Rainfall pulse response of carbon fluxes in a temperate grass ecosystem in the semiarid Loess Plateau.

Yakun Tang,Chen Chen,Jun Jiang,Yunming Chen,Xu Wu

doi:10.1002/ece3.4587

Abstract

Rainfall pulses can significantly influence carbon cycling in water‐limited ecosystems. The magnitude of carbon flux component responses to precipitation may vary depending on precipitation amount and antecedent soil moisture, associated with nonlinear responses of plants and soil microbes. The present study was carried out in a temperate grass ecosystem during 2013–2015 in the semiarid Loess Plateau of China, to examine the response of carbon fluxes to precipitation using the “threshold‐delay” model. The unique contribution of environmental variables such as precipitation amount and antecedent soil moisture before rainfall (SWC_antecedent) to carbon fluxes in response to rainfall was also investigated. The lower threshold of effective rainfall was 6.6 mm for gross ecosystem production (GEP), 8.5 mm for net ecosystem production (NEP), and 4.5 mm for ecosystem respiration (RE); and the upper threshold of effective rainfall was 21.4 mm for GEP and NEP, and 16.8 mm for RE. Rainfall amount was positively affected the relative rainfall responses of GEP, NEP, and RE. However, SWC_antecedent at 20 cm soil depth offset the response of GEP to rainfall pulses, and SWC_antecedent at 5 cm soil depth offset the response of NEP and RE to rainfall pulses, with corresponding partial slopes of linear regressions of −0.50, −0.40, and −0.52. These results indicated that NEP was more sensitive to rainfall pulses and RE was more sensitive to SWC_antecedent. These results demonstrate the importance of rainfall events of <10 mm and that the negative effect of SWC_antecedent should also be considered when estimating ecosystem carbon fluxes in this semiarid region.

Highlights

Precipitation in arid and semiarid ecosystems is highly variable and discontinuous, with periods of low water availability separated by ep‐ isodic events (Zeppel, Macinnis‐Ng, Ford, & Eamus, 2008; Zhao & Liu, 2010)
Understanding the responses of carbon fluxes to rainfall pulses is essential for comprehending the car‐ bon cycles in the context of climate change, especially in water‐limited ecosystems (Peng et al, 2013; Tang et al, 2016)
Precipita‐ tion manipulation experiments suggested that the effect of small rainfall events (e.g.,

Summary

| INTRODUCTION

Precipitation in arid and semiarid ecosystems is highly variable and discontinuous, with periods of low water availability separated by ep‐ isodic events (Zeppel, Macinnis‐Ng, Ford, & Eamus, 2008; Zhao & Liu, 2010). The temporal variability of rainfall pulses, associated with the nonlinear responses of plants and soil microbes, constrains ecosystem carbon processes through the influence on water availability (Huxman et al, 2004; Tang et al, 2016). Different responses of car‐ bon fluxes to rainfall pulses can be attributed to differing meta‐ bolic responses of plant and soil microbes (Hao et al, 2010; Potts et al, 2006). The threshold‐delay model has been widely use in shrub (Jian, Wu, Hu, & Zhang, 2016; Zhao & Liu, 2010) and grass (Zeppel et al, 2008) ecosystems to identify the lower effective rainfall amount that affects vegetation. Hao et al (2010) indicated that high SWC_antecedent can offset the sensitivity of GEP to rainfall pulses in a semiarid grass ecosystem. The objectives were to (a) determine the rainfall threshold that significantly in‐ creases GEP, NEP, and RE, and (b) identify the unique contribution of rainfall amount and SWC_antecedent to these carbon fluxes in response to rainfall

| MATERIALS AND METHODS

| CONCLUSIONS

Findings

CONFLICT OF INTEREST