Prior to the

establishment of the study, the stand was burned, and resprouting vegetation (that is, it was a young, developing stand) in the ambient CO2 treatment took up a considerable quantity of N from the soil (−58 kg ha−1 yr−1), redistributing it to tree biomass (+11 kg ha−1 yr−1) and forest floor (+15 kg ha−1 yr−1). The average net ecosystem change of -32 kg ha−1 yr−1 was statistically significant (P = 0.09, Student’s t-test), and presumed to be a residual effect of the fire. In the elevated CO2 treatment, soil change of −39 kg ha−1 yr−1 was approximately balanced by the net increases in forest floor (21 kg ha−1 yr−1) and vegetation (15 kg ha−1 yr−1), for a net, non-significant ecosystem change of -3 kg ha−1 yr−1. Johnson and Todd (1998) reported unaccountably large increases in soil N over a period of 15 years after sawlog (+55 kg ha−1 yr−1) and whole-tree (+83 kg ha−1 yr−1) harvesting in developing stands. Combined with increments GSK1210151A nmr in vegetation (6 kg ha−1 yr−1 in sawlog, 5 kg ha−1 yr−1 in whole-tree) and changes in forest floor + logging residues (−8 kg ha−1 yr−1 in sawlog, −1 kg ha−1 yr−1 in whole-tree), the ecosystem level N increments (+53 kg ha−1 yr−1 in sawlog, +87 kg ha−1 yr−1 in whole-tree) were inexplicably large compared to known N inputs (10 kg ha−1 yr−1 in atmospheric deposition to nearby Walker Branch Watershed). The

authors took great care to sample and analyze samples consistently, but nevertheless remain www.selleckchem.com/products/Everolimus(RAD001).html skeptical about these results. A second resampling of these sites 32 years after the harvest is now underway. Johnson et al. (2006) reported on the third resampling of soils on Walker Branch Watershed, Tennessee.

Between the first and second resampling (1982–1993), ecosystem soil total N had significantly declined by an average of −73 kg ha−1 yr−1, mostly because of a decline in soil N (−70 kg ha−1 yr−1) but also because of tree mortality and associated loss of N. Between the second and third resampling (1993–2004), however, ecosystem N had increased by 43 kg ha−1 yr−1, with about half of the increase occurring in the soil (23 kg ha−1 yr−1). Over the entire 22 year sampling period, ecosystem N declined by −15 kg ha−1 yr−1, mostly a result of decreases in these soil N (−24 kg ha−1 yr−1) offset somewhat by net increases of 5 and 4 kg ha−1 yr−1 in vegetation and forest floor, respectively. The authors of that paper concluded that while large and sometimes statistically significant changes in soil N can occur, these are not unidirectional and can fluctuate over a period of a decade. The authors used only current (2005) analyses on soils from all sampling times, precluding laboratory bias. They also calculated the potential errors due to variations in sampling depth (2 cm out of 15 cm increments) and found that this could not account for the observed changes. Johnson et al.