Nonlinear carbon cycling responses to precipitation variability in a semiarid grassland
Graphical abstract
Introduction
Arid and semiarid regions cover ca. 40% of global land area (Safriel and Adeel, 2005), 87% and 54% of which are occupied by grasslands, respectively (UN, 2011). Overall, annual mean precipitation will likely decrease but precipitation variability will likely intensify in many arid and semiarid regions (Stocker, 2014). This combination of changing precipitation mean and temporal variability has great and complex impacts on carbon (C) cycling processes in arid and semiarid grasslands (Gherardi and Sala, 2015b; Knapp et al., 2002). Thus, quantifying the magnitude of arid and semiarid grassland responses to changing precipitation is crucial for accurately evaluating future global C budgets.
Impacts of precipitation variability on terrestrial C cycling have received much less attention than effects of changes in precipitation amount. Nonetheless, in some cases, increased precipitation variability characterized by larger event size and longer dry intervals may have more implications on terrestrial C cycling than that of decreased precipitation amounts (Fay et al., 2000). However, the majority of such studies manipulated only two levels of precipitation variability (e.g. ambient control and prolonged drought treatment; Thomey et al., 2011; Wilcox et al., 2015) rather than multiple levels (see Heisler-White et al., 2008; Gherardi and Sala, 2015a). Thus, our understanding of relationships between C cycling and precipitation variability was largely limited, especially with respect to nonlinear responses to precipitation variation (Knapp et al., 2017; Y. Luo et al., 2017).
Ecosystem attributes are controlled by resources availability such as water and nutrients, which are also the vital regulating factors for ecological responses to climate change. According to Liebig's Law of the Minimum, the least available resource is the limiting factor at a particular time (von Liebig and Gregory, 1842). A growing body of evidence suggested that many ecological processes and functions are co-limited by multiple resources (Harpole et al., 2011). Furthermore, the primary resource limitation could shift from one toward another, or from single-resource limitation to multiple-resources co-limitation or conversely, for a specific ecosystem by climate change (Falkenberg et al., 2013; Ren et al., 2017). Determining causality between observable ecological responses (e.g. biomass production, CO2 fluxes) and the dominant resource drivers (e.g. water availability or nutrient content) under climate change is of great value for understanding and forecasting ecosystem dynamic and feedback in the future.
It is a common view that better soil water status would enhance C cycling, as water is the primary limited resource for arid or semiarid ecosystems. Changed precipitation variability would be expected to directly alter soil water availability and subsequently, ecosystem C cycling processes. Past two-level precipitation variability studies suggested that the direction and magnitude of alteration in water availability caused by high precipitation variability varied depending on ecosystem type (e.g. xeric vs. mesic system; Heisler-White et al., 2009; Knapp et al., 2008) and the annual precipitation amounts (e.g. dry vs. wet year; Thomey et al., 2011). Additionally, another potential explanation for the inconsistent results from the past two-level experiments may relate to nonlinear responses of soil moisture to precipitation variability. Soil water availability may reach its peak at moderate precipitation variability, because small precipitation events are just able to wet the soil surface and are quickly evaporated, while large precipitation events increase potential water losses from the ecosystem through runoff, deep soil water percolation or relative high evaporation (Huxman et al., 2004; Yin et al., 2018). But these potential mechanisms have not been well-examined due to a lack of experiments with multiple-level precipitation variability gradient. Furthermore, in contrast to the common view, altered water availability may not always result in expected C cycling responses under changed precipitation variability. For example, increased precipitation variability did not change seasonal average soil moisture, but largely promoted productivity in a shortgrass prairie; in contrast, increased precipitation variability increased soil moisture but did not affect the productivity in a tallgrass prairie (Wilcox et al., 2015). Taken together, above-mentioned studies suggested that predicting changes in C cycling, mediated by soil water availability, is not straightforward.
Nitrogen (N) availability, another important resource regulating ecosystem C cycling (Wieder et al., 2015), may also be profoundly impacted by precipitation variability in complex and counteracting ways. First, if altered precipitation variability increases (decreases) seasonal average soil moisture, we may expect increased (decreased) mineralization of organic N to plant-available inorganic N (Guntiñas et al., 2012; Knoepp and Swank, 2002; Larsen et al., 2011). Second, higher seasonal average soil moisture would increase soil N loss through N2O emission (Li et al., 2019b). Third, mineralization of microbial biomass killed by water stress may promote the accumulation of inorganic N during the prolonged drought intervals of an extreme precipitation regime (Liu et al., 2014). Finally, the subsequent, large precipitation pulse may stimulate abnormally high N2O emissions and soil inorganic N loss (Li et al., 2016). Collectively, changes in dry-wet cycling and average soil moisture may result in diverse and even opposite responses of N availability. Ultimately, the direct effects of precipitation variability on C cycling may even be overwhelmed by the impacts of altered N availability (Ren et al., 2017) on processes including photosynthesis (Reich et al., 2009), plant biomass production (Wang et al., 2019) and ecosystem CO2 flux (Zong et al., 2013). However, the responses of N availability and their effects on ecosystem C cycling are often overlooked in precipitation variability studies.
To quantify the relationships between C cycling and precipitation variability, here we experimentally imposed four levels of precipitation frequency and magnitude without changing the total growing season precipitation amount over a semiarid grassland from 2014 to 2016. Furthermore, to examine the above relationships at more precipitation alterations and a larger timescale, a process-oriented biogeochemistry model, DNDC, which is calibrated and validated at this site (Kang et al., 2011), was employed to model biomass carbon and CO2 fluxes under six precipitation variability simulation scenarios during growing seasons from 2000 to 2011. Specifically, we hypothesized that 1) seasonal mean soil water content (SWC) will have convex relationships with increasing variability of precipitation. 2) Accordingly, CO2 fluxes as well as seasonal biomass production will peak at moderate precipitation variability, controlled primarily by the water availability. 3) Changed precipitation variability induced various in N availability may exert measurable control on CO2 fluxes and biomass accumulation for conditions when water availability is not limiting or similar across treatments.
Section snippets
Study site
This study was conducted in a semiarid temperate steppe at the Inner Mongolia Grassland Ecosystem Research Station, within the Xilin River Basin (43°20′ N, 116°400′ E, 1200 m a.s.l.), Inner Mongolia, China. Long-term (1953–2017) mean annual temperature is 2.5 °C and mean annual precipitation is 281 mm with approximately 86% (~242 mm) occurring in the growing season (from May to September) (Fig. S1). The soil is classified as dark chestnut in Chinese soil classification and Calcic Chernozem in
Responses of soil properties to precipitation variation
The precipitation treatments significantly changed SWC, with higher seasonal mean SWC in P10 and P24 than in P6 and P16 treatments (Fig. S2a), resulting in a trinomial relationship between mean SWC and Pcv with two tipping points: 3 and 4.3 (Figs. 2a and S3a). Soil temperature was not affected by precipitation variation (Fig. S2b). SIN changed significantly with the highest values observed in the P16 treatment each year (Fig. S3c), showing a hump-shaped relationship with Pcv, peaking at 3.5 (
Nonlinear responses of SWC, SR and BGB to precipitation variability
Bucket models predicted that xeric systems may experience less seasonal water stress under extreme precipitation regimes with larger rainfall event magnitudes and lower frequencies, because larger precipitation events can more sufficiently recharge the water storage, increasing the period of time when SWC is above stress thresholds (Knapp et al., 2008). Past manipulative precipitation variability experiments suggested extreme precipitation regimes that led to greater soil water content in arid
Conclusion
A 3-year field experiment and a 12-year model simulation were conducted to explore the relationships between C cycling processes and precipitation variability in a semiarid grassland. Unexpectedly, these two methods showed completely different patterns. The manipulative experiment suggested the multiple nonlinear responses of ecosystem CO2 fluxes and belowground biomass to changing precipitation variability, which were regulated by both individual or combined soil water and N availability. In
CRediT authorship contribution statement
Linfeng Li: Conceptualization, Methodology, Formal analysis, Investigation, Data curation, Writing – original draft, Writing – review & editing, Visualization. Xiaoming Kang: Methodology, Formal analysis, Data curation, Writing – review & editing. Joel A. Biederman: Conceptualization, Methodology, Formal analysis, Writing – review & editing, Visualization. Weijin Wang: Conceptualization, Methodology, Formal analysis, Writing – review & editing, Visualization. Ruyan Qian: Methodology,
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
This project was funded by the CAS Strategic Priority Research Programme (A) (Grant No. XDA20050103, XDA19030202) and the International Cooperation and Exchange of National Natural Science Foundation of China (Grant No. 31761123001, 31761143018). J. Biederman's contributions were supported by the U.S. Department of Agriculture, Agricultural Research Service. USDA is an equal-opportunity employer. Great thanks for the help of the Inner Mongolia Grassland Ecosystem Research Station.
References (60)
- et al.
Effects of moisture and temperature on net soil nitrogen mineralization: a laboratory study
Eur. J. Soil Biol.
(2012) - et al.
Seasonal timing regulates extreme drought impacts on CO2 and H2O exchanges over semiarid steppes in Inner Mongolia
China Agric. Ecosyst. Environ.
(2018) - et al.
Soil organic matter and texture affect responses to dry/wet cycles: effects on carbon dioxide and nitrous oxide emissions
Soil Biol. Biochem.
(2013) - et al.
Responses of greenhouse gas fluxes to climate extremes in a semiarid grassland
Atmos. Environ.
(2016) - et al.
Contribution of root respiration to total soil respiration in a semi-arid grassland on the Loess Plateau, China
Sci. Total Environ.
(2018) - et al.
Repackaging precipitation into fewer, larger storms reduces ecosystem exchanges of CO2 and H2O in a semiarid steppe
Agric. For. Meteorol.
(2017) - et al.
Increased temperature and precipitation interact to affect root production, mortality, and turnover in a temperate steppe: implications for ecosystem C cycling
Glob. Chang. Biol.
(2010) - et al.
Stability of ecosystem CO2 flux in response to changes in precipitation in a semiarid grassland
Sustainability
(2019) - et al.
Reappraisal of drying and wetting effects on C and N mineralization and fluxes in soils
Glob. Chang. Biol.
(2009) - et al.
The response of sub-Mediterranean grasslands to rainfall variation is influenced by early season precipitation
Appl. Veg. Sci.
(2016)