13–15 Nov 2024
Leipziger KUBUS Helmholtz-Zentrum für Umweltforschung – UFZ
Europe/Berlin timezone
Welcome to the iDiv Conference 2024! Registration is now closed

Leveraging Deep-Learning for an holistic understanding of soil moisture dynamics as affected by plant diversity

Not scheduled
1m
Leipziger KUBUS/2-AB - Hall 2 (Leipziger KUBUS)

Leipziger KUBUS/2-AB - Hall 2

Leipziger KUBUS

100
Poster Flexpool Poster Flash Talks

Speaker

Gideon Stein

Description

As recent studies suggest, soil moisture seems to be positively affected by plant diversity. Since soil moisture is an essential part of an ecosystem's microclimate, driving plant and microbial interactions, this effect is likely an important mechanism that drives the diversity-ecosystem stability relationship. However, complicated interactions between soil moisture, soil temperature, and plant diversity are still not yet fully causally understood. Such research questions require sophisticated statistical methods and computational tools, especially since the relationship between plant diversity and soil moisture may vary depending on the scale of observation.
Here, novel Deep Learning approaches provide potential to accurately model these complicated and intertwined mechanisms given that enough data is available. By learning a joint soil moisture model based on 20 years of Jena Experiment data, we not only gain insights into the relationship between moisture and other variables such as soil temperature and plant diversity, but also receive proper soil moisture estimates where no measurements were taken, providing an effective interpolation strategy that takes into account all information sources available.

Status Group Doctoral Researcher

Primary author

Co-authors

Dr Maha Shadaydeh Yuanyuan Huang Anne Ebeling Nico Eisenhauer (German Centre for Integrative Biodiversity Research (iDiv) Halle-Jena-Leipzig) Joachim Denzler (Computer Vision Group, Friedrich Schiller University Jena)

Presentation materials