommended SeMNPV dosages determined for cultivation practice. An early timing of virus applications, soon after egg batch deposition, appeared to be essential for effective control. UV-protection agents may contribute only marginally to effective biological control in the glasshouse. The decay rates of SeMNPV and AcMNPV already ensure a period of exposure to active virus that is long enough for caterpillars to acquire a lethal dose under practical conditions. The effect of genetic improvement towards shortening the survival time of infected larvae depends on the situation in which such a virus is used. When used in a situation with constant immigration of pest insects, a greater reduction of feeding injury is achieved with a faster killing virus than with a wild-type virus. When, however, immigration of pest insects occurs by way of sudden influxes, resulting in synchronous insect populations in the crop, it appears that early timing of virus applications is more critical to reduce crop injury than using a fast-killing virus. In such a situation, good monitoring followed by virus application is indispensable, and may even alleviate the need for a fast killing virus.Besides the use of BACSIM as a tool to assist in the determination of effective spraying regimes of (genetically modified) baculoviruses or formulations under varying conditions, the model can also be used as a tool to gain insight in the insect-baculovirus-crop system. For example, BACSIM may be used to address fundamental ecological questions or used to evaluate the relative importance of viral characteristics for crop protection purposes. In addition, BACSIM may contribute to the assessment of risks associated with the use of genetically modified baculoviruses by the generation of quantitative information of the population dynamics of recombinant baculoviruses in crops. Finally, the methodology described in this thesis may form a basic concept for process-based modeling of more complex systems, such as the control of bollworm with baculoviruses in cotton.