[Study on Estimation of Deserts Soil Total Phosphorus Content by Vis-NIR Spectra with Variable Selection]

Area of Science:

• Soil Science
• Spectroscopy
• Chemometrics

Context:

• Desert soils in Xinjiang's Ebinur Lake Wetland Nature Reserve were analyzed.
• Visible/near-infrared spectra and chemical analysis data were utilized.
• Soil samples were pretreated using Savizky-Golay smoothing, standard normal variation transformation, and first-order differential methods.

Purpose:

• To extract characteristic wavelengths for total phosphorus content in desert soil.
• To develop and compare partial least squares regression models for predicting soil total phosphorus.
• To evaluate the effectiveness of combined ant colony optimization interval partial least squares (ACO-iPLS) and genetic algorithm interval partial least squares (GA-iPLS) methods.

Summary:

• ACO-iPLS identified key spectral bands (500-700, 1101-1300, 1501-1700, 1901-2100 nm).
• GA-iPLS further selected 13 effective wavelengths with minimal collinearity.
• The combined ACO-iPLS and GA-iPLS model demonstrated superior prediction accuracy (RMSEP=0.108 mg/g, R(p)=0.5559) compared to other methods.

Impact:

• A robust and accurate model for estimating total phosphorus in desert soils was established.
• The findings offer a simplified and effective approach for soil nutrient monitoring.
• This method can aid in the assessment and management of desert ecosystems.