Aridity Trends in Central America: A Spatial Correlation Analysis
Published in Atmosphere, 2020
This paper presents a comprehensive analysis of hydroclimate trends across Central America from 1970-1999, with particular focus on aridity patterns and their spatial correlations. The study addresses a critical methodological challenge in climate trend analysis: accounting for spatial correlation between observation stations when assessing statistical significance. Using data from 199 meteorological stations, the research examines trends in five key variables: aridity index, modeled runoff, precipitation, near-surface temperature, and potential evapotranspiration (PET).
The methodology employs Monte Carlo simulations and empirical variogram modeling to identify extreme trend outliers that stand out beyond expected regional climate signals. Results show widespread significant warming trends in temperature and PET when analyzed independently, but when spatial correlation is considered, only the most extreme trends remain statistically significant. The analysis reveals concerning patterns in Costa Rica and Nicaragua, where some locations show significant aridity increases and runoff decreases, particularly in critical watersheds like the Tempisque catchment. The work demonstrates how spatial correlation corrections can help distinguish between broad regional climate signals (such as global warming) and localized extreme changes that may be driven by additional factors like urbanization or land use changes.
Recommended citation: Alfaro-Córdoba, M., Hidalgo, H. G., & Alfaro, E. J. (2020). "Aridity Trends in Central America: A Spatial Correlation Analysis." Atmosphere, 11(4), 427. DOI: 10.3390/atmos11040427
Download Paper