Understanding soil nitrous oxide (N2O) emissions responses from terrestrial ecosystems to future CO2 enrichment and warming is critical for the development of mitigation and adaptation policies. The continuous increase of elevated CO2 (EC) and elevated temperature (ET) effects on N2O emissions are not fully known. We synthesized 209 measurements from 70 published studies and carried out a meta-analysis to examine individual and interactive effects of EC and ET on N2O emissions from grasslands, croplands, and forests. On average, a significant increase of 23% in N2O emission was observed under EC across all case studies. A sharp increase in the N2O emissions was noticed in the response to EC under less than 150 ppm enrichment levels. Nonetheless, a significant but weak negative response of N2O emission to EC was examined. Compared with non-significant responses of grassland and forest, EC resulted in the highest increase (38%) in N2O emission of cropland. The extent of ET effect on N2O emissions was nonsignificant and there was no significant difference in N2O emissions responses among three terrestrial systems. ET only promoted the N2O emissions in the forest about 32% when ET was less than 2°C. The interactive effect of EC and ET on N2O emissions was significantly synergistic, ergo, showed a greater increase than the sum of effects caused by EC and ET alone. Our findings suggest that future studies on soil N2O need more long-term experiments with multifactor combination to better predict ecosystem responses.

We made an extensive search of articles published in peer-reviewed journals before October 2019. All published data were extracted from the “Web of Science” and “Google Scholar”. Every combination of searching keywords including “CO2”, “temperature”,“warming”, “N2O emissions” or “N2O fluxes”, “nitrous oxide emissions” was used for data extraction. Any further synthesis cited in papers of review, meta-analysis or synthesis were also considered. The following selection criteria were used to avoid bias in selecting publications: (1) experiments must be conducted in growth chambers, open-top chambers, free-air CO2 enrichment facilities, and greenhouses. Measurements from the soil incubation performed in laboratory studies or modeling results were excluded (2) data of soil N2O flux must be present in experimental results. Because effect size (natural log of the response ratio, Equation 1) could not be calculated for those observations with zero or negative values under either control or treatments, we excluded the observation if means soil N2O emission were less than zero; (3) ambient and elevated values of CO2 or temperature must have been provided; (4) means and replication number had to be reported for the treatments of ambient CO2, EC, ambient temperature and ET. A total of 209 observations from 1990 to 2019 on N2O response to EC and ET from 70 peer-reviewed articles were collected, including 86 effect sizes from 36 ET experiments, 106 effect sizes from 43 EC experiments, and 17 effect sizes from 8 combinations of co-elevated temperature and CO2 experiments. The dataset also included site location, experimental duration, N fertilization and land use types (e.g., grassland, forest and cropland). Observations from different years within the same study were considered and its average was considered as one observation averagely. Units were converted to μg N m-2 day-1 for all observations. The CO2 concentration increase in the EC treatment ranged from 35 to 400 ppm (254 ppm on average). The temperature increase in the ET treatment ranged from 0.39°C to 4°C.

Based on observations, land use types were categorized into: (1) grassland; (2) forest; and (3) cropland. Croplands were mainly planted wheat, maize and other food crops. The annual nitrogen application rate of croplands was ranged from 10 to 40 g m-2. Tillage mainly included traditional tillage and conservation tillage. Grasslands cover plants included a variety of herbs and shrubs. Most grasslands use nitrogen fertilization less than 10 g m-2 per year except for a few experiments with different levels of nitrogen application. And forests were mainly planted with perennial trees. Compared with croplands, grasslands and forests do not have much human cultivation activities. To address whether the response of soil N2O emissions to EC and ET might be affected by the change in the magnitude of the experimental factors, we categorized the EC into three groups: (0, 150), (150, 300) and 132(300, 450). Whereas, ET was categorized into two groups: ≤2°C and ＞2°C.

See attached README. Methods contain a more detailed description of the data.