Spatial variability of organic layer carbon stocks —implications and suggestions for sampling designs

Accurate field measurements from inventories across fine spatial scales are critical to improve sampling designs and to increase the precision of forest carbon cycling modeling. By studying soils undisturbed from active forest management, this paper gives a unique insight in the naturally occurring variability of organic layer C and provides valuable references against which subsequent and future sampling schemes can be evaluated. We found that the organic layer C stocks displayed great short-range variability with spatial autocorrelation distances ranging from 0.86 up to 2.85 m. When spatial autocorrelations are known, we show that a minimum of 20 inventory samples separated by ∼5 m is needed to determine the organic layer C stock with a precision of ±0.5 kg C m−2. Our data also demonstrates a strong relationship between the organic layer C stock and horizon thickness (R 2 ranging from 0.58 to 0.82). This relationship suggests that relatively inexpensive measurements of horizon thickness can supplement soil C sampling, by reducing the number of soil samples collected, or to enhance the spatial resolution of organic layer C mapping.

Full text: http://link.springer.com/article/10.1007%2Fs10661-015-4741-x

Omnidirectional Mantel correlogram showing the strength of the spatial correlation between organic layer C and horizon thickness for soil cores within different distance classes (at 1000 permutations). Points are plotted at the midpoint of each distance class. Bins with less than 25 pairs were excluded from the analysis. Solid circles represent significant correlations, while open circles show correlations that are not significantly different from zero. At the distance where the plotted lines approach the x-intercept (red dotted line) objects are no more similar than that expected by-chance-alone across the plot.

Omnidirectional Mantel correlogram showing the strength of the spatial correlation between organic layer C and horizon thickness for soil cores within different distance classes (at 1000 permutations). Points are plotted at the midpoint of each distance class. Bins with less than 25 pairs were excluded from the analysis. Solid circles represent significant correlations, while open circles show correlations that are not significantly different from zero. At the distance where the plotted lines approach the x-intercept (red dotted line) objects are no more similar than that expected by-chance-alone across the plot.

Advertisements

About TeeYay

Ph.D. in Assessment, Monitoring, and Geospatial Analysis (applied statistics) from the University of Minnesota, USA.
This entry was posted in Statistics and tagged , , , , , , , , , , , , , , , , . Bookmark the permalink.