Towards a standardized, ground-based network of hyperspectral measurements: Combining time series from autonomous field spectrometers with Sentinel-2
Bibliographic record
Abstract
Sentinel-2 satellite data enables multispectral monitoring of the earth at a high temporal revisit rate. Combining this information with a network of optical ground measurements enables a more detailed and a more complete understanding of terrestrial ecosystems . However, independent optical ground measurements often lack consistency, especially when comparing different sites in geographically remote locations. Using the very high temporal and spectral resolution offered by the automated field spectrometer systems FloX and RoX (Fluorescence Box and Reflectance Box, respectively, JB-Hyperspectral Devices GmbH, Duesseldorf, Germany), we investigated continuous time series ranging over three years and in ten different locations across Europe, Africa, America and Asia . The continuous records of ground-measured reflectance were first validated against Sentinel-2 top of canopy (TOC) reflectance to evaluate the consistency of the in-situ network. Our results suggest a good agreement of ground-measured reflectance with Sentinel-2 TOC reflectance in vegetation and snow with R 2 around 0.79 in the 833 nm band and R 2 up to 0.94 in the bands around 559 nm and 492 nm, demonstrating good consistency across the network. Spatial misalignment of Sentinel-2 pixel-sizes with respect to the different footprint sizes of the ten automated spectrometers on the ground, atmospheric uncertainties, sub-optimal instrument setup and spatial-temporal variable landscape heterogeneity were identified as the most relevant sources of uncertainties in the network. Comparing the Normalized Difference Vegetation Index (NDVI), Transformed Chlorophyll Absorption in Reflectance Index (TCARI) and Enhanced Vegetation Index (EVI) between ground and satellite revealed a decreasing agreement with increasing complexity of index formulation. The best agreement between satellite and ground was exhibited by NDVI with R 2 around 0.96 and relative error of 4.3% investigating vegetation and snow across all ten sites. Furthermore, we identified a seasonal pattern in residuals of NDVI between ground and satellite in an alpine ecosystem in northern Italy, which was associated with increased spatial heterogeneity due to the effects of diverse vegetation phenology and snowfall. In contrast, a random distribution of residuals was recognized in a rather uniform oak forest canopy in southern France. Clustering Sentinel-2 pixels with respect to their temporal patterns in NDVI identified similar areas seen as homogenous in the canopy of Torgnon, Italy, and Observatoire de Haute-Provence (OHP), France, each. The very high temporal resolution of NDVI measured on the ground confirmed overlap with matched homogenous areas, but must consider seasonal landscape heterogeneity. Using well-standardized and globally homogenous Sentinel-2 TOC reflectance enabled the assessment of uncertainties in ten field spectrometer sites around the world. The standardization of the automated field spectrometers, their data products and data annotation were essential prerequisites that enabled joint validation against Sentinel-2. Harmonizing optical ground measurements with respect to a satellite is promising for future research to ensure the valid intercomparison and transfer of data products across different sites in a network worldwide.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.009 | 0.005 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.003 | 0.003 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.002 | 0.003 |
| Open science | 0.002 | 0.003 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.001 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".