N cessation
We observed an evident convergence in plant N and soil inorganic N concentrations between controls and previous N addition plots, implying a rapid recovery in response to cessation of N inputs, a common observation in cessation of N addition studies (O’Sullivan et al. 2011; Shi et al. 2014; Storkey et al. 2015). Plant N uptake follows a positive relationship with soil mineral N concentration (Hajari et al. 2014). Declines of soil inorganic N pools resulted in the recovery in aboveground biomass and tissue N following cessation of N inputs. These findings suggest that plant and soil inorganic pools are not long-term sinks for added N (Pilkington et al. 2005). Microbial biomass N was significantly higher than untreated controls after 7 years of N cessation. Although we did not measure microbial biomass N during the fertilized period, research in other grasslands indicated that N addition generally decreased microbial biomass and heterotrophic respiration (Liu & Greaver, 2010; Sillen & Dieleman, 2012; Riggs et al. 2015). Moreover, the N input adverse effects were more pronounced in grasslands in long-term trials (>5yr) (Geisseler et al. 2016). Our findings imply that stopping N input for many years would eliminate nitrogen-induced soil microbial biomass inhibition.
Although N addition effects on plant and soil inorganic N pools may be relatively transient, AOB abundance and nitrification rate remain high even though a reduction in inorganic N pools occurs following nutrient inputs cessation. Few studies examined the legacy effect of N addition on on nitrifiers in field soils. Most long-term investigations report the lack of recovery in soil mineralization and nitrification, such as 20 years in a short grass steppe community (Vinton & Burke, 1995), 14 years in a Swedish pine forest (Chen & Hogberg, 2006), 12 years in a Minnesota prairie-like acid grassland (Clark et al. 2009). This suggests that the legacy effect of N addition on internal cycling may persist many years after the cessation of N inputs. Changes in soil microbes are possibly closely related to the functional process. Our SEM models (Fig. 5b) implied that soil microbial biomass N predominantly controlled mineralization variability and soil nitrogen availability for ammonia oxidizers. Other studies have confirmed the positive relationship between soil nitrification rate and microbial biomass nitrogen (Baldos et al. 2015; Wang et al. 2018; Li et al. 2019). The potential mechanism is that higher soil microbial biomass stimulates soil AOB-induced nitrification through increased nitrogen mineralization. A large portion of the mineralizable nitrogen (55 – 89%) is derived from microbial biomass nitrogen (Bonde et al. 1988). High soil microbial biomass stimulates soil nitrogen mineralization (Li et al. 2019) and mineralization ammonium is the substrate of soil autotrophic nitrifiers. Therefore, soils with additional substrates enhanced soil ammonia oxidizers, which could expedite nitrification.
Legacy effects of the microbial community due to N addition have been recorded previously (Johnson, 1993; Bradley et al. 2006). Previous studies in the same sites have shown that N addition acts as a deterministic filtering factor favoring AOB but against AOA (Xiao et al. 2014). Even after 7 years of N cessation, high AOB suggests that long-term impacts of nutrient addition on the AOB community may be detrimental to nitrification, further triggering a series of co-occurrence network changes in other groups. AOA was more likely to co-occur with comammox Nitrospira than AOB in control plots, indicating that AOA might share more similar niches or have more biotic interactions with comammox Nitrospira . Ammonia oxidation kinetics revealed that AOA and comammox Nitrospira possessed transporters with a higher affinity for ammonia than AOB (Kits et al. 2017; Koch et al. 2019), which facilitated their prevalence in oligotrophic environments. The AOB community dynamic responses enhance the robustness of microbial co-occurrence networks, and the linked ecosystem functions could be less vulnerable for the future ecosystem. Many of these species, such as competitive AOB, may need elevated N to become established within a typically low-nutrient community. This creates a situation where the less competitive species cannot compete sufficiently to re-establish themselves or become dominant again.