Repackaging precipitation into fewer, larger storms reduces ecosystem exchanges of CO2 and H2O in a semiarid steppe

- Reduced precipitation frequency led non-significant monotonic decline in ANPP.
- Reduced precipitation frequency negatively affected CO2 uptake and H2O exchange.
- Lower CO2 fluxes were influenced by reduced abiotic and biotic factors.

General circulation models predict that precipitation will become more extreme, i.e. rainfall events of larger size but reduced frequency. Studies in North American grasslands have shown that such repackaging of precipitation into fewer, larger events enhanced above ground net primary productivity (ANPP), likely due to deeper soil moisture infiltration favoring plant water use over evaporation. However, ANPP responses in other regions remain poorly understood, and responses of carbon and water exchanges with the atmosphere remain unknown. Here we manipulated rainfall in a steppe ecosystem of northern China over 4 years to investigate how temporal packaging of precipitation impacts ANPP, evapotranspiration (ET), net ecosystem CO2 exchange (NEE) and the component fluxes gross primary productivity (GPP) and ecosystem respiration (RE). Experimental plots received precipitation equivalent to the 60-year growing-season average of 240 mm, variously packaged into 6, 10, 16, or 24 events representing extreme (P6) to historical average (P24) rainfall frequency. Extraordinarily extreme frequency (6 large events) reduced NEE, GPP, RE, ET and water use efficiency (WUE = |NEE|/ET). The average NEE, GPP and RE declined 35%, 45% and 48% respectively in the P6 treatment as compared to P16, which showed maximum ET and CO2 exchange. After peaking in the 16-event treatment, GPP and WUE in P24 were not distinguishable from P6. These peaks suggest that P16 was optimal for photosynthesis, with sufficiently frequent rain to maintain unregulated plants and adequately deep soil moisture infiltration to favour transpiration, with associated carbon uptake, over evaporation. Path analysis indicated the lower CO2 fluxes were influenced by reduced soil water content and leaf area index and higher soil temperature, with ET regulating the effects of these microclimatic drivers. ANPP showed a monotonic but non-significant decline with decreasing precipitation frequency, consistent with reduced CO2 fluxes. We found an increase in ANPP of xerophyte plants partially compensated for the ANPP decline in the dominant eurytopic xerophyte plants. Our results suggest that extreme temporal repackaging of precipitation into few events with correspondingly long dry intervals may reduce the capacity of steppe ecosystems to assimilate atmospheric CO2, although community diversity may moderate impacts.