Methane and nitrous oxide fluxes from cattle excrement on C3 pasture and C4 native rangeland of the shortgrass steppe

摘要

Grazers play a major role in nutrient cycling of grassland ecosystems through the removal of biomass and the deposition of excrement in the forms of liquid, urine and solid feces. We studied the effects of cattle excrement patches on methane (CH4) and nitrous oxide (N 2O) fluxes using semi-static chambers on cool-season (C3), Bozoisky-select pasture, and warm-season (C4-dominated) native rangeland on the shortgrass steppe. Trace gas measurements were conducted over a 2 year period from cattle urine (43 g N m-2) and feces (94 g N m-2) patches within replicated exclosures on each plant community. Cumulative N2 O emissions for the 2 year experimental period, on a per area basis, were 55% greater from feces relative to urine patches on native rangeland (1.81 and 1.17 kg N2O-N ha-1) and 25% greater on Bozoisky-select pasture (1.66 and 1.25 kg N2O-N ha-1). While the cumulative N2O emissions were similar within treatments across plant communities, the magnitude of seasonal fluxes were different. Emissions from the excrement treatments were greater on the Bozoisky-select pasture the summer following treatment application, while emissions were greater on the native rangeland the following fall and spring. The emission factors for urine and feces did not differ for urine and feces on native rangeland (0.13 and 0.13%) and Bozoisky-select pasture (0.14 and 0.11%), but these emission factors were substantially less than the IPCC Tier 1 default factor (2%) for manure deposited on pasture, indicating that N2O emissions from these plant communities are currently overestimated. These findings suggest that the IPCC Tier 1 Default N2O emission factor of 2% for manure deposited on pasture is not representative of N2O emissions from cattle excrement on shortgrass steppe. Nitrous oxide emissions from the control plots on native rangeland and Bozoisky-select pasture were similar, 0.61 and 0.65 kg ha-1, respectively. Methane uptake was significantly less from cattle excrement compared to control plots for both plant communities. Cumulative net CH4 uptake rates were 68% greater for urine compared to feces patches on native rangeland (-2.73 and -0.88 kg CH4-C ha-1) and 86% greater on Bozoisky-select pasture (-2.16 and -0.30 kg CH4-C ha-1). Methane uptake rates were also 14% less for the control plots on Bozoisky-select pasture (-3.15 kg CH 4 ha-1) compared to native rangeland (-3.60 kg CH 4 ha-1). Future research should focus on CH4 and N2O fluxes from pasture `hotspots', where nitrogen loading and soil compaction are commonly present.;We tested the capacity of the biogeochemical model DAYCENT to simulate N2O and CH4 fluxes from control plots and cattle excrement amended soils of the shortgrass steppe for both plant communities. Cumulative N2O emissions from the urine treatment were overestimated using the DAYCENT model by a factor of 4 for native rangeland and by a factor of 5 for the Bozoisky-select pasture. While the measured and modeled cumulative emissions agreed reasonably well for the feces, water, and blank plots, the model did not accurately simulate the magnitude of seasonal N2O emissions from these plots, overestimating emissions during periods of high fluxes during the growing season and underestimating during periods of low fluxes such as the winter. The cause for the poor agreement between measured and modeled N2O emissions may be attributed to an overestimation of total system N, an overestimation of the proportion of nitrified-N emitted as N2O, and the possibility that a substantial amount (> 20%) of the urine-N was rapidly volatilized as NH3 due to the extremely dry conditions at the time of treatment application. Additional model validation for shortgrass steppe soils is needed using data sets that include extensive soil N data to accompany the trace gas data to determine if the model is accurately simulating nitrification rates, the proportion of nitrified-N emitted as N 2O, and the proportion of N immobilized in microbial biomass. The model strongly overestimated CH4 uptake rates for the control plots by a factor of 3 for native rangeland and 2 for Bozoisky-select, while the excrement plots were overestimated by a factor of 2 for both plant communities. The model underestimated the optimum water content for maximum CH4 uptake by approximately 5%, which led to an overestimation of CH4 uptake by a factor of 2 to 4 during periods of biological limitation when soils were extremely dry. The agriculture reduction factor, which accounts for fertilization and cultivation events, reduced CH4 uptake from the urine and feces plots, but the uptake rates were still overestimated by a factor of 2 since the modeled failed to capture reduced uptake rates under low soil water content (< 0.15 volumetric water content). The overestimation of CH4 uptake may partly be resolved by increasing the optimum water content at which maximum CH4 uptake occurs, allowing the model to capture biological limitation on CH4 uptake.