%0 Journal Article %J Ecology %D 2004 %T Vegetation-mediated changes in microclimate reduce soil respiration as woodlands expand into grasslands %A Smith, D.L. %A Johnson, L.C. %X This study compared annual and growing-season in situ soil respiration and soil C cycling in paired juniper woodland and C4-dominated grassland sites in eastern Kansas (USA) to determine if, under similar edaphic and regional climate conditions, vegetation change alters soil CO2 dynamics. We found marked differences in soil respiration related to vegetation: Growing season mean woodland soil respiration rates (4.6 μmoles·m−2·s−1) averaged 38% less than paired grassland sites. Soil moisture did not explain the difference in soil respiration between woodlands and grasslands. Soil temperatures at the 10-cm depth were 5°C cooler in woodlands during the growing season and significantly different between woodlands and grasslands throughout the year, explaining most of the low soil respiration in woodlands. However, there were subtle intrinsic differences in the response of soil respiration to temperature between woodlands and grasslands: Woodland and grassland respiration response was significantly different at the P = 0.05 level, also indicated by a grassland Q10 of 2.4 compared to 2.2 in woodlands. We found no significant differences between woodlands and grasslands in long-term (82-week) laboratory incubations of potentially mineralizable soil C and short-term incubations for microbial biomass C. Similarly, root biomass did not differ between woodlands and grasslands and could not explain the lower in situ woodland soil respiration. Thus, vegetation-mediated reduction in soil temperature under the canopy explained much of the lower in situ woodland soil respiration. Lower soil respiration in the woodland resulted in an annual flux of 533.6 (±21.7) compared to 858.4 (±14.5) g C·m−2·yr−1 in grasslands, nearly 38% lower in the woodland (means ± 1 se). Assuming root respiration is 50% of soil respiration, we estimate that the turnover of woodland soil C stocks may be slowed by 15 years relative to grassland. This suggests that, if juniper expansion (now occurring across nearly 5 million hectares in the Great Plains) proceeds to canopy closure, annual soil C flux may be potentially reduced by as much as 19 × 106 Mg of C below C flux rates that occurred historically from tallgrass prairie soils. %B Ecology %V 85 %P 3348 -3361 %G eng %M KNZ00961 %R 10.1890/03-0576