Project Objectives

The Copernicus LSTM mission will provide high spatio-temporal resolution thermal observations. The IPL-LSTM project has developed algorithms, products and validations that enable the transformation of satellite data into useful information for agriculture, water management, climate change, wildfires and urban environments, consolidating the position of the Global Change Unit (UCG) of the University of Valencia in collaboration with ESA.

Main Objectives

  1. Data Curation & Preprocessing: compilation and cleaning of databases and thermal imagery.
  2. LST Retrieval: development of surface temperature algorithms (TES, SW, SW-TES, ANN).
  3. Products: generation of derived indices such as CWSI, evapotranspiration (S-SEBI) and SUHI.
  4. Calibration & Validation: field campaigns and permanent station use.
  5. Demo Applications: applications in water, agriculture, wildfires, urban and coastal waters.
  6. Management: project management, coordination and dissemination.

Phases and Algorithms

Phase 1: Databases and Simulations

Review of campaigns, simulations with MODTRAN and IDL, preprocessing of scenes and atmospheric corrections.

Phase 2: LST Algorithms

Implementation of physical algorithms and artificial intelligence models.

Split Window (SW)
LST = a₀ + a₁·(T₁ + T₂)/2 + a₂·(T₁ - T₂)/2 + a₃·((T₁ - T₂)/2)² + a₄·(1 - ε) + a₅·Δε

T₁, T₂ are brightness temperatures in two thermal channels, ε is the mean emissivity, and Δε is the emissivity difference.

Temperature & Emissivity Separation (TES)
εi = a + b·MMDi
LST = f(Lc, εi)

TES estimates emissivity from the Minimum-Maximum Difference (MMD) of the thermal spectrum, then derives LST by correcting the surface radiance (Lc) with the retrieved emissivity (εi).

SW-TES
LSTSW-TES = α·LSTSW + β·LSTTES

Hybrid combination of SW and TES algorithms to improve accuracy under different atmospheric conditions.

Artificial Neural Networks (ANN)
LST = f(W · X + b)

X are spectral and atmospheric inputs, W the trained weights, b the biases, and f an activation function.

Phase 3: Advanced Products

Crop Water Stress Index (CWSI)
CWSI = (Tc - Twet) / (Tdry - Twet)

Tc is canopy temperature, Twet is temperature under maximum transpiration, and Tdry is temperature under minimum transpiration.

Evapotranspiration (S-SEBI)
Λ = (Rn - G - H) / (Rn - G)

Λ is the evaporative fraction, Rn the net radiation, G the soil heat flux, and H the sensible heat flux.

Phase 4: Calibration and Validation

Intensive campaigns in Seville, Grosseto, Lleida and Valencia, with permanent stations in Barrax, Doñana, Cabo de Gata and Albufera.

Phase 5: Demonstration Applications

  • Water management (daily ET estimation with Sentinel and in situ data).
  • Urban comfort (SUHI maps of Valencia).
  • Agricultural prediction (wheat yield model using GDD + satellite imagery).
  • Coastal monitoring (thermal plumes in waters with Landsat).
  • Solid Earth (fire analysis and integration in hotspot geoportals).

Phase 6: Dissemination

More than 9 indexed publications, 28 national and international conferences, ongoing doctoral theses, and the development of the project website and open data repository.

Research Team

  • José Antonio Sobrino (PI)
  • Juan Carlos Jiménez (Co-PI)
  • Guillem Pau Sòria Barrés (Senior Researcher)
  • Ana Belén Ruescas Orient (Senior Researcher)
  • Belén Franch Gras (Senior Researcher)
  • Susana García Monteiro (FPU, algorithm development and processing)
  • Yingwei Sun (Researcher, simulations and radiative transfer)
  • Drazen Skokovic (Researcher, algorithms and in situ measurements)
  • Rafael Llorens Company (Researcher, processing and campaigns)
  • Daniel Salinas González (Researcher, calibration/validation and processing)