Integrating Phenological Data from observational networks, Citizen Science, and a Land Surface Model in Central European Forests

Phenology plays an important role in mediating ecosystem responses to global change, linking plant physiology and ecosystem processes. However there is not one single data source that has both the temporal and spatial coverage needed to fully capture phenology at regional and global scales. This influences the accuracy of our future predictions of phenology and implicitly the carbon cycle, as the performance of the model relies on the type and quality of data used for validation. Given that accurate representation of phenological processes in land surface models (LSMs) is crucial for predicting ecosystem responses to environmental changes, integrating data across multiple scales and sources is essential for improving model accuracy and reliability. Focusing on central European forest sites, we combine the growing season simulated with the QUINCY LSM with phenological information from eddy covariance tower networks (Fluxnet, ICOS, European Eddy Fluxes Database Cluster), phenocam imagery from the Phenocam network, and remote sensing products.For the first time, we include citizen science observations collected through the Flora Incognita app, complemented with observations from the GBIF and iNaturalist platforms. We compare growing season metrics derived from these datasets to i) evaluate the advantages and limitations of each individual data source for data-model integration and ii) determine how data sources can be integrated to better represent phenological stages in temperate deciduous forests. This work underscores the importance of network efforts to provide datasets useful for bridging information gaps, and highlights how citizen science participation can meaningfully contribute to traditional datasets and foster a broader understanding of ecosystem dynamics in the face of global change.