Near real-time estimation of daytime and nighttime evapotranspiration using GOES-R observations and machine learning models
7 affiliations
- 1Department of Biological Systems Engineering, University of Wisconsin – Madison, Madison, WI, USA; Department of Earth and Planetary Sciences, Yale University, New Haven, CT, USA
- 2Department of Geography, University of Colorado Boulder, Boulder, CO, USA
- 3Department of Biological Systems Engineering, University of Wisconsin – Madison, Madison, WI, USA
- 4Department of Atmospheric and Oceanic Sciences, University of Wisconsin – Madison, Madison, WI, USA
- 5Space Sciences and Engineering Center, University of Wisconsin – Madison, Madison, WI, USA
- 6USDA-ARS Hydrology and Remote Sensing Laboratory, Beltsville, MD, USA
- 7NASA Marshall Space Flight Center, Earth Science Branch, Huntsville, AL, USA
Abstract
Evapotranspiration (ET) is a critical component of the water cycle, influencing climate, agriculture, and water resource management. However, most satellite-derived ET products are limited to daily or coarser temporal resolutions, despite the strong diurnal variability of ET processes. Existing satellite-based ET retrievals are largely restricted to daytime conditions, when nighttime ET is a small but often non-trivial flux. In this study, we introduce the Advanced Baseline Imager Live Imaging of Vegetated Ecosystems ET (ALIVEET), a near real-time, 5 min ET estimation framework, leveraging geostationary satellite observations from the GOES-R Advanced Baseline Imager (ABI) and machine learning models under both clear and cloudy conditions. We test Gradient Boosting Regression (GBR) and Long Short-Term Memory (LSTM) models to assess their ability to estimate ET variations across the diurnal cycle. GBR captures daytime ET with an R2 of 0.74 (normalized RMSE of 0.91) while maintaining low computational cost. For nighttime ET, LSTM models trained on time-series observations perform better, achieving an R2 of 0.24 (nRMSE of 1.29) by leveraging temporal dependencies in land surface temperature (LST) and past ABI observations. Comparisons against daily ET estimates from the physically-based ALEXI remote sensing model demonstrates good agreement but opportunities for improvement. This study demonstrates the potential of integrating machine learning with geostationary remote sensing to advance high-temporal-resolution ET estimation.
Related
- Concurrent daytime and nighttime heatwaves in the late 21st century over the CORDEX-East Asia phase 2 domain using multi-GCM and multi-RCM chains
- Evaluating hv-block cross-validation for dependence-aware model selection in reference evapotranspiration estimation
- Horizontal Strain Rate estimation from InSAR Observations and the Strapdown Method
- Observations-Based Approach to the Estimation of the Oxidizing Capacity of the Atmosphere of Central London over the 2000-2024 Period.
- Hydroclimatic controls and interaction-driven sensitivity of reference evapotranspiration across Indian climate regimes
BibTeX
@article{ranjbar2026near,
title = {{Near real-time estimation of daytime and nighttime evapotranspiration using GOES-R observations and machine learning models}},
author = {Sadegh Ranjbar and Danielle Losos and Sophie Hoffman and Yafang Zhong and Jason A. Otkin and Ankur R. Desai and Martha C. Anderson and Christopher R. Hain and Paul C. Stoy},
journal = {Hydrology and Earth System Sciences},
year = {2026},
month = aug,
doi = {10.5194/hess-30-4927-2026},
url = {https://doi.org/10.5194/hess-30-4927-2026},
note = {CC BY},
}
RIS
TY - JOUR
TI - Near real-time estimation of daytime and nighttime evapotranspiration using GOES-R observations and machine learning models
AU - Sadegh Ranjbar
AU - Danielle Losos
AU - Sophie Hoffman
AU - Yafang Zhong
AU - Jason A. Otkin
AU - Ankur R. Desai
AU - Martha C. Anderson
AU - Christopher R. Hain
AU - Paul C. Stoy
PY - 2026
DA - 2026/08/01
JO - Hydrology and Earth System Sciences
DO - 10.5194/hess-30-4927-2026
UR - https://doi.org/10.5194/hess-30-4927-2026
AB - Evapotranspiration (ET) is a critical component of the water cycle, influencing climate, agriculture, and water resource management. However, most satellite-derived ET products are limited to daily or coarser temporal resolutions, despite the strong diurnal variability of ET processes. Existing satellite-based ET retrievals are largely restricted to daytime conditions, when nighttime ET is a small but often non-trivial flux. In this study, we introduce the Advanced Baseline Imager Live Imaging of Vegetated Ecosystems ET (ALIVEET), a near real-time, 5 min ET estimation framework, leveraging geostationary satellite observations from the GOES-R Advanced Baseline Imager (ABI) and machine learning models under both clear and cloudy conditions. We test Gradient Boosting Regression (GBR) and Long Short-Term Memory (LSTM) models to assess their ability to estimate ET variations across the diurnal cycle. GBR captures daytime ET with an R2 of 0.74 (normalized RMSE of 0.91) while maintaining low computational cost. For nighttime ET, LSTM models trained on time-series observations perform better, achieving an R2 of 0.24 (nRMSE of 1.29) by leveraging temporal dependencies in land surface temperature (LST) and past ABI observations. Comparisons against daily ET estimates from the physically-based ALEXI remote sensing model demonstrates good agreement but opportunities for improvement. This study demonstrates the potential of integrating machine learning with geostationary remote sensing to advance high-temporal-resolution ET estimation.
SN - 1027-5606
SN - 1607-7938
ER -