Worldwide concerns with global change and its effects on our future environment require an improved understanding of the impact of land cover changes on the global C cycle. Overgrazing causes a reduction in plant cover with accepted consequences on soil infiltration and soil erosion, yet the impact on the loss of soil organic carbon (SOC) and its associated processes remain unaccounted for. In this study performed in South Africa, our main objective was to evaluate the impact of plant cover reduction on (i) SOC erosion by water in both particulate (POC) and dissolved (DOC) forms, and (ii) soil CO2 emissions to the atmosphere. The study performed under sandy-loam Acrisols investigated three proportions of soil surface coverage by plants (Cov), from 100% (Cov100) for the "non-degraded" treatment to 25-50% (Cov50) and 0-5% (Cov5). POC and DOC losses were evaluated using an artificial rainfall of 30 mm h(-1) applied for a period of 30 min on bounded 1 x 1 m(2) microplots (n = 3 per treatment). CO2 emissions from undisturbed soil samples (n = 9) were evaluated continuously at the laboratory over a 6-month period. At the "non-degraded" treatment of Cov100, plant-C inputs to the soil profile were 1950 +/- 180 gC m(-2) y(-1) and SOC stocks in the 0-0.02 m layer were 300.6 +/- 16.2 gC m(-2). While soil-C inputs by plants significantly (P<0.05 level) decreased by 38.5 +/- 3.5% at Cov50 and by 75.4 +/- 6.9% at Cov5, SOC, the losses by water erosion of 0.75 gC m(-2) at Cov100 increased from 66% at Cov50 (i.e. 3.76 +/- 1.8 gC m(-2)) to a staggering 213% at Cov5 (i.e. 7.08 +/- 2.9 gC m(-2)). These losses were for the most part in particulate form (from 88.0% for Cov100 to 98.7% for Cov5). Plant cover reduction significantly decreased both the cumulative C-CO2 emissions (by 68% at Cov50 and 69% at Cov5) and the mineralization rate of the soil organic matter (from 0.039 gC-CO2 gC(-1) at Cov100 to 0.031 gC-CO2 gC(-1) at Cov5). These results are expected to increase our understanding of the impact of land degradation on the global C cycle. Further in-situ research studies, however, need to investigate whether or not grassland degradation induces net C-emissions to the atmosphere.
