To clarify seasonal and inter-annual variations in the contribution of heterotrophic respiration to soil respiration and influences of such changes on the carbon balance in a forest ecosystem, relationships between respiration rates, soil temperature and soil water content, and the contribution ratio of heterotrophic respiration, were estimated over almost 3 years. Heterotrophic respiration was separated from soil respiration by the trenching method; CO2 emissions from the soil surface were measured along with soil temperature and volumetric soil water content in a cool-temperate deciduous forest from November 2009 to September 2012. Also, the root bag method and two decay models were used to accurately measure CO2 emissions from decomposing dead roots. The soil and heterotrophic respiration responded differently to changes in soil temperature and soil water content. An increase in soil temperature caused increases in respiration, and the response of heterotrophic respiration was lower than that of soil respiration. Also, an increase in the soil water content caused a slight increase in soil respiration and a decrease in heterotrophic respiration. These responses of the respirations to environmental factors caused a decrease in the contribution ratio of heterotrophic respiration with an increase in soil temperature and soil water content. These results suggest that the contribution ratio could experience complex changes when both factors change simultaneously in field conditions. Annual contribution ratios of heterotrophic respiration were estimated at 62％ in 2010 and 2011, and the contribution ratio showed seasonal variation ranging from 60％ to 100％. The estimated annual heterotrophic respiration rate in 2010 without such seasonal variation ranged from 1.50 to 2.51 kg CO2 m -2 yr -1, and these rates were 96％ to 161％ of that with seasonal variation （1.56 kg CO2 m -2 yr -1 ） . Also, the annual contribution ratio in 2010 estimated without the effect of soil water content （using respiration and soil temperature curve） was 80％. The resulting annual heterotrophic respiration rate was 2.01 kg CO2 m -2 yr -1, and this rate was 128％ of that with the effects of soil water content （using respiration, soil temperature and soil water content curve） . In contrast, the effect of inter-annual variation in the contribution ratio on annual heterotrophic respiration rate was small. Therefore, it is important to take into account the seasonal variation in the contribution ratio and the effects of the soil water content on the contribution ratio for more accurate estimation of heterotrophic respiration and net ecosystem production.

The aim of this study is to clarify changes in the ecosystem carbon cycle in response to predicted global warming in various ecosystems including semi-natural grassland. To clarify responses of the whole ecosystem to warming in a semi-natural cool-temperate grassland, we conducted an in situ warming experiment and examined plant growth and CO2 flux responses. Five pairs （control and warmed plots） of Zoysia japonica plots were established. Warmed plots were warmed using infrared heaters from June to November 2009. Once a month, aboveground biomass （AGB） of Z. japonica was estimated using the point frame method. Net ecosystem production （NEP） and ecosystem respiration （Re） were determined from CO2 flux measured using the closed chamber method. Each month, relationships were obtained between photosynthetic photon flux density （PPFD） and NEP （PPFD-NEP curve） and soil temperature （ST） and Re （ST-Re curve） . Monthly cumulative NEP, Re, and gross primary production （GPP, sum of NEP and Re） were calculated using these relationships and continuously recorded PPFD and ST data. Although there were some mechanical problems when using infrared heaters, the soil temperature in the warmed plots where infrared heaters worked well was an average of 2.3℃ higher than the control plots. This suggests the heating method using infrared heaters is applicable to grassland ecosystems. AGB in the warmed plots tended to be higher （by a maximum of 70％） than in the control plots throughout the experimental period, suggesting that experimental warming affected the phenology and extended the growth period of Z. japonica. Initial slope and light compensation point of PPFD-NEP curve were significantly affected by the warming. Monthly cumulative GPP in the warmed plots tended to be higher （by a maximum of 32％） than in the control plots. This is partly explained by the increased biomass and changed photosynthetic characteristics in the warmed plots. Although not all parameters of the ST-Re curve were affected by warm-＊ 早稲田大学先進理工学研究科

Soil organic layer samples of two different forest types were observed using compact MRI to visualize internal structure and clarify physical properties of forest soil. Soil pores and organic materials were distinguished by differences in proton mobility and visualized with a spatial resolution of 234 µm. Soil pore ratios and water mobility were calculated by image processing, and their differences between the two forest soils were detected. Our results suggest that compact MRI has potential for non‐destructive analysis of soil physical properties and is expected to have significant applications in ecological studies.

To understand the role of forest ecosystems in the global carbon cycle, it is important to clarify the factors affecting the carbon balance of forest ecosystems. However, little is known about the direct effect of forest types, especially dominant species, on their different carbon dynamics. To clarify the effect of difference in forest types, an experiment was conducted in three forests, which were located in the same place and exposed to the same climate conditions. These forests were middle-aged (40-45 years) and dominated by Quercus serrata (Q forest), Larix kaempferi (L forest) and Pinus densiflora (P forest). Net primary production (NPP), heterotrophic respiration (HR) and net ecosystem production (NEP) were estimated in each forest, using a biometric method over one year. For NPP estimated from the annual growth of tree biomass (ΔB) and amount of litter (LF), P forest NPP (5.3 MgC•ha −1 •yr −1) was higher than Q and L forest NPP (4.6 and 3.2 MgC•ha −1 •yr −1). The difference was affected by a significant difference in ΔB (p = 0.032) and LF (p < 0.001) mainly because of leaf biomass. The HR in Q forest (4.1 MgC•ha −1 •yr −1) was higher than L and P forest (2.3 and 2.1 MgC•ha −1 •yr −1). This difference could result from the amount of litter (respiration substrate) and chemical properties of litter (lability of decomposition). The NEP, which was calculated from the difference between NPP and HR, varied widely among the forest types (0.5, 0.9 and 3.2 MgC•ha −1 •yr −1 in Q, L and P forests, respectively). The range of values among the forest types was comparable to those among age sequences and climate zones in previous studies. These results suggest that the difference in forest types (especially dominant species) can potential-How to cite this paper: Tomotsune, M.,