Impact of land use change on greenhouse gases emissions and consequent climate change has received greater attention in recent years. Forest and soils are important reservoirs of terrestrial carbon (C) and can be a source or sink of atmospheric C depending upon its management. A study was conducted in Pokhare Khola, a mid hill watershed in Nepal, to determine the historic land use change dynamics and its relation to vegetation and soil organic carbon (SOC) pools; to estimate the existing C pools in soil and vegetation, to find aggregateand particle-size-associated SOC; to evaluate the quality of soil organic matter (SOM) and to simulate the changing SOC pool through time and to predict the future pools under changing climate scenarios. Mainly four types of forest [Managed dense Shorea forest (DS), degraded forest/grazing land (DF), Pine mixed forest (PS) and Schima-Castanopsis (SC) forest] and two types of cultivated lands [rain fed upland (Bari) and irrigated low land (Khet)] were identified in the watershed. Soil samples and fine root samples were collected from each land use types down to the bed rock in an incremental depth of 0-10, 10-20, 20-40 and 40-70 cm. Soil bulk density (BD), SOC and major nutrient contents in soil and plant roots were determined. Forest vegetations were measured categorizing them into Trees, Shrubs, and Grasses. Biomass of trees and shrubs were estimated using allometric equations after measuring their diameter and height, while grass biomass was estimated by harvesting method. Land use change and its effect on C pool were determined by analyzing satellite image of 1976, 1989 and 2003 and relating it with C pools of corresponding year. Aggregate fractionation was done by wet sieving method, while mechanical fractionation was done ultrasonically and the fractions were analyzed to determine the associated SOC. Quality of SOM from surface (0-20 cm) and subsurface (20-40 cm) soil depth was determined by solid state 13C nuclear magnetic resonance (NMR) spectroscopy for DS, SC, Khet, anciently reclaimed (Bari 1) and recently reclaimed (Bari 2) upland. The bulk SOM of these soils was fractionated into three pools: easily labile (LP I), moderately labile (LP II) and recalcitrant pool (RP) by acid hydrolysis method. Simulation of SOC pools was done by using Century C model after its parameterization based on biophysical and weather data of the watershed. The time blocks were divided into 1950-1970, 1971-1990 and 1991-2004 based on land use change history in the study area. Prediction was done for the period of 2005-2050 under prevailing and changing climate scenario.

In the period of 1976-2003, there was a net increase in agricultural area (84%) and decrease of total forest area (24%) relative to the base year 1976. However, DS forest area was significantly increased by 174% but degraded forest was decreased by 35% relative to the base year, suggesting a remarkable contribution of the community forestry program. Conversion of forests to cultivated soils resulted into net increment of SOC pool (highest 2.7 kg C m-2), reflecting the effect of fertilizer and manure input but the same change resulted in the loss of vegetation carbon pools (highest 22.7 kg C m-2) from the system. The existing vegetation carbon pool was found to be highest in DS (219 ± 34 Mg ha-1) and lowest in the SC forest (36 ± 5 Mg ha-1). The total SOC pool in the watershed was estimated to be 59 815 Mg, of which 36, 32 and 32% were in 0-20, 20-40 and below 40 cm soil depths, respectively.

Forest soils showed higher proportion of macroaggregates than microaggregates but the opposite was true in cultivated soils. Soil aggregates in pine mixed forests were more stable than aggregates from other land uses. Among the cultivated soils, Bari soils showed more stable aggregates than the Khet soils. Macroaggregates in the surface layer of cultivated soils had higher SOC concentration than those in forest soils. Microaggregates of cultivated soils contained 77-80%, while macroaggregates of forest soils contained (~60%) of total SOC. Pine mixed forest showed the highest sand fraction and SC forest the highest clay fraction. Clay associated SOC was highest (44 ± 4 g C kg-1) in SC forest. Bulk SOC concentration was correlated better with silt plus clay contents than clay contents only. A negative relationship between clay content and clay-associated carbon showed the dilution effect.

NMR analysis showed a prevalence of O-alkyl C over other C forms in all land use types, and only small variations in alkyl C and aromatic C for the different soils and horizons of the same soil. Total area of chemical shift between 0-200 ppm shared by O-alkyl C was highest (51%) in DS and Bari 1 while SC forest shared only 44%. But, in the lower depth SC shared the highest area (49%) compared to other land uses, showing similarity of SOM quality in both depths. The DS soil showed marked differences with higher content of alkyl C in the subsurface indicating advanced stage of humification at depth. Acid hydrolysis showed
that bulk SOM is dominated by aromatic carbon. The Bari 2 soil contained only 50% of labile pool than SC forest indicating the effect of plowing. The Khet soil contained more labile C than DS forest showing the effect of water logging. There was a higher C/N ratio in the surface layers of the forest soils compared to the cultivated soils indicating dominance of plant derived SOM in it. The effect of land use change on SOM quality was distinct when pristine forest was reclaimed. The same conversion led to prevalent loss of hydrolysable C, which is the most prone to biochemical degradation and effect of global warming.

The simulation result revealed that there was a loss of SOC pool in the first temporal block (1950-1970) from equilibrium value of 6.8 kg C m-2 to 3.9 kg C m-2 both in the cultivated and forest soils. The model predicts increases in SOC in response to better management of forest and intensified cultivation with the use of higher agricultural input. In the projection period, the SOC pool in the managed forest was increased from 3.70 to 4.28 kg C m-2 under prevailing climate scenario, while it reached to only 4.21 kg C m-2 under the climate change scenario. A similar effect was observed in the cultivated soil indicating that under climate change scenario SOC sequestration will be reduced.

This study revealed that SOC was affected quantitatively and qualitatively by land use changes. Significant loss of vegetation carbon and SOC pool occurred due to land use change. However the SOC pool was increased under better management of forest and cultivated soil. Less stable aggregates having higher associated carbon in cultivated soil suggests a need of judicious management of these soils. Increased area under managed dense forest showed higher potential of C sequestration in forest vegetation, a significant contribution of the community forestry program. Further comprehensive studies at longer temporal scales and differing spatial scales are needed to draw precise and representative conclusions at the regional level. Additional components of carbon dynamics for further studies can be C oxidation due to soil respiration and loss due to leaching and runoff.