A5054250328- Remote Sensing in Agriculture
- Distributed and Parallel Computing Systems
- Remote Sensing and LiDAR Applications
- Methane Hydrates and Related Phenomena
- Forest ecology and management
- Advanced Computational Techniques and Applications
- Plant Ecology and Soil Science
- Scientific Computing and Data Management
- Meteorological Phenomena and Simulations
- Synthetic Aperture Radar (SAR) Applications and Techniques
- Tropical and Extratropical Cyclones Research
- Remote Sensing and Land Use
- Tree-ring climate responses
- Time Series Analysis and Forecasting
- Soil Geostatistics and Mapping
- Computational Physics and Python Applications
- Species Distribution and Climate Change
- Geological Modeling and Analysis
- Landslides and related hazards
- Ecology and Vegetation Dynamics Studies
- Neural Networks and Applications
- Flood Risk Assessment and Management
- Geochemistry and Geologic Mapping
- Ecosystem dynamics and resilience
- 3D Surveying and Cultural Heritage
Max Planck Institute for Biogeochemistry
2023-2024
Deutsches Zentrum für Luft- und Raumfahrt e. V. (DLR)
2019-2023
Max Planck Institute of Biochemistry
2022-2023
Friedrich Schiller University Jena
2018-2020
University of Twente
1995
Information on the spatial distribution of aboveground biomass (AGB) over large areas is needed for understanding and managing processes involved in carbon cycle supporting international policies climate change mitigation adaption. Furthermore, these products provide important baseline data development sustainable management strategies to local stakeholders. The use remote sensing can spatially explicit information AGB from global scales. In this study, we mapped national Mexican forest...
Abstract. Climate variables carry signatures of variability at multiple timescales. How these modes are reflected in the state terrestrial biosphere is still not quantified or discussed global scale. Here, we set out to gain a understanding relevance different vegetation greenness and its covariability with climate. We used >30 years remote sensing records normalized difference index (NDVI) characterize across timescales from submonthly oscillations decadal trends using discrete Fourier...
Abstract Recent advancements in Earth system science have been marked by the exponential increase availability of diverse, multivariate datasets characterised moderate to high spatio-temporal resolutions. System Data Cubes (ESDCs) emerged as one suitable solution for transforming this flood data into a simple yet robust structure. ESDCs achieve organising an analysis-ready format aligned with grid, facilitating user-friendly analysis and diminishing need extensive technical processing...
Progress in Earth system science is accelerating rapidly, due to the increasing availability of multivariate datasets, often global, with moderate high spatio-temporal resolutions. Turning these data into knowledge presents interoperability, technical, analytical, and other challenges. System Data Cubes (ESDCs) have surfaced as essential tools, offering analysis-ready, cloud-optimised solutions. Coupled advancements Artificial Intelligence (AI), solutions potential release a wealth...
There is no doubt that unmanned aerial systems (UAS) will play an increasing role in Earth observation the near future. The field of application very broad and includes aspects environmental monitoring, security, humanitarian aid, or engineering. In particular, drones with camera are already widely used. capability to compute ultra-high-resolution orthomosaics three-dimensional (3D) point clouds from UAS imagery generates a wide interest such systems, not only science community, but also...
Information on the spatial distribution of forest structure parameters (e.g., aboveground biomass, vegetation height) are crucial for assessing terrestrial carbon stocks and emissions. In this study, we sought to assess potential merit multi-temporal dual-polarised L-band observations height estimation in tropical deciduous evergreen forests Mexico. We estimated using a machine learning approach. used airborne LiDAR-based model training result validation. split into test data two different...
Abstract. Climate variables carry signatures of variability at multiple time scales. How these modes are reflected in the state terrestrial biosphere is still not quantified, nor discussed global scale. Here, we set out to gain a understanding relevance different vegetation greenness and its co-variability with climate. We used > 30 years remote sensing records Normalized Difference Vegetation Index (NDVI) characterize across scales from sub-monthly oscillations decadal trends using...
Abstract. In this study, we analyze Sentinel-1 time series data to characterize the observed seasonality of different land cover classes in eastern Thuringia, Germany and identify multi-temporal metrics for their classification. We assess influence polarizations pass directions on backscatter profile. The novelty approach is determination phenological parameters, based a tool that has been originally developed optical imagery. Furthermore, several additional multitemporal are determined...
Today, very dense synthetic aperture radar (SAR) time series are available through the framework of European Copernicus Programme. These require innovative processing and preprocessing approaches including novel speckle suppression algorithms. Here we propose an image transform for hypertemporal SAR stacks. This proposed relies on temporal patterns only, therefore fully preserves spatial resolution. Specifically, explore potential empirical mode decomposition (EMD), a data-driven approach to...
The REDD+ framework requires accurate estimates of deforestation. These are derived by ground measurements supported methods based on remote sensing data to automatically detect and delineate deforestations over large areas. In particular, in the tropics, optical is seldom available due cloud cover. As synthetic aperture radar (SAR) overcomes this limitation, we performed a separability analysis two statistical metrics Sentinel-1 SAR backscatter forested deforested We compared range between...
The application of automatic differentiation and deep learning approaches to tackle current challenges is now a widespread practice. biogeosciences community no stranger this trend; however, quite often, previously known physical model abstractions are discarded.In study, we the ecosystem dynamics vegetation, water, carbon cycles adopting hybrid approach. This methodology involves preserving representations for simulating targeted processes while utilizing neural networks learn spatial...
Global forest ecosystems face unprecedented challenges, such as fire, wind, drought, and insect outbreaks, resulting in rapid decline. Analyzing these disturbances on a large scale requires the use of remote sensing techniques, but spatial temporal uncertainty disturbance reference data poses significant obstacle. In this study, we validate refine existing labels U.S. Forest Service Health Protection [1] Dataset USDA by using change detection algorithm [2] based radar from Sentinel-1. To...
Recent advancements in Earth system science have been marked by the exponential increase availability of diverse, multivariate datasets characterised moderate to high spatio-temporal resolutions. System Data Cubes (ESDCs) emerged as one suitable solution for transforming this flood data into a simple yet robust structure. ESDCs achieve organising an analysis-ready format aligned with grid, facilitating user-friendly analysis and diminishing need extensive technical processing knowledge....
In this work we estimated Mexican forest aboveground biomass (AGB) using Synthetic Aperture Radar (SAR) and optical remote sensing data two different reference data: (1) extensive national inventory (NFl) (2) airborne Light Detection Ranging (LiDAR) data. For the second modelling scenario, applied a two-stage upscaling approach: firstly AGB for LiDAR transects were used then to calibrate satellite imagery. Furthermore, propagated uncertainties from field measurements LiDAR-derived...
With the amount of high resolution earth observation data available it is not feasible anymore to do all analysis on local computers or even cluster systems. To achieve performance for out-of-memory datasets we develop YAXArrays.jl package in Julia programming language. provides both an abstraction over chunked n-dimensional arrays with labelled axes and efficient multi-threaded multi-process computation these arrays. In this contribution would like present lessons learned from scaling...
<p>Preprint of a paper that is submitted to the "Journal Selected Topics in Earth Observation and Remote sensing" with following abstract:</p> <p>It widely assumed C-Band Synthetic Aperture Radar (SAR) signal do not reach forest floor dense forests, hence SAR cannot be used for sub-canopy flood mapping tropical forests. Indeed, flooded non-flooded forests are distinguishable single acquisitions. The question whether long-term seasonal dynamics C- Band time series data...
<p>Preprint of a paper that is submitted to the "Journal Selected Topics in Earth Observation and Remote sensing" with following abstract:</p> <p>It widely assumed C-Band Synthetic Aperture Radar (SAR) signal do not reach forest floor dense forests, hence SAR cannot be used for sub-canopy flood mapping tropical forests. Indeed, flooded non-flooded forests are distinguishable single acquisitions. The question whether long-term seasonal dynamics C- Band time series data...