Comparison of spatial association approaches for landscape mapping of soil organic carbon stocks

• Main Authors: B. A. Miller, S. Koszinski, M. Wehrhan, M. Sommer
• Format: Article
• Language: English
• Published: Copernicus Publications 2015-03-01
• Series: SOIL
• Online Access: http://www.soil-journal.net/1/217/2015/soil-1-217-2015.pdf
• ISSN: 2199-3971; 2199-398X

The distribution of soil organic carbon (SOC) can be variable at small analysis scales, but consideration of its role in regional and global issues demands the mapping of large extents. There are many different strategies for mapping SOC, among which is to model the variables needed to calculate the SOC stock indirectly or to model the SOC stock directly. The purpose of this research is to compare direct and indirect approaches to mapping SOC stocks from rule-based, multiple linear regression models applied at the landscape scale via spatial association. The final products for both strategies are high-resolution maps of SOC stocks (kg m^&minus;2), covering an area of 122 km², with accompanying maps of estimated error. For the direct modelling approach, the estimated error map was based on the internal error estimations from the model rules. For the indirect approach, the estimated error map was produced by spatially combining the error estimates of component models via standard error propagation equations. We compared these two strategies for mapping SOC stocks on the basis of the qualities of the resulting maps as well as the magnitude and distribution of the estimated error. The direct approach produced a map with less spatial variation than the map produced by the indirect approach. The increased spatial variation represented by the indirect approach improved <i>R</i>² values for the topsoil and subsoil stocks. Although the indirect approach had a lower mean estimated error for the topsoil stock, the mean estimated error for the total SOC stock (topsoil + subsoil) was lower for the direct approach. For these reasons, we recommend the direct approach to modelling SOC stocks be considered a more conservative estimate of the SOC stocks' spatial distribution.