Integrative Hyperspectral Approaches for Advanced Soil Property Analysis and Environmental Monitoring

Hyperspectral imaging has emerged as a powerful tool for enhancing the analysis of diverse soil properties and environmental monitoring. This study refines the precision of soil biogeophysical analyses by improving the estimation of soil moisture content (SMC), soil organic matter (SOM), total carbon (C), and nitrogen (N) through hyperspectral remote sensing techniques. The research examines the capabilities of two prominent moisture retrieval models: the multilayer radiative transfer model (MARMIT) and the modified soil water parametric (SWAP)-Hapke model. These models are evaluated using hyperspectral imagery derived from unmanned aerial systems (UAS) and goniometric data collected across various environmental settings. The findings indicate that the MARMIT model performs well in field applications, particularly with UAS-derived imagery, and identify opportunities to optimize the SWAP-Hapke model for greater accuracy. Additionally, by integrating the PROSAIL, MARMIT, and SWAP-Hapke models with advanced statistical methods such as Elastic Net Regression and Gradient Boosted Regression Trees, the study improves the predictive modeling of soil organic matter, total carbon, and nitrogen in wetland ecosystems. Overall, the results highlight the potential of combining hyperspectral imaging with sophisticated modeling techniques, offering a more comprehensive understanding of soil properties and contributing to more effective environmental management and monitoring practices across diverse ecosystems.