Biodiversity technology and modelling

Malefix - Machine Learning aided ForecastIng of related eXtremes

In this project, we will collect and compile data about physiological constraints of different species from the literature to predict the risk of Swiss biodiversity from sub-seasonal forecasts of extreme weather events. Extremes in water availability and air temperature will be investigated depending on the characteristics of the species. We will use a physiological modeling approach to assess the impact of extreme temperatures on species fitness under climate change scenarios.

Metaweb - Swiss Interaction Networks

This project aims to generate a metaweb, a network of all possible interactions between species from a regional pool, with local interaction networks serving as distinct subsets of the regional metaweb. This allows to quantify the robustness of local networks to global changes to develop novel biodiversity change indicators to function alongside ongoing monitoring programmes.  

Satellite4Biodiversity - Large scale fast biodiversity observation from space

In this project, we combine environmental DNA data from Vigilife and remote sensing images to build the large-scale biodiversity observatory for the freshwater system. We aim at developing improved monitoring approach for biodiversity in rivers using a combination of remote sensing, eDNA metabarcoding data and machine learning driven model to monitor biodiversity temporally in a global scale and build a routine biodiversity observatory.

Swiss Catchment - The Correspondence of Biodiversity Gradients in Blue and Green Ecosystems
The goal of this project is to investigate biodiversity on the basis of 1200 catchments and sub-catchments in Switzerland. We will consider multiple taxonomic groups in terrestrial and freshwater ecosystems, and relying on mostly existing data from biodiversity monitoring, better understand the drivers of biodiversity in green and blue systems as well as their historical changes that have affected Swiss ecosystems.

X-Prize - Remote Biodiversity assessments

Although rainforests are the most diverse ecosystem, the knowledge of the species living in these iconic environemnts is still very limited. The XPrize Rainforest competition was designed to shed light into the rainforests of our planets which harbour many unidentified animals, plants and various chemical compounds that can potentially be used as drugs against human diseases.

Processing of environmental DNA using artificial intelligence for ecosystem monitoring

In this project, we propose to harness a combination of machine learning approaches that transforms eDNA metabarcoding data into informative ecological indicators that improves ecosystem monitoring and decision making. As interpretation from eDNA signal is traditionally generated from the association of the DNA reads to taxonomic labels, we will first develop a machine learning method to associate sequences with a reference database and identify the taxonomic composition of eDNA in samples.  

Robot-based automated collection of eDNA

In this project, we innovate the design, perception and operation of drones to advance applications of eDNA technologies for biodiversity monitoring. By leveraging these automated systems, we can overcome the limitations of manual sampling, reduce cost, and expedite data acquisition.

GRiP – Genetic and Robotic technologies for Pest detection in vineyards

An interdisciplinary research project aiming at improving the sensitivity and timeliness of pest detection and surveillance. The primary application of this study is the detection of two invasive species, Scaphoideus titanus (which causes Flavescence dorée disease) and Popillia japonica Newman, in Swiss vineyards.