The undefined N source might be overestimated by ¹⁵N tracer trials

Abstract

Based on data from previous 15N-tracer trials, Yan et al. (Global Change Biology, 26, 191–199; 2020) accounted the sources of cereal crop nitrogen (N) from fertilizer inputs (from both current and previous seasons) and possible uptake of nonfertilizer N inputs, leaving a significant part of N source undefined. We argue that the undefined N source does not reflect the net change in soil N stock and might be overestimated for three reasons: (a) higher yield in 15N tracer trials; (b) a low residue return rate is assumed; (c) underestimated contribution by fertilizer inputs from previous seasons due to the lack of long-term trials. Recently, Yan et al. (2020) reported that the major source of cereal crop nitrogen (N) is not the current-year mineral or organic fertilizer but something else not yet well understood (i.e., crop N derived from other sources—Ndfo in Figure 1). Based on data from previous 15N tracer trials, the authors demonstrated that N uptake from nonfertilizer inputs and previous years' fertilizer only contributed a “small” or “modest” part to the Ndfo, leaving a significant N source (approximately 53 kg N ha−1 year−1) not yet accounted for and speculated to be from soil N turnover. While we agree with the authors that current-year N fertilizer inputs only account for a fraction of crop N uptake and soil plays an essential role in providing N to meet crop N demand, we argue that the undefined N source in Yan et al. (2020) does not reflect the net change in soil N stock and might be overestimated. While being discussed as part of the undefined N source, soil N turnover is not clearly defined in Yan et al. (2020) and could be interpreted as gross N mineralization or net N mineralization, resulting in different implications for soil N stock. According to mineralization-immobilization turnover theory (Schimel & Bennett, 2004), the “turnover rate” of soil organic N (SON) is gross N mineralization (total N released from SON stock), while the “annual loss” of SON is likely a result of net N mineralization (gross N mineralization minus gross N immobilization). The latter is much lower than the former because microbial immobilization will recycle a large part of released N back to SON (Figure 1a). As Yan et al. (2020) attempted to compare the soil turnover rate with SON stock change (section 4.5) and illustrated SON turnover as a one-way arrow (figure 5 in Yan et al., 2020), we consider that the authors interpreted soil N turnover as net N mineralization and consequently indicated the net loss of soil N stock. However, a detailed account of N flows in the soil-cereal crop system presented in Yan et al. (2020) does not suggest a significant depletion of N stock (Figure 1a). Assuming the nonfertilizer N inputs follow the same in-season, uptake-to-loss ratio as fertilizer N, there will be approximately 83 kg N ha−1 year−1 (= 57 + 26) added to the soil, which can potentially compensate for the 78 kg N ha−1 year−1 (= 95 – 17) uptake in the Ndfo category (Figure 1b). Given the uncertainties around flow estimates and potential subsequent loss from previous N inputs (“sl” in Figure 1a), a net loss of soil N stock is still possible, but a continuous depletion rate at the level of 53 kg ha−1 year−1 for all study plots examined is unlikely. In addition, several recent studies have suggested stable or even increasing trends of soil N stock in China (Yan et al., 2014) and the world (Ladha et al., 2016). The estimates of defined N sources in Yan et al. (2020) are associated with large uncertainties and are likely underestimated due to the following considerations (Figure 1b): (a) The management in 15N tracer trials at the plot or micro-plot scales was intensive and might result in higher grain yield and lower 15N retention than that of local farmers (Cassman et al., 2002; Dobermann & Cassman, 2005). Consequently, results from plot-based trials may overestimate the Ndfo on regional or global scale. (b) The nonfertilizer N input might be underestimated because a low residue return rate was assumed in Yan et al. (2020). However, the return rate of crop residue is generally high in countries where those 15N tracer trials are conducted (e.g., United States, United Kingdom, Canada, and Australia). (c) The “subsequent uptake of previous N inputs” (“su” in Figure 1a) might be a gradual process, and its quantity could be potentially underestimated by 15N trials. According to the only long-term tracer study (Sebilo et al., 2013), the cumulative 15N recovery by crops could still maintain a stable increasing trend even 28 years after the 15N fertilization. This result indicates that the “su” measured by 15N trials can potentially be higher if the trial continues. In addition, site-specific conditions in many 15N trials may lead to underestimation of “su.” For example, in some trials with low fertilizer N rate (e.g., 0–60 kg ha−1 year−1), small plot size (e.g., <1 m2), or shallow soil sampling depth (e.g., <40 cm), the 15N retention and its subsequent uptake were usually underestimated. The work was financially supported by the National Key Research and Development Program of China (grant number 2018YFC0213305), the National Natural Science Foundation of China (grant number 41701309), the Open Research Project of Shouguang Facility Agriculture Center in Institute of Applied Ecology (grant number 2018SG-B-03), and the National Science Foundation (grant number CNS-1739823).

Topics & Concepts