# When Should a Satellite Estimate Be Changed? Stress-Testing Neural Corrections for Evapotranspiration

arXiv: https://arxiv.org/abs/2609.35314 · HTML: https://mlsanity.com/paper/2609.35314/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 9 of its 18 references that are in the map.
- First posted on arXiv 28 Sep 2026.

## Builds on (most cited first, 9 of 9)

- [Simple and Scalable Predictive Uncertainty Estimation using Deep Ensembles](https://mlsanity.com/paper/1612.01474/) (2016)
- [How Neural Networks Extrapolate: From Feedforward to Graph Neural Networks](https://mlsanity.com/paper/2009.11848/) (2020)
- [Predicting into unknown space? Estimating the area of applicability of spatial prediction models](https://mlsanity.com/paper/2005.07939/) (2020)
- [Regression with Multi-Expert Deferral](https://mlsanity.com/paper/2403.19494/) (2024)
- [Deep Anti-Regularized Ensembles provide reliable out-of-distribution uncertainty quantification](https://mlsanity.com/paper/2304.04042/) (2023)
- [Conformalized Selective Regression](https://mlsanity.com/paper/2402.16300/) (2024)
- [Extrapolability Improvement of Machine Learning-Based Evapotranspiration Models via Domain-Adversarial Neural Networks](https://mlsanity.com/paper/2406.00805/) (2024)
- [Knowledge-Guided Machine Learning Models to Upscale Evapotranspiration in the U.S. Midwest](https://mlsanity.com/paper/2510.11505/) (2025)
- [GeoQ: Geometry-Aware Conditional Quantile Error Estimation for Scientific Surrogate Models](https://mlsanity.com/paper/2608.21652/) (2026)

## Abstract (arXiv)

Neural residuals can improve satellite evapotranspiration (ET) estimates, but selectors must predict when a correction helps and reject unsupported inputs. We evaluate ten-member models on 16,366 flux-tower observations from 151 stations paired with OpenET, across nine rolling years and five spatial folds. At one held-out station, Gain accepted corrections on all 32 physically invalid records: it predicted a mean benefit of 0.83 mm/day, but the corrections increased mean absolute error by 21.6 mm/day versus OpenET. On spatially held-out unit errors, SupportGain reduced station-macro MAE versus Gain by 0.148 mm/day under wind x3.6 (simultaneous 95% interval, 0.070 to 0.226), with 9.3% acceptance versus Gain's 51.8%; on clean inputs, its 0.006 mm/day advantage had an interval that includes zero. These fault analyses are exploratory; none of 40 preplanned temporal comparisons passed Holm correction, while a separate predeclared cropland contrast found 0.041 mm/day lower station-macro MAE with crop-only training (95% interval, 0.009 to 0.079).