aiqbench analysis
The paper tests whether TimesFM, a 500M-parameter time-series foundation model, can forecast soil moisture and temperature without per-site training. It beats simple statistical baselines at sub-daily horizons but falls behind beyond a day, and a CatBoost model trained on the same data never wins overall. For ML readers, the useful takeaways are the horizon-dependent trade-off and the cold-start limits of zero-shot forecasting on short deployments, not a general claim that foundation models outperform domain-specific models.
Zero-Shot Forecasting of In-Situ Soil Moisture and Temperature with a Time-Series Foundation Model
arXiv:2610.30971v1 Announce Type: new Abstract: Soil moisture and temperature govern plant-water availability and microbial activity, making them key indicators for evaluating land-restoration efforts on degraded landscapes. However, the physical and data-driven models normally used to forecast them both need information that a newly instrumented restoration site does not yet have: hydraulic parameters for the former and months of per-site training data for the latter. We evaluate TimesFM 1, a 500M-parameter time-series foundation model, zero-shot on 15-minute telemetry from 27 sensor stations across three land-management regimes (Control, Treatment, Scald) at a semi-arid restoration site, forecasting 5 cm and 30 cm soil moisture and enclosure temperature at seven lead times from 1 h to 168 h. Using gap-aware rolling-origin evaluation across sensors, we find that TimesFM beats the best rule-based statistical forecasting baselines (persistence, seasonal-naive, or 7-day climatology) by 21-76% at sub-daily lead times but loses by 6-79% beyond a day, a pattern explained by a diurnal-alignment mechanism that holds for both soil-moisture depths and temperature. A CatBoost baseline trained walk-forward on the same origins with full history and weather covariates never wins overall, and we show that its weakness is partly a cold-start effect specific to short deployments. We additionally quantify the inference cost of zero-shot forecasting on commodity CPU hardware. Our results highlight a practical trade-off between zero-shot forecasting capability, predictive performance, and computational efficiency.
Abstract:Soil moisture and temperature govern plant-water availability and microbial activity, making them key indicators for evaluating land-restoration efforts on degraded landscapes. However, the physical and data-driven models normally used to forecast them both need information that a newly instrumented restoration site does not yet have: hydraulic parameters for the former and months of per-site training data for the latter. We evaluate TimesFM 1, a 500M-parameter time-series foundation model, zero-shot on 15-minute telemetry from 27 sensor stations across three land-management regimes (Control, Treatment, Scald) at a semi-arid restoration site, forecasting 5 cm and 30 cm soil moisture and enclosure temperature at seven lead times from 1 h to 168 h. Using gap-aware rolling-origin evaluation across sensors, we find that TimesFM beats the best rule-based statistical forecasting baselines (persistence, seasonal-naive, or 7-day climatology) by 21-76% at sub-daily lead times but loses by 6-79% beyond a day, a pattern explained by a diurnal-alignment mechanism that holds for both soil-moisture depths and temperature. A CatBoost baseline trained walk-forward on the same origins with full history and weather covariates never wins overall, and we show that its weakness is partly a cold-start effect specific to short deployments. We additionally quantify the inference cost of zero-shot forecasting on commodity CPU hardware. Our results highlight a practical trade-off between zero-shot forecasting capability, predictive performance, and computational efficiency.
Artificial Intelligence (cs.AI)
arXiv:2610.30971 [cs.AI]
(or arXiv:2610.30971v1 [cs.AI] for this version)
https://doi.org/10.48550/arXiv.2610.30971 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 314 words
Related coverage
Learning Continuous Regional Temperature Fields with Lead-Time and Resolution Queries
arXiv · Aug 27, 2026
PCA-Enhanced Adaptive NVAR Framework for High-Resolution Sea Surface Temperature Forecasting in the East Sea
arXiv · Jun 12, 2026
RESCAST-100K: A Comprehensive Dataset for Cross-Domain Residential Load and Indoor Temperature Forecasting
arXiv · Jun 3, 2026
Sparse-Observation Atmospheric Thermal Forecasting with Physics-Informed Neural Networks for Climate-Aware Digital Twins
arXiv · Sep 24, 2026