EFFECT OF NITROGEN FERTILIZATION ON GREENHOUSE GAS BALANCE IN A BOREAL GRASSLAND
Zheng Yu MSc Thesis CARBO-NURMI project University of Eastern Finland, Department of Environmental and Biological Sciences
July 2020
CARBO-NURMI project, Biology of Environmental Change
Zheng Yu: Effect of Nitrogen Fertilization on Greenhouse Gas Balance in A Boreal Grassland MSc thesis 67 pages, 1 appendix (5 pages)
Supervisors: Marja Maljanen, Saara Lind July 2020
________________________________________________________________________
keywords: greenhouse gas, grassland, nitrogen, yields, fertilization, carbon dioxide, methane, nitrous oxide, carbon balance
ABSTRACT
Grass farming is paramount part in Finnish agriculture, grasslands could meanwhile behave as a C sink under boreal climate. The aim of this study is to investigate the relation of GHG budget and yields under various N- fertilizer amounts of an agricultural grassland during summer. CO2, CH4
and N2O fluxes from a mineral agricultural grassland in eastern Finland were measured during June, July and August 2019. Mixture of timothy and meadow fescue were cultivated as animal folders with three nitrogen fertilizer rates of 0 (N0), 150 (N150) and 300 (N300) kg ha-1 year-1. Static chamber measurement was employed wee seasonal continuous fluxes were model calculated. The GHG budgets were significantly enhanced about 104% and 94.7% under fertilizer rates of 300 and 150 kg N ha-1 year-1, respectively. NEEs consisted the overwhelming part of GHG budgets whereas contribution of CH4 and N2O were small. N2O emissions were significantly boosted across fertilizer amounts while CH4 uptake was not significantly differentiated. But N300 showed a stronger potential in CH4 sink capacity. N300 and N150 had the significantly higher annual yields which were 224% and 150% more than that of N0, respectively. Whereas N0 had the best ratio of GHG net balance over yields, which was five times and twice more than those of N300 and N150, respectively. The most GHG net balance amount came from N150 with 591 g m-2 eq CO2.
I would like to express my sincere sense of gratitude to my supervisor Marja Maljanen, who offered me the valuable opportunity of the research topic, who also provided essential suggestions and help in completing my thesis.
I am profoundly grateful to my supervisor Saara Lind, who provided responsible and timely guidance. It would be impossible to complete the job without her crucially technical support.
I acknowledge with thank the University of Eastern Finland for both laboratory and financial support together with the field practice support from Natural Resources Institute Finland, Maaninka during June, July and August 2019.
This thesis attempts to investigate the relation between GHG budget and yields in boreal grass farming based on three different nitrogen fertilizer amounts during the summer 2019. The results are used as a part of the investigation in the optimum nitrogen fertilizer practice for carbon-neutral food production. The study belongs to the project of CARBONURMI funded by BusinessFinland and operated in collaboration with Natural Resources Institute Finland, Maaninka.
AFOLU: agriculture, forest and land-use C: carbon
DM: dry matter GHG: greenhouse gas
GPP: gross primary production GWP: global warming potential LAI: leaf area index
N: nitrogen
NEE: net ecosystem respiration
PAR: photosynthetically active radiation, SOC: soil organic carbon
SOM: soil organic matter
TER: total ecosystem respiration
1. INTRODUCTION ... 7
2. LITERATURE REVIEW ... 9
2.1. GREENHOUSE GASES AND GLOBAL WARMING POTENTIAL ... 9
2.2. GHG EMISSIONS ... 10
2.3. NITROUS OXIDE ... 12
2.3.1 Properties and roles in agriculture ... 12
2.3.2 Nitrification and denitrification ... 14
2.3.3 Influential factors to nitrification and denitrification ... 15
2.4. METHANE ... 17
2.4.1 Emission overviews and biological processes ... 17
2.4.2 Fertilization on methane flux ... 19
2.4.3 Other influential factors ... 19
2.5. CARBON DIOXIDE ... 20
2.5.1 Situation and ecological pathways ... 20
2.5.2 Fertilization on agricultural carbon dioxide exchange ... 21
2.5.3 Other influential factors ... 22
2.6. GHG EMISSIONS AND GRASS FARMING IN FINLAND ... 22
2.6.1 Mitigation commitment and grass farming in Finland ... 22
2.6.2 Fertilization on grass farming ... 24
3. MATERIAL AND METHOD... 25
3.1. EXPERIMENTAL SETUP ... 25
3.2. GAS FLUX MEASUREMENTS ... 27
3.3. SUPPORT MEASUREMENTS ... 28
3.4. STATISTIC ANALYSIS ... 29
3.4.1 Nitrous oxide and methane data ... 29
3.4.2 Carbon dioxide data ... 30
3.4.3 Normality, correlation and significance test ... 31
4. RESULTS ... 32
4.1. WEATHER AND PLANT VARIABLES ... 32
4.2. YIELDS ... 32
4.3. NITROUS OXIDE EMISSIONS ... 34
4.4. MEATHANE UPTAKE ... 36
4.5. CARBON DIOXIDE EXCHANGE ... 38
4.5.1 Measured data ... 38
4.5.2 Modelled data... 40
4.6. THE NET GHG EFFECT ... 42
5. DISCUSSION ... 45
5.1. WEATHER, YIELDS AND PLANT VARIABLES ... 45
5.4. CARBON DIOXIDE EXCHANGE ... 48
5.5. CARBON BALANCE ... 48
6. FUTURE STUDIES ... 51
7. CONCLUSIONS ... 52
8. REFERENCE ... 53
APPENDIX
1. INTRODUCTION
The earth surface has been warming up by 1 ± 0.2°C by 2018 from the preindustrial time and it is projected to reach 1.5°C in the middle of this century. Global warming will cause a series of economic and environmental consequences such as more frequent and severe weather, higher death rate, higher wildlife extinction rate, sea-level lifting and ocean acidification. In order to keep the temperature rise well below 2°C at the end of this century, international efforts and agreements were made to strengthen the ability of countries to deal with the impact of climate change.
The global warming attributes to the anthropogenic greenhouse gas emissions. The most significant anthropogenic greenhouse gases are carbon dioxide, methane and nitrous oxide because of their strong radioactive forcing capacity and long atmospheric life span. Atmospheric carbon dioxide, methane and nitrous oxide are 147%, 259% and 123% of their preindustrial level. Agricultural sector contributes the second most (24%) of the anthropogenic emissions on a global level, among which synthetic fertilization consists of 14% of the agricultural emissions. Nitrous oxide is the primarily greenhouse gas resulted from fertilization, which has almost 300 times stronger global warming potential than that of carbon dioxide in 100-year time horizon. It is not viable to simply cut the usage of nitrogen fertilization to reduce nitrous oxide emissions in agriculture because it will lead to other problems such as soil fertility degradation, land use change from natural forest or wetland to cultivation area. On the contrary, fertilizer usage is expected to grow continuously owing to population growth and arising threat of food security. Concurrently, carbon dioxide exchange in agriculture is promoted by fertilization due to enhanced both total respiration (soil and plant) and biomass production. Although carbon dioxide is taken as a natural being other than an anthropogenic identity in agriculture. The exchange amount is gigantic compared to other greenhouse gases. Upland soil is usually a methane sink spot whereas studies of fertilizer promotion on methane sink capacity are contradictory concerning either refraining or facilitating.
Above all else, microbial activities play the key role governing nitrous oxide, methane and part of carbon dioxide fluxes, thus environmental conditions that impact microbial activities such as temperature, soil moisture and available carbon and etc. would influence the gas fluxes. All in all, greenhouse budget in the upland agricultural system is the balanced complex of carbon dioxide net ecosystem exchange, net nitrous oxide emission and the net result of methane producing and consuming in aerobic soils.
In Finland, grass farming area accounts for 30% of its land area. The related milk and beef industry contribute one half of the country’s agricultural income. Besides, grass cultivations on boreal mineral soils could also be a carbon sink. Thus, grass farming does not play a paramount role as economic propeller. but also a feasible climate change mitigation measure for the country. Timothy and meadow fescue are the most common fodder grasses in Finland. The optimum fertilizer amount for these grasses concerning both yields and feed qualities such as crude protein, digestibly and other features have been studied. However, carbon balance based on fertilizer amount and dry matter yields has not been properly studied. This thesis attempts to investigate the relation between GHG budget and yields in a boreal grassaland based on three different nitrogen fertilizer amounts (0, 150 and 300 kg ha-1 year-1) during the summer 2019. The results are used as a part of the investigation in the optimum nitrogen fertilizer practice for carbon-neutral food production (CARBO-NURMI project).
2. LITERATURE REVIEW
2.1. GREENHOUSE GASES AND GLOBAL WARMING POTENTIAL
The earth surface receives sun radiation and reradiates infrared radiation back into the atmosphere.
The gases in the earth atmosphere are transparent to incoming solar radiation but some of the atmospheric gases tend to trap the outgoing infrared radiation. Those gas molecules are heated and vibrate radiating energy (thermal radiation) to all directions. A part of the reradiated energy travels downwards back to the earth surface thus warms up the earth surface. This additional heating apart from the heating absorbed directly from solar radiation is greenhouse effect and the gases involved in are termed as greenhouse gases (GHGs) (Jones and Henderson-Sellers, 1990). In nature, the greenhouse effect is imperative as it keeps the planet surface mean temperature at a habitable + 15°C for current life forms. Otherwise, the earth surface would be - 18°C (Jones and Henderson- Sellers, 1990). Greenhouse gas effect attributes to mere atmospheric trace gases, among which carbon dioxide (CO2), methane (CH4) and nitrous oxide (N2O) are the top three anthropogenic long-lived greenhouse gases Their atmospheric life span are nearly permanent for CO2, 14 years for CH4 and 121 years for N2O, respectively (Jones and Henderson-Sellers, 1990).
Different GHGs have different greenhouse gas effect due to various gas lifetime and radiation absorbing capacity (Jones and Henderson-Sellers, 1990). Hence, uniformed metrics or climate parameters are necessary to quantify and compare the potency of different GHGs. The most well- known one is global warming potential (GWP), which is a measure of how much energy one ton of the gas in question absorbs over a given time period in comparison to that of one ton CO2 (IPCC, 2013). As a result, various greenhouse gas effects are normalized into comparable scales as equivalent CO2 emissions. It is somewhat “exchange rate” in a multi-component context and does not result in equivalence in temperature or any other climate indicating parameter. The GWP value of same GHG varies due to different choices of time horizons. The most often mentioned time horizons are a constant time period of either 20, 100 or 500 years and there is not any scientific argument about the choices of time horizons (IPCC, 2013). GWP offers possibilities to add up emission estimates of different gases straightforwardly such as compiling national GHG inventory.
It also provides opportunities to find mitigation strategies across different emission sectors and gases.
2.2. GHG EMISSIONS
It was reported that human activities had caused 1 ± 0.2°C global warming above the preindustrial level by 2018 and was likely to reach 1.5°C before the middle of this century under current increasing rate (IPCC, 2018). In order to keep the temperature increase below 2°C above the preindustrial scale by 2100, a vigorous reduction in GHG emissions over this century and negligible or even negative emissions at the end of the century are required (Moss et al., 2010).
Global temperature rising will cause a series of consequences such as glacier receding, extreme weathers, sea level rising and etc and the warming caused by anthropogenic emissions is irreversible unless further measures are taken (IPCC, 2014a).
Economic and population growth are the primary driving force of GHG emissions since preindustrial time. The atmospheric concentration of CO2, CH4 and N2O reach their historical summit than ever in the past 800,000 years (IPCC, 2014b). By 2018, the average atmospheric concentration of CO2, CH4 and N2O were 407.8 ppm, 1869 ppb and 331.1 ppb each, which were 147%, 259% and 123% in composition to those of preindustrial time (WMO, 2019).
According to IPCC (2014a) (see Figure 1), the sector of agriculture, forest and land-use (AFOLU) contributed the second most to the anthropogenic equivalent CO2 emissions in 2010 following energy supply sectors. CO2 was the primary GHG emission consisting 76% of the total anthropogenic GHG emissions in 2010. While CH4 emissions accounted for 16% and N2O made up of 6.2%. It is noteworthy that AFOLU sector was the only sector whose GHG emissions did not increase during the decade from 2000 to 2010.
Figure 1 Constituents of anthropogenic greenhouse gas (GHG) emissions by economic sectors in 2010. The unit of total emission is 49 Gt eq CO2 (gigaton of equivalent CO2 per year). The circle shows the percentage of five economic sectors of the total anthropogenic GHG emissions in 2010. The pull-out shows the contributions of indirect CO2 emissions (in % of total anthropogenic GHG emissions) from each sector to electricity and heat production (IPCC, 2014a).
In agricultural sector, only non- CO2 sources are considered as anthropogenic emissions. Non- CO2
gases refer to CH4 and N2O emissions due to microbial activities in agriculture such as cropland and grassland decomposition, livestock digestive exhausts and other pathways (IPCC, 2014a). CO2
emissions are excluded as a natural being that takes part in the carbon cycle by photosynthesis and respiration (IPCC, 2014a). Besides, energy combustion and transportation in agricultural practice are attributed to indirect emissions in IPCC’s report and agricultural machinery also contributes a considerable amount of CO2 emissions. (Schneider and Smith, 2009; Snyder et al., 2009).
Agricultural emissions accounted for 50% of the AFOLU emissions from 2001 to 2010 (Tubiello et al., 2013) and is the foremost contributor to global anthropogenic non- CO2 emissions (IPCC,
2014a). In 2010, the agricultural land-use was 1.6 billion hectares covering 12% of the terrestrial surface (FAO, 2011). In accordance, non-CO2 emissions were estimated as 5.2~5.8 GtCO2 eq year- 1 accounting for 10~12% anthropogenic emissions (Tubiello et al., 2013). Fertilization derived emissions were the third most contributor to agricultural sector consisting of 13% in the decade of 2001 to 2010 (IPCC, 2014a). Emissions of synthetic fertilizer increased at a rate at 3.9% year-1
since 1961 to 2010, from 0.07 to 0.68 Gt eq CO2 year-1 (Tubiello et al., 2013). Agricultural grassland is closely related to both non-CO2 and CO2 fluxes (Tubiello et al., 2015). Emissions of N2O and CH4 compensate to net ecosystem GHG balance offsetting CO2 sequestration (Liu and Greaver, 2009; Schulze et al., 2009). In general, fertilization of grassland contributes to CO2 sink more than CO2 and N2O emissions (Gomez‐Casanovas et al., 2016).
2.3. NITROUS OXIDE
2.3.1 Properties and roles in agriculture
Nitrous oxide is a colorless water-soluble and non-toxic gas. It is commonly known as laughing gas due to its euphoric effects while inhaling. It is a significant dissociative anesthetic in surgery and dentistry (Thomson et al., 2012). Apart from this, N2O is a detrimental greenhouse gas and ozone depleting gas. 90% of the atmospheric N2O is removed by photolysis in stratosphere (see equation (1)) (Reay et al., 2007). While the remaining N2O undergoes oxidation reactions with oxygen (O) producing NO which will further participate in the ozone depleting reaction (see equation (2) and (3)) (Ravishankara, Daniel and Portmann, 2009). Both sink pathways are slow and lead to its long atmospheric lifetime of 121 years (IPCC, 2014b). Therefore, N2O is 298 times stronger in Global Warming Potential (GWP) than that of CO2 in 100-year time horizon (IPCC, 2013). Ozone depleting induces a series of health problems such as sunburn, genetic mutation, cataracts and other health problems and N2O is likely to play the dominant role in ozone depletion in future (Portmann, Daniel and Ravishankara, 2012).
𝑁2𝑂 →
ℎ𝑣𝑁2+ 𝑂 (1) 𝑂 + 𝑁2𝑂 → 𝑁𝑂 + 𝑂2 (2) 𝑁𝑂 + 𝑂3→ 𝑂2+ 𝑁𝑂2 (3)
𝑁𝑒𝑡: 𝑂 + 𝑂3→ 2𝑂2 (4)
Since Haber–Bosch process was available to artificially fix nitrogen as ammonia in 1908, intensive farming based on synthetic fertilizer has been thriving and N2O emissions drastically increased due to the enrichment of reactive nitrogen in soils (Thomson et al., 2012). Nitrogen is an essential element for all forms of life to biosynthesize molecules such as amino acids, proteins, nucleic acids and other substances. All the nitrogen circulated in biosphere ends up in the atmosphere as dinitrogen (N2) that is the most abundant and stable atmospheric molecule with a percentage of 78.08 (Thomson et al., 2012). The triple bonds in dinitrogen need 10-3 KJ M-1 to break thus effective catalysts are needed to accelerate the molecule division (Thomson et al., 2012). All accessible redox states are involved in nitrogen cycles entailing -3 (NH3), -2(NO), -1(NH2OH), 0(N2), +1(N2O), +3(NO2-) to +5(NO3-) (Thomson et al., 2012).
N2O is biologically generated from specific classes of bacteria, archaea and fungi residing both in soil and ocean as a metabolite of respiration and energy producing (Thomson et al., 2012).
Processes included are nitrification, denitrification, nitrate ammonification (dissimilatory nitrate reduction to ammonia, DNRA), anaerobic ammonia oxidation (anammox) and other pathways (Shoun and Tanimoto, 1991; Wrage et al., 2001; Prendergast-Miller, Baggs and Johnson, 2011;
Marcel, Marchant and Kartal, 2018). Bacteria denitrification and ammonia oxidation from both agricultural and natural soils contribute to roughly 62% of N2O emissions around 6 and 4.2 Tg N year-1, respectively, the other one-third of the emissions is from the ocean dominated by archaea (Thomson et al., 2012). Besides. a tiny amount of N2O is produced from non-biological process such as chemical decomposition of NO2- (chemidenitrification) and hydroxylamine (NH2OH) oxidation (Bremner, Blackmer and Waring, 1980; Bremner, 1997).
Agricultural soil plays the largest anthropogenic N2O source (Del Grosso, 2010). 80% of atmospheric N2O increase attributes to food production (IPCC, 2013). The usage of nitrogen fertilizer in agriculture is still increasing (WMO, 2019). The world population increased by nearly 90% from 3.6 to 6.9 billion from 1970 to 2010 and productivity increased 230% due to additive usage of nitrogen fertilizer from 32 Mt to 106 Mt (IPCC, 2014a). Consequently, N2O emission accounted for 38% of the agricultural GHG emissions in 2010, which had increased by 73% since 1970 (IPCC, 2014a). The atmospheric N2O concentration was increasing averagely at a rate of 0.93 ppb per year, reaching 329.9 ± 0.1 ppb by 2017 (WMO, 2019).
In agriculture, the most significant biological N2O emissions are nitrification and denitrification from soils (Thomson et al., 2012). Ammonia oxidizing bacteria and nitrite oxidizing bacteria dominates autotrophic aerobic nitrification while denitrifying bacteria mediate anaerobic denitrification (Hayatsu, Tago and Saito, 2008). Other microbes that involve in nitrification and denitrification are anammox bacteria that convert NH4+ and NO2- to N2 under anaerobic conditions.
Some fungi also process denitrification and co-denitrification that produces N2O and N2 (Hayatsu, Tago and Saito, 2008). Archaea also promote soil denitrification (Santoro et al., 2011).
2.3.2 Nitrification and denitrification
Atmospheric nitrogen (N2) is fixed as ammonia (NH3) naturally only by free-living and symbiotic bacteria and archaea (diazotrophs). The triple bond is divided under the catalyst of nitrogenase generating ammonia. Nitrogenases is the only family that known catalyze this paramount step in nitrogen fixation. There are three variants of nitrogenase enzymes, each of which contains a metal ion of molybdenum, iron or vanadium and they possess complex, unique iron and Sulphur clusters (Thomson et al., 2012).
The ammonium ion (NH4+) is oxidized to nitrate (NO3-) in nitrification with three steps. Firstly, ammonium is oxidized to hydroxylamine (NH2OH) catalyzed by ammonia monooxygenase (AMO), during which N2O is produced. AMO is a transmembrane copper protein. It is an endergonic reaction under the presence of O2 (see equation (5)). The second step is hydroxylamine oxidized to nitrite (NO2-). It had been believed that this step was catalyzed by octaheme hydroxylamine oxidoreductase (HAO). However, it was found that HAO produces nitric oxide (NO) instead of nitrite. Nitrite is derived from nitric oxide by further unknown process (see equation (6)) (Caranto and Lancaster, 2017). The last step is nitrite oxidized to nitrate under the catalyst of nitrite oxidoreductase (NXR) that is a membrane-associated iron-sulfur molybdoprotein (see equation (7)) (Meincke et al., 1992; Spieck et al., 1998). Both ammonia oxidizing bacteria (AOB) and ammonia oxidizing archaea (AOA) participate in the ammonia oxidation in soils while AOA dominate N2O production in marine environment (Leininger et al., 2006; Wuchter et al., 2006; Hatzenpichler, 2012).
𝑁𝐻3 + 𝑂2 + 2𝐻++ 2𝑒− →
𝐴𝑀𝑂𝑁𝐻2𝑂𝐻 + 𝐻2𝑂 (5) 𝑁𝐻2𝑂𝐻 + 𝐻2𝑂 →
𝐻𝐴𝑂+𝑢𝑛𝑘𝑛𝑜𝑤𝑛𝑁𝑂2−+ 5𝐻++ 4𝑒− (6)
𝑁𝑂2−+ 𝐻2𝑂 →
𝑁𝑋𝑅𝑁𝑂3−+ 2𝐻++ 2𝑒− (7)
Nitrate and nitrite ion generated from nitrification is readily reduced by denitrification which is a stepwise reduction from nitrate to N2 with the presence of four enzymes (Zumft, 1997; Bothe, Ferguson and Newton, 2006). The first step is the reduction from nitrate to nitrite catalyzed by nitrate reductase (Nar) (see equation (8)). The second step is nitrite reduced to nitric oxide catalyzed by nitrite reductase (Nir) (see equation (9)). nitric oxide is further reduced to N2O catalyzed by Nitric oxide reductase (Nor) (see equation (10)). N2O is converted into N2 as a termination of denitrification catalyzed by Nitrous oxide reductase (Nos) (see equation (11)). Nitrous oxide reductase is the only known enzyme that governs the reaction from N2O to N2 (Zhang et al., 2019).
Thus, failure of the enzyme leads to the termination as N2O rather than N2. This is a crucial implication in respect of N2O emissions in agriculture concerning N fertilizer application.
𝑁𝑂3−+ 2𝐻++ 2𝑒− →
𝑁𝑎𝑟𝑁𝑂2−+ 𝐻2𝑂 (8) 2𝑁𝑂2−+ 4𝐻++ 2𝑒− →
𝑁𝑖𝑟 2𝑁𝑂 + 2𝐻2𝑂 (9) 2𝑁𝑂 + 2𝐻++ 2𝑒− →
𝑁𝑜𝑟𝑁2𝑂 + 𝐻2𝑂 (10) 𝑁2𝑂 + 2𝐻++ 2𝑒− →
𝑁𝑜𝑠𝑁2+ 𝐻2𝑂 (11)
The majority of denitrifying bacteria is anoxic or facultative aerobic heterotrophs. nitrate acts as an electron acceptor when oxygen is limited (Zumft, 1997). Both nitrification and denitrification are pathways for denitrification bacteria to generate ATP (Bothe, Ferguson and Newton, 2006).
2.3.3 Influential factors to nitrification and denitrification
Both processes of nitrification and denitrification are governed by complex conditions respecting microbes, soil properties and plant species. The related factors include e.g., N availability, aeration, temperature, moisture, organic carbon content, C:N ratio, texture, pH, soil management, metal cofactors (Signor and Cerri, 2013), soil type (Stevens and Laughlin, 1998), earthworm activities (Karsten and Drake, 1997; Borken, Brumme and Xu, 2000; Speratti, Whalen and Rochette, 2007) and other factors.
Fertilization is a crucial part of field management that influence both nitrification and denitrification thus further impact N2O emissions. N Fertilizer modifies soil N content by
increasing available ammonia (NH4+) and nitrate (NO3-). It is reported that the higher amount of NH4+-N, the greater nitrification process (Mosier, 2001; Khalil, Mary and Renault, 2004; Liu et al., 2005) Thus, more N2O is converted from NH4+ through various intermediate reactions of nitrification. Secondly, when nitrate amounts and other conditions are favorable, denitrification releases N2O subsequently to nitrification (Carmo et al., 2005; Ruser et al., 2006; Zanatta et al., 2010). In addition, lack of oxygen during ammonia oxidation would encourage NO2- as an electron acceptor instead of O2 to produce N2O or NO (Signor and Cerri, 2013). Last, the promoted biomass production by fertilization increases leftovers on field, which facilitates N2O emission in a long term (Hellebrand, Scholz and Kern, 2008). N addition might lead to soil organic C depleting because of the promotion of both soil mineralization and microbial activities (Signor and Cerri, 2013). As soil organic carbon is a basic feedstock for microbial growth and activities, soil organic C positively correlated with N2O emissions (Brentrup et al., 2000), the higher available soil organic carbon, the higher N2O emissions when N and moisture are not limiting factors (Ruser et al., 2006).
The N2O emission is mostly from the rhizosphere because root leachate promotes heterotrophic activities and denitrifiers compete with ammonifiers for this carbon. As a result, N2O emissions are influenced by both the organic carbon availability and the competence of the two microbial groups (Kim et al., 2004; Morley and Baggs, 2010).
There are other field management concerning fertilization that would affect N2O emissions such as N in-depth application and splitting application. There is not uniformed result about the influence of those applications since the studies are highly case specialized including specific crop types, meteorological conditions and other detailed environmental conditions. But in general, the lower available oxygen and nitrogen, the more N2O ends up as N2 (Brentrup et al., 2000; Yang and Cai, 2007). The longer N2O remains in soil, the more chance it is reduced to N2 as an electron acceptor regardless of soil properties (Chapuis‐Lardy et al., 2007). Therefore, in-depth application prevents N2O emission than surface application. Both in-depth and splitting application avoid nitrogen surface runoffs by precipitation and irrigation, which prevent N2O emissions (Signor and Cerri, 2013).
Cutting N addition to deduce the N2O emission is not feasible since the insufficient N supplement leads to SOC decomposition and poor productivity (Jaynes and Karlen, 2005). It is found that N2O emissions from agriculture could be reduced with no or little yield penalty by reducing N fertilizer inputs to levels that just satisfy crop needs (McSwiney and Robertson, 2005). Enhanced and
sustained soil productivity optimize crop production per hectare and reduce overall GHG emissions by avoiding land use change from other natural lands, such as wetlands, peatlands and forests, into farmlands (Jaynes and Karlen, 2005).
Temperature and moisture are crucial factors for both nitrification and denitrification since they govern microbial activities. They also influence N2O diffusion from soil to the atmosphere (Davidson and Swank, 1986). A close relation between N2O emission and air as well as soil temperature was reported (Wolf and Brumme, 2002; Zhang and Han, 2008). Increasing soil temperature will stimulate soil respiration thus creating anaerobic environment for denitrification (Signor and Cerri, 2013). Soil N2O emissions exponentially increase in response to the increase of soil temperatures (0-50ºC) (Liu et al., 2011). Soil moisture also intervenes in both nitrification and denitrification (Davidson and Swank, 1986). The extremely high moisture prohibits N2O emissions due to the inhibition of microbial activity while alternating dry and moist period might promote N2O emissions (Brentrup et al., 2000). When soil moisture is high, fewer soil pores are filled with air so that N2O emitted from denitrification is thriving (Brentrup et al., 2000). It also complies to non-fertilized field as non-fertilized field has a baseline N2O emission derived from mineralization of SOM (Del Grosso et al., 2006). Thus, rain and irrigation event might increase N2O emissions.
Perdomo observed N2O emission promotion after rain during high soil temperature period and Liu found the N2O diffusion clearly increased after rain and returned back to normal after three days (Liu et al., 2006; Perdomo, Irisarri and Ernst, 2009).
2.4. METHANE
2.4.1 Emission overviews and biological processes
Methane is a critical GHG with a GWP of 34 times stronger than that of CO2 in 100-year horizon (IPCC, 2013). CH4 contributes to 17% radioactive forcing by LLGHGs (WMO, 2019). The atmospheric CH4 nowadays is 259% of that in preindustrial time and the increment mainly attributes to anthropogenic emissions (WMO, 2019). Anthropogenic sources account for 60% of the atmospheric CH4 abundance including cattle farming, rice agriculture, fossil fuel exploitation, landfills and biomass burning (WMO, 2019). In the decade of 2001 to 2011, emissions from synthetic fertilizers containing both CH4 and N2O accounts for the third most agricultural emissions, among which synthetic fertilization induced CH4 emissions grew faster (3.5%) than
other agricultural contributors (FAO, 2014). It is projected that synthetic fertilizer would be the second largest CH4 source in the next decade (FAO, 2014).
The methanogenic archaea are the main CH4 biological source (Peters and Conrad, 1995; Angel, Claus and Conrad, 2011). In the soil, CH4 is generated from methanogenesis in anaerobic layer and travels from soil to atmosphere by diffusion, aerenchyma structures and ebullition (Serrano-Silva et al., 2014; Marushchak et al., 2016). Diffusion is the primary gas translocation pathway in well aerated agricultural soils. The amount of methanogens in aerobic soil is often less than that in anoxic soils (Peters and Conrad, 1995; Angel, Claus and Conrad, 2011).
The primary biological process that consume CH4 and integrates carbon cycle is CH4 oxidation (Dunfield et al., 2003). It is an aerobically enzymatic reaction catalyzed by monooxygenase (MMO) (Dunfield et al., 2003). Methane is predominantly oxidized by anaerobic bacteria (methanotrophs) in both aerobic and anaerobic soil conditions during its translocation (Curry, 2009; Knittel and Boetius, 2009). Methane oxidizing bacteria (methanotrophs) are the most well- known biological CH4 sink (Shrestha et al., 2010). They contain two types and consume CH4 as a carbon and energy source (Shrestha et al., 2010). Type I methanotrophs are more dynamic, rapidly proliferate and be active under favorable conditions such as in rhizospheres and even more on roots (Shrestha et al., 2010). Type II are more adapted methanotrophs that fit less favorable conditions such as in soils other than on roots, where is less abundant with CH4 (Shrestha et al., 2010). Both types of methanotrophs are found in various soils including forest, rice, landfill and agriculture (Jensen et al., 1998; Wise, McArthur and Shimkets, 1999; Eller and Frenzel, 2001; Reay et al., 2001). The abundance of either type is the result of the competition of available CH4. Type I methanotrophs dominate low CH4 and high O2 conditions while those of type II prevail the opposite situations (Graham et al., 1993). Besides, studies found that atmospheric CH4 oxidizer can oxidize CH4 at lower concentration (<1 nM) than cultured methanotrophs (Henckel, Friedrich and Conrad, 1999). Apart from methanotrophs, ammonia oxidizing bacteria are also able to oxidize CH4 due to its enzymatic system, but they are not able to gain energy from the process (Seghers et al., 2003).
It is also reported that ammonia oxidizing bacteria is enriched in minimally fertilized soils (Hermansson and Lindgren, 2001).
2.4.2 Fertilization on methane flux
Nitrogen fertilizer is found a prohibitor to CH4 oxidation as soil CH4 sink ability is constrained under fertilization (Hütsch W., Webster and Powlson, 1993; Hütsch W., 1996; Kravchenko et al., 2002; Seghers et al., 2003). The inhibition mechanisms differ in respect of short and longterm N fertilization. In short term fertilization, the inhibition of CH4 oxidation is caused by the interference of NH4+ which impact the methanotrophic enzyme system (Boeckx and Van Cleemput, 1996;
Tlustos et al., 1998). NH4+ is a competitive inhibitor of CH4 monooxygenase due to its low substrate specificity (Seghers et al., 2003). It is also an inhibitor for high affinity CH4 oxidation methanotrophs while NO3- does not affect high affinity CH4 oxidation (Seghers et al., 2003). Long- term N fertilization leads to a shift of microbial community composition (Seghers et al., 2005) and the influence is significantly pronounced on roots other than that of rhizospheres’ soils (Shrestha et al., 2010). It is reported that addition of N fertilizer promotes activities of ammonia oxidizing bacteria (Hermansson and Lindgren, 2001) thus leading to the competition for CH4 between methanotrophs and the ammonia oxidizing bacteria (Seghers et al., 2003). The microbial shifting diminishes atmospheric CH4 oxidizing capacity of agricultural soils (Seghers et al., 2003). The chronical effect is observed in an agricultural grassland that annually fertilized with NH4NO3
(Mosier et al., 1991). It also indicates that low affinity methanotrophs do not present or enrich in fertilized mineral soils (Seghers et al., 2003). However, N fertilizer addition meanwhile improves N substrates for methanotrophs’ growth and development, which might contribute to CH4 oxidation (Hahn, Arth and Frenzel, 2000; Paul et al., 2000; Schimel, 2000; Dan et al., 2001).
2.4.3 Other influential factors
Various factors influence methane generating and consuming including soil temperature, soil moisture, PH, agricultural management and other factors (Topp and Pattey, 1997; Le Mer and Roger, 2001; Kammann et al., 2009). Soil temperature impact both methanogenic and methanotrophic activities due to the sensitivity of underlying enzymatic process (Steinkamp, Butterbach-Bahl and Papen, 2001; Butterbach-Bahl and Papen, 2002). In general, the optimum CH4 production temperature ranges from 35 to 40℃ by slurry incubation experiment (Fey, Chin and Conrad, 2001; Conrad, 2002) while temperature around 25℃ and low salinity create a favorable environment for CH4 consumption (Serrano-Silva et al., 2014). Noteworthy, temperature influence is more pronounced under 15 °C because higher temperature leads to gas diffusion limits
and draught effects rather than temperature sensitivity (Steinkamp, Butterbach-Bahl and Papen, 2001)
Gas diffusion which is a major limiting factor coordinating activities of methanogens and methanotrophs was governed by soil moisture (Khalil and Baggs, 2005). Hence, soil moisture conditions control CH4 flux rate under both high and low soil moisture due to either physiological water stress or O2 and CH4 diffusion burdens (Khalil and Baggs, 2005). Higher soil water contents are favored by methanogens since it maintains a habitable environment (Andersen et al., 1998;
Peng et al., 2008; Prem, Reitschuler and Illmer, 2014). While CH4 consumption is ambiguous under the combined influence of both temperature and soil moisture. Some studies found that increasing water content at higher temperature results in a negative effect on CH4 consumption than lower temperatures (1-10℃). In other words, CH4 consumption respond to temperature declines with moisture abundance (King and Adamsen, 1992; Khalil and Baggs, 2005; Shukla, Pandey and Mishra, 2013). Moreover, declination of soil moisture may result in an increased evapotranspiration that promotes the enzyme activity of CH4 oxidizing bacteria thus increasing CH4 consumption (Blankinship et al., 2010).
Agricultural soils reside both methanogens and methanotrophs, the CH4 flux from the agricultural soil is the net result of both methane production and methane oxidation (Peters and Conrad, 1995;
Angel, Claus and Conrad, 2011) while upland soils are usually CH4 sinks rather than sources (Maljanen et al., 2010).
2.5. CARBON DIOXIDE
2.5.1 Situation and ecological pathways
Carbon dioxide is the single most significant anthropogenic GHG (IPCC, 2014b). The atmospheric CO2 was 147% of that in preindustrial time. It makes up 66% of the radiative forcing of long-lived greenhouse gases (LLGHGs) and attributes to 82% of the radiative forcing increase over the past decade (WMO, 2019). It is reported that every trillion tons of carbon emitted in the form of CO2
would lead to the peak surface temperature rise about 0.8 to 2.5°C (IPCC, 2014b).
Biological respiration and fossil fuel combustion is the main source of CO2 while photosynthesis and ocean are the main carbon sink (IPCC, 2014a). Terrestrial ecosystems sequester nearly 30% of anthropogenic carbon emissions offering the most effective natural means of climate change
mitigation (Quéré et al., 2009). The photosynthesis and respiration dominate the net CO2 exchange of carbon cycle by fixing atmospheric CO2 into biomass and exhaling CO2 from both plants and soils. While soil respiration is a primary contributor of 60~90% to ecosystem respiration (Kuzyakov, 2006). CO2 emissions from agricultural soils are strongly related to environmental factors such as soil temperature, soil texture, snowpack, pH, soil management, available carbon and nitrogen, C:N ratio (Yadav and Wang, 2017).
2.5.2 Fertilization on agricultural carbon dioxide exchange
Gross primary production (GPP) is usually promoted by fertilizer addition (Niu et al., 2010) because plants’ photosynthesis, light and water use efficiency are usually improved under fertilization (Evans, 1989; Sinclair and Horie, 1989; Muchow and Sinclair, 1994). Noteworthy, nitrogen addition stimulate C storage primarily occurs in plant pools but little in soil pools (Lu et al., 2011) since mere belowground organic matter inputs significantly increase soil C stock (Norby et al., 2004; Ruser et al., 2006; Russell et al., 2007). Fertilization reduces the root to shoot ratio leading to most stimulation being allotted to aboveground biomass production (Lu et al., 2011).
Apart from this, harvest would further decrease C allocation to belowground as a result of removing photosynthesis materials (Craine, Wedin and Chapin, 1999; Kuzyakov, 2006).
Improved GPP is accompanied by promoted net ecosystem respiration (NEE) since more substrates are available for both soil microbes and plants (Reich et al., 2008; Niu et al., 2010). The majority of carbon uptake from photosynthesis is released back to the atmosphere by ecosystem respiration (Luo and Zhou, 2006). Improved respiration also promotes microbial activates as root exudate secretion and root turnover are stimulated by higher CO2 concentration (Luo and Zhou, 2006).
Accordingly, soil carbon decomposition is affected and the plant-derived carbon is either converted to SOC or CO2 as a response of microbes (Franzluebbers, Hons and Zuberer, 1994, 1995).
Microbial SOC decomposition significantly affects soil CO2 flux (Hanson et al., 2000). While the response of SOC decomposition to N addition involves complex microbial processes (Carreiro et al., 2000; Hobbie, 2000). Generally, SOC decomposition is encouraged by fertilization under high C:N soil (>15) since microbial activities are limited by available N and pone to decompose SOC as a response to N addition. While in low C:N soil (<15), microbial activities might be refrained by overloaded N leading to decreasing decomposition of SOC (Lu et al., 2011). Grassland ecosystem is known as CO2 net uptake spots (Ciais et al., 2010; Chang et al., 2015; Gomez‐
Casanovas et al., 2016; Lind et al., 2020) Therefore, fertilized grassland produces more biomass accompanied by higher CO2 emissions (Lam et al., 2011; Yadav and Wang, 2017). However, constant N application might acidify the soils and cause ammonium toxicity (Nihlgård, 1985; Niu et al., 2010) shifting microbial community structure and vegetation compositions (Menge and Field, 2007; Xia, Niu and Wan, 2009).
2.5.3 Other influential factors
Plant respiration presents an exponential response to temperature increase (Eckhardt et al., 2019;
Lind et al., 2020) while the influence of temperature on soil respiration remain uncertain respecting both heterotrophic and autotrophic respirations (Carey et al., 2016). Experimentally higher temperature does not differentiate plant respiration rate from that of cooler temperature and the plant response to temperature shift tends to be less sensitive in a long term (Smith et al., 2018).
Precipitation is a key factor (Niu et al. 2010b) to influence carbon cycle in the grassland by either increase or decrease carbon uptake or release (Gerten et al., 2008; Niu et al., 2010; Wilcox et al., 2015).
Carbon ecosystem exchange is sensitive to precipitation regimes in terms of amount and temporal distribution (Chen et al., 2009; Gerten et al., 2008; Niu et al., 2009; Weltzin et al., 2003; Wilcox et al., 2015; Yan et al., 2011). Conclusions are controversial depending on water availability and vegetation demands as well as distinct environmental conditions that interact with soil moisture, such as climate, soil type, dominant vegetation species (Niu, Liu and Wan, 2008; Koerner and Collins, 2014).
2.6. GHG EMISSIONS AND GRASS FARMING IN FINLAND 2.6.1 Mitigation commitment and grass farming in Finland
International efforts and agreements are made to strengthen countries’ ability to deal with the adaptation and mitigation to global warming such as “the EU Climate and Energy Package 2020”
“Paris Agreement” and the “Kyoto Protocol” (IPCC, 2014a). Finland is committed to the GHG emission reduction target of 39% in 2030 compared to that of 2005 under the joint target of EU (Mnistry of Environment, 2019). According to the fourth report from Finnish Mnistry of Environment (2019), the total emissions of greenhouse gases were 56.5 MtCO2 eq in 2018 which was 21% lower than that of 1990. In 2017, the estimated agricultural emissions were 6.5 MtCO2
eq which was 12% of Finland’s total greenhouse emissions and the number is 13% lower than that of 1990. N2O was mostly from agricultural soils and contributed to 53% of total agricultural emissions while CH4 that was mostly from cattle enteric fermentation consisted 32%. The N2O emission from agricultural soils in 2017 was 8% lower than that of 1990. The primary reason attributed to the 40% declination of fertilizer usage in the interim.
Finland has an agricultural area of 22750 km2 by June of 2020, covering 7.49% of Finland land area (Trading Economics, 2020). In 2019, grassland that was cultivated with hay, silage, green fodder, pasture and hay seeds accounted for 7862 km2 consisting 34.6% of total agricultural land area of Finland (Luke, 2020a). Grass farming covers one-third of Finland agricultural area and the related milk and beef production consist of one half the national agricultural income. Thus, grass farming plays a key role in the profitability and competitiveness of Finnish agriculture (Luke, 2020b).
Timothy (Phleum pratense) and meadow fescue (Festuca pratensis) are substantial boreal forage grass species and usually cultivated together in Finland (Lind et al., 2020; Termonen et al., 2020).
Timothy is a native species of northern Europe (Casler and Kallenbach, 2007). while meadow fescue is native in temperate and northern areas (Lind et al., 2020). Both timothy and meadow fescue are C3 plants, whose growth cycles are observed in previous studies (Kemp and Williams, 1980; Maragni, Knapp and Mcallister, 2000). C3 plants flourish in spring and autumn (Niu, Liu and Wan, 2008) owing to lower temperature optima of photosynthesis (Kemp and Williams, 1980).
They are perennial species with possible two or three harvest times per season (Virkajärvi et al., 2015) with yields from 630 to 830 g m-2 DM according to different environmental and management measures concerning fertilizer amounts and harvest frequencies (Nissinen and Hakkola, 1994). The cultivation rotation time is three to four years (Virkajärvi et al., 2015). The yield response of perennial cool season grasses to N fertilizer ranges from 20 to 25 kg DM ha-1year-1kg-1 N (Hopkins, 2000). It is found that perennial cropping systems usually sequester more carbon than annual ones(Christensen et al., 2009; Schjønning, Heckrath and Christensen, 2009; Müller-Stöver et al., 2012), because perennial crops have high root biomass and the slow root derived carbon decomposition (Kätterer et al., 2011) due to longer belowground biomass accumulating and less tillage (DuPont et al., 2010).
2.6.2 Fertilization on grass farming
N application influences plant nutrient levels of forages such as crude protein (CP) contents (Shingfield, Jaakkola and Huhtanen, 2001) and mineral concentration (Pelletier et al., 2008).
Higher N rates result in high NO3= concentration in biomass which is harmful to ruminants (Maryland et al., 2007). Higher N rates would also lead to environmental problems such as NO3=
leaching (Wachendorf et al., 2004) and N2O emissions (Shcherbak, Millar and Robertson, 2014).
Thus, N fertilizer amount is regulated by Finnish authorities roughly under 250 kg N ha−1
year−1(Termonen et al., 2020). A fertilizer amount of 300 kg N ha-1 year-1 for boreal timothy and meadow fescue farming was recommended concerning feed quality, N balance and other features (Termonen et al., 2020). However, carbon balance based on fertilizer amounts and dry matter yields have not been properly studied.
The aim of this study is to investigate the relation of GHG budget and dry matter yields under three nitrogen fertilizer rates (0, 150 and 300 kg N ha-1 year-1) of a mineral agricultural soil in Eastern Finland during summer. Meanwhile, the results will contribute to the investigation of the optimum nitrogen fertilizer practice for carbon-neutral food production. The study belongs to CARBONURMI project funded by Business Finland and operated in collaboration with Natural resources Institute Finland, Maaninka .
3. MATERIAL AND METHOD
3.1. EXPERIMENTAL SETUP
The study site is 6.3 ha agricultural field located in Maaninka (63°08'44.0"N, 27°18'47.3"E, 89 m above the sea level) in eastern Finland. The annual mean temperature in 30-year scale (1981-2010) is 3.2℃, monthly mean temperatures in June, July and August are 14.1, 17.0, 14.5℃ each (Pirinen et al., 2012). The annual mean precipitation in 30-year scale (1981-2010) is 612 mm, monthly precipitation amounts in June, July and August are 66, 77, 75 mm each (Pirinen et al., 2012).
The experimental site has been used for either grain or grass fodder cultivation in the past ten years since 2009. Mineral fertilizers and herbicides were employed in between. The site setting for this experiment was established on 12 May 2018.
The experiment contains 5 replicates. Each replicate includes three treatments concerning different amount of nitrogen fertilizer 0, 150, 300 kg N ha-1 year-1 marked as N0, N150 and N300, respectively. Thus, 15 subplots are involved in. The area of each subplot is 12 m-2 with 8 m long and 1.5 m wide. One protection area of three subplots is set between each replicate in order to prevent disturbance from treatments. One aluminum collar (60 cm*60 cm*15-20 cm) with water groove was permanently installed into each subplot for gas flux measurements.
Timothy (Phleum pratense cultivar Nuutti) and meadow fescue (Festuca pratensis cultivar Valtteri) were sowed in a ratio of 70:30 mixed with barley (Hordeum vulgare L. cultivar Toria). The initial fertilizer was applied accompanying seeding with a rate of 200 kg ha-1 (64 kg N ha-1, 8 kg P ha-1, 22 kg K ha-1) via a plot-scaling sowing machine. The site was fertilized and barley harvested only once in the year 2018. Only the mixture of meadow fescue and timothy has been growing on the site onwards. Since 2019, fertilizer was surface applied to subplots by a farm scale fertilizing machine while the inside collar area was hand applied to avoid fertilizer misplacement. The fertilization information is shown in Table 1. Herbicide (Primus XL, Tuotekoodi: HM59319) was applied on 23rd May 2019 with amount of 2 kg ha-1 to prevent weeds
Table 1 Fertilizer dates and amounts of N0, N150 and N300 in 2019. Nitrogen fertilizer is from YaraBela Suomensalpietari 27-0-1 (700g bag-1); Phosphorus fertilizer is from YARA SUPERPHOSPHATE P20 0-20-0 (750kg bag-1); Potassium fertilizer is from Kaliumsuola 0-0-50 (700g bag-1).
N
(kg ha-1)
P
(kg ha-1)
K (kg ha-1) 9.5.2019
N0 0 20 25
N150 66 20 27
N300 132 20 27
13.6.2019
N0 0 0 25
N150 54 0 27
N300 108 0 27
25.7.2019
N0 0 0 0
N150 30 0 1
N300 60 0 1
In Total
N0 0 20 50
N150 150 20 56
N300 300 20 56
Biomass harvest date was determined by D-value (the concentration of digestible organic matter in dry matter) model of 680~700 g kg -1 DW (KARPE, 2019). Grass was cut to 7 cm by a farm scale plot harvester (model: Haldrup 1500 plot harvester, Løgtør, Denmark). Inside collar area and 30 cm outside the collar area was hand harvested to the same height by scissors to avoid collar damage by the harvester. The hand clipped biomass was packed into paper bags and transferred from field to lab at the same day. Fresh biomass weight was recorded and stored at 4°C at the same day. Dry mass was measured after oven dried at 65°C till constant weight (at least 72 hours). The harvest information is shown in Table 2.
Table 2 Harvest dates and Yields within collar area of N0, N150 and N300 in 2019
Mean Fresh weight/Area (g m-2)
Mean Dry weight/Area (g m-2)
Mean Dry % 13.Jun
N0 606 170 28.2 %
N150 1680 392 23.4 %
N300 2020 433 21.5 %
25.Jul
N0 172 88.5 54.2 %
N150 701 272 39.0 %
N300 1239 367 29.7 %
3.Sep
N0 125 53.3 42.9 %
N150 483 138 28.8 %
N300 1013 239 23.8 %
3.2. GAS FLUX MEASUREMENTS
N2O and CH4 measurements were implemented weekly using a static chamber method employing opaque chambers (Nykanen et al., 1995). 25 ml gas sample was taken via syringe and stored in a 12 ml vial (Labco Exetainers®, UK) separately at four incubation time points of 5, 15, 25, 35 minutes. Gas samples were analyzed by a Gas Chromatography (GC model: Aglient Technologies 7890B CN13423196) with an auto sampler in lab. The GC equipped with an Electron Capture Detector (ECD) for N2O concentration, Flame Ionization Detector (FID) for CH4 concentration and a Thermal Conductivity Detector (TCD) for CO2 concentration (Silvennoinen et al., 2008).
The gas flux is calculated by the concentration linear change in the head space of the chamber as equation (12).
𝐹 =𝑝0 ∗ 𝑘 ∗ 𝑉 ∗ 𝑀
𝑅 ∗ 𝑇 ∗ 𝐴 ∗ 60 ∗ 24 (12)
F represents for the flux rate (CO2 and CH4 (mg m-2 d-1), N2O (ug m-2 d-1))), P0 is the standard atmosphere pressure 101.325 Kpa, k is slope of the linear regression, V is the chamber volume (m3), M is the gas (CO2/ CH4/ N2O) molar mass, R is ideal gas constant 8.314 J mol-1 K-1, T is chamber temperature (K), A is the chamber area (m2).
Carbon dioxide exchange was measured weekly using a static chamber method employing transparent and opaque chambers (Alm et al., 1997). Three measurements were conducted for each subplot (light, shadow and dark measurements). Each measurement collected data for at least two minutes with two-second interval. The measurement gear includes a transparent chamber (60cm*60cm*30cm) equipped with an infrared gas analyzer measuring CO2 and H2O concentrations inside the chamber (model: LI-850-1 CO2/H2O Gas Analyzer), a PAR sensor (model: SKP215, Skye Instruments), a temperature sensor (model: 109, Campbell Scientific). Data was transmitted to a data logger (model: CR-850, Campbell Scientific). An ice-water cooling system with a pump is connected with the chamber aiming to keep the temperature difference inside and outside the chamber within 2°C.
3.3. SUPPORT MEASUREMENTS
Air temperature and precipitation was collected from Kuopio Maaninka observation station of Finnish Meteorology Institute (FMI, 2019). Photosynthetically active radiation (PAR, 400-700 nm) (Ross, 1975) was determined from a quantum sensor installed on field (model: QSO-E, Apogee instruments inc.), which collected PAR values on 60-second interval. A data logger (model ZL6, METER Group, Inc. USA) was connected and process the PAR information into hourly means.
Instantaneous soil temperature was measured via a thermometer (model: TM-80N, Tenmars) equipped with a temperature probe (model: ATT-50, Comark Instruments) at 5 and 10 cm depths concurrently with each chamber measurement. Instantaneous soil moisture was also determined between each chamber measurement with a moisture meter integrating 7 cm probes (model: HH2 equipped with ThetaProbe ML3, Delta-T Devices Ltd.).
Besides, in replicate D (subplots N0D, N150D, N300D), a soil water content reflectometer (CS655- VS, Campbell Scientific) was buried at 10 cm depth collecting both soil moisture and temperature.
Another temperature sensor (model: 107, Campbell Scientific) was buried at 20 cm depth. Both sensors were connected to a datalogger (CR200, Campbell Scientific) which collected both moisture and temperature mean values at 30-minute interval.
Plant height within the collar of each subplot was measured concurrently with chamber measurements. Leaf area index (LAI) was measured on a weekly basis by a plant canopy analyzer (model LAI -2000, LiCor) with a 180° view cap. Measurement that fulfill the both criteria of standard errors under 0.3 and successful measurement pairs (above and under vegetation) more
than 7 were accepted. As LAI measurements were conducted on different days from CO2 flux measurements. Daily plant variables used in CO2 correlation analysis including plant height, LAI and canopy index were collated by specific quadratic polynomial regression while the plant heights used in CH4 and N2O flux correlation analysis were measured data.
Soil samples were collected in spring 2019 with depth 0~20 and 20~40 cm separately and oven dried at 35℃ grounding to pass 2 mm sieve before analysis. The particle size distribution was determined with the pipette method (Elonen, 1971). Soil total C and N content were analyzed by dry combustion using a Leco® analyzer. Soil PH and electrical conductivity were measured in soil water suspension (1: 1.25 v/v). The soluble P and exchangeable K were extracted with acid ammonium acetate at pH 4.7 (Vuorinen and Mäkitie, 1955).
The soil was classified as a dystric regosol, medium textured (WRB, 2015). The topsoil (0~20 cm) profile contains mainly clay 11 ± 1.0%, silt 53 ± 3.1% and sand 35 ± 4.0%. The average soil characteristics of the topsoil were as follows: PH (H2O) 6.4, electrical conductivity 8 ± 0.1 mS cm−1, soil total carbon 2.0 ± 0.07%, total nitrogen 0.1±0.01%, C:N ratio 13.7 ± 0.7, the ammonium acetate extractable K 107.3 ± 19.6 mg L-1, soil P 9.4 ± 0.5 mg L-1.
3.4. STATISTIC ANALYSIS
3.4.1 Nitrous oxide and methane data
Gas emissions from soil and vegetation to the atmosphere is defined as positive while gas sink from the atmosphere to soil and vegetation is negative.
The data quality of each measurement was checked by R2 from the regression of concentration changes against incubation time. R2 higher than 0.8 was taken as successful measurement. While the small mean daily fluxes (N2O: 0 ± 100 µg m-2 d-1; CH4: 0 ± 50 mg m-2 d-1) were accepted regardless R2 since the variation between small concentration might be big and cause the poor fit of linear regression, whereas the measurement itself is trustworthy. Apart from this, the CO2
concentration derived from GC is an indicator for the chamber seal that R2 higher than 0.95 was deemed as good seal during the measurement. Seasonal continuous N2O and CH4 fluxes were interpolated based on measured daily mean fluxes by linear regression.
There were 13 whole weeks and extra two days of the entire experimental period from the 1st June to the 31st August in 2019. As both N2O and CH4 measurements were on weekly basis, there should
be 13 measurements for 15 subplots. Whereas one each extra chamber measurements after both fertilizer application dates of 13th June and 25th July were conducted in order to capture N2O flux boost. Hence, the total measurements were ideally 15 with 225 original flux data for each of CH4
and N2O fluxes. While after data quality check, available data of N2O was 197 (N0:68; N150: 66;
N300: 63), of CH4 fluxes was 185 (N0:60; N150: 65; N300: 60).
3.4.2 Carbon dioxide data
CO2 fluxes including net ecosystem exchange (NEE) and total ecosystem respiration (TER) were calculated from the concentration changes in chamber headspace over time using different regression models by MATLAB® (version: 2019b Mathworks) (Eckhardt et al., 2019) (see Figure A1for an example). First 10s of each measurement was discarded to avoid possible perturbations at the beginning while the left 55 data points were used for flux model calculation. Exponential non-linear model was chosen from various models because of the better performance evaluated by statistical criteria such as adjusted R2, root mean squared error (RMSE) and etc. (Pihlatie et al., 2013). Data was visually checked to avoid any turbulence caused by reasons such as maloperation or vapor disturbance. RMSE values that exceed 4 were considered as failure match to the exponential non-linear model. The misfit data would be re-inspected if the turbulence happened at the beginning or end of the measurement, otherwise, the measurement would be discarded. If any dark measurement was rejected, the corresponding light and shadow measurements would be rejected in consistency. In addition, instantaneous PAR values were model collated together with the NEE and TER calculation. Variation of PAR indicates the shift of cloud cover during the measurement, if the standard deviation of modelled PAR was larger than ± 100 mol m-2 s-1, it would be risky for a potentially poor fit, but it depends on specific cases. After data quality check, 93.5%
(547 of 585 are valid) of CO2 flux data was remained for further analysis. Corresponding gross primary production (GPP) is calculated by NEE=GPP+TER.
Seasonal continuous TER and GPP fluxes of each subplot were interpolated by corresponding non- linear exponential model (equation 13) or non-linear rectangular hyperbolic model (equation 14).
Model parameters were determined by plugging in model calculated TER and measured temperature values or model calculated GPP and PAR values. Each subplot got a specialized seasonal flux equation of TER against temperature and GPP against PAR, respectively (see Table
A1). Ultimately, seasonal continuous fluxes of TER and GPP of each subplot were calculated by plug in continuous seasonal temperature and PAR values (see Figure A2 for an example).
TER = R10× Q(10
TA T10)
(13)
where TA ishourly mean air temperature, T10 is 10°C, Q10 is temperature sensitivity at 10°C, the fitted parameter R10 is base respiration (mg m-2 h-1) (Shurpali et al., 2009).
GPP =GPmax× PAR × α
GPmax+ PAR × α (14)
where GPmax (umol m-2 h-1) is the theoretical maximum photosynthetic rate at infinite PAR, α is a model parameter representing initial canopy quantum efficiency (µg m-2 s-1 or µmol m-2 s-1) 3.4.3 Normality, correlation and significance test
Measured flux normality was checked by Kolmogorov-Smirnov test (Yap and Sim, 2011). N2O fluxes from all treatments, CH4 and TER fluxes from N0 were not normally distributed (see Table A2). Since measured data is not continuous, the Spearman’s rank correlation test (Glasser and Winter, 1961) was employed between gas fluxes and either environmental variables such as volumetric water content, air temperature and soil temperatures etc., or plant variables such as plant height , leaf area index (LAI) and canopy (see Table A3 and A4). Correlation coefficient higher than 0.6 (P<0.05) is a criterion as correlated (software: IBM®SPSS®statistics, version 25). As N2O emission was boosted by the first fertilization. N2O daily mean fluxes higher than 1000 µg m-2 d-1
was blocked in correlation test. Difference between treatments was tested by linear mixed model (software: IBM®SPSS®statistics, version 25).
