Long‐Term Stimulatory Warming Effect on Soil Heterotrophic Respiration in a Cool‐Temperate Broad‐Leaved Deciduous Forest in Northern Japan

To evaluate the long‐term response of soil organic carbon decomposition to global warming in Asian monsoon forests, we established a multichannel automated chamber system combined with infrared carbon‐filament heat lamps in a 70‐year‐old cool‐temperate broad‐leaved deciduous forest in northern Japan in September 2011. We designed control plots to measure total soil respiration (Rs), root exclusion (trenched) plots to measure heterotrophic respiration (Rh), and warmed trenched plots to measure Rh under warmed conditions (+2.5°C soil temperature at 5 cm depth, Rhw). Annual effluxes ranged from 12.02 to 14.15 tC ha−1 in Rs, 7.78 to 11.49 tC ha−1 in Rh, and 8.76 to 15.25 tC ha−1 in Rhw over five years (2012–2016). During the summer season (between July and September), the daily warming effect was negatively related to soil temperature. In comparison, the relationship between soil moisture and the daily warming effect varied in each year depending on soil moisture levels. The annual warming effect exhibited large interannual variation, ranging from 6.2 to 17.7% °C−1; however, the five year average (10.9% °C−1) was close to the estimated value (10.2% °C−1) based on the annual Q10 of Rh (2.66). Interannual variation was positively related to the number of rainy days (p = 0.013). Our results indicate that existing global terrestrial models underestimate the strength of the feedback of Rh to global warming in Asian monsoon forests, because most global terrestrial models used relatively low Q10 values (usually below 2.0). Key Points Five consecutive years of stimulatory warming effect on heterotrophic respiration (Rh) was confirmed in a cool‐temperate deciduous forest The observed mean annual warming effect (10.9% °C−1) was close to the value (10.2% °C−1) estimated from the annual Q10 (2.66) of Rh Feedback of SOC decomposition to global warming might be stronger than predicted by global terrestrial model simulations using Q10