[Insert Figure 4 here]
We then applied the empirical model to estimate the CH4sink strength of global grasslands (Fig. S8 ), under four
treatments: ambient (i.e. pre-industrial) levels of N and P
depositions, elevated N deposition (contemporary N deposition and
pre-industrial P deposition), elevated P deposition (contemporary P
deposition and pre-industrial N deposition), and concurrently elevated N
+ P depositions (contemporary N and P depositions) (Fig. 4 ).
Simulated global grassland CH4 sinks amounted to 4.43 ±
0.20 Tg C-CH4 y-1 for the ambient
scenario, 3.92 ± 0.16 Tg C-CH4 y-1 for
the elevated N scenario, 4.60 ± 0.22 Tg C-CH4y-1 for the elevated P scenario, and 4.18 ± 0.18 Tg
C-CH4 y-1, for the N + P scenario
(Fig. 4 ). Addition of N-only thus suppressed the global
grassland CH4 sink by ~0.50 Tg C
(~11.4%), while concurrent P deposition alleviated this
suppression by more than half (~5.8%).
Conceptual Model for P Alleviation of
N-suppressed CH4Sink
Based on our field experiments, the global meta-analysis, and the
empirically-derived insights of N and P impacts on CH4uptake, we here propose a conceptual framework that summarizes the
possible mechanisms underlying the interactive impacts of N and P
additions on CH4 uptake (Fig. 5 ). In this
conceptual model, MMO represents the whole group of enzymes responsible
for CH4 oxidation under aerobic conditions (Dunfield &
Knowles 1995) (Fig. 5a ). Under ambient conditions (e.g.
pre-industrial N and P deposition), grassland productivity is generally
limited by low soil N availability (Ladwig et al. 2012) and by
low soil P availability in more than half of the grasslands (Fayet al. 2015) and grassland plant species have therefore optimized
their N uptake and allocation processes during ecological succession
(Bai et al. 2004). Soil N does not leach under these low-N
conditions and available N is either assimilated by plants or
immobilized by soil microbes. Because soil mineral N (particularly
NH4+) is maintained at a low level,
competition with CH4 for the MMO enzyme is weak
(Fig. 5a ). In contrast, sustained or high N addition will push
the system out of N limitation and result in
NH4+ accumulation in the soil
(Fig.5b ). This in turn strengthens the competition with
CH4 for the MMO enzyme, thereby suppressing the
oxidation of atmospheric CH4 in grassland soils
(Fig. 5b ). If N and P are concurrently added, the added P
stimulates vegetation growth and uptake of mineral N, especially
NH4+, and thus alleviates the
N-induced suppression of CH4 oxidation. (Fig.
5c ). This theoretical framework emphasizes the substrate competition
theory when explaining the P alleviation of N-suppression on
CH4 uptake.