Environmental Tradeoffs between Nutrient Recycling and Greenhouse Gases Emissions in an Integrated Aquaculture–Agriculture System

The unlimited nitrogen (N) availability that has characterized crop production in the last few decades is accompanied by environmental burdens, including the greenhouse gas (GHG) emissions associated with fertilizer production, post-application nitrate (NO3–) pollution of water bodies, and emissions of reactive gaseous N forms into the atmosphere. Here, we quantified the environmental tradeoffs of replacing mineral N fertilizer with NO3– and ammonium (NH4+) originating from effluent water of aquaculture in a cucumber (Cucumis sativus) cultivation system. While the yield, nitrogen use efficiency (NUE), and NO3– leaching were similar between the cucumbers fertilized and irrigated (fertigated) by aquaculture effluent water containing 100 mg of NO3–-N L–1 (AN), by aquaculture effluent water supplemented with NH4+ (AN+), or by tap water with NO3– and NH4+ added (FN+), there were significant differences in the nitrous oxide (N2O) emissions between the systems. The N2O emissions peaked after each irrigation event followed by an exponential decline. The cumulative N2O emissions were between 60 and 600 g N2O-N ha–1, smaller than predicted based on a fertilizer application rate of 600 kg N ha–1 and were in the order AN+ ≫ FN+ > AN.


Estimation of N 2 O production pathway
The N 2 O emissions were assigned to different pathways based on the environmental requirements of each pathway (i.e. oxygen, N and C available in the system; Table S1) and the similarity of root zone oxygen concentration (aeration status) and NO 3 concentration for all treatments for the whole season. The contribution of the irrigation water degassing to the total emissions was considered to be the total amount of N 2 O dissolved above the concentration of N 2 O-N at equilibrium with the atmospheric concentration (i.e. 0.27 µg L -1 ) per amount of irrigated water over time (  (Table S1). Anaerobic or microaerophilic conditions required for the denitrification potentially existed lower down in the root zone where the soil was saturated, as the bottom boundary condition was atmospheric. It was assumed that the same amount of denitrification taking place in the AN treatment could also take place in the FN+ treatment, although it could also be less due to the lower availability of C (see the Carbon addition experiments section, Table 1 in the main text and Table S1). S1 The remainder of the N 2 O emissions in the FN+ treatment were assumed to occur via autotrophic nitrification, potentially including incomplete hydroxylamine oxidation S2 and nitrifier denitrification. S3 The denitrification in the AN+ treatment was assumed to be the same as that of the AN treatment. As similar amounts of NH 4 + and NO 3 were present in the AN+ and FN+ treatments, autotrophic nitrifier pathways were assumed to account for the same amount of N2O emissions in these treatments. The N2O emissions, beyond what could be accounted for by degassing, denitrification, and the autotrophic nitrifier pathways in the AN+ treatment, were presumed to occur via the heterotrophic nitrification pathway (Table S1).

N2O fluxes over time and plant growth curve
Over the length of the growth season more than 450 individual measurements of soil N2O fluxes were performed. The fluxes exhibited very high temporal variability ( Figure S1), explained by postfertigation peaks with a rapid decline in emissions with time ( Figure 3 in the main text and Figure   S1).

S3
The growth curve for the different parts of the cucumber plant can be seen in Figure S2. Differences in the growth rate and yield between the treatments were not found ( Figure 1 in the main text and Figure S2). The identical growth rates and yields were expected as the same amount of water and fertilizers was applied to the crops (Table 1 and Figure 1 in main text and Table S1).

Estimation of N 2 O production pathway
Degassing contributed between 10 and 4% to the total N 2 O emitted from the AN and AN+ systems, respectively, and none for FN+ ( Figure S3). The anaerobic conditions required for denitrification were not prevalent in the upper soil layer, as the oxygen concentration there fluctuated between 15 and 20%, with an average value for all treatments of 18.5% (± 0.8%), but did exist in the lower water-saturated soil layer. As the TOC and the level of anaerobicity in the AN and AN+ treatments was similar (Table S1) but the N 2 O emissions of AN+ were much higher, the limit of denitrification in AN was assumed to be due to the lack of anaerobic conditions and not due to limitation in C.
According to Table S1, autotrophic nitrification was the only N2O emission pathway for the FN+ treatment. However, any C that becomes available in the root zone, from root exudates, for example, could lead to denitrification. A similarity in root zone aeration would limit the denitrification pathway to producing no more N2O than in the AN treatment, so that the rest of the N2O emissions from FN+ would be due to an autotrophic nitrification pathway. The assumption that this autotrophic nitrification pathway would account for the same amount in both the AN+ and FN+ treatments is based on the reasoning that the differences in available C do not affect this pathway. The remaining N2O emissions from the AN+ treatment are attributed to heterotrophic nitrification and the potential aerobic denitrification associated with it. According to this reasoning, heterotrophic nitrification accounts for more N2O emissions than does autotrophic nitrification, a phenomenon that has been demonstrated to be possible in pure culture studies. S4 The substrate for heterotrophic nitrification is thought to be mineral NH4 + , as the availability of NH4 + was significantly higher in the AN+ treatment than in the AN treatment, while the TOC levels were similar. Heterotrophic nitrification has often been associated with acidic conditions which limit autotrophic nitrification, S5-7 but acidic conditions are not obligatory for heterotrophic nitrifiers. S8,9 Heterotrophic nitrification was found in different environmental niches, S10 and using wide variety of C and N sources. S8,11-15 As anammox does not produce N2O it was not considered here.
Chemodenitrification was also not expected to play a role as the pH was around 6.5. S16   Table 2 (main text), and an estimation of the different pathways contributing to this production for each treatment according to Table S1. The hashed colors indicate ambiguity with regard to which pathway the N2O emission should be assigned.