Estimating Frost Safety Margins from Temperature Patterns Using Markov Chains
Journal of Applied Meteorology and Climatology · 2026-07 · 65 (7) · 1101-1115
DOI: 10.1175/jamc-d-25-0150.1 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Applied Meteorology and Climatology, 2026-07. Verfasst von Eduardo Pérez Sosa und 3 weiteren Autor:innen. Die Arbeit nennt 38 Literaturangaben. Gefördert von: Ministerio de Ciencia e Innovación, Banco de España.
Abstract (Originalsprache)
Frost safety margins (FSMs) have traditionally been defined as the difference between critical damage temperatures and observed minima, requiring detailed phenological and physiological data that are often unavailable at regional scales. This study reconceptualizes FSM as a regime-based agroclimatic proxy derived solely from temperature records. FSMs were defined using the timing of the last two annual frost events and their temporal separation, allowing classification into discrete frost regimes. Temporal dynamics of FSM were analyzed through discrete-time Markov chains to estimate transition probabilities, persistence, and recurrence across Aragón (northeastern Spain) using daily minimum temperature records from 1950 to 2020. In 98% of sites, empirical transition frequencies showed no statistically significant differences from the fitted first-order Markov chain representation. Spatial patterns revealed a strong altitudinal gradient: Higher elevations exhibited fewer FSM types, higher persistence, lower effective transition diversity, and substantially improved predictive performance. While model skill at low elevations was moderate and characterized by greater regime diversity, areas above 1500 m MSL approached near-deterministic behavior. Incorporating elevation as a covariate in the transition model improved probabilistic skill, particularly at higher altitudes. An illustrative application combining FSM with simulated flowering dates demonstrated how the framework can contextualize potential frost exposure without relying on direct damage observations. Overall, FSMs provide a structurally consistent and spatially interpretable proxy for characterizing frost exposure regimes, offering a complementary approach for agroclimatic risk assessment in data-limited regions and under changing climatic conditions.
Förderung
- Ministerio de Ciencia e Innovación
- Banco de España
Zitieren
APA
Sosa, E. P., Serrano-Notivoli, R., Saz, S. M. Á., & Hernández-Navarro, M. L. (2026). Estimating Frost Safety Margins from Temperature Patterns Using Markov Chains. Journal of Applied Meteorology and Climatology, 65(7), 1101-1115. https://doi.org/10.1175/jamc-d-25-0150.1
Vancouver
Sosa EP, Serrano-Notivoli R, Saz SMÁ, Hernández-Navarro ML. Estimating Frost Safety Margins from Temperature Patterns Using Markov Chains. Journal of Applied Meteorology and Climatology. 2026;65(7):1101-1115. doi:10.1175/jamc-d-25-0150.1
BibTeX
@article{Sosa2026,
author = {Sosa, Eduardo Pérez and Serrano-Notivoli, Roberto and Saz, S. Miguel Ángel and Hernández-Navarro, María Luz},
title = {Estimating Frost Safety Margins from Temperature Patterns Using Markov Chains},
journal = {Journal of Applied Meteorology and Climatology},
year = {2026},
volume = {65},
number = {7},
pages = {1101--1115},
doi = {10.1175/jamc-d-25-0150.1}
}
EUVIMED