BOWEN RATIO/ENERGY BALANCE AND SCALED LEAF MEASUREMENTS OF CO2FLUX OVER BURNEDPROSOPISSAVANNA

Abstract

Prescribed fire is used as a management practice to maintain grassland dominance and reduce woody plant encroachment on grasslands and rangelands. Little is known regarding effects of these fires on CO2 fluxes and their potential contribution to atmospheric CO2. The objectives of this study were (1) to quantify the effect of fire on net ecosystem CO2 flux above Prosopis glandulosa Torr. (honey mesquite) mixed-grass savannas using the Bowen ratio/energy balance (BREB) method, and (2) to compare these fluxes to fluxes determined by an empirical model that scaled measurements of leaf photosynthesis and leaf area of dominant species and soil CO2 fluxes to the ecosystem level. Measurements were made during a wet year (1995) and a dry year (1996) with different savanna areas burned in late winter (February–March) each year. In the wet year, BREB-estimated daily (24-h) CO2 flux averaged −0.1, 14.6, and 0.8 g/m2 on the unburned plot, and 15.7, 19.7, and 15.2 g/m2 on the burned plot during spring, summer, and fall periods, respectively, during the first growing season following the fire (positive values mean net CO2 uptake). In the drought year, mean daily CO2 flux during spring, summer, and fall was −9.2, 4.1, and 4.8 g/m2 on the unburned plot, and 8.1, −0.9, and 4.6 g/m2 on the burned plot, respectively. The increased carbon uptake following a fire when compared to the unburned plot was estimated to offset the initial carbon loss from combustion within 28 growing-season days in the wet year, while it took 82 d in the dry year. Empirically determined daylight (12-h) CO2 fluxes, measured on two midsummer days each year, averaged 17.9 g/m2 when determined by BREB, and 12.0 g/m2 when calculated from scaled-up leaf-level measurements. By the end of the first growing season following the burn, it is likely that the amount of carbon emitted during the burn had been taken up by the regrowing vegetation, in both a wet and a drought year.