Relating soil carbon fractions to land use in sloping uplands in northern Thailand

Abstract In this study from Khun Samun Watershed in northern Thailand we investigate land use and soil carbon relationships in 99 sloping upland soils that pertain to soil degradation/conservation and carbon storage/loss. Approximately 35% of the variability in total soil carbon could be explained by the available quantitative and semi-quantitative information, primarily clay content and CEC, and to a lesser degree by factors concerning land management. When categorical information such as land-use class and location was included, the explanatory power increased to 65% for total soil carbon, whereas for particulate and soluble carbon fractions the model predicted 71–95% of the total variability. The water soluble carbon fractions were high in systems with input of high C-to-N ratio material, whereas the permanganate oxidizable fraction was somewhat increased in the predominantly anoxic wetland rice site, and in fruit tree plantations. Coarse particulate organic matter carbon was strongly affected by land use. A causal relationship between land use and soil carbon or carbon fractions cannot be firmly established since historical data from these soils are lacking, but the significance of the various carbon fractions is discussed. It is proposed that the communities have selected land for particular purposes based on suitability, i.e. perceived inherent quality, accessibility and possibly other factors. Thus, the best land is apt to be used for the cultivation of fruit trees, teak and annual crops, leaving the poorest land for secondary forest. This study did not indicate that the cultivation of land is rapidly degrading the soil quality, either by carbon content, carbon fractions, or the profile gradients of total carbon or particulate carbon through the topsoil (0–20 cm). Profile carbon gradients were significantly explained by land-use class, but not by the position in the landscape which indicates that erosion is not an overriding force. While the secondary forest had the strongest profile gradient of total carbon, it was not much higher than fallow and fruit tree plantations, and only significantly different from wetland rice, which had a very small gradient due to the frequent ploughing of the soil. The similarities in profile carbon gradients is ascribed to the actual land management, where planting of crops is commonly done by punching holes in the otherwise undisturbed topsoil, and subsequent weeding is done by manual slashing or with herbicides, thus effectively forming a zero-till practice.