Microplastic fibers affect dynamics and intensity of CO2 and N2O fluxes from soil differently

Microplastics may affect soil ecosystem functioning in critical ways, with previously documented effects including changes in soil structure and water dynamics; this suggests that microbial populations and the processes they mediate could also be affected. Given the importance for global carbon and nitrogen cycle and greenhouse warming potential, we here experimentally examined potential effects of plastic microfiber additions on CO2 and N2O greenhouse gas fluxes. We carried out a fully factorial laboratory experiment with the factors presence of microplastic fibers (0.4% w/w) and addition of urea fertilizer (100 mg N kg−1). The conditions in an intensively N-fertilized arable soil were simulated by adding biogas digestate at the beginning of the incubation to all samples. We continuously monitored CO2 and N2O emissions from soil before and after urea application using a custom-built flow-through steady-state system, and we assessed soil properties, including soil structure. Microplastics affected soil properties, notably increasing soil aggregate water-stability and pneumatic conductivity, and caused changes in the dynamics and overall level of emission of both gases, but in opposite directions: overall fluxes of CO2 were increased by microplastic presence, whereas N2O emission were decreased, a pattern that was intensified following urea addition. This divergent response is explained by effects of microplastic on soil structure, with the increased air permeability likely improving O2 supply: this will have stimulated CO2 production, since mineralization benefits from better aeration. Increased O2 would at the same time have inhibited denitrification, a process contributing to N2O emissions, thus likely explaining the decrease in the latter. Our results clearly suggest that microplastic consequences for greenhouse gas emissions should become an integral part of future impact assessments, and that to understand such responses, soil structure should be assessed.


Introduction
As a result of human activities, the load of reactive nitrogen compounds (NH 3 /NH 4 , NO 3 -, NO x , N 2 O) on the earth has more than doubled in recent decades 1,2 . This was accompanied by a doubling of the intensity of the global nitrogen cycle. A main driver of this development are intensified agricultural practices entailing increased application of synthetic nitrogen fertilizers since the end of the Second World War [3][4][5] . While ensuring food security for an ever-growing world population 6 , agriculture has developed globally into a major source of climate-relevant trace gases. This applies in particular to nitrous oxide. Agriculture accounts for 60% of the total man-made nitrous oxide release 7 . The continuing increase in N fertilization or N surplus in production also appears to be the main reason for the unexpectedly strong acceleration in atmospheric N 2 0 concentration in recent times 8,9 . Consequences for the role of soils as sources and sinks of the important greenhouse gas CO 2 can also be expected, since the carbon and nitrogen budgets of soils are closely linked. This is especially true for the mineralization of soil organic matter as a source of CO 2 release from soils 10 . A prerequisite for the development of effective strategies to reduce land-use-related greenhouse gas emissions is comprehensive knowledge of the relevant processes and their regulation by internal and external drivers [11][12][13] . An important process that has been under-researched in microplastic-affected soils is the emission of greenhouse gases 14 . Despite its potential importance, compared to other factors of global change, we have so far only scratched the surface in terms of assessing microplastic impacts on soil properties and processes in general [15][16][17][18] .
Microplastic pollution is becoming increasingly recognized as a factor of global change, affecting not only aquatic but also terrestrial ecosystems and the soil 17,19 . Microplastics occur as primary microplastic or secondary microplastic and in a wide variety of sizes, shapes, chemistries and with a huge diversity of additives. Microplastic particles are expected to arrive in most ecosystems via aerial deposition 20,21 , but in agroecosystems there are also other input pathways including addition of sewage sludge or compost, which have been estimated to represent rather large input fluxes 16 . Once they have arrived in agroecosystems, there are a range of plausible pathways (including plowing) that lead to a transport of such particles into the soil 22 , where effects upon soil properties, processes and biodiversity can then unfold. Previous studies on the effects of microplastics have shown effects on soil organisms, especially microorganisms, and chemical conversion processes in soils [23][24][25] . Initial evidence also pointed to soil physical properties being altered by microplastics 26 . We have evidence that microplastic can affect basic parameters including soil structure and bulk density 15 , and that the performance of biota can be altered, which has been shown, for example for earthworms 27 , microbes 28 , and for plant growth [29][30][31][32] . However, the only study to date on the effect of microplastics on the emission of climate-relevant trace gases from soils does not address the impact of soil physical properties on the greenhouse gas emission 28 .
However, the fact that the intensity of N 2 O and CO 2 release is very strongly determined by soil physical properties, irrespective of the amount of N fertilization, has been shown in numerous studies. Parameters of the soil structure such as air permeability, aggregate size distribution, and size and design of pore space seem to play an important role. On the one hand, they have a direct influence on the movement of the gases via mass flow and diffusion in the soil in a variety of ways and thus also the emission of greenhouse gases [33][34][35][36][37][38][39][40][41] . On the other hand, they act indirectly by controlling the availability of oxygen in the soil, which in turn influences the processes of CO 2 and N 2 O formation and N 2 O consumption in many ways [42][43][44][45] .
In view of this, it is quite possible that the changes in soil structure caused by microplastics could indeed have an impact on the release of climate-relevant trace gases. The aim of our investigations here was thus to contribute to the clarification of the effect of microplastic addition on the emission of the greenhouse gases CO 2 and N 2 O in interaction with N fertilization. We wished to test two main hypotheses: (i) addition of microplastics leads to significant changes in important soil physical properties, most notably soil aggregation; and (ii) effect of N fertilization on the release of N20 and CO2 are therefore altered by the addition of microplastics.

Materials and Methods
Soil material. The soil material investigated was taken from the Ap horizon of a non-eroded Sample preparation and treatments. The investigations were carried out using soil samples filled into steel cores with a volume of 250 cm 3 and a bulk density of 1.4 g dry soil cm -3 . In preparation for incubations, the dry soil was sieved to 2 mm and then moistened to a water-filled pore volume of 48 %. To stimulate the activity of soil microorganisms similar to the conditions in an intensively N fertilized soil, diluted biogas digestate (dry matter 2.5%, pH 8.3) was used for soil moistening. In this way, the substrate was provided with 290 mg total N respectively 170 mg ammonium N per kg dry soil for all treatments at day zero of incubation (Table 1).
To investigate the effect of N fertilization and soil contamination with microplastics alone and in their interaction, four variants were established, each comprising four of these soil cylinders (Table 1). To test the N effect, 35 mg urea-N per core (100 mg N per kg dry soil) was added at day 22 of the incubation.
For contamination of the soil with microplastics we used microplastic fibers, since fibers have repeatedly been shown to affect soil structure 26,[47][48][49] , possibly due to their linear shape 50 . We ). Fibers were briefly microwaved to minimize microbial loads, following a previous protocol 30 .
The amount of microplastic fibers mixed in was 1.4 g per core (0.4% w/w) at day zero of the incubation.
The microplastic fibers were distributed homogeneously on the surface of the soil substrate. The soil substrate, the diluted biogas digestate and, depending on the variant, also the microfibers were then carefully mixed together and filled into the stainless steel cylinders in layers at day zero of incubation. In each case one treatment without urea fertilization and without microplastic contamination served as control ( Table 1). All samples received the same amount of mixing disturbance.

55%
Incubation experiments. First, the gas emission from the soil cores was monitored over 18 days. On day 19 of incubation, 35 mg urea-N was applied to four samples with or without microplastic contamination. The amount of urea N was dissolved in 10 ml water and injected into the soil using a syringe. Four samples each with and without microplastic contamination were used as controls, into each of which 10 ml of water was injected using a syringe. As a result of this measure the water-filled pore space in all samples increased from 48 to 55%. Subsequently, the gas emissions from the cores was investigated for another 22 days (Table 1).
To determine the CO 2 and N 2 O emissions, the soil samples were transferred to an incubation facility developed by us (Fig. 1). It works as a flow-through steady-state system corresponding to Livingston and Hutchinson 51 . The system contains 16 airtight, cylindrical incubation vessels (diameter and height of 13 cm, made from commercially available KG DN sewer pipes and accessories, Marley, Germany), each filled with one soil core. A temperature of 20 °C degrees was maintained in the incubation vessels by means of a climate box. Ambient air flows (32 mL min -1 ) continuously through the headspace of the incubation vessels via channels connecting the pressure vessel and the gas analyzer. In parallel, there is a control channel through which ambient air passes the incubation vessels with the same flow rate directly from the pressure vessel to the gas analyzer. To prevent the soil cores from drying out, the air was saturated to 100% relative humidity before passing through the incubation vessels. Each channel is directly connected to the gas analyzer via a multiplexer and a special circular channel for 7 minutes each. This results in a frequency per measurement and channel of 119 minutes. Gas concentration measurements were performed using cavity ring-down spectroscopy technology in a Picarro G2508 gas concentration analyzer (PICARRO, INC., Santa Clara, USA). Air was circulated between the incubation unit headspace and the CRDS analyzer at 250 mL min -1 using a low-leak diaphragm pump (A0702, Picarro, Santa Clara, CA, USA). The air from each of the 17 measuring channels was fed into this circuit via a connecting channel from the multiplexer at a slight overpressure. The overpressure was reduced via another opening on the other side of the circuit, from which the air then flowed out into the environment. In this way the continuous flushing of the circuit with the air from the respective channel to be measured was ensured.
The gas flux rates are calculated from the current gas concentration in the channel, which is connected to the outlet of a specific vessel, and the temporally corresponding concentration in the control channel, which represents the vessel inlet so to say, over time according to Equation

1
:  Samples were carefully reconstituted and mixed after measuring the MWD before taking 4.0 g of soil. These were placed on a small sieve with 250 µm mesh size, allowed to capilarrily re-wet with deionized water and placed into a sieving machine (Agrisearch Equipment, Eijkelkamp, Giesbeek, Netherlands). During the procedure, the samples were moved vertically for 3 min in metal bins filled with deionized water to experience a disintegrating force. The resulting slaking of the treated soil aggregates caused a separation into a water-stable and water-unstable fraction with a size > 250 µm. From the water-stable fraction, debris (sand particles and organic matter) were extracted to allow calculation of the water-stable aggregate fraction: WSA= (water stable fraction -debris) / (4.0g -debris).
For each sample two technical replicates were tested which were later merged into one mean value for the statistical analysis.

Statistics.
For the statistical analysis, we used a generalized least square model of the "nlme" package 57 with implemented varIdent function to account for heterogeneity in the applied treatment (i.e. control, microplastic, urea, microplastic : urea dual application). Control samples were set as reference level. We tested model residuals for normality and heteroscedasticity.  (Table 2). Related to the total incubation period, the emitted CO 2 was slightly promoted by microplastics and strongly by urea application. The  (Table 1).

Physical and chemical soil properties. The two test factors microplastics and urea addition
also influenced the physical and chemical soil properties in different ways. The presence of microplastics caused a significant increase in the proportion of water-stable soil aggregates (WSA) and in the mean weight diameter (MWD), and a slight increase in air permeability ( Table   2). In contrast, the addition of urea caused a significant increase in the content of cold water soluble carbon and ammonium N in the soil (Table 3, 4). The content of cold water soluble N and of nitrate N was promoted by urea as well as by microplastics. Therefore, the highest values for both parameters again occurred when combining microplastics with urea. The addition of microplastics also resulted in a slight increase in the total and organic carbon content of the soil, likely because microplastic-carbon was co-detected (Table 3, 4).

The analysis of the relationships between soil properties and the cumulative CO 2 and N 2 O fluxes
showed that in the case of CO 2 only closer relationships could be detected to the ammonium N content of the soil and in the case of N 2 O to the ammonium N and cold-water soluble carbon content of the soil (Figure 3).   DM dry matter, TC total carbon, TOC total organic carbon, C cwc cold water soluble carbon, N t total, N cwn cold water soluble nitrogen  Relationship between content of cold water extractable soil C and cumulative N 2 O emission.
(AVG = average per treatments combination). Treatment combinations are indicated by different colors (red = no N; blue with N added) and symbols.

Discussion
We here present clear evidence that microplastic fibers affect the dynamics and intensity of trace gas fluxes, in particular of CO 2 and N 2 O, from a sandy, intensively fertilized agricultural loam soil. It is interesting that microplastics mitigate the promotion of N 2 O release during intensive N fertilization. In the following we discuss these findings and the degree to which they are generalizable.

Experimental approach allows clear determination of the effect of the test factors.
In the first phase of the experiment, a clear effect of microplastic on gas flux dynamics was also observed, but over a period of about 18 days these effects were largely masked by the type of experimental approach, i.e. rewetting of the dry soil. This was probably due to a temporarily increased supply of microbially easily degradable C and N compounds as a result of soil disturbance and the addition of the diluted biogas digestate. By applying the urea only after this phenomenon had subsided, it was possible in the second phase of the experiment to clearly separate the effects of the test factors microplastics and N-fertilization. The renewed increase in CO 2 and N 2 O flux rates is certainly due to the N fertilization and partially to the slight increase in WFPS to the final value of 55% (Fig. 3).  (Table 2). To find out if this is the case and to elucidate the real proportion of various processes contributing to N2O release, further investigations are required, especially on the basis of isotopic approaches in combination with quantification of gas movement into the soil 33,65,66 .

Impact of N fertilization on CO
Recommendations for a more comprehensive assessment of the effect of microplastics on greenhouse gas fluxes. Our monitoring approach entailed high-frequency measurements; this is necessary to obtain precise results for trace gas emissions 28,[68][69][70][71][72] . Future endeavors aimed at quantifying trace gas flux responses to microplastic addition should also rely on such measurements.
We carried out this experiment to study effects of microplastic fibers on soil under highly controlled conditions, excluding the role of plant roots or larger soil animals such as earthworms.
In agricultural systems, plants may modify dynamics and trace gas fluxes. It is not clear which direction such modifications would take, because plants can affect outcomes in complicated ways in terms of their effects on rhizodeposition, competition for N, or changes in soil moisture [73][74][75][76][77] . It is thus a high priority to include plant responses in assessments of trace gas fluxes when soils are exposed to microplastic.
Our study used microplastic fibers, which is a common shape of microplastics in the environment, but microplastics come in a wide variety of shapes 50 , chemistries, and with many different additives present in consumer products 78 . The only other experiment to test greenhouse gas effects used PE particles with a much higher concentration than in our investigations 28 . Other shapes to examine include films, which have been shown to affect formation of cracks and water fluxes 79 ; this is relevant in agricultural systems owing to the prevalent use of mulching films. Different chemistries, including non-intentional additives and other compounds, may also affect different microbial players in the nutrient cycles leading to greenhouse gas emissions 28 , possibly also leading to effects diverging from the ones observed here. Clearly, examining a broader parameter space of microplastic properties should be a priority for future research.

Conclusions
Our laboratory study has clearly shown that microplastic fibers can influence trace gas emissions, and that soil structure effects are key to understanding such responses. Many studies of microplastic focus on a more classical ecotoxicological perspective, but our results suggest that microplastic should not be ignored in future estimates of greenhouse gas emissions and in assessing the actual risk to the environment from excessive N fertilization. Given the widespread presence of microplastic, especially in agricultural fields, such findings are relevant for understanding potential Earth system feedbacks of microplastic contamination 14 . It is clear that ecosystem-level feedbacks should be included as well to achieve a more complete assessment of impacts.

Data availability statement
Data supporting the findings of this study are freely accessible at the ZALF open-research-data repository (http://doi.org/104228/ZALF.DK.152).