Nitrous oxide emission from a flooded tropical wetland across a vegetation and land use gradient

Climate change is currently one of the most critical global environmental concerns, which science-based evidence attributes to the increasing emission of greenhouse gases (GHGs) into the atmosphere. Addressing climate change and its impacts is currently no longer an option but a must for the survival of humanity and ecosystems. According to the IPCC (2013), the main GHGs implicated in climate change are carbon dioxide (CO₂), methane (CH₄) and nitrous oxide (N₂O), which account for close to 90% (WMO 2019) of the radiative forcing by long-lived GHGs. However, even though N₂O has the least atmospheric concentration among the three GHGs, its global warming potential (GWP) is 265 times that of CO₂ (Liu et al. 2020), far higher than that of CH₄, whose GWP is only 28 times that of CO₂ on a 100-year timescale (IPCC 2014). This makes N₂O an important greenhouse gas in the global climate change equation and has since captured the attention of climate change scientists and policy makers worldwide.

Wetlands are recognized by the Millennium Ecosystem Assessment (2005) as vital ecosystems due to their various ecosystem services. They are among the most important natural ecosystems in climate change mitigation through carbon capture and storage. Natural wetlands, in an undisturbed state, are vital carbon sinks, where their photosynthetic CO₂ uptake through primary production exceeds CO₂ losses due to ecosystem respiration (Mitsch et al. 2013; Were et al. 2019). However, despite being carbon sinks, a number of studies indicate that wetlands are also important sources of N₂O emitted into the atmosphere (Wu et al. 2009; Liengaard et al. 2013; Audet et al. 2014; Yang et al. 2019).

Unlike CO₂, N₂O is primarily biogenic (IPCC 2007), with denitrifiers and nitrifiers being key players in the regulation of its global sinks and sources. Denitrification and nitrification processes in wetland soils are affected by changes in environmental conditions, which can occur either naturally due to temporal (e.g., diurnal and seasonal) variations, or due to wetland management practices (Ajwang'Ondiek et al. 2021). Intermittent wetting and drying cycles in wetlands in relation to natural changes in hydrological regimes or due to soil and water management practices (such as in rice paddies) alter soil parameters which regulate denitrification and nitrification processes (Liengaard et al. 2013). These parameters are even expected to vary more under future climate scenarios, due to alterations of temperature, rainfall and nutrient regimes (Tian et al. 2015). Besides, environmental conditions in wetlands also vary spatially since soil conditions, vegetation characteristics and soil and water management practices change even at small spatial scales (Butterbach-Bahl et al. 2016). Consequently, N₂O emission can also vary across small spatial scales, necessitating measurements in all wetlands situated across different climatic and geographic locations, and under different management practices, e.g., natural, agricultural and wastewater treatment wetlands.

Although there has been an increase in knowledge of the sinks and sources of N₂O, the global N₂O budget remains less understood (IPCC 2007; Liengaard et al. 2013). Tropical ecosystems are vital in understanding global N₂O balance because several studies have indicated that the high productivity of these ecosystems is likely to translate into rates of accumulation, recycling and loss of nitrogen far higher than those of temperate ecosystems (Hedin et al. 2009; Liengaard et al. 2013; Pärn et al. 2018; Liu et al. 2020). Further, with a coverage of about 30% of the global wetland area (Marín-Muñiz et al. 2015), tropical wetlands have a substantial influence on the global atmospheric N₂O budget. Despite these, studies on N₂O emission from wetlands have mainly focused on temperate wetlands (Liengaard et al. 2013; Ajwang'Ondiek et al. 2021). This has limited a proper understanding of the magnitudes and controlling factors of N₂O emission from tropical wetlands, hence hindering their inclusion in global climate models. Indeed, recent studies by Butterbach-Bahl et al. (2016) and Boateng et al. (2017) have noted with concern the reliance of field measurements and estimation of GHG fluxes from tropical wetlands using emission factors calibrated from temperate wetlands, despite significant differences between the two climate types. Further, with specific reference to Uganda, whereas a few studies have investigated controls and magnitudes of GHG fluxes from the country's wetlands (e.g., Were et al.2021a, 2021b), they have mainly focused on CO₂ and CH₄, at the expense of N₂O.

The aims of this study were to investigate the impact of: (1) vegetation community type in a natural tropical freshwater marsh wetland on N₂O emission, and (2) converting a natural tropical freshwater marsh wetland into a rice paddy wetland, on N₂O emission. We hypothesized that N₂O emission in a natural tropical freshwater marsh wetland varies based on the type of vegetation community, and that conversion of the wetland into a rice paddy wetland enhances N₂O emission.

The natural wetland can be categorized as a marsh and exhibits unaltered hydrology and vegetation characteristics dominated by perennial sedge and grass communities. In view of the dominant vegetation community types, the wetland can be further subdivided into three different sections: Typha latifolia (Typha), Phragmites mauritianus (Phragmites) and Cyperus papyrus (Papyrus). The Papyrus vegetation community occurs downstream of the Typha and Phragmites vegetation communities, and its morphology displays two growth forms; emergent and floating (Were et al. 2020a), unlike the Typha and Phragmites vegetation communities that grow only emergent.

The rice paddy wetland was previously part of the natural wetland, with similar conditions as described above for the natural wetland, before it was altered to enable the cultivation of rice. Rice is cultivated on a seasonal basis because it relies on natural flooding, which itself is dependent on seasonal rainfall. The cultivation is done on smallholder scales, where rice is planted through the broadcasting of rice seeds by hand into the fields or by transplanting rice seedlings from nursery beds into the main fields. There is no standard planting density, and therefore, plant density varies from field to field. Rice cultivation is done under no fertilization conditions but rather depends on the natural fertility of the soil.

Surface and ground water level dynamics in these wetlands and their riparian areas have been described by Kayendeke et al. (2018). Seasonal flooding and drying cycles have been reported at wetland edges of the natural wetland in relation to the wet and dry seasons, respectively. However, wetland areas away from the edges are usually flooded throughout the year, but the flooding level varies seasonally. In the rice paddy wetland, intermittent flooding and drying cycles also occur with respect to the wet and dry seasons. However, compared to the natural wetland, both the flooding level and duration in the rice paddy wetland are manually regulated depending on the rice growth/cultivation stage. During the entire sampling course for this study, both the natural and rice paddy wetlands were continuously flooded (with water above the soil surface).

The study area temperature can generally be described as warm. Air temperature data for the study area during the sampling period were obtained from the Uganda National Meteorological Authority (UNMA) and showed wet season air temperature of 21.3 ± 0.2 °C (mean ± standard error; SE), in comparison to 23.4 ± 0.1 °C during the dry season.

Soil-atmosphere N₂O exchange was measured in each of the wetlands/vegetation community types described in section 2.1. Gas samples were collected using steady-state flux chambers (Minamikawa et al. 2015). The chamber technique is preferred and has been widely applied in the measurement of gas fluxes from wetlands because it is less costly and allows for measurements at fine scales (Butterbach-Bahl et al. 2016).

Chambers included the following components: a thermometer to monitor internal air temperature, a vent tube to stabilize inside air pressure, and a gas sampling port, which allowed for manual extraction of the gas samples from the chambers using a syringe and a needle (Minamikawa et al. 2015). Because of the strong solar radiation that increases the heating up of the chambers during daytime sampling (the study area is located along the equator), white-coloured chambers were used whose external surfaces were further covered with a reflective aluminum tape. Chamber bases were sunk up to 10 cm into the soil (Butterbach-Bahl et al. 2016), onto which chamber lids were firmly fixed to provide a gas-tight enclosure. The headspace of the chambers had the following average dimensions: height = 25 cm and basal area = 490.63 cm², amounting to a volume of 10 L.

In each sampling area, three chambers were deployed and sampled consecutively to increase the spatial representativeness (Minamikawa et al. 2015; Butterbach-Bahl et al. 2016). To minimize artificial gas ebullition associated with physical soil disturbance during gas sampling, wooden walking platforms were installed. The wooden walking platforms also eased movement during sampling in these usually waterlogged environments. However, for the floating growth form of the Papyrus vegetation community, it was realized that the walking platforms could not control shaking and artificial gas ebullition during sampling due to the suspended nature of the root mat over the water column. This occurrence was noticed from preliminary measurements that showed abnormally high gas concentrations. Consequently, the final sampling plan considered only the emergent growth form of the Papyrus vegetation community.

Sampling was done over a period of 12 months, as follows: Dry season; February, March and April 2019, and February, March and April 2021; and Wet season; August, September and October 2019, and August, September and October 2021. During each sampling event, the duration of gas sampling for each chamber was limited to 30 min, at intervals of 10 min, i.e., 0, 10, 20 and 30 min (Butterbach-Bahl et al. 2016). Further, prior to the start of each sampling event, ambient air samples outside the chamber were collected for quality control (Butterbach-Bahl et al. 2016). Gas samples in the chambers were collected using 60 mL luer lockable syringes attached to needles. The collected gas samples were stored in 10 mL glass vials under high pressure, after being evacuated using 40 mL of the syringe gas volume. Vial tops were covered with parafilms to prevent contamination of samples prior to analysis. The frequency of sampling in each wetland/vegetation community type was twice a month, on a fortnightly basis. Gas samples were analyzed by gas chromatography (SRI 8610C gas chromatograph, USA), at the International Livestock Research Institute (ILRI) in Nairobi, Kenya.

Environmental (soil and hydrological) parameters were measured during the sampling period. In each wetland/vegetation community type, soil samples at the top (0–10 cm) soil layer (Inglett et al. 2012) were collected, from which soil physico-chemical characteristics; salinity, pH, total nitrogen (TN) and organic carbon (OC) were determined. Soil temperature was also measured in situ, using a digital soil thermometer. Soil moisture content was not investigated since the wetlands were under continuous flooding throughout the sampling period.

Composite soil samples that were obtained after mixing three samples in each sampling plot were placed in labeled ziploc bags and transported for analysis at the Soil Science Laboratory, Makerere University, Kampala, Uganda. Soil samples were air dried at room temperature in the laboratory for 21 days, ground and sieved through a 2 mm nylon sieve.

Soil OC was obtained from soil organic matter, following the loss on ignition procedure and using van Bemmelen's index of 0.58 as in Were et al. (2020a). Soil TN content was determined following the Kjeldahl digestion procedure. Soil pH and salinity were obtained using a portable multi-parameter meter (CyberScan PC 300), after mixing soil and deionized water using a soil : water ratio of 1 : 5 (Were et al. 2021a, 2021b).

The wetland hydrological parameter considered in this study was water level, and specifically surface water level since the water was above the soil surface for the entire sampling period. Surface water level was measured at the gas sampling location during each sampling event, using a cm-marked wooden stick.

Data were sorted and statistically analyzed using the Microsoft Excel (2016) and R programming software (version 4.0.5). The data were first tested for normal distribution and homogeneity of variance. Whereas data for soil and surface water level showed normal distribution, N₂O flux data did not satisfy conditions for normal distribution. Consequently, parametric one-way ANOVA alongside the Tukey HSD post hoc tests were used to examine the significance of mean values of soil and water level characteristics between the wetlands/vegetation community types and seasons, at p < 0.05 significance. On the other hand, the Kruskal–Wallis H test was used to examine the significance of mean N₂O fluxes between the wetlands/vegetation community types and seasons, at p < 0.05 significance. The effect of environmental conditions on N₂O fluxes was investigated by determining the Spearman's rank-order correlations between soil and surface water level characteristics and N₂O flux, at p < 0.05 significance. Unless stated otherwise, all values presented from the analysis are mean ± SE.

Soil physico-chemical characteristics did not vary significantly either between the wetlands/among vegetation community types or between seasons, except OC and C: N ratio (Table 1). Whereas mean OC contents of the three vegetation communities did not differ significantly (p > 0.05) during both the dry and wet seasons, they were over two-fold higher than was recorded in the rice paddy wetland. Mean C:N ratios of the Typha vegetation community (dry season = 33.2:1 and wet season = 33.3:1) were significantly higher (p < 0.05) than those measured in the other two vegetation communities of the natural wetland and rice paddy wetland. Mean pH values during the wet and dry seasons in both wetlands were slightly acidic, ranging from 5.98 to 6.54 and 6.00 to 6.87 during the wet and dry seasons, respectively. Temperature had mean values in both wetlands ranging from 26.3 to 26.9 and 26.6 to 27.0 during the wet and dry seasons, respectively.

Generally, fluxes of N₂O from both the natural and rice paddy wetlands were very low (close to zero) during the two sampling seasons. Considering the impact of vegetation community type (in the natural wetland) on N₂O fluxes, no significant variation (p > 0.05) in N₂O flux was noted among the three vegetation communities during both seasons (Table 2). Mean N₂O fluxes from the three vegetation communities of the natural wetland ranged from 0.5 to 0.6 μg m⁻² h⁻¹ during the dry season, and from 0.4 to 0.5 μg m⁻² h⁻¹ during the wet season.

In view of the impact of land use change (from natural to rice paddy wetland) on N₂O fluxes, the difference between N₂O fluxes from the natural and rice paddy wetlands was insignificant (p > 0.05), with the latter recording an average flux value of 0.7 μg m⁻² h⁻¹. Further, in both wetlands, seasonal changes (dry vs wet) did not affect N₂O fluxes significantly (p > 0.05). Cumulatively, mean annual N₂O fluxes from the natural and rice paddy wetlands ranged from 4.4 to 6.1 mg m⁻² yr⁻¹, with annual carbon dioxide equivalents (CO₂e) ranging from 1.2 to 1.6 g m⁻² yr⁻¹ (Table 2).

The relationship between soil physico-chemical and surface water level characteristics and N₂O fluxes from the wetlands is depicted in Table 3. As shown by the Spearman's rank-order correlation coefficients, soil physico-chemical parameters temperature, pH and TN positively correlated with N₂O flux, unlike OC, salinity and C: N that showed negative correlations. Similarly, the correlation between surface water level and N₂O flux was negative. In terms of the magnitude of the correlations, neither soil physico-chemical characteristics nor surface water level significantly correlated (p > 0.05) with N₂O.

Emission of N₂O did not vary significantly among the vegetation community types in the natural section of the wetland. In natural wetlands, OC and nitrogen are some of the most important factors controlling the production of N₂O, which are influenced by the type of plant communities (Marín-Muñiz et al. 2015). Detailed discussions on these factors have been made by Maucieri et al. (2017) and Smith et al. (2018), while Huang et al. (2013) discussed in detail the mechanisms of N₂O emission from wetlands. OC in soils is important for N₂O emission since it provides electrons for heterotrophic denitrification (Marín-Muñiz et al. 2015), while organic nitrogen is a substrate for nitrification and denitrification processes. Most of the greenhouse gas production in wetland soils occurs in the top 5 cm (Were et al. 2021a), and it is in this layer where accumulation of aboveground residues from plants occurs. Therefore, the observation of no significant difference in N₂O emission among the three different vegetation communities in this study could be explained by the fact that OC and nitrogen contents did not differ significantly among the vegetation communities (Table 1). These findings are consistent with those of other studies. Marín-Muñiz et al. (2015) reported no significant difference in N₂O emission from swamp and marsh wetlands with varying vegetation community types in Mexico. In Canada, Baskerville et al. (2021) reported similar N₂O emission rates across different vegetation communities in riparian zones along the Washington Creek.

However, several other studies have reported different results. Liu et al. (2014) showed that emissions of N₂O from two vegetated coastal wetland zones, one dominated by Spatina spp. and the other dominated by Phragmites spp., differed significantly, with the Phragmites spp. zone showing higher N₂O flux. This observation was explained by differences in biomass productivity, which affected the resultant organic matter and nitrogen input into the soil. Similarly, Piñeiro-Guerra et al. (2019), who studied several wetland sites across Argentina, showed that N₂O emission increased with biomass productivity. Wang et al. (2008) while investigating the impact of plant species on N₂O emission noted significantly higher N₂O fluxes from a Zizania latifolia-dominated wetland compared to those dominated by Phragmites australis and Typha latifolia. They attributed the results to differences in the root structure of the plant species, where the root structure of Zizania latifolia was favored by ammonia-oxidizing bacteria for N₂O formation.

This study found no significant difference in N₂O emission between the natural and rice paddy wetlands. However, it had been expected that N₂O emission would be higher in the rice paddy wetland compared to the natural wetland. Conversion of a wetland from its natural state into a rice paddy wetland enhances N₂O emission mainly due to: (i) water table drawdown through drainage favors oxygen availability, thus increasing the mineralization of soil organic nitrogen (Liengaard et al. 2013; Ajwang'Ondiek et al. 2021), and (ii) fertilizer application in rice paddies increases nitrogen availability in soil. Liengaard et al. (2013) reported that partial soil wetting in a tropical South American freshwater wetland resulted in high N₂O emission compared to long-term soil waterlogging following heavy rain. Kritee et al. (2018) established that intermittently flooded Indian rice paddies emitted 30–40% more N₂O than those under continuous flooding. In China, Yang et al. (2013) observed that N₂O emission from a natural marsh wetland increased by 120% as the water level reduced from +14 to −11 cm (below the soil surface). A study on European peatlands by Liu et al. (2020) has reported a N₂O emission factor of 19.3 kg N ha⁻¹ year⁻¹ in agriculture-drained peatlands, higher than in grasslands (17.4 kg N ha⁻¹ year⁻¹) and forests (3.4 kg N ha⁻¹ year⁻¹). The study further recommended that rewetting of all drained European peatlands could cut the cumulative N₂O emissions by 70%. Similarly, a recent synthesis of several studies on the effect of soil moisture content (as affected by water table depth) on N₂O emission from freshwater sediments has reported N₂O pulses following sediment drying and rewetting events, with exposed sediments being active spots for N₂O emissions during dry phases (Pinto et al. 2021). In view of the impact of fertilization, nitrogen addition into rice paddies enhances N₂O emission. Owino et al. (2020) reported that the conversion of Papyrus wetlands into nitrogen-fertilized rice paddies in Kenya significantly increased N₂O emission (4.37 ± 3.18 μg m⁻² h⁻¹ from the fertilized fields against −3.59 ± 2.56 18 μg m⁻² h⁻¹ from the unfertilized fields). Zhang et al. (2014) also reported a positive correlation between the amount of fertilizer application and N₂O emissions from rice paddies, irrespective of the rice growth stage. Therefore, in this study, the lack of a significant difference in N₂O fluxes between the natural and rice paddy wetlands could be explained as follows: (i) both the natural and rice paddy wetlands were continuously flooded throughout the sampling period, so the influence of water level on N₂O fluxes was similar across both wetlands, and (ii) artificial fertilization of the studied rice paddy fields was not being done, as rice cultivation was only reliant on the natural fertility of the soil. Undoubtedly, the soil nitrogen contents in the natural and rice paddy wetlands were not different, as was seen in Table 1.

Seasonal variations showed no impact on N₂O emission during this study, in both wetlands. The influence of seasonal changes on N₂O emissions from wetlands has been previously reported, especially in temperate wetlands (Czóbel et al. 2010; Jørgensen & Elberling 2012). Warmer soil temperatures during summer seasons enhance microbial activities, resulting in increased mineralization of organic nitrogen in soils. In tropical regions, however, Sjögersten et al. (2014) indicated that temperature is unlikely to exert a significant control on GHGs fluxes from wetlands because temperatures are relatively stable, irrespective of season. Bernal & Mitsch (2013) instead showed that seasonal fluxes of N₂O from tropical wetlands are more likely to be driven by changes in water table depth (or soil moisture content). In the present study, seasonal changes (dry vs wet) were not associated with drying and wetting cycles as water level was above the soil surface in both seasons, and neither did soil temperature significantly vary between the dry and wet seasons (Table 1). These, thus, could explain why N₂O emissions did not vary between the dry and wet seasons.

The average N₂O emissions measured in this study were generally low, about four times lower than the average emission reported for tropical wetland systems (van Lent et al. 2015). However, it was noted that the average N₂O emission reported by van Lent et al. (2015) included fertilized wetland systems, and those under different degrees of disturbance, which could have accounted for the high N₂O flux value. This study's average N₂O flux values are within the range reported for undisturbed natural tropical systems (Matson & Vitousek 1987; Marín-Muñiz et al. 2015) and unfertilized rice paddy wetlands (Owino et al. 2020).

This study showed great variations in individual N₂O fluxes even within the same wetland/vegetation community type or season, with values ranging from negative (indicating N₂O consumption) to positive (indicating N₂O emission) (Figures 3 and 4). This has also been reported by other studies (Audet et al. 2014; Owino et al. 2020; Ajwang'Ondiek et al. 2021), where it is attributed to wetland soils acting as both N₂O sinks and sources. Wu et al. (2013) explained that soils act as sinks of N₂O when N₂O is consumed during either nitrification or denitrification, largely under conditions of limited nitrate (⁠⁠). However, in most cases, as has been reported from instantaneous and annual estimates, soils act as net sources of N₂O (Schlesinger 2013). Indeed, a review by Chapuis-Lardy et al. (2007) on soil N₂O uptake made a conclusion that most field values were small and unlikely to contribute to a large N₂O sink globally, while Schlesinger (2013) has suggested that soil N₂O uptake is as low as 5% of the estimates of global net flux from soils to the atmosphere. In this study, N₂O emission and uptake could be attributed to soil oxygen level, with anoxic soil conditions favoring N₂O emission, while suboxic soil conditions enhancing N₂O consumption (Schlesinger 2013). This is supported by the fact that there was an inverse relationship between N₂O flux and water level (Table 3), and the water table also displays an inverse correlation with soil oxygen content (Bernal & Mitsch 2013).

In Uganda, no study has evaluated the totality of N₂O emissions from the country's natural and rice paddy wetlands. Therefore, to provide a basis for future studies, we used our study findings to roughly estimate total annual N₂O emissions from Uganda's natural and rice paddy wetlands. To achieve this, we made a simple assumption that other natural and rice paddy wetlands in the country present more or less similar conditions to our study wetland. Natural wetlands in Uganda cover about 26,165 km², compared to 150 km² occupied by rice paddy wetlands (Were et al. 2021b). We obtained mean annual N₂O emission from the natural and rice paddy wetlands as 4.4 ± 13.1 and 6.1 ± 17.8 mg m⁻² yr⁻¹ (Table 2). Therefore, total N₂O emissions from Uganda's natural and rice paddy wetlands are estimated at 115.1 ± 342.8 T yr⁻¹ (CO₂e = 30,501.5 ± 90,842 T yr⁻¹) and 0.9 ± 2.7 T yr⁻¹ (CO₂e = 242.5 ± 707.6 T yr⁻¹), respectively. However, due to the rapid rate of conversion of natural wetlands into rice paddy wetlands, rice paddy wetlands are expected to be the main sources of N₂O emitted from the country's wetlands in the near future.

To mitigate climate change and its impacts, the IPCC (2014) has emphasized increasing carbon sequestration while at the same time minimizing the emission of GHGs into the atmosphere. Some of the suggested measures to increase carbon sequestration include protection and conservation of natural ecosystems such as wetlands, forests and grasslands. Wetlands are unique ecosystems, where water is the terminal parameter that influences the development of soil and plant characteristics that are different from other ecosystems. These unique characteristics have a great influence on carbon and nitrogen cycling in these ecosystems (Mitsch et al. 2013). Several studies have acknowledged that conversion of natural wetlands into farmed wetlands compromises climate change mitigation by enhancing carbon and nitrogen emission (Owino et al. 2020; Were et al. 2020b; Ajwang'Ondiek et al. 2021). As earlier explained, water drawdown in rice paddy induces carbon and nitrogen emissions from wetlands (Were et al. 2019; Owino et al. 2020), with fertilized rice paddies being hotspots for N2O emission (Owino et al. 2020).

With the increasing human population globally, increased food demand implies that natural tropical freshwater wetlands will remain under high pressure for conversion into rice paddies. Indeed, Davidson & Finlayson (2018) have reported an average increase in global rice paddy acreage of 0.62% per year. As result, there is need to a strike a balance between rice production and climate change. This would involve exploring ways of optimizing carbon sequestration while minimizing emission of other GHGs such as N₂O from rice paddy wetlands.

In this study, it was noted that N₂O emission from the rice paddy wetland was not significantly higher than that emitted from the natural wetland, an indication that rice cultivation under continuous flooding and without fertilizer application can minimize N₂O emissions associated with rice paddies, hence enhancing climate change mitigation. However, this finding may be limited due to several factors:

• (a)

  It was observed that SOC content in the rice paddy wetland was less than half that obtained in any of the three vegetation communities of the natural wetland (Table 1). This indicates that the conversion of the natural wetland into the rice paddy wetland resulted in the decomposition of organic matter, and the consequent release of carbon (either as CO₂ or CH₄) into the atmosphere. Certainly, in the same wetland, Were et al. (2021b) reported significantly high carbon dioxide emission in the wetland section under rice cultivation, compared to the natural section. Therefore, whereas undertaking rice cultivation under continuous flooding and no fertilization may minimize N₂O emission associated with rice paddies, it enhances the release of soil carbon into the atmosphere.

• (b)

  Different rice varieties have differing water requirements, with some requiring drying and wetting cycles. As a result, continuous flooding may not be feasibly applied across all rice paddies due to differences in rice varieties grown. Indeed, studies have found higher rice yields in rice cropping systems subjected to intermittent flooding (Pascual & Wang 2016; Isnawan et al. 2022). Yet, intermittently flooded systems have also been reported to be hotspots for N₂O emission (Liengaard et al. 2013).

• (c)

  In agricultural systems such as rice paddies, soil fertility decreases as the period of crop cultivation increases, implying that unfertilized rice cultivation may not be sustainable in the long run. Long-term studies show that nitrogen fertilization in rice farms significantly increases rice yields ranging from 81.9 to 92.3% (Liao et al. 2010; Lu et al. 2015). As a result, in a bid to maintain or increase rice yields, the application of fertilizers in rice paddies may be inevitable as a response to declining soil fertility.

Based on these factors, rice cultivation under continuous flooding and no fertilizer application is not a sustainable option for enhancing climate change mitigation in the long run. Therefore, wetland management, in view of climate change mitigation should give priority to the conservation/protection of existing natural wetlands. It has already been shown that undrained wetlands represent net carbon (Mitsch et al. 2013) and nitrogen (Tangen & Bansal 2022) sinks. Alongside climate change mitigation, wetland conservation/protection will guarantee the availability of several other ecosystem services that are necessary for adaptation to climate change impacts.

In a continuously flooded natural tropical freshwater marsh wetland, N₂O emission does not vary spatially based on changes in the type of vegetation community. Further, it was also found that N₂O emission did not vary between the natural and rice paddy wetlands. Therefore, the conversion of tropical freshwater marsh wetlands into continuously flooded and unfertilized rice paddy wetlands does not affect N₂O emission. Nevertheless, the observation of significantly lower SOC content in the rice paddy wetland (less than half that measured in the natural wetland) implies that rice paddy wetlands even under continuous flooding and no fertilizer application can be significant sources of carbon. This, therefore, calls for prioritizing the conservation/protection of existing natural wetlands to enhance climate change mitigation by wetlands.

Temporal variations, due to changes in season (dry vs wet) have no effect on N₂O emission from continuously flooded tropical wetlands. Environmental parameters showed no significant correlation with N₂O flux from the wetlands, indicating that N₂O emission in flooded tropical wetlands is controlled by an interplay between several factors, with no single factor exerting a dominant influence.

We thank the University of Natural Resources and Life Sciences (BOKU), Vienna, Austria, and the International Foundation for Science (IFS) for the financial support during this study. We also acknowledge the immense support rendered by the local community members in the study area during the period of field sampling.

This work was supported by the University of Natural Resources and Life Sciences (BOKU), Vienna, Austria, through Mag. Gerold Winkler and Prof. Thomas Hein, and the International Foundation for Science (IFS).

The study was conceptualized and designed by DW. DW carried out data collection and analysis, and wrote, edited and reviewed the first draft manuscript. TH and FK edited, reviewed and contributed to the discussion of the manuscript. Funding acquisition was by TH and DW. The study was supervised by TH and FK. All authors read and approved the final manuscript.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.