Predicting forested wetland soil carbon using quantitative color sensor measurements in the region of northern...

Forested wetland soils within the Piedmont and Coastal Plain physiographic provinces of Northern Virginia (NOVA) were investigated to determine the utility of a handheld colorimeter, the Nix Pro Color Sensor ("Nix"), for predicting carbon contents (TC) and stocks (TC stocks) from on-site color measurements. Both the color variables recorded with each Nix scan ("Nix color variables"; n = 15) and carbon contents significantly differed between sites, with redder soils (higher a and h) at Piedmont sites, and higher TC at sites with darker soils (lower values of L, or lightness; p < 0.05). Nix-carbon correlation analysis revealed strong relationships between L (lightness), X (a virtual spectral variable), R (additive red), and K_K (black) and log-transformed TC (Ln[TC]; |r| = 0.70; p < 0.01 for all). Simple linear regressions were conducted to identify how well these four final Nix variables could predict soil carbon. Using all color measurements, about 50% of Ln(TC) variability could be explained by L, X, R, or K_K (p < 0.01), yet with higher predictive power obtained for Coastal Plain soils (0.55 < R² < 0.65; p < 0.01). Regression model strength was maximized between Ln(TC) and the four final Nix variables using simple linear regressions when color measurements observed at a specific depth were first averaged (0.66 < R² < 0.70; p < 0.01). While further study is warranted to investigate Nix applicability within various soil settings, these results demonstrate potential for the Nix and its soil color measurements to assist with rapid field-based assessments of soil carbon in forested wetlands.