4.1.1 Nitrogen fertilization impacts on soil biodiversity

Bacterial diversity—Nitrogen fertilization can affect soil bacterial diversity directly by changing soil chemistry. When applied at high rates, ammonium and urea fertilizers can inhibit soil microorganisms due to ammonia toxicity and increases in ionic strength (Eno et al., 1955; Omar and Ismail, 1999). Inorganic N fertilizers are often supplied as NH₄⁺ which releases H⁺ ions upon oxidation and reduces soil pH (Magdoff et al., 1997). This typically reduces soil microbial diversity (Fierer and Jackson, 2006; Zhang et al., 2015). Despite a plethora of studies showing that high rates of synthetic N inputs negatively impact soil bacterial communities (Fierer et al., 2012; Frey et al., 2014; Treseder, 2008; Wallenstein et al., 2006), our meta-analysis showed that high N fertilization rates had no significant negative effect on bacterial biodiversity. In addition, low N fertilization rates increased bacterial diversity. While the significant reduction in diversity with higher relative to lower N inputs confirms that fertilizer inputs can negatively affect bacterial diversity (e.g. Fierer et al., 2012; Poulsen et al., 2013; Suleiman et al., 2016; Wood et al., 2015), our data underscore that responses are highly variable. This variability can often be explained by among-study differences in site conditions (Fierer et al., 2012), where fertilization may differentially affect other factors that mediate soil biodiversity such as soil synthetic properties, plant productivity, plant diversity, and soil organic matter content.

Our results suggest that fertilization-induced changes in soil organic matter (SOM) content are particularly important in mediating the response of bacterial diversity to N additions. Namely, we found that bacterial diversity increased (~ 6%) when N was applied as an organic fertilizer or as a combination of inorganic and organic N fertilizers. In addition, N fertilizer inputs increased microbial diversity when studies were conducted over a period longer than 5 years even when N was applied at high rates. In both cases, N fertilization likely led to SOM accumulation either directly through the application of organic materials or through fertilizer-induced increases in plant-derived C inputs to the soil (Belay-Tedla et al., 2009; Chen et al., 2018; Rasse et al., 2005; Zhang et al., 2017b). SOM increases resource availability to soil microbes (Hao and Chang, 2002; Mooleki et al., 2002), it improves soil physical properties like structure, aeration, drainage and water-holding capacity (Miller et al., 2002; Reynolds et al., 2003; Whalen and Chang, 2002), and buffers against fluctuations in pH. Our findings suggest that management techniques that enhance SOM input and retention may work to retain or promote soil bacterial biodiversity. Further research assessing the role of SOM amendments in buffering biodiversity losses from synthetic N inputs, and particularly an understanding of the amount of SOM needed to avert negative impacts will be useful in informing management decisions.

Fungal diversity—Fungal diversity consistently increased with N input. Positive impacts of N addition on fungal taxa other than ecto- and arbuscular mycorrhizal fungi have been observed across a variety of temperate ecosystems, including pine forests (Weber et al., 2013), mixed hardwoods (Morrison et al., 2016), and grasslands (Chen et al., 2018). It appears that this increase in diversity may be driven, in particular, by increased diversity and richness of specific functional and/or taxonomic groups, in particular, ascomycetes, saprotrophs, and yeasts (Morrison et al., 2016; Weber et al., 2013). The mechanisms that lead to increased saprophytic diversity with N additions are uncertain. However, it appears that similar to the responses of the bacterial community, the fungal community may have increased number of species with a more copiotrophic lifestyle, such as yeasts, which concomitantly have a greater genetic capability for inorganic N uptake than other groups of fungi (Treseder and Lennon, 2015). Further, P availability may mediate the response of the fungal community to N addition (Lauber et al., 2008). Enhanced fungal diversity may be caused, at least in part, by N fertilizer-induced reduction in pH. This change can enhance weathering processes that release P and increase phosphatase activity (Chen et al., 2018; Marklein and Houlton, 2012; Vitousek et al., 2010). However, mechanistic links between soil fertility and diversity of saprophytic fungi remain to be explored more.

Another possible mechanism for the positive impacts of N fertilizer on fungal diversity is indirect change in plant community composition. Fifty percent of the fungal diversity studies in our analysis were conducted in rangelands, where N inputs typically reduce plant community diversity by promoting fast-growing, nutrient-acquisitive species (Chen et al., 2018; Clark and Tilman, 2008; Roth et al., 2013). Some studies have linked loss of plant biodiversity to loss of soil biodiversity (Fanin et al., 2018; Kowalchuk et al., 2002; Wagg et al., 2011), thus suggesting that N inputs to uncultivated ecosystems may lead to soil biodiversity losses. However, losses in plant biodiversity from N additions appear to not predictably reduce soil biodiversity (Chen et al., 2018; Fierer and Jackson, 2006). This suggests that the mechanisms operating on aboveground diversity do not similarly translate to the belowground system (Chen et al., 2018; Tedersoo et al., 2014). Additional work is required to link above- and belowground ecosystem structure and changes therein following environmental perturbations (Kardol and De Long, 2018).

AMF diversity—Application of N fertilizer at quantities that exceeded 150 kg N ha^− 1 year^− 1, or fertilization exceeding 5 years significantly reduced AMF diversity (− 20% and − 10%). This may be caused by a reduction in AMF species abundance following N fertilization (Egerton-Warburton and Allen, 2000; Johnson et al., 1991), owing to changes in soil chemistry and pH (Dumbrell et al., 2010; Liu et al., 2012; Qin et al., 2015) or decreased plant investment in resources for AMF under high N conditions (Treseder, 2004). Alternatively, N inputs may promote proliferation of AMF taxa that outcompete taxa that do not thrive under N-rich conditions, hence reducing diversity (Johnson, 2010; Johnson et al., 2015; Liu et al., 2015). The mechanism by which N fertilization reduces AMF diversity appears to be site specific. For example, diversity in C4 grasslands was decreased due to proliferation of competitive glomeromycetes species, while in C3 grasslands there was a general loss of species richness (Egerton-Warburton et al., 2007). There was no effect of N fertilizer application on AMF diversity when N was applied at low quantities or in short-term studies, suggesting that the magnitude and direction of AMF community diversity responses is governed by a threshold of N application rate or cumulative N amount. This threshold may be determined by stoichiometric relations of C:N and N:P in the systems. Low C availability in combination with high N availability can significantly change competitive relations in AMF communities, leading to the proliferation of copiotrophic species and reducing evenness (Leff et al., 2015; Verbruggen and Kiers, 2010). Moreover, in low P soils N fertilization tends to exacerbate host plant dependence on AMF for P uptake, thereby promoting AMF species diversity (Egerton-Warburton et al., 2007). In conclusion, our results indicate that N fertilizer applied at large quantities and for a prolonged duration is likely to reduce AMF community diversity, and we speculate that these effects will be particularly pronounced in systems with low soil C and high P concentrations. This makes agroecosystems, especially those that are conventionally managed with repeated N additions at large quantities, particularly susceptible to loss of AMF diversity.

Soil fauna diversity—Soil faunal diversity was negatively affected by synthetic N inputs, but only if they were supplied at rates not exceeding 150 kg N ha^− 1 year^− 1. In addition, negative effects of N fertilization on soil fauna diversity were found for studies with a duration of less than 5 years (but not for longer-term studies). Soil nematodes are often negatively affected by N inputs (Wei et al., 2012). However, responses to N addition typically vary among nematode feeding groups (Liang et al., 2009; Sarathchandra et al., 2001) with fungal-feeders linearly decreasing in response to N addition (Hu et al., 2010; Liang et al., 2009) and bacterial-feeders being stimulated at low N doses, i.e., showing humped relationships (Wei et al., 2012). In addition, diversity responses of plant-feeding nematodes often vary with time after fertilizer application (Liang et al., 2009), and are strongly dependent on the crop species or shifts in plant species composition upon fertilization. Other soil fauna, such as collembolan, may not be particularly sensitive to N fertilization (e.g., Coulibaly et al., 2017). This variability in responses among and within groups of soil fauna may explain why only low synthetic N inputs and short-term studies negatively affected biodiversity. Namely, the studies included in this analysis comprised a highly diverse group of organisms (i.e., nematodes, collembola, and mites), and stochasticity of organisms included in the categories may strongly affect results. Alternatively, the disappearance of negative effects of N at higher N fertilization rates and in longer-term studies may be related to N-induced increases in plant C inputs. Increased plant C inputs tend to promote soil fauna diversity by increasing food availability (i.e., SOM, microbial communities or other soil fauna; Wang et al., 2016). Taken together, our results suggest that soil fauna tends to be negatively impacted by N fertilization, but these negative responses may be counteracted by positive responses to SOM accumulation. More studies for individual taxonomic and functional groups of soil fauna are needed to clarify the patterns and mechanisms of responses.