Exploring the use of thermal neutron counts to track orchard phenological development

Regular monitoring of plant development and soil moisture variations is essential for managing orchard systems and optimizing irrigation. Cosmic Ray Neutron Sensors (CRNS) are increasingly used for reliable, non-invasive soil moisture estimation. However, the potential of CRNS for monitoring plant development remains largely uninvestigated. The objective of this study is to assess the response of thermal ( N th ) and epithermal ( N epi ) neutron intensities to the seasonal changes in tree structure and water content. In particular, we aim to investigate whether the observed neutron responses can be used as an indicator of plant development in commercial orchard settings. A CRNS was installed at a cherry orchard site in southeastern France and operated continuously for 10 months in 2022. Observations were compared to several proxies for tree canopy characteristics. First, neutron intensity values were compared with monthly plant area index (PAI) estimates derived from images collected with a light detection and ranging (LiDAR) sensor mounted on an unmanned aerial vehicle (UAV). PAI in (m 2 m −2 ) is defined as the total surface area of all above-ground canopy components, including leaves, stems, and branches per unit horizontal ground surface area. Second, N th was compared with commonly used vegetation indices derived from multispectral satellite images acquired by PlanetScope and Sentinel-2. The results show a strong correlation between N th and UAV-derived PAI with R 2 = 0.86. N th increased linearly by approximately 4.5% per 1 m 2 m −2 increase in PAI. Of the vegetation indices, the Normalized Difference Red Edge (NDRE) index derived from PlanetScope images showed the highest correlation ( R 2 = 0.69) with N th . The corresponding R 2 with NDRE from coarser-resolution Sentinel-2 data was lower ( R 2 = 0.51). The correlation between N th and PAI was higher than that between N th and SM ( R 2 = 0.61). Results suggest that variations in N th are potentially valuable for vegetation monitoring, provided the confounding effect of soil moisture can be taken into account.