CR 28:163-173 (2005)  -  doi:10.3354/cr028163

ABSTRACT: Changes in climate and land use affect soil respiration rates (SRR) significantly, but studies of these effects across entire landscapes are rare. We simulated responses of landscape mean SRR (LMSRR) to such changes from May to October over a 30 yr period in a managed, predominantly forested landscape in northern Wisconsin, USA, using: (1) 6 satellite-derived land-cover maps (1972, 1978, 1982, 1987, 1992, and 2001); (2) monthly air temperature data in the corresponding years of the cover maps; and (3) SRR models driven by soil temperature (T_s) at 5 cm depth. LMSRR seemed to increase linearly by 77% from 0.625 in May to 1.104 g CO₂ m^–2 h^–1 in July, and then decreased at an increasing rate to 0.411 g CO₂ m^–2 h^–1 in October. LMSRR was more sensitive to an increase of minimum temperature than that of mean or maximum temperature, suggesting that future climate change might impact SRR in high-latitude forests more than other biomes. LMSRR in September over the study period was similar to that of June but with 92% higher variation, while both landscape mean air temperature and precipitation in September had lower variation than in June. This indicates that the topsoil layer functions differently during soil warming and cooling phases. Changes in land cover composition from 1972 to 2001 increased LMSRR by 2.8 to 3.1% while 2°C differences in growing season mean air temperature increased the SRR by 6.7 to 7.0%. The combined effects of both variables on the SRR are more complex, varying from 3.8 to 10.0%.

KEY WORDS:Landscape-scale soil respiration · Temperature variation · Landscape composition · Empirical model · Carbon flux