Taimur Khan

Contact / Address

Taimur Khan
Data Scientist

Department Community Ecology (Biozönoseforschung)
Helmholtz-Zentrum
für Umweltforschung - UFZ
Theodor-Lieser-Str. 4
06120 Halle, Germany

Tel: +49 341 6025 4363
taimur.khan@ufz.de

LinkedIn | Github | Gitlab


Taimur Khan

Main Focus

I specialize in building data-driven digital twins of ecological systems by combining deep learning, remote sensing, and high-performance computing (HPC). My work sits at the intersection of environmental science and machine learning. I develop scalable research software and data pipelines that power large-scale biodiversity modeling, forest monitoring, and ecosystem analysis.

My core expertise lies in deep learning for environmental image understanding—with a focus on tree crown detection, segmentation, and forecasting from remote sensing imagery. Through projects like DeepTrees, I’ve advanced methods for automated tree inventorying and habitat assessment at continental scales. I also contribute to European initiatives like BioDT and TwinEco, where I help design foundation model–driven biodiversity digital twins. My pipelines run on supercomputing platforms like LUMI and JSC, enabling rapid processing of high-resolution datasets from satellites, UAVs, and ground sensors.

As a certified drone operator, I conduct regular flights to capture multispectral and hyperspectral imagery, integrating site-specific data into broader workflows. These efforts support applications from phenology monitoring to resilience modeling.

Ultimately, I aim to bridge cutting-edge AI and ecological science—developing tools that are not only technically robust but also transparent, reproducible, and accessible. My goal is to empower researchers, institutions, and conservation efforts with insights grounded in data and designed for impact.

Kooperationen / Projekte | Co-operations / Projects

Current Projects:

PhenoEmbed

Biodiversity Meets Data

WALDRESILIENZ

Biodiversity Digital Twin (BioDT)

DeepTrees






Selected Works

torchgbif: FAIR PyTorch DataLoaders and DataSets for GBIF data
IASDT-Workflows
: Data workflows for the Invasive Alien Species Digital twin
DeepTrees: Deep-Learning based spatiotemporal tree inventorying and monitoring from public orthoimages.
IASDT-Dataserver: Data server for interfacing data stored in LUMI-O
Halle Treecrowns: Deep-learning based modeled tree counts in the city of Halle (Saale) (w/ web interface)
Metalabel: Semantic labels for tabular data
OPeNDAP Data Catalog: Containerized template for serving grid and sequence datasets with OPeNDAP
Snakemake Cookiecutter: Workflow template for Snakemake




Publications

  • Khan, T., Koning, de K., Endresen, D., Chala, D., Kusch, E. (In-Review) TwinEco: A Unified Framework for Dynamic Data-Driven Digital Twins in Ecology. Ecological Modeling. Pre-print DOI: https://doi.org/10.1101/2024.07.23.604592
  • Taubert, F., Rossi, T., Wohner, C., Venier, S., Martinovič, T., Khan, T., ... & Banitz, T. (2024). Prototype Biodiversity Digital Twin: grassland biodiversity dynamics. Research Ideas and Outcomes, 10, e124168. DOI: https://doi.org/10.3897/rio.10.e124168
  • Khan, T., El-Gabbas, A., Golivets, M., Souza, A., Gordillo, J., Kierans, D., & Kühn, I. (2024). Prototype Biodiversity Digital Twin: Invasive Alien Species. Research Ideas and Outcomes, 10, e124579. DOI: https://doi.org/10.3897/rio.10.e124579
  • Khan, T., Banitz, T., Golivets, M., Grimm, V., Groeneveld, J., Kühn, I., Taubert, F. (2022). Prototyping a Biodiversity Digital Twin. Helmholtz-UFZ Science Days 2022. DOI: https://doi.org/10.5281/zenodo.8079131
  • Morche, D., Baewert, H., Schuchardt, A., Faust, M., Weber, M., & **Khan, T.** (2019). Fluvial sediment transport in the proglacial Fagge river, Kaunertal, Austria. Geomorphology of Proglacial Systems (pp. 219-229). Springer, Cham. DOI: https://doi.org/10.1007/978-3-319-94184-4_13
  • Khan, T. (2017). DC Resistivity: Estimating pore moisture distribution and mapping permafrost content in Kaunertal, Austria. Department of Geosciences. Skidmore College.