When starting a new year it always is good to look back at past achievements. Although the shifting geopolitical circumstances have soured the research climate, BlueGreen Labs’ output remained strong throughout 2025. In the spirit of the end-of-year lists we have created our own.
Publications
The past year was a successful year when it came to published results. In total, we contributed to 15 publications. Our work on Banana and Enset disease management and food security was particularly strong. Highlights include a machine learning mapping exercise to assess Banana Bunchy Top Disease across Rwanda (Gaidashova et al. 2025), guidance on disease management of Fusarium wilt (Nguyen et al. 2025), and causal inference on the importance of both insect vectors and management on disease spread through field trials for Xanthomonas on Enset (Shara et al. 2025). Most of this work was supported by Dr. Ir. Kearsley.
Significant contributions were made to vegetation modelling frameworks by Dr. Hufkens. In particular, the refactoring and updating of the PT-JPL_{SM} model was key in calculating the transit times of water through vegetation in Felton et al. (2025). Major initial refactoring was also covered by Dr. Hufkens for the R implementation of the sofun model in {rsofun} which led to the formal publication of the package on CRAN and as peer-reviewed software (Paredes et al. 2025).
BlueGreen Labs continues to support phenocam based phenology research both across the US and European networks (as integrated part of the Integrated Carbon Observation System, or ICOS). As such, 2025 has seen the migration from the old Stardot SC5 to the new Stardot Live2 cameras. With this came software updates (see below), which now include encrypted data transfers and a more consistent installation routine. A comparison between the new and old hardware is described in Javadian et al. (2025). Similarly, we contributed to the updated PhenoCam dataset description paper, bringing new improvements to the data workflow (Young et al. 2025). Expert advice was also provided within the context of Handwritten Text Recognition of historical climate observations in a collaboration with the Water and Climate Research Group at the Free University Brussels (VUB) and the Center for Digital Humanities at Ghent University (Vercruysse et al. 2025).
You can find a full list of our publications on our publications page.
Software & Hardware
On the software front we continued the support for all our R packages, with new versions for the {ecmwfr} package, which now sees over 80K downloads. Updates to the {appeears} package, and demands for features, hint towards increased uptake of this tool. The {skylight} package has seen a rewrite to C++ to improve speed in support of our new {skytrackr} package. The {skytrackr} package implements geolocation by light to support our ornithology research and workflows. Additional contributions were made to the software workflows for Handwritten Text Recognition at the the Water and Climate Research Group at VUB. Our own independent workflow has seen updates to in the form of the WeaHTR framework. The continued development of the PhenoCam networks in the US and Europe has led to new software to support the required hardware. In particular, the original Phenocam Installation Tool (PIT) was updated to version 2 (in a separate project).
Talks & Teaching
In 2025 we contributed to two main outreach efforts. Together with Prof. Dr. Monika Barget, as part of the HERMES (Humanities Education in Research, Data, and Methods) project, we taught best practices within the context of handwritten text recognition. The workshop outlined potential frameworks to use as well as the pitfalls of common GenAI (multimodal LLMs) models within this context. Dr. Hufkens also contributed to the ECDPM panel discussion regarding the ethical and responsible use of machine learning and artificial inteligence across Africa, drawing from experiences through previous collaborations in Africa in this context with the International Food Policy Research Institute (IFPRI). The full panel discussion is available online.
Funding
BlueGreen Labs was also successful in securing funding across various projects. Most notable, together with a large consortium of partners we successfully secured funding for the Root2Fork project within the Horizon Europe agroecology partnership, a European partnership on accelerating farming systems transition. In this project we provided key editorial and management services, and will continue to do so, in addition to field survey work, for the duration of the project. After a long administrative delay the FieldValAI project together with IFPRI and Wageningen University got funded through the Climate Change AI framework. In this project we will build analysis ready AI/ML datasets within the context of smallholder crop insurance. As such, we will report more on these projects going forward.
