Spatial variation of soil organic carbon in a northeastern U.S. watershed

Increasing the accumulation of organic carbon (C) in agricultural soils provides one means to reduce atmospheric carbon dioxide (CO₂) concentrations, but detection of the relatively small changes in soil organic C is complicated by spatial variability. Soil organic C variation was assessed at various scales within a small (40 ha; 98 ac), mixed-use watershed in central Pennsylvania to determine sampling requirement for possible C credit programs. Composite soil samples (0 to 5 cm; 0 to 2 in deep) were collected on 30-m (98-ft) grid intervals across the watershed and at 10- and 0.6-m (33- and 2-ft) intervals at selected locations, and descriptive- and geo-statistical analysis utilized. Concentrations of soil organic C in pasture and forest soils were approximately two times greater than cultivated fields, where means ranged from 15 to 24 g C kg^-1 (1.5 to 2.4 percent) and coefficients of variation were typically 15 to 20 percent. Soil organic C was spatially dependent over a range of approximately 200 m (66ï ft) when sampled at 30-m (98-ft) intervals, and high nugget variances indicated spatially-dependent variability over lag distances shorter than 30 m (98ft). However, sampling at 10-m (33 ft) intervals appeared to adequately describe variation. Estimates of sample size requirement showed that, with the observed coefficient variances for individual fields, two- to five-fold increases in sample numbers would be required to verify statistically significant soil organic C changes ≤ 10 percent. Given the large number of samples required to provide representative measurements and the concurrent cost for labor and analysis, participation by farmers in a C credit program could be low if measured verification of soil organic C increases are required. Basing payments on modeled, rather than measured C sequestration rates, should be considered.