David Hafezi Rachti

BG Biogeosciences

The 2024 Outstanding Student and PhD candidate Presentation (OSPP) Award is awarded to David Hafezi Rachti for the poster/PICO entitled:

Interpretable Machine Learning to Understand Multi-Scale Meteorological Impacts on Ecosystem Carbon Uptake (Hafezi Rachti, D.; Reimers, C.; Ament, F.; Winkler, A. J.)

Click here to download the poster/PICO file.

David Hafezi Rachti is a Master student enrolled in the programme Earth System Data Science and Remote Sensing at Leipzig University. His main research interests are atmosphere-biosphere interactions, particularly the terrestrial carbon cycle and the use of machine learning. David works as a student research assistant at the Max Planck Institute for Biogeochemistry in the Atmosphere-Biosphere Coupling, Climate and Causality (ABC3) Group led by Dr. Alexander J. Winkler, and the ERC Synergy Grant “Understanding and Modelling the Earth System with Machine Learning” (USMILE), co-supervised by Dr. Christian Reimers.

At this year's EGU24 conference, David presented a poster highlighting results from his Bachelor thesis at the University of Hamburg, along with his ongoing work as a research assistant at ABC3. In his project, he utilises an innovative interpretable machine learning framework to quantify the impact of multi-scale meteorological events on carbon uptake in forest ecosystems. This framework combines wavelet transforms and convolutional neural networks to connect model predictions with past weather events, considering their timing and time-scale. This approach allows for the identification of legacy effects in ecosystems linked to past meteorological anomalies, such as droughts and heatwaves.