Answers are grounded in this paper and cite their sources.
Mangroves constitute a globally critical coastal ecosystem, whose ecological functions are fundamentally governed by key soil nutrients (SOC, TN, and TP) and their stoichiometric ratios (C:N, C:P, and N:P). Their hyperspectral remote sensing retrieval remains challenging due to inherently weak spectral signals and significant regional habitat heterogeneity, which limit model generalizability. To address these issues, we systematically investigated the spectral response mechanisms of mangrove soil nutrients using 227 soil samples from two representative mangrove regions in China. We developed a continuous wavelet spectral similarity (CWSS) method to quantify the coupling between nutrient variables and spectral responses, and developed an improved two-dimensional correlation spectroscopy approach to identify diagnostic spectral bands for each for SOC (390–800 and 1369–1394 nm), TN (350–900, 1180–1600, and 2150–2400 nm), and TP (350–600 and 1660–2250 nm). The mechanism-guided competitive partial least squares (CPLS) mode outperformed conventional PLS across 27 inversion schemes, with R2 improvements of 0.06–0.65. The highest inversion accuracy was achieved for SOC, followed by TN and TP. Integrating fine-tuned transfer learning enhanced cross-regional applicability (ΔR2 = 0.218–0.403), confirming strong model transferability. A structural equation model identified soil depth and habitat quality as primary drivers of nutrient spatial variation.
No takes yet. Share an insight, caveat, or question.
Huang et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: