Using Color(imetry) to Predict Soil C in California Forest Soils.

Soils house the globes largest fraction of terrestrial C. Although we can all agree that ‘a lot’ of C is stored in soil, pool size estimates and treatment induced change detection at plot or soil map unit scale lack confidence and limits soil C’s viability as a tradable commodity in current and future C credit systems. Intrinsic variability, limited sample size, (due to cost, handling and analysis) and lack of systematic approach (often site specific analysis) reinforce this ‘problem’ and support the consistent reuse of existing estimates with no clear path to greater precision or accuracy. Can we reduce variability without increasing sample size? Perhaps a more rapid, accurate and low cost assessment method for soil C is needed.  
Color is the most common metric in soil science but Munsell color book observations lack quantitative robustness to predict soil C at meaningful levels. Colorimetry takes seconds and coupled with traditional soil C assessment holds great promise to generate predictive models of soil C content.
In this research, we employ soils from soil surveys and managed forest stands to develop relationships between components of measured color and soil C content for C rich surface soils (0 – 80 cm). Our simple bivariate model between % soil C and the L component of the LAB color space (L is analogous to Value in the Munsell system) explained > 85 % of the variability for forest soils around Northern California across a range of treated and untreated soil conditions. Future efforts will utilize randomly generated/collected samples from across the production forest landscape in California for model testing and validation. We feel that this approach provides a simple low cost reproducible method to assess soil color/C in forest soils of California.