[1] Stable and radiocarbon isotopes were used to investigate the role of soil clay content in the storage and dynamics of soil carbon in tropical forest soils. Organic matter in clay-rich Oxisols and Ultisols contains at least two distinct components: (1) material with light d 13 C signatures and turnover times of decades or less; and (2) clay-associated, 13 C-enriched, carbon with turnover times of decades at the surface to millennia at depths >20 cm. Soil texture, in this case clay content, exerts a major control on the amount of slowly cycling carbon and therefore influences the storage and dynamics of carbon in tropical forest soils. Soils in primary tropical forest have been proposed as a potentially large sink for anthropogenic carbon. Comparison of carbon stocks in Oxisols sampled near Manaus, Brazil, shows no measurable change in organic carbon stocks over the past 20 years. Simple models estimating the response of soil carbon pools to a sustained 0.5% yr À1 increase in productivity result in C storage rates of 0.09 to 0.13 MgC ha À1 yr À1 in soil organic matter, with additional potential storage of 0.18 to 0.27 MgC ha À1 yr À1 in surface litter and roots. Most storage occurs in organic matter pools with turnover times less than a decade. Export of carbon in dissolved form from upland terra firme Oxisols likely accounts for <0.2 MgC ha À1 yr À1 , but more work is required to assess the export potential for periodically inundated Spodosols.

In the Brazilian Amazon, selective logging is second only to forest conversion in its extent. Conversion to pasture or agriculture tends to reduce soil nutrients and site productivity over time unless fertilizers are added. Logging removes nutrients in bole wood, enough that repeated logging could deplete essential nutrients over time. After a single logging event, nutrient losses are likely to be too small to observe in the large soil nutrient pools, but disturbances associated with logging also alter soil properties. Selective logging, particularly reduced-impact logging, results in consistent patterns of disturbance that may be associated with particular changes in soil properties. Soil bulk density, pH, carbon (C), nitrogen (N), phosphorus (P), calcium (Ca), magnesium (Mg), potassium (K), iron (Fe), aluminum (Al), δ13C, δ15N, and P fractionations were measured on the soils of four different types of logging-related disturbances: roads, decks, skids, and treefall gaps. Litter biomass and percent bare ground were also determined in these areas. To evaluate the importance of fresh foliage inputs from downed tree crowns in treefall gaps, foliar nutrients for mature forest trees were also determined and compared to that of fresh litterfall. The immediate impacts of logging on soil properties and how these might link to the longer-term estimated nutrient losses and the observed changes in soils were studied. In the most disturbed areas, roads and decks, the authors found litter biomass removed and reduced soil C, N, P, particularly organic P, and δ13C. Soils were compacted and often experienced reducing conditions in the deck areas, resulting in higher pH, Ca, and Mg. No increases in soil nutrients were observed in the treefall gaps despite the flush of nutrient-rich fresh foliage in the tree crown that is left behind after the bole wood is removed. Observed nutrient losses are most likely caused by displacement of the litter layer. Increases in soil pH, Ca, and Mg occur in areas with reducing conditions (decks and roads) and may result from Fe reduction, freeing exchange sites that can then retain these cations. Calculations suggest that nutrient inputs from crown foliage in treefall gaps are probably too small to detect against the background level of nutrients in the top soils. The logging disturbances with the greatest spatial extent, skids and gaps, have the smallest immediate effect on soil nutrients, while those with the smallest spatial extent, roads and decks, have the largest impact. The changes observed 3–6 months after logging were similar to those measured 16 yr after logging, suggesting some interesting linkages between the mechanisms causing the immediate change and those maintaining these changes over time. The direct impacts on soil properties appear less important than the loss of nutrients in bole wood in determining the sustainability of selective logging. Medium-to-low intensity selective logging with a sufficiently long cutting cycle may be sustainable in these forests.

Figure 2. Radiocarbon in organic matter, reported as D 14 C (%) from bulk soil after it has been sieved to <2 mm and hand-picked to remove root detritus and charcoal.

No abstract