Spectral signatures discrimination of Norway spruce trees under experimentally induced drought and acute thermal stress using hyperspectral imaging

Abstract

Climate change intensifies drought and thermal stress in forests, weakening tree vitality and increasing susceptibility to Ips typographus (L.) infestations. This study evaluates the potential of hyperspectral remote sensing to detect physiological stress in Norway spruce (Picea abies (L.) Karst.) induced by chronic drought and acute thermal stress. An in situ manipulation experiment was conducted within a mature spruce stand, establishing precipitation-exclusion roofs (chronic drought), stand edges (acute thermal stress), and control plots. High resolution hyperspectral imagery data was collected across visible, NIR, and SWIR wavelengths (380–2500 nm) during six flights (2022–2023) using CASI-1500 (0.5 m) and SASI-600 (1.25 m) hyperspectral cameras.

Spectral discrimination revealed significant differences between stressed and control trees, particularly in the red (669.58 nm), NIR (854.96 nm), and SWIR (1557.5 nm, 2082.5 nm) wavebands, while the green region (498.56 nm) separated better in reflectance than in derivative analyses. These wavebands effectively identified physiological changes, including pigment depletion, reduced water content, and nitrogen fluctuations. Spectral separations were strongest in late summer and autumn, coinciding with intensified stress responses. Acute stress showed variable NIR reflectance trends, increasing in autumn but declining in August 2023. Chronic stress exhibited higher SWIR reflectance, indicating reduced needle water content. Statistical analyses, including the Kruskal-Wallis test and post hoc Dunn and Mann-Whitney U tests, confirmed significant spectral separability between stress types.

These spectral markers provide a framework for monitoring tree stress under variable environmental conditions. By integrating remote sensing with climatic data, this study demonstrates how hyperspectral imaging enables early stress detection, supporting proactive forest management against climate-induced threats.