Assessment of soil quality using VIS–NIR spectra in invaded coastal wetlands.

Assessment of soil quality is one of the most important issues concerning changes in coastal ecosystem services resulting from the establishment of Spartina alterniflora (S. alterniflora) marshes. Conventional recognition of soil quality involves many soil analyses that can be costly and time consuming; by contrast, visible and near-infrared spectroscopy (VNIRS) for the prediction of soil quality may offer an inexpensive and rapid approach. This study explored the suitability of using VNIRS and partial least squares analysis for predicting a global soil quality index (SQI) and its categories along a chronosequence of S. alterniflora invasion. Soil properties that were significantly correlated with S. alterniflora age were selected for calculating the SQI. Two types of SQIs (productive function and salt-driven) were considered in this study. The SQI significantly (P < 0.05) increased for both types during 17 years of establishment. There were spectral differences in the SQI category, especially in the range of 700–2500 nm. Samples from different SQI categories and different invasion ages could be discriminated from spectral reflectance. 'Moderate to good' prediction models could be obtained for the SQI and its categories based on refined spectra that were identified as important responses between the SQI and spectra. These results demonstrate the application of VNIRS as an efficient approach for estimating soil quality, as well as for the discrimination of soil state associated with the establishment phase. The results suggest that VNIRS has a high potential for monitoring the soil quality of S. alterniflora soils, which can be used for soil development assessments in coastal ecosystems. [ABSTRACT FROM AUTHOR]