Denitrification, carbon and nitrogen emissions over the Amazonian wetlands

In this paper, we quantify CO2 and N2O emissions from denitrification over the Amazonian wetlands. The study concerns the entire Amazonian wetland ecosystem with a specific focus on three focal locations: the Branco Floodplain, the Madeira Floodplain and the floodplains alongside the Amazon River. We adapted a simple denitrification model to the case of tropical wetlands and forced it by open water surface extent products from the Soil Moisture and Ocean Salinity (SMOS) satellite. A priori model parameters were provided by in situ observations and gauging stations from the HyBAm 5 observatory. Our results show that the denitrification and emissions present a strong cyclic pattern linked to the inundation processes that can be divided into three distinct phases: activation stabilization deactivation. We quantify the average yearly denitrification and associated emissions of CO2 and N2O over the entire watershed at 17.8 kgN/ha/yr, 0.37 gC/m/yr and 0.18 gN/m/yr respectively. When compared to local observations, it was found that the CO2 emissions accounted for 0.01% of the integrated ecosystem, which emphasis the fact that minor changes to the land cover may induce strong impacts to the 10 Amazonian carbon budget. Our results are quite consistent with the state of the art global nitrogen models with a positive bias of 28%. When compared to other wetlands in different pedo-climatic environments we found that the Amazonian wetlands have close emissions of N2O to the tropical Congo wetlands and lower emissions than the tropical and temperate anthropogenic wetlands of the Garonne river, the Rhine river, and south-eastern Asia rice paddies. In summary our paper shows that a data driven approach can be successfully applied to quantify N2O and CO2 fluxes associated with denitrification over the Amazon 15 basin. In the future, the use of higher resolution remote sensing product from sensor fusion or new sensors like the SWOT mission will permit the transposition to other large scale watersheds in tropical environment. 1 https://doi.org/10.5194/bg-2020-3 Preprint. Discussion started: 12 February 2020 c © Author(s) 2020. CC BY 4.0 License.


Introduction
Inland waters play a crucial role in the carbon and nitrogen cycle. In particular, wetlands are known to sequester the atmospheric and fluvial carbon (Abril and Borges, 2018). This phenomenon is intimately linked to nitrous oxide (N 2 O) and carbon dioxide (CO 2 ) emissions to the atmosphere (Borges et al., 2015). In wetlands, during inundation periods denitrification processes nitrates into atmospheric dinitrogen that sustains emissions of N 2 O and CO 2 . These processes are controlled by biogeochem- 5 ical reactions linked to micro-organisms activity and pedoclimatic conditions (soil characteristics, nutrients availability and water content). Moreover the alternations between terrestrial and aquatic phases in wetlands promotes carbon and nitrogen mineralization and denitrification in soils (Koschorreck and Darwich, 2003). Our understanding and capacity to quantify the mechanisms involved in N 2 O and CO 2 emissions over wetlands are limited which leads to uncertainties in estimating them at large scales. 10 During the last decade, process-based models have become paramount tools in estimating carbon and nitrogen budgets in the context of global multi-source changes. Recent studies presented a review of existing models capable of quantifying N 2 O and CO 2 fluxes over continental ecosystems (Tian et al., 2018;Lauerwald et al., 2017). They are mainly used to characterize the part of greenhouse gases (GHGs) emissions due to natural and anthropogenic/agricultural activities at different spatial-temporal scales. The estimation of N 2 O emissions from natural sources are still subject to large uncertainties (Ciais and Coauthors.,15 2013) while N 2 O emissions from anthropogenic activities are under investigations. Assessing N 2 O budget for wetlands at large scale currently constitutes a knowledge gap. In terms of denitrification, the relatively sparse and shot-term observations limits our capability to estimate the carbon and nitrogen recycling in terrestrial ecosystems, especially over wetlands. Since in situ measurements constitute the main source of data, few studies assess N 2 O and CO 2 emissions from denitrification at large scale and are usually limited to field scale or small scale watersheds (Russell et al., 2019;Johnson et al., 2019;Korol et al., 20 2019).
In the case of the Amazon basin, the total amount of CO 2 emission reaches 0.3 PgC/yr for both natural and agricultural sources. Scofield et al. (2016) pointed out over the Amazonian wetlands the disproportionally high CO 2 out-gassing may be explained by the abundant amount of podzols for the Negro Basin. Those types of soils are likely to slow organic matter decomposition and increase leaching of humus. Elsewhere over the Amazon, floodplain soils are mainly Gleysols (Legros,25 2007) which are characterized by a high microbiological activity. CO 2 emissions from the river are mainly due to organic matter respiration as well as exports from the wetland system. In wetland, root respiration and microbial activities are a major source of CO 2 emissions (Abril et al., 2014). Ultimately CO 2 outgassed from the Amazon River is about 145 ± 40 TgC/yr (de Fatima F. L. Rasera et al., 2008) and tops at 470 TgC/yr when extrapolated to the whole basin (Richey et al., 2002). However, considering the carbon budget, it is not clear to which extent the Amazon basin acts as a sink of carbon. Some studies 30 show that the Amazon basin is more or less in balance and even acts as a small sink of carbon at the amount of 1GtC/yr (Lloyd et al., 2007). Remote sensing have emerged as an essential tools for GHGs quantification, either via assimilation into physically-based models (Engelen et al., 2009) or as a direct observation (Bréon and Ciais, 2010). For wetlands the monitoring of water extents is crucial for the denitrification processes. Water surface monitoring has been done with a variety of spectral bands (Martinez and Le Toan, 2007;Pekel et al., 2016;Birkett et al., 2002) in active and passive remote sensing. Recently L-Band microwave remote sensing showed advanced capabilities to monitor water surfaces in tropical environment because of all-weather capabilities, providing soil signal under vegetation .
This study aims at delivering an enhanced understanding and quantification of the denitrification process over Amazonian wetlands with their associated fluxes of N 2 O and CO 2 using modelling and microwave remote sensing. We constrained and 5 adapted denitrification process-based set of equations by L-Band microwave water surface extents from the SMOS satellite and a priori information from in situ. The specific objectives of the study are to highlight the main key factors controlling the denitrification and to provide the hot-spots and hot-moments of denitrification over wetlands.

10
The Amazon basin ( Fig.1) is the world largest drainage basin with an area of 5.5 × 10 6 km 2 and an average water discharge of 208 000 m 3 s −1 (Callode et al., 2010) representing 20% of all freshwaters transported to the ocean. The watershed spans across Bolivia, Colombia, Ecuador, French Guiana, Peru, Suriname, and Guyana and 68% of the basin pertains to Brazil. The Amazon hydrology is governed by three main sources: the Andes, the Brazilian and Guyana shields and the lowlands. Devol et al. (1995) described the hydrology of the main stream as the aggregation of the water originating from Andean regions, from the main 15 tributaries and from "local sources" corresponding to smaller streams draining local lowlands. The contribution of each water body differs in time. For example from November to May the contribution of Andean waters reaches 60% and declines during the dry season to 30%. Wetlands are paramount in the watershed functioning : 30% of the Amazon discharge has once passed through the floodplain (Richey et al., 1990) alongside a 2010 km long reach between São Paulo de Olivença and Òbidos (Bourgoin et al., 2007). The Amazon watershed can be divided into 8 major sub-basins: (1) the Negro basin, (2) the Branco basin, (3) the Solimoes River and its tributaries, (4) the Madeira basin, (5) the Purus basin, (6) the Tapajos basin, (7) the Xingu 5 basin and (8) the section between Manaus and the mouth of the Amazon River. We used that delineation when we designed our model (Fig. 3) as it represents the main hydrosystems of the Amazon basin.
The Amazon basin contains several floodplains. Here we consider the three main floodplains: the Branco floodplain in the northern part, the Madeira floodplain in the southern part and the floodplain between Odidos and Manaus which is called Obidos -Manaus floodplain (O-M FP) in the following. The O-M FP covers an area of 2.5 × 10 5 km 2 whereas the Madeira 10 floodplain covers 3.7 × 10 5 km 2 . The Branco floodplain is the widest of the three floodplains with a covered area of 6.7 × 10 5 km 2 .

In situ data from the HyBAm observatory
In situ data and gauging stations data were obtained from the HyBAm international network (www.ore-hybam.org). Hybam stems from the precedent PHICAB and HiBAm projects in Bolivia and Brazil. The main objective of this network is to monitor 15 the hydrology, geochemistry and sediment load of the main Amazonian rivers with associated quality and uncertainty. The   (Kerr et al., 2010). SMOS orbits at a 757 km altitude and provides brightness temperature emitted from the Earth over a range of incidence angles (0°to 55°) with a spatial resolution of 35 to 50 km. Parrens et al. (2017) showed the capability of SMOS to retrieve the water fraction under dense forests over the Amazon basin. One of the main upsides of SMOS is its sensitivity to soil signal under vegetation in all-weather conditions thanks to the L-Band frequency. The SWAF data was averaged each month over the sampling period (2011 -2015) within the Amazon basin. Fig.2 shows the aver-

Assessing denitrification and out-gassing
In this study, we modified the denitrification rate proposed by Peyrard et al. (2010) to fit tropical wetland conditions. Denitrification is a consumption of dissolved organic carbon (DOC), particulate organic carbon (POC) and nitrate (N O − 3 ) in soil limited by dioxygen (O 2 ) and ammonium (N H + 4 ) availability. Here, because denitrification is regarded as occurring during 5 flooding events while the soil is saturated and extremely low oxygen is available, we consider that O 2 is not a limiting factor (Dodla et al., 2008). Furthermore, as there is only one flooding event over the watershed per year over the particularly active Amazonian ecosystem, it is supposed that all the ammonium is processed into nitrate between two consecutive floods. Thus, we also supposed that N H + 4 is not limiting neither. The fact that nitrate stocks are reconstituted under aerobic conditions, e.g when soils are no longer flooded, is a reasonable assumption in the case of the Amazon basin and more particularly for the 10 5 https://doi.org/10.5194/bg-2020-3 Preprint. Discussion started: 12 February 2020 c Author(s) 2020. CC BY 4.0 License. wetland parts (Hulme, 2005;Brettar et al., 2002). Besides, many studies consider denitrification as a combine consumption of nitrates and carbon (Scofield et al. (2016); Dodla et al. (2008); Goldman et al. (2017)). Taking into consideration the above statements, the denitrification rate can be expressed as: where R N O3 is the denitrification rate in µmol L −1 d −1 , 0.8 ·alpha represent the stoichiometric proportion of nitrate consumed 5 in denitrification compared to the organic matter used with alpha = 4 as mentioned in Peyrard et al. (2010) , ρ is the dry sediment density kg dm −3 , φ is the sediment porosity, k P OC is mineralization rate constant of P OC (d −1 ), P OC is the Particultate Organic Carbon content in the soil and the aquifer sediment (1 per thousand), MC is the carbon molar mass g mol −1 , DOC is the concentration of Dissolved Organic Carbon (DOC) in the aquifer water µmol L −1 , k DOC is the mineralization rate constant of DOC (d −1 ), k N O3 is the half-saturation for nitrate limitation in µmol L −1 and N O 3 is nitrate concentration 10 in the aquifer in µmol L −1 .
Estimation of out-gassing CO 2 is based on the denitrification equation where gaseous CO 2 is formed. It is supposed that neither nitrates nor organic matter are limiting factors for the reaction which is considered total (eq. 2) (de Freitas et al., 2001). Abril and Frankignoulle (2001) showed that denitrification tends to raise the alkalinity. In order to take into account this phenomenon, the formation of HCO − 3 from dissolved CO 2 (eq. 3) was coupled to the denitrification (eq. 2). Overall, in this study, 15 denitrification was modelled using: The equation of the chemical reaction of denitrification (eq. 4) is used to determine the generated amount of CO 2 by relating 20 it to the amount of denitrified N O3. Finally, N 2 O production is indirectly estimated as a result of N 2 formation and the ratio was set to 0.1.

Parametrization of dissolved/particulate organic carbon and nitrate concentrations
The model parameters for the denitrification are taken from references studies and in situ measurements. The sediment porosity 25 φ was set to 25%. k P OC , k DOC and k N O3 were set to 1.6 × 10 −5 d −1 , 1 × 10 −3 d −1 and 1.0 × 10 −6 µmol L −1 respectively.
Nitrate concentrations (N O − 3 ) were considered constant over the period. As the Amazon is one of the most active region of the world (Legros, 2007) in term of microbial soil dynamic, it was assumed that on the one hand, during non-flooding period, mineralization of nitrogen was sufficient to compensate nitrate loses by plant assimilation and leaching. On the other hand, during flooding period and for inundated soils, when denitrification is triggered, nitrate inputs from streams were abundant 30 enough to sustain this process (Sánchez-Perez et al., 1999). For P OC concentration, according to the studies performed by Moreira-Turcq et al. (2013), it was considered constant over the whole watershed and for the global period of the simulation (2011 -2015) to 1 ‰. The HyBAm database provides measurements of DOC both over time and for few streams. Dissolved organic carbon concentration in streams is highly correlated to discharge (Ludwig et al., 1996). The stable seasonality over time of the Amazonian streams was demonstrated by prior studies (Paiva et al., 2013). Considering these two properties, streams were regrouped regarding the main sub-basins of the Amazon watershed mentioned in Section 2.1. For each sub-basin, the 5 average monthly discharge was calculated using the HyBAm's gauging station's records. We then used those discharge tendencies to extrapolate the average monthly DOC concentration based on in situ records (figure 3). It was supposed that there is no difference for the DOC value between the different years.

Denitrification computation
The methodology focuses on modelling the denitrification process that occurs in the first 30 cm of water-saturated soils in 10 wetland. Thereby, only the nitrates included in that layer were considered undergoing denitrification. Nitrates brought by streams are supposed not to modify significantly the amount of nitrates contained in the soil solution. Indeed, the concentration of nitrates in the river is negligible to the concentration of riverine aquifers (Sánchez-Pérez et al., 2003). We consider that the DOC in the soil is directly brought by streams so the amount of DOC included in soils is set up to the streams values. Most of the organic carbon is transported from alluvial sediments or brought by streams during flooding events (Peter et al., 2012). We 15 consider that the gases produced during the denitrification are entirely emitted to the atmosphere regarding the supersaturation of pCO 2 in groundwater (Davidson et al., 2010). Overall, denitrification was calculated as: where D N O3 is the net denitrification in mole/month, R N O3 is the denitrification rate in mole/month/L, SWAF is the fraction of land covered with open waters and Q wa is the water storage capacity for each type of soil (L). Soil data were determined from the Food and Agricultural Organization (FAO) database. The soil description file, constituted of thousands of soils, was 5 summed up into 36 categories accordingly to soil textures and an associated nitrate concentration limit was given ( figure 3).
In summary the model requires the inputs and parameters for : (1) the types of soil and their nitrogen contents in order to assess the base pool of nitrates, (2) the dissolved organic carbon concentrations of the streams that overflow and (3)

Focus on the main three Amazonian floodplains
The temporal patterns of the processes over the entire basin and throughout the whole period are unique in each floodplain. In fact, the three floodplains do not become active/ inactive at the same time and do not reach their maximum potential activity at 5 the same moment either. Figure 6 shows the monthly behaviour of denitrification, CO 2 and N 2 O emissions for each floodplain together with the respective one over the entire basin. The following comments can be given:  -The Branco floodplain is the less constant of the three floodplains even though a general pattern can be observed.

5
The floodplain becomes active in January but the activation is slow and the denitrification is low until April (less than     Table 1 depicts the yearly emissions of CO 2 and N 2 O over the Amazon basin and the three main floodplains. Emissions of CO 2 from denitrification are twice as much higher than N 2 O emissions over the basin. Averagely, flooded areas emits 2.76 × 10 9 kg C-CO 2 per year and 1.03 × 10 9 kg N-N 2 O per year by denitrification from the natural nitrate pool of the watershed.

Denitrification and gazes emissions anomalies
During the period of the study, major meteorological events were recorded over the Basin. On one hand, year 2011 was a "la   Months undergoing the El Niño episode show a reduction of 27.7% than average values.
It appears that extreme events do not have an uniform impact on the whole basin. Table 2 sums up the spatial denitrification for the Amazon basin and the three floodplains at a yearly scale. Overall denitrification may not be impacted at watershed scale. The average denitrification rate for the whole basin shows little inter annual variations. However, in 2015 simulated denitrification for the Branco FP was twice as low than for the year 2011. As so, it can be assumed that this floodplain has been 25 drying off during the 2011 -2015 period and thus is much more sensitive to drier conditions than other parts of the watershed.

Determining key factors of the denitrification
A sensitivity analysis of the parameters of the denitrification equation (1) was performed. k poc can range from 0.15 × 10 −6 to 1.1 × 10 −4 which leads to a yearly denitrification 46% lower and 18% higher than the initial values respectively. k doc range from 1.0 × 10 −4 to 1.22 which leads to values of denitrification 94% lower and 1.333 × 10 5 % higher respectively. It follows 5 that for the Amazon Basin k doc is evaluated as more sensitive than k poc . Also, the nitrates related part of the denitrification equation was analysed analytically. Nitrates are relatively abundant in the watershed's soils and it is noticeable that k N O3 is negligible compared to N O 3 though lim N O3→∞   Overall, the actual denitrification rate (equation 1) may be improved as a combination of a potential rate function (provided by DOC and P OC) and limitation functions provided by the peculiar environmental conditions.

Comparing to physically-based models
The N 2 O emissions at large scale are compared to results to the NMIP project (Tian et al., 2018) model, more particularly DLEM (Xu et al., 2017), VISIT (Ito and Inatomi, 2012) and O-CN (Zaehle and Friend, 2010).These models consider the N 2 O emissions from nitrification and denitrification, where in our case only denitrification during flooding is considered as nitrification is mainly used for refiling the nitrogen pool. Also the aggregate impact of temperature, water saturation of the soil, nitrogen contents, soil pH and micro-organisms activity, explicitly modelled in the physically based models, is accounted for in our approach through the parameters k P OC and k DOC and mineralisation rate included in equation (1) ( (Peyrard et al., 2010;Sun et al., 2017). x 10 11 m² wetlands system represent 0.81 ± 0.02 gN/m²/yr. We can observe that our model gets a global higher estimation of the emissions of N 2 O at a rate of 28% than the other models with 80% of them (0.14 gN/m²/yr) originates from the three main floodplains; the Obidos -Manaus, the Madeira and the Branco. In term of input data, our model as well as DLEM, VISIT 25 and O-CN use climate data, soil types and inundated fractions/surfaces. A divergent point is how nitrogen pool is calculated.
We consider it as being produced by the organic matter mineralization and maximum nitrification. Whereas the other models compute it from nitrogen deposition. Moreover, they also take natural vegetation, swamps delineation (O-CN) and land cover as input data while we only focus on wetland types. These models assess N 2 O emissions based on the processes of the nitrogen cycle such as denitrification. Our model apprehends denitrification as a function of carbon and nitrate contents (DOC, P OC and N O 3 ) and inundated surfaces (SWAF). As a result, these models do not fully distinguish the alluvial floodplain from other 5 lands (Xu et al., 2017) and underestimate its effects (Ito and Inatomi, 2012). Thus our results bring us to conclude that current physically-based N 2 O emissions models are likely to slightly underestimate the contribution of wetlands in the global budget.

Wetlands and integrated ecosystem emissions
In this section, our model outputs for wetlands emissions are compared to local in situ measurements of the ecosystem N 2 O 10 and CO 2 emissions. Table 4 summarizes the different results from in situ measurements for N 2 O and CO 2 and the closest simulation node from our simulation. When comparing the N 2 O with in situ campaigns performed by (Koschorreck, 2005) and (Keller et al., 2005) at Manaus plateau and Santarem, the wetlands emissions from this study are roughly 1/200 of the integrated ecosystem observed emisisons. CO 2 emissions at local in situ measurements (Keller et al., 2005) as well as to broader measurements (Richey et al., 2002) are compared to our models outputs. Our wetlands estimations are critically lower 15 (10 4 ) than integrated ecosystem observations. As expected, even though CO 2 emissions from wetland denitrification are about 2.16 × 10 9 kgC-CO 2 per year over the Amazon basin, these emissions are negligible when compared to the full ecosystem Carbon emissisons (Cole et al., 2007;Davidson et al., 2010). Overall, CO 2 emissions from denitrification over the whole Amazon basin participate to 0.01% of the carbon emissions of the watershed. Most of the CO 2 emissions over the Amazon are attributed to processes such as organic matter respiration from biomass. Confirming previous studies, this result means that 20 even a small change in the distribution of wetlands cover over the Amazonian basin may drastically modify the carbon budget.
It constitutes a topical subject for the Amazonian basin.

Limitations of the current approach
The findings of this study have to be seen in light of some limitations. First, the sampling resolution of input data can induce bias. The SWAF product tends to underestimate water surface extents variability and land cover identification due to the coarse resolution of 25 km x 25 km. Second, the use of uniform k P OC and k DOC values limits the capabilities of the model to fully consider the impact of the spatial variability of both geophysical and biological variables. Third, as highlighted by the 25 present study, the lack of in situ measurements of N 2 O emissions over tropical wetlands specifically increases the uncertainties and equifinalities for the calibration of model parameters and validation. Future studies should concentrate on: adding more remotely sensed geophysical variables at the adapted spatial resolution (Parrens et al., 2019), taking into account the fact that flooding actually sustains the different processes.

30
The main objective of the study is to quantify and assess CO 2 and N 2 O emissions over the Amazonian wetlands during flooding periods. To achieve these goals we design a data driven methodology that relies on modelling and remote-sensing products. It aims to estimate emissions linked to denitrification at large scale. The model parametrisation was justified by results from several published papers. It appears that denitrification mainly relies on DOC contents in the watershed. The study also contributes to better understand the functioning of the major floodplains of the Amazon Basin and their respective involvement in the Amazon Carbon and Nitrogen budget. It transpires that the most active floodplain is the Òbidos-Manaus, which is responsible for the majority of processes. It may also be pointed out that each floodplain possesses its own functioning 5 that depends on rainfalls and the hydrology of the floodplain's river. Overall, the results appear quite close to other large scale models; especially for N 2 O emissions. Key factors of denitrification for the Amazon Basin were identified in the study. Future studies will concentrate in extending the current approach to other tropical basins, needless to say that local observations will be paramount for the validation of such exercise and preferably over the same period of analysis. Data from future missions like SWOT will deliver water heights at 21 days global coverage, which will improve the results of such studies through the 10 integration of surfaces and volume information.