Pedotransfer Functions for Carbon Methods and Bulk-Density Estimation in Florida Soils.

Two primary problems exist to combine historic and emerging datasets for the estimation of soil organic carbon (SOC) stocks. First is the variety of methods that have been used to measure SOC including total organic carbon by combustion and gas analysis (TOC), Walkley-Black dichromate oxidation (WB), and loss on ignition (LOI). Second is the lack of bulk-density (BD) data needed to quantify SOC in a specific volume of soil. The objective of this research is to develop pedotransfer functions (PTFs) to combine three spatial datasets with differing soil carbon analyses and incomplete bulk-density information for a statewide analysis of soil carbon stocks within Florida, USA. We capitalize on the existence of a statewide dataset of fully characterized soil profiles (n=1288) with WB, LOI and BD measurements, a large dataset of profiles from the Santa Fe River Watershed (n=141) with TOC, LOI, and BD measurements, and another dataset of profiles from a smaller landscape unit nested within the watershed (n=152) with LOI measurements. A stratified random technique was used to select 144 samples from the statewide dataset for reanalysis using the TOC and LOI methods. Initial findings are that TOC and LOI are highly correlated for O, A, and E horizons, but not for B and C horizons because of the confounding effect of heat induced water loss from mineral components in deeper horizons. Additionally WB and TOC measurements are highly correlated and robust C method PTFs could be developed for surface soil horizons but ones with higher uncertainty in the subsurface. Class and continuous PTFs are compared to estimate BD. Given these PTF functions it will be possible to pool these datasets for use in a variety of geospatial and mechanistic SOC modeling projects.