The impact of land use on carbon sequestration and CO₂ emissions were investigated in 1980-2000's on gray forest soils (Dystric Podzoluvisol) in Pushchino, and on sod-podzolic soil (Humic Podzol) in Prioksko-Terrasny Biosphere Reserve, Moscow Region, Russia (54^o50'N, 37^o35'E). 18 years grassing increased substantially the Ñ_org and N_total accumulation in soil. Mean annual rates of C sequestration in soil organic matter (SOM) ranged from 60 to 90 g C m^-2yr^-1 and from 98 to 182 g C m^-2yr^-1 for 0-20 cm and 0-60 cm soil layers respectively (Larionova et al, 2003). N accumulation comprised 5-8 g N m^-2yr^-1 in top soil and 7-18 g N m^-2yr^-1 in the layer of 0-60 cm. Hay cutting led to increase of C and N accumulations in soil compare to uncut plots. Statistically significant extension of Ñ è N pools, microbial biomass and intensity of CO₂ emission in soil after moderate application of mineral fertilizers (N₆₀P₆₀K₆₀) was revealed only in cut treatments, while in uncut grasslands fertilization decreased the rate of SOM accumulation. Humification of plant residues corresponded to the rates of SOM accumulation: the lowest humification was observed in uncut fertilized, and the highest – in cut fertilized grassland. The measurement of C and N storages after 24 years grassing showed the decrease in SOM accumulation rates for the last 5 years (between years 18 and 24 after grassland establishment).

Mean annual C accumulation in sandy sod-podzolic soil under 53 year grassland was two times lower than in clay gray forest soil. The difference between the rates was mainly explained by the different clay content between soil types rather than by ecosystem age. C sequestration in the sod-podzolic soil was close to the world-wide mean value of 33 g C m^-2yr^-1 calculated by Post and Kwon (2000), while for the gray forest soil the higher value of C sequestration was revealed. The majority of data review by Post and Kwon (2000) were for SOM accumulation in top soils only. The high rates of SOM accumulation in the gray forest soil in our study were revealed due to accumulation in both the top and deeper soil horizons, while the significant SOM increase in soddy-podzolic soil was found only in top soil. No significant loss or accumulation of SOM was found in the soils under mature forests or croplands for the last 15 years.

Decomposition of SOM was studied during long-term (1-2 years) incubations of soil samples at 22^oC and 70% WHC. Though application of mineral fertilizers as such led to SOM accumulation in grasslands, total CO₂ emissions from fertilized soils was higher than that from unfertilized control. Intensive soil respiration and high decrease of microbial carbon evidenced the enhanced degradability of the new formed humus in fertilized grasslands in comparison with unfertilized treatments.

An approximation of cumulative C losses by double exponential function reflecting the degradation of labile and recalcitrant SOM pools demonstrated higher decomposability of labile SOM and lower degradation of recalcitrant fraction after the addition of NPK fertilizers to the grasslands in comparison with unfertilized control.

The CO₂ emissions of 8.4 g C m^-2yr^-1 from fossil fuel use for producing and transporting of 60 kg ha^-1 mineral fertilizers significantly decrease positive effect of NPK application and result in equal C accumulation in top soil of fertilized and unfertilized cut grasslands. Hence, grassing combined with hay cutting is the optimal agricultural treatment that enables us to improve sufficiently SOM balance in ecosystems without excluding them from the agricultural use.

The constants of double exponential equation depended on clay content in the soil: higher decomposability of both SOM pools was found in sandy sod-podzolic soil in comparison with clay gray forest soil. The lowest values of constants were found for the cropland soils while the highest – for the soils under mature forest. The results of mathematical simulations of cumulative C losses by more sophisticated functions including an approach used in ROMUL model (Chertov et al., 2001) will be discussed.

This study was supported by Russian Foundation for Basic Researches and Presidium of Russian Academy of Sciences.