To clarify the effects of long‐term warming on ecosystem matter cycling, we conducted an in situ 7‐year experimental warming (2009–2015) using infrared heaters in a cool temperate semi‐natural grassland in Japan. We measured plant aboveground biomass, soil total C and N, soil inorganic N (NH4+‐N and NO3−‐N), and soil microbial biomass for 7 years (2009–2015). We also measured heterotrophic respiration for 2 years (2013–2014) and assessed net N mineralization and nitrification in 2015. We found that warming immediately increased plant aboveground biomass, but this effect ceased in 2013. However, the soil microbial biomass was continuously depressed by warming. Soil inorganic N concentrations in warmed plots substantially increased in the later years of the experiment (2013–2015) and the potential net N mineralization rate was also higher than in the earlier years. In contrast, heterotrophic respiration decreased with warming in 2013–2014. Our observations indicate that long‐term warming has a contrasting effect on plants and soil microbes. In addition, the warming could have different effects on subterranean C and N cycling. To enhance the accuracy of estimation of future climate change, it is essential to continuously observe the warming effects on ecosystems and to focus on the change in subterranean C and N cycling.

The chamber method with plant clipping has been widely used for measuring soil respiration (SR) in grassland ecosystems. However, plant clipping may cause overestimation of SR by changing the environmental factors and injuring the plants. To solve these problems, we developed a new non‐destructive method using multiple‐microchambers (3 cm diameter, 8 cm height), which enables measurement of SR without plant clipping by installing chambers into gaps among the grasses. The new method was compared with the conventional method at various flow rates in vitro to assess the accuracy of SR measurement. The new method overestimated the SR rate; however, the ratio of overestimation to the conventional method was constant for each flow rate. These ratios fitted the logarithmic curve, indicating the potential for correction of the SR rate measured by the new method using the logarithmic equation. The corrected SR rate obtained by the new method was equal to the rate by the conventional method. This suggests that accurate measurement of SR in grassland ecosystems is possible using the multiple‐microchambers method. We then compared the non‐destructive method and the destructive method in situ on summer season and found that the destructive method overestimated SR rate in the grassland ecosystem by about 276% on average. There were two possible reasons for this overestimation; first, the clipping treatment may change environmental conditions such as soil temperature and soil water content, and second, it may directly increase plant respiration.