A model for the effect of Sphecophaga vesparum vesparum as a biological control agent of the common wasp in New Zealand

1. A model of intermediate complexity is described to predict and understand the reasons for the ultimate impact of Sphecophaga vesparum vesparum (Hymenoptera: Ichneumonidae), introduced as a classical biological control agent of Vespula vulgaris (Hymenoptera: Vespidae) in New Zealand. The model was parameterized as far as possible from independent field data then fitted to the observed performance of the parasitoid over the first 5 years after release. 2. Wasp nest densities monitored over 5 years and seven sites in beech (Nothofagus spp.) forests averaged 12.2 ha(-1), with a maximum of 33 ha(-1). These are among the highest Vespula densities in the world. A Ricker model accounted for changes in nest density from year to year, giving a maximum ratio of increase in nest density of 3.3 per year and overcompensating density dependence at high densities, probably caused by queen competition and nest usurpation in spring. 3. The parasitoid has established at low levels in two sites, one of which has been studied in detail. Here parasitism levels (% autumn nests parasitized) have remained around 5% for 5 years, with a slight suggestion of an upward trend. 4. The model suggests that ultimate parasitism levels depend almost entirely on the parasitoid's effective ratio of increase, R, defined as the maximum number of spring adults produced per spring adult (spanning several intermediate summer generations). Ultimate suppression of wasp nest densities depends on R, the rate of increase in parasitism within a year, and the mortality of parasitized early spring nests. The initial rate of build-up of parasitism additionally depends on the pattern of emergence of parasitoid cocoons, which may extend over 4 years. 5. R for S. vesparum vesparum at the site where it has established appears to be about 1.3-1.6, which is close to the lower limit of 1 for persistence. Such values suggest an ultimate suppression of wasp nest density and level of parasitism of about 10% and 25%, respectively. 6. The low value of R at this site and the parasitoid's limited likely impact, appear to be due to a combination of delayed emergence of overwintering cocoons, low overwintering cocoon survival, and low production of cocoons per parasitized autumn nest. Additionally, a reduced or variable attack rate may contribute to the parasitoid's lack of establishment at other sites, possibly due to poor synchrony between spring emergence of wasp queens and adult parasitoids. 7. A more successful parasitoid species or ecotype would need to have a higher R value, which requires cocoon emergence after 1 year rather than 2 or more, and/or higher values for the above parameters, particularly overwintering survival of cocoons. In addition, it would have to: act after the spring host density dependence; cause a high (> 80%) mortality of parasitized spring nests; and ideally also reduce queen output from autumn nests. An alternative to an agent causing mortality to spring nests would be a microbial one which significantly reduced queen fertility without impairing her competitive ability. 8. A 50% reduction in queen output from autumn nests, due to parasitism by S. vesparum vesparum, contributes little to host suppression because of its timing relative to host density dependence. If an alternative agent reduced wasp densities to a greater extent, the additional contribution of reduced queen output would become relatively more significant as density declined and the density dependence became less intense.