Semiquantitative color profiling of soils over a land degradation gradient in Sakaerat, Thailand

In this study, we attempted multivariate color profiling of soils over a land degradation gradient represented by dry evergreen forest (original vegetation), dry deciduous forest (moderately disturbed by fire), and bare ground (severely degraded) in Sakaerat, Thailand. The soils were sampled in a dry-to-wet seasonal transition. Values of the red-green-blue (RGB), cyan-magenta-yellow-key black (CMYK), L*a*b*, and hue-intensity-saturation (HIS) color models were determined using the digital software Adobe Photoshop. Land degradation produced significant variations (p<0.05) in R, C, Y, L*, a*, b*, S, and I values (p<0.05). The seasonal transition produced significant variations (p<0.05) in R, G, B, C, M, K, L*, b*, and I values. In discriminating the soils, the color models did not differ in discriminatory power, while discriminatory power was affected by seasonal changes. Most color variation patterns had nonlinear relationships with the intensity of the land degradation gradient, due to effects of fire that darkened the deciduous forest soil, masking the nature of the soil as the intermediate between the evergreen forest and the bare ground soils. Taking these findings into account, the utilization of color profiling of soils in land conservation and rehabilitation is discussed.