Plant phenological modelling
Climate change is a reality, and thousands of scientists have declared that we are living in a climate emergency. Climate change is altering plant growth patterns and the timing of life-cycle stages such as flowering and leafing. Phenology is the science that studies these timings and their role as key indicators of global environmental change.
Phenological modeling stands as a vital tool to quantify the influence of environmental changes on plants. We propose using Machine/Deep Learning approaches to model phenology.
The diversity of data sources complicates integration and particularly data preparation for modeling. Moreover, traditional methods have not been designed to capture the underlying spatial, temporal, and spatio-temporal correlations inherent in phenological datasets. The main objective of this thesis is to develop tailored-based workflows and a benchmark study for phenological modeling using observations from phenology networks and open-access geospatial datasets.
[1] M. Khodadadzadeh, P. Kalverla and R. Zurita-Milla, "Harmonizing Machine Learning Based Phenological Modeling: A Unified Workflow for Comparative Analyses," IGARSS 2024 - 2024 IEEE International Geoscience and Remote Sensing Symposium, Athens, Greece, 2024, pp. 5333-5336, doi: 10.1109/IGARSS53475.2024.10641356.
[2] https://springtime.readthedocs.io/en/latest/
[3] R. Zurita-Milla, R. Goncalves, E. Izquierdo-Verdiguier and F. O. Ostermann, "Exploring Spring Onset at Continental Scales: Mapping Phenoregions and Correlating Temperature and Satellite-Based Phenometrics," in IEEE Transactions on Big Data, vol. 6, no. 3, pp. 583-593, 1 Sept. 2020, doi: 10.1109/TBDATA.2019.2926292.