Innovation Fellowship 

Optimization of agricultural pest management strategies by combining modelling and digital insect monitoring.

Monitoring is the foundation of effective integrated pest management (IPM) (Barak et al. 1990), as both the implementation and evaluation of IPM programs require accurate information on pest and natural enemy populations (Hassell et al. 1991, Norton and Mumford 1993, Qureshi and Stansly 2007, Castle and Naranjo 2009). Since collecting every individual in a population is impractical, representative samples must be taken. These samples are then used to estimate pest and natural enemy populations over time, resulting in the population dynamics estimations used to inform management practices such as pesticide applications and biological control releases. Precise targeting of pests with respect to timing or spatial location of susceptible developmental stage or reduces the extreme use of agrochemicals and thus the downstream negative effects on health and the environment (Nicolopoulou-Stamati et al. 2016). Due to the importance of these estimations, there is a high demand for affordable, practical, and accurate insect monitoring methods for pests and natural enemies across a wide range of agricultural systems (Binns and Nyrop 1992, Cushing 1998, Bianchi et al. 2013, Roubos et al. 2014).

The existing art discloses an insect monitoring system that aim laser light (LIDAR) at flying insects, detecting the reflective signature of their wings and their body size, and thus identifying the insect to species (Gebru et al. 2018, Kirkeby et al. 2016). Existing devices monitor insect populations to determine efficacy of control measures and location of insects, supplying unprecedented data on insect immigration and presence. However, there are a few shortcomings: 1) Adults insects are exclusively detected, due to reliance on wing beat frequency, 2) lack of sampling discrete individuals may result in overestimation from double counting, 3) abiotic interferences (ex. from rain) may obstruct monitoring, and 4) device use outside of biologically relevant time period increasing operating cost. The incorporation of pest models may help mitigate these difficulties.

Models have long been used to forecast population dynamics. Insects are poikilothermic and both their behavior and development is temperature/dependent. Accumulated heat units, called Growing Degree Days (GDD), are paired with pest biology to predict insect pest populations (Pruess 1983). Stage structured temperature-driven models of insect development can be used to predict adult emergence of insect pests (Sigsgaard 2000; Esbjerg and Sigsgaard 2019) and guide the timing of field monitoring of insects. A model is only as predictive as its starting information allows. While models often use precise biological and regional weather station data implemented in software framework (Holst 2013), they also rely on assumptions about the initial population and immigration. LIDAR integration will challenge the need for assumptions – the model’s initial population and immigration rely on visual data rather than assumptions and weather data will be field-specific. We hypothesize these data will result in more accurate population predictions.

We propose testing the integration of models and LIDAR using Oil Rapeseed as a model system and its world-distributed economic pest Cabbage Stem Flea Beetle (Psylliodes chrysocephala) (CSFB) (Bonnemaison 1965; Nadein 2010). Adult densities were shown to be poor predictors of larval densities (Mathiasen 2015; Mathiasen et al. 2015a, 2015b), the economically damaging stage. University of Copenhagen (UCPH) research found temperature driven CSFB models are better predictors of initial pest presence and field immigration timing (Mathiasen 2015) but also identified a need for more precise pest forecasts. The use of this well studied agricultural system and pest allows this proposal’s collaboration to capitalize on the current research while advancing the field through the addition of stage structured and spatially explicit models. LIDAR technology can provide high quality data about adult activity in space and time. The integration of updated models and LIDAR will result in more accurate estimations of pest population dynamics and greater ability to target pests. The successful implementation of this project advances the field of precision agriculture and furthers the attainability of sustainable agriculture.

This project is under the supervision of Dr. Lene Sigsgaard (University of Copenhagen), Dr. Niels Holst (Aarhus University), and Dr. Jesper Lemmich (FaunaPhotonics Aps.). It is being executed in close collaboration with Dr. Sam Cook (Rothamsted Research). The project is funded by Danmarks Innovationsfond.

In the news:


Barak, A., W. Burkholder, and D. Faustini. 1990. Factors affecting the design of traps for stored-product insects. Journal of the Kansas Entomological Society: 466-485.

Bonnemaison, L. and Jourdheuil, P. 1954. L'altise d'hiver du colza (Psylliodes chrysocephala L.). Ann. Épiphyties. 5: 345-524.

Bianchi, F., A. Ives, and N. Schellhorn. 2013. Interactions between conventional and organic farming for biocontrol services across the landscape. Ecological Applications 23: 1531-1543.

Binns, M., and J. Nyrop. 1992. Sampling insect populations for the purpose of IPM decision making. Annual review of entomology 37: 427-45

Castle, S., and S. E. Naranjo. 2009. Sampling plans, selective insecticides and sustainability: the case for IPM as ‘informed pest management’. Pest management science 65: 1321-1328.

Cushing, J. M. 1998. An introduction to structured population dynamics, SIAM.

Esbjerg, P, Sigsgaard, L. 2019. Temperature dependent growth and mortality of Agrotis segetum Insects 10(1), 7;

Gebru, A., Jansson, S., Ignell, R., Kirkeby, C., Prangsma, J.C. and Brydegaard, M., 2018. Multiband modulation spectroscopy for the determination of sex and species of mosquitoes in flight. Journal of biophotonics, 11(8).

Hassell, M. P., H. N. Comins, and R. M. May. 1991. Spatial structure and chaos in insect population dynamics. Nature 353: 255-258.

Holst, N., 2013. A universal simulator for ecological models. Ecological informatics, 13, pp.70-76.

Kirkeby, C., Wellenreuther, M. and Brydegaard, M., 2016. Observations of movement dynamics of flying insects using high resolution lidar. Scientific reports, 6.

Mathiasen, H., 2015. Biological aspects for forecasting of the cabbage stem flea beetle, Psylliodes chrysocephala L (Doctoral dissertation, Department of Plant and Environmental Sciences, Faculty of Science, University of Copenhagen).

Mathiasen, H.; Sørensen, H.; Bligaard, J.; Esbjerg, Peter 2015. Effect of temperature on reproduction and embryonic development of the cabbage stem flea beetle, Psylliodes chrysocephala L., (Coleoptera: Chrysomelidae). Journal of Applied Entomology 139, 8, 600-608.

Mathiasen, Helle; Bligaard, J.; Esbjerg, Peter. 2015. Survival of cabbage stem flea beetle larvae, Psylliodes chrysocephala, exposed to low temperatures.Entomologia Experimentalis et Applicata 157, 2, 220-226.

Nadein KS, 2010. A review of the genus Psylliodes Latreille (Coleoptera, Chrysomelidae) of the fauna of Russia and neighbouring countries. II. Annotated list of species. Entomol. Obozr. 89, 845-889.

Nicolopoulou-Stamati, P., Maipas, S., Kotampasi, C., Stamatis, P. and Hens, L., 2016. Chemical pesticides and human health: the urgent need for a new concept in agriculture. Frontiers in public health, 4, p.148.

Alan Turning Institute - UK AI Conference

AI Pollen Beetle.mp4