PhD Thesis

Research Invasive species H. armigera outbreaks in cotton fields Landscape pollen index and hoverflies Fields hedgerows and ground beetles Dwelling beetles in North China grasslands My main skills Where I worked

Abundance and trophic origin of the Cotton Bollworm (Helicoverpa armigera) in cotton producing farmland of North Benin Keywords: Crop protection – landscape diversity - agricultural practices – pest sustainability – gossypol – tomatine – stable Carbone isotopes – maize – sorghum – partial least square The main challenge of integrated pest management is to design and implement measures that promote both productivity and sustainability in agrosystems. For this purpose, the effects of agricultural practices and landscape factors on pest population dynamics should be studied. In northern Benin (study area), cotton farmers are part of national programs. Chemicals application is managed based on a predefined schedule for pest control. However, these measures raise several problems concerning the impact of pesticide for the environment as well as for human population (both farmers and consumers). In addition, studies found insecticide (pyrethroids) resistance genes in cotton pest populations even for transgenic cotton (Udamale et al. 2013). In this context, it is necessary to move towards sustainable pest management strategies independent of pesticides or at least towards strategies which can slow the resistance of pests to pesticides. This will improve the quality of West African cotton and livelihood of families of small cotton growers in this region. The purpose of my thesis was to improve our ability to control the cotton bollworm, Helicoverpa armigera by non-pesticide methods via habitat conservation. I investigated agricultural practices and landscape factors driving H. armigera abundance in cotton fields in the northern Benin (Tsafack et al., 2013, 2016) I investigated on one hand, the drivers of an isolated H. armigera outbreaks (Study 1: larvae infestation) and on another the drivers of a delayed infestation (Study 2: abundance and trophic origin of adults H. armigera). Four questions were guidelines of my thesis research work. Where are the H. armigera populations outbreaking in cotton fields each September coming from? Are individuals’ local or migrant, from the southern Benin? (2nd part of the thesis)  Do local factors like agricultural practices impact H. armigera outbreaks? Which landscape factors influence the abundance and trophic origin of H. armigera? Which landscape factors determine the trophic origin of adult H. armigera trapped in a landscape? To answer to these questions, a large part of my thesis work was to develop tracking methods for H. armigera  well adapted to the study site. Do landscape context and agricultural practices influence the abundance of cotton bollworm Helicoverpa armigera in cotton fields? Tsafack et al, 2013 Based on the available literature on landscape effects on pest and the biology of H. armigera we constructed landscape cover factors that we believed could influence larvae abundance. These were landscapes containing high proportions of: i) cotton crops, given that cotton crops around a target cotton field might increase the abundance of H. armigera larvae; ii) maize (first cultivated host plant); iii) tomato; and, iv) natural vegetation. We considered natural vegetation as a factor with two potential roles, first contributing to the very early development of H. armigera with the presence of a native host plant (Cleome viscosa), and then providing a habitat for H. armigera predators and parasitoids which could reduce the abundance of larvae. Given the polyphagous diet of H. armigera, the landscape diversity of the crop mosaic was considered. Then, at the field scale, we addressed the hypothesis that a manually weeded field should have lower infestations, because any pupae diapausing in the soil would be destroyed. We also considered the sowing date and the previous crop of the studied cotton field as factors that could influence the abundance of H. armigera larvae. Due to the lack of available and current remote sensed data or aerial photographies for the studied area, digitalisation of landcover recorded around the selected cotton field was based on a georeferenced Google Earth satellite image (Google Earth 2011 and 2012). Land cover data were then integrated in a GIS using ArcGiS. We then extracted the proportion of each landcover type and estimated for each field the surrounding landscape diversity using Shannon diversity index. We interviewed cotton farmers for the agricultural practices. We used the theoretical information approach developed by Burnham and Anderson (2002) to investigate the effects and determine the relative importance of agricultural and landscape factors on the abundance of H. armigera in cotton fields. We found that agricultural practices are the most important variables that determine H. armigera larvae in cotton fields. This study also highlights the role of the previous landcover in the infestation of a cotton field: A previous tomato landcover increased infestation three times more than a previous maize landcover. Some cotton fields following on from tomato can suffer from heavy local infestations caused by the last diapausing generation as showed by Nibouche (1994) that H. armigera can undergo a diapause of three months in Burkina Faso. At landscape scale, we found a strong and positive effect of the proportion of cotton and the proportion of tomato on the abundance of H. armigera larvae. This result is consistent with the resource concentration hypothesis (Root 1973). Nor the proportion of maize in the landscape nor the Shannon diversity index of crops showed significant relationships with the abundance of H. armigera larvae Relative importance of predictor variables included in the multiple logistic regression models performed to predict H. armigera incidence. H. armigera infestation rate (number of infested cotton plants per number of observed cotton plants) discriminated by previous land cover type. Letters indicate statistical differences of Mann-Whitney test among previous land cover types. Implication for H. armigera outbreaks management. This study highlights the importance of agricultural practices and landscape composition for H. armigera control. We found a strong effect of agricultural practices on H. armigera larvae abundance. Delays in the sowing date and increased weeding frequency reduced the abundance of H. armigera in cotton fields; whereas the proportion of cotton and tomato in the landscape increased it. This study also highlights the role of the previous landcover in the infestation of a cotton field: A previous tomato landcover increased infestation three times more than a previous maize landcover. Do landscape factors influence the abundance and larval diet Helicoverpa armigera in cotton fields in northern Benin? Tsafack et al, 2016 In this part of my thesis, I focused on adult Helicoverpa armigera. I investigated the impact of landscape composition and host crop diversity on the abundance of H. armigera and natal host plant use. H. armigera is a facultative migrant species, able to travel long distances of thousands of kilometers (Feng et al., 2009) but also local smaller scale movements (Ramaswamy 1988) in favourable environmental conditions like in tropical regions. Little is known on its small-scale movements. To understand these movements, we investigated the effect of landscape composition and H. armigera plant use on three spatial scales: 100 m, 250 m and 500 m diameters. Three Buffer scales of diameter 100 m, 250 m et 500 m centered on the cotton field selected. We selected 37 cotton fields with the same criteria as the previous study. The cotton fields were selected according to a diversity gradient of the landcover: ≥50% cotton, ≥50% maize, ≥50% natural vegetation and the presence of tomato. First, we found that the 500 m scale was the best scale at which abundance and larval diet of H. armigera should be investigated. Second, host plant diversity was the main landscape factor explaining the abundance of H. armigera. Landscapes with great diversity of H. armigera host crops harboured more H. armigera than the other landscapes. The effect of the host plants (cotton, corn or sorghum) considered individually, is less strong and varies according to the landscape scale and the year considered. Unlike our expectations, we did not observe any effect of natural vegetation on the abundance or larval diet of H. armigera. As suggestion for H. armigera management, landscape composition and sequential availability of host plants should be considered as keys factors. See Tsafack et al, 2016 for more details. Bibliographical references Nibouche S. 1994. Cycle évolutif de Helicoverpa armigera (Hubner, 1808) (Lepidoptera, Noctuidae) dans l'Ouest du Burkina Faso : biologie, écologie et variabilité géographique des populations. PhD Thesis, Ecole Nationale supérieure Agronomique, Montpellier, France. Root RB. 1973. Organization of a plant-arthropod association in simple and diverse habitats: the fauna of collards (Brassica oleracea). Ecological monograph, 43:95–124. Udamale SK, Moharil MP, Ugale TB, Mankar JM. 2013. Isolation of proteinase inhibitors from okra for inhibiting the Helicoverpa armigera (Hubner) gut proteinases. Indian journal of agricultural research, 47(3). Ramaswamy, S.B. 1988. Host finding by moths: sensory modalities and behaviours. Journal of Insect Physiological 34: 235-249 Tsafack, N., Alignier, A., Head, G.P., Kim, J.H., Goulard, M., Menozzi, P., Ouin, A., 2016. Landscape effects on the abundance and larval diet of the polyphagous pest Helicoverpa armigera in cotton fields in North Benin. Pest Manag. Sci. 72, 1613–1626. https://doi.org/10.1002/ps.4197 Tsafack, N., Menozzi, P., Brevault, T., Soti, V., Deconchat, M., Ouin, A., 2013. Effects of landscape context and agricultural practices on the abundance of cotton bollworm Helicoverpa armigera in cotton fields: A case study in northern Benin. Int. J. Pest Manag. 59, 294–302. https://doi.org/10.1080/09670874.2013.852270