Estimating Organic Carbon from Loss-On-Ignition in Northern Arizona Forest Soils

Many studies in ecology, soil science, and global climate change require accurate estimates of soil organic C (SOC). When calibrated with direct SOC determinations, loss-on-ignition (LOI) has been proposed as a rapid, inexpensive, and accurate method for estimating SOC. We collected 0- to 15- and 15- to 50-cm mineral soil samples from 102 plots within a 110000-ha ponderosa pine (Pinus ponderosa P & C. Lawson) landscape to develop regression equations between LOI and SOC measured with an elemental C analyzer. We tested nine LOI temperature-duration combinations ranging from 300 to 600 C and 2 to 6 h to discern optimal combinations for estimating SOC, used the optimal combination to develop regressions for 100 samples each of 0- to 15- and 15- to 50-cm depths, and assessed whether stratifying samples into ecosystem types improved LOI-SOC equations. Pearson r(2) values between LOI and SOC did not exceed 0.74 for any LOI temperature-duration combination. These values showed no consistent trend to change with increasing duration, but tended to be slightly higher at the lowest temperature (300 degrees C). Multiple regressions, including LOI and clay concentration, explained only 78 (0-15 cm) and 64% (15-50 cm) of the variation in SOC. Relationships between LOl and SOC found in this study are among the weakest reported in the soil literature, and it remains unclear precisely why observed relationships were weak. Our results suggest that LOI may be useful for roughly estimating SOC in this region, but other methods or modifications to LOI are needed when more accurate SOC measurements are required.