Bulgin et al. (2025) Spatial Sampling Uncertainty for MODIS Terra Land Surface Temperature Retrievals

Identification

• Journal: Remote Sensing
• Year: 2025
• Date: 2025-10-15
• Authors: Claire E. Bulgin, Darren Ghent, Mike Perry
• DOI: 10.3390/rs17203435

Research Groups

• Department of Meteorology, University of Reading, Reading, UK
• National Centre for Earth Observation, University of Reading, Reading, UK
• National Centre for Earth Observation, Space Park Leicester, Leicester, UK

Short Summary

This study develops a spatial sampling uncertainty model for MODIS Terra Land Surface Temperature (LST) products when coarsening from 0.01° to 0.05° and 0.1° resolutions, revealing that uncertainty is dependent on land cover and solar zenith angle.

Objective

• To develop and apply a spatial sampling uncertainty model for MODIS Terra Land Surface Temperature (LST) products when re-gridding from 0.01° (approximately 1 km) to coarser resolutions of 0.05° and 0.1°, specifically adapting the methodology for the greater heterogeneity of land surfaces compared to oceans.

Study Configuration

• Spatial Scale: Global coverage, with target resolutions of 0.05° and 0.1° derived from native 0.01° (approximately 1 km) satellite data.
• Temporal Scale: One year of data from 2011, including both day and nighttime retrievals, focusing on spatial sampling uncertainty at the time of satellite overpass.

Methodology and Data

• Models used: Spatial sampling uncertainty derivation based on a previously established methodology for sea surface temperature, adapted for land. Fourth-order polynomial fits are used to parameterize the spatial sampling uncertainty curves.
• Data sources: Moderate-Resolution Imaging Spectroradiometer (MODIS) aboard the Terra satellite, specifically Level 3 (L3) LST products at 0.01° resolution. Land cover classifications are based on ESA Land Cover CCI biome definitions (42 classes, simplified to 8 dominant biomes). Realistic cloud masks were derived from Advanced Along-Track Scanning Radiometer (AATSR) data.

Main Results

• A spatial sampling uncertainty model was developed for coarsening 1 km MODIS Terra LST products to 0.05° and 0.1° resolutions.
• Sampling uncertainty for LST is dependent on both the underlying land cover (biome) and the solar zenith angle at the time of observation.
• The largest sampling uncertainties occur in regions of mixed land cover at 0.05° (maximum 0.99 K) and for urban areas at 0.1° (maximum 2.5 K).
• The shape of the spatial sampling uncertainty curve with clear-sky fraction differs significantly from previous parameterizations, which showed an inconsistent trend.
• For most biomes, spatial sampling uncertainty is larger when the sun is directly overhead (solar zenith angle 0–10°), with bare soil being an exception where uncertainty is lowest under these conditions.
• For mixed pixels, sampling uncertainty increases as the percentage of the dominant land cover decreases and as the number of different land cover classes within the grid cell increases.

Contributions

• Development of a robust spatial sampling uncertainty parameterization specifically for land surface temperature (LST) data, adapting a methodology previously applied to sea surface temperature (SST) for the more heterogeneous land surface.
• Quantification of LST sampling uncertainty as a function of clear-sky fraction, underlying land cover (biome), and solar zenith angle for 0.05° and 0.1° resolutions.
• Demonstration that the current sampling uncertainty approach used in ESA LST CCI products has limited utility and proposing a more accurate, physically-based model.
• Provision of polynomial coefficients for direct application of the developed spatial sampling uncertainty models by data providers and users.
• The methodology is applicable to any infrared LST products from a native 0.01° resolution, irrespective of retrieval algorithm or satellite, assuming instrument noise can be parameterized and removed.

Funding

• European Space Agency (ESA) within the framework of the Land Surface Temperature project under the Climate Change Initiative (LST_cci), grant number 4000123553/18/I-NB.
• National Centre for Earth Observation from the Natural Environment Research Council through award NE/R016518/1.

Citation

@article{Bulgin2025Spatial,
  author = {Bulgin, Claire E. and Ghent, Darren and Perry, Mike},
  title = {Spatial Sampling Uncertainty for MODIS Terra Land Surface Temperature Retrievals},
  journal = {Remote Sensing},
  year = {2025},
  doi = {10.3390/rs17203435},
  url = {https://doi.org/10.3390/rs17203435}
}