Multiple approaches are needed to decrease greenhouse gas emissions and reduce the pace of climate change. Wetlands are among the most carbon-rich ecosystems on the planet, and when disturbed, they can generate disproportionately high emissions. Invasive hoofed mammals, such as feral pigs (Sus scrofa), cause significant soil disturbances by trampling, grubbing, and digging in wetlands. We tested whether soil disturbances by feral pigs would increase greenhouse gas emissions (carbon dioxide, CO2, methane, CH4, and nitrous oxide, N2O) as a result of reduced soil oxygen and plant cover, and increased nitrogen. Six paired sites were sampled in Kakadu National Park, northern Australia, a site of immense cultural and natural significance. Fluxes of CH4 were significantly different among treatments, with emissions being higher at disturbed plots (median (IQR) of 663 (845) µg m−2 h−1) compared to reference plots (10 (342) µg m−2 h−1). The most notable differences were observed for N2O, with significantly higher emissions from disturbed plots (155 (270) µm−2 h−1), which were up to an order of magnitude higher than those from the reference plots (13 (11) µg m−2 h−1). Soil surface soil Eh values were correlated with emissions in plots disturbed by pigs, with low values associated with high CH4 emissions. The highest emissions were found in recently disturbed sites. This study provides a compelling example of how animal populations can significantly impact greenhouse gas fluxes at the landscape scale. It also provides evidence for the viability of a carbon methodology to reduce greenhouse gas emissions through feral pig management, supporting both culture and nature.
Australian Governmentn/aAustralian Research CouncilFT40100621Introduction
As climate change and other human impacts on landscapes intensify, multiple approaches are needed to decrease greenhouse gas emissions. First and foremost, reducing fossil fuel use is required; additionally, nature-based solutions that reduce emissions while increasing ecosystem and societal benefits are possible (Matthews et al., 2022). One potential solution is the restoration and improved management of wetlands. Wetlands are one of the most carbon-rich ecosystems on the planet; their high soil carbon content and low oxygen conditions favour sequestration but can also lead to potentially high methane (CH4) and nitrous oxide (N2O) emissions. These potent greenhouse gases are emitted in large quantities in wetlands, mostly when they are polluted (Rosentreter et al., 2021; Saunois et al., 2020). Emissions from wetlands increase when soils are disturbed, for example, by the trampling, grubbing, and digging of invasive ungulates or hoofed mammals (O'Bryan et al., 2021; Treby and Grover, 2023). Thus, managing invasive species could decrease greenhouse gas emissions and improve wetland carbon sequestration (Rowland and Lovelock, 2024; Schmitz et al., 2014).
In tropical Australia, the management of invasive ungulates, such as water buffaloes (Bubalus bubalis), horses (Equus caballus), and feral pigs (Sus scrofa), is a huge challenge (Bradshaw et al., 2007). Pigs are especially abundant, with the latest estimates of around 3.2 million pigs across Australia and a mean density of one pig per square kilometre (Hone, 2020). Feral pigs are common in wetlands, where they cause extensive soil damage, disperse weeds, and prey on native species (Ballari and Barrios-García, 2014; Waltham and Schaffer, 2018). Introducing feral pigs into Australia has had catastrophic effects on biodiversity and ecosystem functions. For instance, pigs consume large numbers of native animals, such as freshwater turtles (Doupé et al., 2009). Meanwhile, feral pigs have become prey to native saltwater crocodiles (Crocodylus porosus), which currently obtain a significant portion of their nutrition from these invasive animals (Adame et al., 2018). Finally, feral pigs are potential carriers of human diseases such as Japanese encephalitis, which is transmitted by mosquitoes (Bradshaw et al., 2007). To date, controlling the numbers of feral ungulates has been intermittent and controversial, as some people consider them an important food source (Bradshaw et al., 2007). Long-term planning with cost-effective strategies is necessary to effectively mitigate their environmental and human impacts (Hamnett et al., 2024).
Carbon markets can provide long-term funding for projects that reduce greenhouse gas emissions, increase carbon sequestration, and protect carbon stocks. In Australia, a national carbon market exists, allowing for the generation of credits (Australian Carbon Credit Units, ACCUs) from projects in agriculture, energy, landfill and waste management, mining, transport, and vegetation (Clean Energy Regulator (CER), Australian Government, 2024). The latter includes nature-based projects such as vegetating landscapes, restoring tidal wetlands, and managing fires (CER, 2024). A methodology that reduces greenhouse gas emissions by managing invasive ungulates could provide cultural and economic benefits while reducing environmental impacts. However, the emissions from pig disturbance of wetlands are still uncertain (median of 4.9 million Mt CO2 per year and a wide range from -30.3 to 94 million Mt CO2), and further investigations assessing the climate benefits of feral pig management are required (O'Bryan et al., 2022).
This study was conducted in Kakadu National Park in northern Australia (Fig. 1). The park has immense cultural and natural values, being one of the last unmodified tropical rivers in the world. Kakadu has been home to First Nations People (Traditional Owners) for more than 65 000 years (Clarkson et al., 2017), being the oldest living culture on Earth with a strong spiritual connection to “Country”, which refers to the lands, waterways, and seas (physical, cultural, and spiritual) with which they are associated. The Traditional Owners of Kakadu are the Bininj in the north and Mungguy in the south, and together, they manage and lead research within the park (Robinson et al., 2022).
(A) Location of Kakadu National Park in northern Australia, and sampling sites (north to south): Munmarlay, Coolaboo, Mamukala in the South Alligator and Ubirr, Djabulukgu and Dja Dja in the East Alligator, (B) floodplain wetlands of the Magela Creek adjacent to the East Alligator River. Grey areas represent the extent of the floodplain (modified from Ward et al., 2014). Picture by M. F. Adame.
We measured the effect of soil disturbance in wetlands by feral pigs on greenhouse gas emissions (carbon dioxide, CO2, CH4, and N2O) by comparing (1) six paired sites distributed across two floodplains with and without soil pig disturbance, (2) variability from time of disturbance (recent and past), and (3) variability within two seasons: Yekke, the cool post-monsoon season, and Kunumeleng, the dry and hot pre-monsoon season. We hypothesised that pig disturbance in wetlands would alter soil physicochemical properties. We predict that nitrogen inputs will increase due to animal excretion (Krull et al., 2013), and that soil oxygen (and thus its reduction-potential) will decrease due to the removal of vegetation and increased microbial activity (Doupé et al., 2009; Liu et al., 2020). We expect these changes to result in increased soil CH4 emissions and incomplete nitrification-denitrification, which increases N2O emissions (Anderson and Levine, 1986). Finally, we predict that emissions would be highest immediately after disturbance (Liu et al., 2020), especially during the hotter Kunumeleng season, as higher temperatures accelerate CH4 production (Yvon-Durocher et al., 2014).
MethodsStudy site and sampling design
Kakadu National Park is bounded by the Van Dieman Gulf to the north, and the Kakadu escarpment to the south; three main rivers flow through the park, Wildman River, and South and East Alligator (Fig. 1). The region has a monsoonal climate with a total mean rainfall of 1547 mm, mostly falling between December and March (1383 mm); annual mean temperature ranges between 22.7 and 34.4 °C (1971–2021; Bureau of Meteorology, BM, Jabiru Airport, Station 14198). The climate is classified into six seasons according to Traditional Knowledge: Kudjewk or monsoon season (December to March), Bangkerreng or storm season (April), Yekke or cool and humid season (May to mid-June), Wurrkeng or cold season (mid-June to mid-August), Kurrung or hot dry season (mid-August to October), and Kunumeleng or pre-monsoon storm season (mid-October to late December). During the Kudjewk, the rivers that run through Kakadu overflow and inundate adjacent low-lying areas where sediments carried by rivers are deposited, and dense vegetation grows (Ward et al., 2016; Erskine et al., 2018). These floodplains remain inundated for about five months per year (Ward et al., 2014). During the drier season, flooding recedes, and permanent deep-water holes remain in the floodplain until the next wet season, when they are once again reconnected to the river channel.
We conducted two field campaigns: one during Yekke (8–14 June 2023), the cool season, and one during Kunumeleng (25–30 October 2023), the hot, pre-monsoon season. Both sampling trips were preceded by dry conditions with no precipitation in the previous 30 d and mean daily temperatures ranging from 33–36 °C during sampling in Yekke and 38–40 °C in Kunumeleng (Jabiru Airport Station, 14198, Australian Bureau of Meteorology). Sampling in other seasons was not possible due to the remoteness of the area, extreme heat, and the dangers posed by saltwater crocodiles (Crocodylus porosus) during the wetter seasons. For our first trip, six sites were selected by the Traditional Owners (co-authors of this study). The selected sites were Munmarlay, Coolaboo, and Mamukala in the South Alligator, and Ubirr, Djabulukgu, and Dja Dja in the East Alligator (Fig. 1).
The vegetation of the sampling sites included floodplain grassland with scattered broad-leaved paperbark (Melaleuca viridiflora) at Dja Dja and Ubirr, paperbark forests at Djabulukgu, Coolaboo, and Munmarlay (M. viridiflora, Melaleuca cajuputi), and freshwater mangroves (Barringtonia acutangula) at Mamukala. Away from the open water, the wetlands were surrounded by savanna or pandanus (Pandanus spiralis) woodland. We sampled two plots at each site: a reference plot, either a paperbark forest or a grass-dominated area (Coolaboo) which was relatively undisturbed and a plot nearby with soil disturbances (Fig. 2). The disturbed plots were either “digging areas” that pigs use for finding roots and other underground food (e.g., turtles) or “wallows” which were deeper and are used by pigs to cover themselves with mud to avoid overheating (Bracke, 2011). Although the soil disturbances were mostly caused by pigs, feral water buffalo (Bubalus bubalis) may have also contributed, as buffalo wallows and tracks were observed at Dja Dja, Ubirr, and Munmarlay.
Examples of (A) a reference or undamaged plot at Djabulukgu, (B) a plot with recent disturbance at Coolaboo, and (C) plot with past disturbance, which was dry and unvegetated (Djabulukgu). Pictures by J. M. Hall.
On our second trip during the Kunumeleng season, we resampled three of the sites (Munmarlay, Djabulukgu, and Mamukala). But this time, targeting not only reference and disturbed plots but also areas that were previously disturbed and were now unvegetated and dry (Fig. 2). The three treatment plots (reference, recent, and past disturbance) were as close to each other as possible, usually less than 50 m away, except for our plot in Munmarlay, where we could not access a recently disturbed plot, and was paired with a reference plot in nearby Coolaboo (Fig. 1).
Soil physicochemical characteristics
At each plot, surface (0–2 cm deep) reduction-oxidation potential (3–5 measurements) and soil temperature were measured (HQ 11d, ORP-meter, and HQ2100 ORP-meter, with a MTC101 probe with an Ag / AgCl reference HACH, Pacific). Redox values are reported relative to a Standard Hydrogen Electrode (SHE; Eh) by adding 207 mV (at 25 °C) as recommended by the manufacturer. Soil Eh from reference and past disturbed sites during the Kunumeleng season could not be measured as the soil was completely dry. Three surface (0–5 cm) soil samples of a known volume were taken at each plot, weighed, and oven-dried at 50 °C. They were then weighed again, from which the water content and bulk density were calculated. The samples were analysed for organic carbon (C, after testing for inorganic carbon, through HCl addition), nitrogen (N), δ13C, and δ15N (from acidified and unacidified samples, respectively; EA-IRMS, Serco System at Griffith University). The particle size distribution of surface sediments (0–5 cm deep) was assessed from three subsamples from each plot, which were homogenised into a single sample and sieved through meshes of the following sizes: <20µm (fine silt and clay), >20µm (medium and coarse silt) ≤, >63µm (very fine sand) ≤, >125µm (fine, medium and coarse sand) ≤, and >2 mm (gravel). The contents of each sieve were weighed, and their contribution to the total sample was calculated. Finally, soil conductivity was measured with a conductivity meter (ProDSS YSI, OH, USA, or HACH) from a soil-to-water (1:5) solution adjusted to the dominant grain size class of each plot.
Greenhouse gas emissions
Instantaneous gas fluxes (n=5 per plot) were measured in dark, closed chambers (Hutchinson and Mosier, 1981) made of polyvinyl chloride (4 L, 15 cm diameter and 30 cm in height; Fig. 2). The chambers were installed right before the measurements, hammered into the soil at least 3 cm deep, acknowledging some disturbance to the soil. After deployment, 20 mL of gas samples were collected from the headspace of the chambers using a plastic 30 mL syringe with a 26 G needle and transferred to 12 mL vacuumed glass containers (flat-bottom, 103 mm length, 15.5 mm diameter, Labco Exetainer). The air samples were collected at 0 and 60 min at all chambers and at 0, 20, 40, and 60 min in one chamber per plot. During each sampling, the headspace was mixed with the syringe three times before the air sample was retrieved.
Back in the laboratory, the CO2, CH4, and N2O concentrations within each sample were measured in a gas chromatograph (Shimadzu GC-2010 Plus, DETSI, Queensland Government). The analyses had a precision of 0.01 ppm and accuracies of 0.2 %, 0.5 %, and 1.5 % for CO2, CH4, and N2O, respectively. From all the samples analysed (n=245), 21 were above the highest standard for CH4 (all in sites disturbed by pigs), but within the ranges expected from tropical wetlands, so were retained in our calculations. Only two plots for the emission value of CH4 in the disturbed site by pigs were excluded from the calculations as they were an order of magnitude higher than the rest of the chambers (Djiru Chamber D during Yekee season and Coolaboo, Chamber D, during Kunumeleng season, see Supplement). Air temperature next to the chambers was monitored throughout the experiment with a handheld thermometer. The chamber with continuous measurements was checked for linearity using a regression of time for each gas within each plot. Coefficients for each regression are shown in Table S1; chambers with poor fit (R2<0.50) were identified at sites with very low N2O fluxes or very low and high CH4 emissions. Because CH4 fluxes in tropical freshwater wetlands are characterised by nonlinearity and spikes due to soil ebullition (Jeffrey et al., 2019), poorly linearly fitted plots were retained in the calculations, acknowledging uncertainty at the upper and lower ends of CH4 emissions. Nevertheless, to remain conservative, we used median values and interquartile ranges, which remove the effect on very low and very high fluxes. Fluxes (F) were calculated per hour from the net changes in gas concentrations (N, mol) during the incubation time (t, hours), per volume of the chamber (V, m3), at a given air temperature (T, °K), assuming a pressure of one atmosphere:
F=NRT×VtA
where R=8.205×10-5 m3 atm K−1 mol−1.
Root respiration
Soil CO2 fluxes are a combination of the decomposition of organic matter and dead roots and the respiration of live roots, the latter contributing up to 79 % of the total (Krauss et al., 2012). To test the contribution of root respiration to soil CO2 emissions, we compared CO2 emissions from the unvegetated and vegetated plots measured during the Kunumeleng season. Additionally, we measured root biomass at each reference and disturbed plot by sampling the surface (top 10 cm) with a sediment corer (7.3 cm diameter). On the same collection day, the sediments were washed and sieved through a 1 mm mesh, and roots were hand-picked, blotted dry, and weighed. The CO2 flux from each pooled sample (coarse and fine roots) was measured in the dark with a portable gas analyser (Picarro, GasScouter, California, USA). The following week, back in the laboratory, the roots were oven-dried at 60 °C and reweighed to measure their biomass. Root respiration was converted to aerial emissions, considering the area of root sampled (22.9 cm2). We acknowledge the limitations of this approach, as root respiration was measured once roots were extracted from the soil, without accounting for the effects of the aboveground vegetation and the soil profile vegetation, which could consume CO2 through photosynthesis and other non-photosynthetic pathways (Rodriguez et al., 2025). Furthermore, direct root respiration cannot be distinguished with certainty from organic matter decomposition without the use of additional tracers, such as 14C (Trumbore et al., 2006). Thus, our estimations on root respiration contributions to CO2 emissions should be interpreted as a first-step calculation.
Emission reduction potential
To estimate the carbon abatement potential of feral pig management, CH4 and N2O fluxes were converted to CO2eq by multiplying them by their 100-year warming potential, which is 27 for CH4 (non-fossil origin) and 273 for N2O (IPCC, 2021). Fluxes of CO2 were included after being corrected for root respiration.
Statistical analyses
Data was tested for normality with Shapiro-Wilk and Kolmogorov-Smirnov tests, for datasets that complied with normality, a two-way nested ANOVA with main effects (disturbance) nested within site was conducted to test differences between disturbed (recent and past) and reference plots across the six sites for gas fluxes (CO2 and N2O) and physicochemical parameters (soil conductivity, temperature, Eh, bulk density, soil C, N, C : N, δ15N and δ13C). For datasets that did not comply with normality (CH4 fluxes), comparisons among treatments were conducted with a non-parametric Wilcoxon matched-pairs signed-rank test. A repeated-measurements ANOVA was used to test for differences in gas fluxes between the two seasons. Regression and non-linear model fits were conducted to assess the response of greenhouse gas fluxes to physicochemical parameters. Statistics were analysed with GraphPrims 10.2.3. Values are mean ± standard error (min–max) unless specified otherwise.
ResultsSoil physicochemical characteristics
During Yekke (cool season), soil conductivity, temperature, and Eh were significantly higher at the reference plots (180±74.3µS cm−1, 27.8±1.7 °C, 413±54 mV) compared to the disturbed plots (56.7±7.6µS cm−1, 25.3±0.8 °C, 275±16 mV; F1,24=580.6, p<0.001, F1,47=424.7p<0.001 and F1,24=26.19p<0.001, respectively; Table 1). In contrast, water content and bulk density were similar between treatments (58.0±4.6 % and 0.46±0.05 g cm3 versus 53.2±5.2 % and 0.45±0.0805 g cm−3p-values for reference and disturbed plots, respectively; Table 1).
Soil physicochemical characteristics of six reference plots paired with six recently disturbed plots by feral pigs during Yekke (cool season). Soil conductivity, temperature, and soil oxido-reduction potential (Eh) were significantly higher at the reference compared to the disturbed plots (F1,24=513.2p<0.01, F1,47=424.7p<0.01, and F1,24=26.19p<0.001). Significant differences are shown as p<0.05* and p<0.001**.
Surface soil C and N, and molar C : N were also similar between reference (8.1±1.1 % C and 0.63±0.09 % N, 9.7±0.9) and disturbed plots (8.5±1.4 % C and 0.68±0.10 % N, 10.8±0.7; p=0.480, p=0.312, and p=0.251, respectively). Soil δ15N did not differ between treatments, with 2.73±0.60 ‰ for reference compared to 2.74±0.17 ‰ for disturbed plots (Table 2). However, δ13C was slightly but significantly lower at the reference plots (-22.6±1.7 ‰) compared to the disturbed ones (-21.4±1.7 ‰; F1,24=8.66, p=0.007; Table 2). All sites had different sediment grain size composition; Mamukala, Dja Dja and Djabulukgu were dominated by sand (fine, medium and coarse), Ubirr was dominated by silt (medium, coarse and fine) and clay, while Munmarlay and Coolaboo were dominated by fine silt and clay (Table 3).
Soil carbon, nitrogen, molar C : N, isotopic composition (δ15N and δ13C) and bulk density of six reference plots paired with six recently disturbed plots by feral pigs. Reference and disturbed plots were similar for all parameters except for δ13C, which was significantly lower in the reference plots (F1,24=8.66, p=0.007). Significant differences are shown as p<0.05* and p<0.001**.
C (%) N (%) C : N ReferenceDisturbedReferenceDisturbedReferenceDisturbedUbirr9.7±0.511.6±2.50.9±0.10.9±0.19.7±0.310.7±0.7Munmarlay7.4±0.88.2±1.80.6±0.10.7±0.19.8±0.29.9±0.7Mamukala6.3±0.84.2±0.50.4±0.00.3±0.015.0±1.413.1±0.6Coolaboo8.3±1.06.1±0.50.7±0.10.5±0.09.6±0.210.1±0.0Dja Dja12.3±1.513.2±0.90.8±0.10.9±0.112.9±0.312.5±0.2Djabulukgu4.3±0.37.8±0.30.4±0.00.8±0.010.3±0.48.6±0.1δ13C (‰) δ15N (‰) Bulk density (g cm−3) ReferenceDisturbedReferenceDisturbedReferenceDisturbedUbirr-17.9±1.5-17.4±0.41.2±0.52.2±0.70.30±0.020.27±0.03Munmarlay-23.4±0.6-21.1±0.9*3.6±0.52.6±0.60.50±0.040.45±0.05Mamukala-28.8±0.5-26.7±0.73.0±0.53.2±0.50.56±0.060.79±0.08Coolaboo-21.1±0.6-18.1±0.8**1.8±0.13.2±0.30.35±0.020.55±0.02Dja Dja-25.3±0.2-26.1±0.41.7±0.62.9±0.10.63±0.020.39±0.03Djabulukgu-19.2±0.5-19.0±0.25.1±0.52.3±0.20.44±0.020.27±0.01
Soil grain size composition (%) of six reference plots (R) paired with six recently disturbed plots by feral pigs (D). Bold values highlight the dominant grain size per plot (reference/disturbed) and site.
Gravel Fine, medium, Very fine Medium and Fine silt coarse sand sand coarse silt and clay RDRDRDRDRDUbirr4.00.93.84.929.026.435.619.927.547.9Munmarlay1.21.28.828.910.56.015.013.764.450.1Mamukala6.96.561.363.711.07.810.86.110.015.9Coolaboo5.30.017.21.825.40.017.22.435.095.8Dja Dja4.21.246.251.615.512.711.017.023.117.5Djabulukgu12.40.555.331.97.518.56.79.918.139.2Greenhouse gas emissionsImpact of pig disturbance
During Yekke or the cool season, CO2 emissions were significantly different among sites (F5,48=3.88p=0.005) and similar between disturbed and reference plots (F1,48=2.90p=0.095; Fig. 3A). Fluxes of CH4 were highly variable, but marginally higher (W=15: n=49, p=0.063) at disturbed plots (1219±740, 54 to 4823 µg m−2 h−1) compared to reference plots (93±90, -8.6 to 454 µg m−2 h−1), at all sites except Coolaboo (Fig. 3B), where the reference plot was not a Melaleuca forest but a grass-dominated wetland. The most significant differences were observed for N2O, with emissions at disturbed plots (F1,48=32.2, p<0.001) of 222±84.1 (26.6 to 549) µg m−2 h−1, which were up to an order of magnitude higher than those for the reference sites with 11.9±3.0 (2.7 to 20.4) µg m−2 h−1 (Fig. 3C).
Greenhouse gas emissions (CO2, CH4 and N2O) from reference plots (blue–left) and plots disturbed by feral pigs (red–right) for six sites across the floodplains during the Yekke season in Kakadu National Park: Ub = Ubirr, Mu = Munmarlay, Ma = Mamukala, Co = Coolaboo, Ja = Djaja, Dj = Djabulukgu. The box plots show the mean and standard error for five chambers; whiskers represent the minimum and maximum values.
Variability from the time of disturbance
The time since soil disturbance, whether the plot was past or recently used by pigs, was assessed during Kunumeleng or the pre-monsoon season. At this time, CO2 emissions were highest at the reference plots, compared to those that were past and recently disturbed (F2,36=6.84p=0.003; Fig. 4A). The time of disturbance significantly affected the CH4 emitted, with higher emissions, although not significantly different, at the newly disturbed plots (F2,35=2.0p=0.15) compared to plots with past disturbance and reference sites (Fig. 4B). Finally, N2O emissions were similar among treatments (F2,36=0.93p=0.40). However, there was a trend of higher N2O emissions at the recently disturbed sites (Fig. 4C).
Greenhouse gas (CO2, CH4; and N2O) fluxes of reference sites (REF – blue), sites recently used by pigs (RECENT – red) and sites previously used by pigs (PAST – orange) Violins represent the distribution of the average of five chambers within three sites (Djabulukgu, Mamukala, and Munmarlay) during Kunumeleng (pre-monsoon) season.
Variability between seasons
The CO2 emissions were significantly lower during the Kunumeleng (pre-monsoon) season compared to the Yekke (cool) season (F1,8=45.1; p=0.0001; Fig. 5A). However, CH4 and N2O emissions were not significantly different (F1,8=3.66; p=0.0 and F1,8=1.85; p=0.21, respectively; Fig. 5B, C). Surface soil Eh of the disturbed plots was lower during Kunumeleng (mean of three sites of 219±251 mV) than during Yekee (mean of five sites of 275±223 mV).
Greenhouse gas (CO2, CH4 and N2O) fluxes for two seasons: Yekke or cool season (blue–left) and Kunumeleng or pre-monsoon season (orange–right), for three sites without (REF-reference) and with feral pig disturbance (PIGS). Violins represent the distribution of the average of five chambers at three plots within three sites (n=3; Djabulukgu, Mamukala, and Munmurlay-Coolaboo).
Physicochemical factors associated with emissions
Two parameters analysed were strongly associated with greenhouse gas fluxes: pig disturbance and soil Eh. Soils disturbed by pigs had lower Eh than reference soils or soils from plots with past disturbance. Low Eh values were correlated with high CH4 emissions, and although the correlation was significant, the regression was poorly fitted to a linear trend (Fig. 6; R2=0.087, p=0.012). Additionally, Eh values approximately between 200–300 mV were associated with increased N2O emissions (Fig. 6B), a response that was best fitted to a Gaussian curve; however, the fit was modest and had low statistical confidence (R2=0.21n=45, Fig. 6). Nevertheless, a clear pattern was observed between soil Eh and emissions; anaerobic soil with low Eh values promoted CH4 emissions and soils with Eh values between approximately 200–300 mV promoted N2O emissions, but only in plots disturbed by pigs (Fig. 6).
Relationship between soil Eh (mV) and (A) CH4 and (B) N2O fluxes (µg m−2 h−1) from reference (blue diamonds) and disturbed plots (red circles) by feral pigs. Low Eh (<200 mV) values were associated with the highest CH4 emissions, and values between 200 and 300 mV were associated with high N2O emissions, but only at plots disturbed by pigs. Shaded areas represent Eh conditions that favoured CH4(<200 mV; A) and N2O (∼200–300 mV; B) emissions.
Root versus soil respiration
Root biomass was variable, but, on average, higher in the reference plots than in those disturbed by pigs (84±63 vs. 17±3 mg cm−3; Table 4). Root respiration was higher than soil CO2 emissions measured directly in the field, suggesting that most of the soil CO2 emissions in the reference plots originated from root respiration (Table 4). This assumption was further supported when comparing vegetated and unvegetated plots during the Kunumeleng season, with 61.5 % higher emissions in the vegetated plots (130±44 (51–204) mg CO2 m−2 h−1) compared to the unvegetated plots (50±52 (-47 to 129) mg CO2 m−2 h−1).
Root respiration and biomass from the top 10 cm during the Kunumeleng in reference and disturbed plots by feral pigs. n.a. = Not available.
The potential reduction in greenhouse gas emissions from feral pig management was estimated per site during Yekke, when the most spatially comprehensive sampling was conducted. We found that avoiding soil disturbance could result in potential CO2 reductions (corrected for root respiration by 61.5 %) at all six sites, CH4 reductions in five out of six sites, and N2O reductions at all sites (Table 5). When transforming the gases by their warming potential (CO2-eq), reductions (propagated error) were highest for CO2 with 378±136, followed by N2O with 39±74 and CH4 with 18±25 mg CO2eq m−2 h−1.
Greenhouse gas emissions (CO2, CH4, and N2O) and emission reduction potential from reference plots compared to disturbed plots by pigs in six sites in Kakadu NP during the Yekke season. Values are mean ± standard error for five chambers within each plot and median (IQR, interquartile range) for the six sites. The soil CO2 emissions were corrected for root respiration (CO2-resp), which contributed approximately to 61 % of the emissions.
The iconic wetlands of Kakadu National Park have been managed for millennia by the Bininj and Mungguy peoples. However, they currently face multiple challenges, including rising sea levels, erratic rainfall patterns, and the impacts of invasive species (Bradshaw et al., 2007; Pettit et al., 2018). In this study, we found that beyond the known impacts of feral pigs on biodiversity (Risch et al., 2021), they also increase soil greenhouse gas emissions, confirming that feral pigs pose a serious threat to soil carbon and other biogeochemical cycles (O'Bryan et al., 2021). This study provides another compelling example of animal populations significantly affecting the carbon cycle at the landscape scale (Schmitz et al., 2014).
The effect of feral pigs was evident and widespread at all sites, where the soil was turned over and mostly devoid of vegetation. The disturbed plots were also colder and filled with fresh groundwater, as pigs used some of these areas to wallow or coat themselves with mud to avoid overheating (Bracke, 2011). The soil Eh potential at all disturbed plots was significantly reduced, suggesting low oxygen conditions. Low Eh soils favour microbial respiration through nitrification-denitrification when nitrate (NO3−) is available. This process produces N2O as a byproduct, especially in soils high in nitrogen, such as those enriched with pig excreta (Krull et al., 2013). However, soil N content in our plots were similar in reference and disturbed plots, suggesting that nitrogen enrichment from feral pigs was sporadic and was rapidly removed, potentially through denitrification, which can result in N2O emissions. Due to potential low oxygen and high nitrogen loads, N2O emissions during the Yekke season were up to an order of magnitude higher than those from the reference plots.
As the dry season progresses, the weather becomes hotter, and the soil becomes drier. Consequently, pigs move closer to the waterhole fringes to exploit resources at the edge of the open water (Froese et al., 2017). During this season, understorey vegetation and litter were scarce, limiting organic carbon inputs into the soil and potentially causing reduced CO2 fluxes compared to Yekke. The wallows left behind had soils with very low Eh values. The lack of oxygen in the soil during this time of the year resulted in high CH4 emissions, but low N2O emissions. This result can be explained by the fact that denitrification is energetically favoured over methanogenesis, but once nitrogen is exhausted and oxygen is depleted, methanogenesis becomes the dominant pathway (Reddy and DeLaune, 2008). These results complement previous studies showing the effect of hoofed animals in increasing CH4 emissions in wetlands (Treby and Grover, 2023).
Once the disturbed site was completely dried up and devoid of vegetation, the C and N from the soil were exhausted, and fluxes decreased (Fig. 7). At this point, the CO2 emissions reached their minimum values because CO2 is the net result of two processes: root respiration and soil organic matter decomposition. During the Yekke season, reference sites were fully vegetated and had large amounts of litter accumulated on the soil surface (see Fig. 2A); thus, root respiration and litter decomposition caused high CO2 emissions. The disturbed plots had low vegetation cover and little standing litter but likely had high decomposition rates, driven by dead root material and exposed soil organic matter, resulting in CO2 emissions similar to or higher than those of the reference plots. During the Kunumeleng season, the plots had dried up and had little vegetation or standing litter, resulting in the lowest CO2 emissions measured. Finally, we expect that during Kudjewk, or the monsoon season, the floodplains will overflow again, providing fresh habitat for feral pigs, which will continue to disturb the soil and release the carbon accumulated during the wetter months (Fig. 7).
Schematic diagram on the greenhouse gas emission changes in a floodplain wetland (A) before pig disturbance; (B) during pig disturbance as vegetation is lost, surface soil Eh decreases, and soil decomposition and nitrogen inputs increase, resulting in very high CO2, CH4 and N2O emissions; (C) as the site is less used as animals move closer to the remaining waterholes during drier months, N2O emissions decrease but CH4 emissions remain high as oxygen is very low; (D) finally, once the site is abandoned, the soil is depleted of carbon and nitrogen, surface soil Eh potential increases, and fluxes are minimal; (E) potential pathway of recovery if feral pigs are managed.
Despite the strong evidence of increased emissions from pig disturbance in this study, there are limitations for extrapolating our results to annual rates. First, greenhouse gas emissions could not be measured during the wettest months due to the hazardous conditions at the location, particularly the presence of saltwater crocodiles. Emissions from Melaleuca wetlands during flooding periods in the region are higher for CH4 but lower for CO2 (Bass et al., 2014), and year-to-year variations in emissions are likely under different pre-flooding conditions, with lower emissions in floods preceded by dry periods (Jeffrey et al., 2019). Second, the contribution of root versus soil respiration remains unresolved, given the confounding effects of CO2 production from root respiration and soil decomposition (O'Bryan et al., 2022). We used an approximate value of 62 % contribution of root respiration in our measurements, consistent with other studies (79 %; Krauss et al., 2012). However, future studies could refine this estimate, facilitating the inclusion of CO2 in carbon emissions from soil disturbance.
We cautiously estimate annual emissions from pig disturbance using our data, supplemented by published estimates of emissions during flooding (Bass et al., 2014). We consider emissions from flooded Melaleuca swamps to be representative of reference sites and open-water areas without vegetation to represent sites damaged by pigs. Seasonal differences were included in our estimations, considering that the wet seasons Kudjewk and Bangkerrent span from 20 December to 30 April, the cool seasons, Yekke and Wurrkeng span from 1 May–15 August (data from this study), and the hot dry seasons, Kurrung and Kunumeleng seasons, span from 16 August to 19 December (data from this study). We estimate that annual emission reductions by comparison (reference vs disturbed) are 13.2 Mg CO2 ha−1 yr−1, 1.1 Mg CH4 ha−1 yr−1, and 0.004 Mg N2O ha−1 yr−1, equivalent to reductions of 43.4 Mg CO2-eq ha−1 yr−1. We can then extrapolate these emissions to the Kakadu and adjacent river floodplains (see Fig 1; 356 800 ha, Ward et al., 2014), considering that about 10 % of the area is disturbed by pigs (unpublished data). We estimate emissions of 0.47, 1.04, and 0.04 Mt CO2-eq yr−1 for CO2, CH4, and N2O, respectively, or a total of 1.5 Mt CO2-eq yr−1. These levels of emissions are equivalent to 5.6 % of the Northern Territory's emissions in Australia, or 57.4 % of the Territory's agricultural emissions. Although we acknowledge the limitations of our calculations and the challenge of removing the thousands of pigs inhabiting these wetlands, it is clear that there is significant potential to reduce greenhouse gas emissions through feral pig control.
In conclusion, we demonstrate that feral pigs pose a significant threat to soil carbon integrity and significantly modify soil biogeochemical processes, increasing greenhouse gas emissions. These effects, measured in this study in tropical floodplains in Australia, are likely to occur in wetlands worldwide. Physical activity in the soil and the increase in nitrogen from pig excreta are significantly increasing emissions, especially of CH4 and N2O. Tackling these two potent greenhouse gases exacerbated by feral pigs has the potential to significantly contribute to climate change mitigation (Ocko et al., 2021), while also improving biodiversity and culture.
Data availability
All greenhouse gas data is provided as supplement, but it will also be stored here: https://research-repository.griffith.edu.au/home (last access: 4 August 2026; DOI: 10.25904/1912/6008, Adame et al., 2026).
The supplement related to this article is available online at https://doi.org/10.5194/bg-23-5781-2026-supplement.
Author contributions
MFA, CS, JN, VH, and CEL designed the study; MFA, AP, JMH, JT, CS, and JN carried out field experiments; MFA and JH analysed the data; MFA wrote the original manuscript, and all co-authors contributed to the editing, interpretation, and conclusions.
Competing interests
The contact author has declared that none of the authors has any competing interests.
Disclaimer
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. The authors bear the ultimate responsibility for providing appropriate place names. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.
Acknowledgements
We acknowledge the Traditional Owners past, present and emerging of the lands in which we worked; the Bininj and Mungguy peoples. We are thankful for their contribution to this research and for allowing us to work together in their country. We are thankful to Kakadu National Park staff and rangers for their logistic support in the field.
Financial support
This research has been supported by the Australian Government and the National Environmental Science Program (NESP) Marine and Coastal Hub, Project 3.8. MF Adame is funded by an Australian Research Council, Future Fellowship (FT40100621).
Review statement
This paper was edited by Perran Cook and reviewed by Aldis Butlers and one anonymous referee.
ReferencesAdame, M., Jardine, T., Fry, B., Valdez, D., Linder, G., Nadji, J., and Bunn, S.: Estuarine crocodiles in a tropical coastal floodplain obtain nutrition from terrestrial prey, PLoS ONE, 13, 6, 10.1371/journal.pone.0197159, 2018.Adame, M. F., Pearse, A., Hill, J. W., Cristopherson, S., Nadji, J., Hall, J. M., Thurman, J., Hagger, V., and Lovelock, C. E.: Soil disturbance in wetlands by feral pigs increases greenhouse gas emissions, Griffith University [data set], 10.25904/1912/6008, 2026. Anderson, I. C. and Levine, J.: Relative rates of nitric oxide and nitrous oxide production by nitrifiers, denitrifiers and nitrate respirers, Appl. Environ. Microb., 51, 938–945, 1986.Ballari, S. A. and Barrios-García, M. N.: A review of wild boar Sus scrofa diet and factors affecting food selection in native and introduced ranges, Mammal Review, 44, 124–134, 10.1111/mam.12015, 2014. Bass, A. M., O'Grady, D., Leblanc, M., Tweed, S., Nelson, P. N., and Bird, M. I.: Carbon dioxide and methane emissions from a wet-dry tropical floodplain in Northern Australia, Wetlands, 34, 619–627, 2014.Bracke, M. B. M.: Review of wallowing in pigs: Description of the behaviour and its motivational basis, Appl. Anim. Behav. Sci., 132, 1–13, 10.1016/j.applanim.2011.01.002, 2011.Bradshaw, C. J. A., Field, I. C., Bowman, D. M. J. S., Haynes, C., and Brook, B. W.: Current and future threats from non-indigenous animal species in northern Australia: A spotlight on World Heritage Area Kakadu National Park, Wildlife Res., 34, 419–436, 10.1071/WR06056, 2007.Clarkson, C., Jacobs, Z., Marwick, B., Fullagar, R., Wallis, L., Smith, M., Roberts, R. G., Hayes, E., Lowe, K., Carah, X., Florin, S. A., McNeil, J., Cox, D., Arnold, L. J., Hua, Q., Huntley, J., Brand, H. E. A., Manne, T., Fairbairn, A., Shulmeister, J., Lyle, L., Salinas, M., Page, M., Connell, K., Park, G., Norman, K., Murphy, T., and Pardoe, C.: Human occupation of northern Australia by 65,000 years ago, Nature, 547, 306–310, 10.1038/nature22968, 2017.Clean Energy Regulator (CER), Australian Government: ACCU Scheme methods, https://cer.gov.au/schemes/australian-carbon-credit-unit-scheme/accu-scheme-methods (last access: 30 July 2026), 2024. Doupé, R., Schaffer, J., Knott, M., and Dicky, P.: A description of freshwater turtle habitat destruction by feral pigs in Tropical North-eastern Australia, Herpetol. Conserv. Bio., 4, 331–339, 2009.Erskine, W. D., Saynor, M. J., Boyden, J. M., and Evans, K. G.: Sediment fluxes and sinks for Magela Creek, Northern Territory, Australia, Mar. Freshwater Res., 69, 1018–1025, 10.1071/MF16107, 2018.Froese, J. G., Smith, C. S., Durr, P. A., McAlpine, C. A., and van Klinken, R. D.: Modelling seasonal habitat suitability for wide-ranging species: Invasive wild pigs in northern Australia, PlosOne, 12, e0177018, 10.1371/journal.pone.0177018, 2017.Hamnett, P. W., Saltre, F., Page, B., Tarran, M., Korcz, M., Fielder, K., Andrews, L., and Bradshaw, C. J. A.: Stochastic population models to identify optimal and cost-effective harvest strategies for feral pig eradication, Ecopshere, 15, e70082, 10.1002/ecs2.70082, 2024.Hone, J.: How many feral pigs are in Australia? An update, Aust. J. Zool., 67, 215–220, 10.1071/ZO20077, 2020.Hutchinson, G. L. and Mosier, A. R.: Improved soil cover method for field measurement of nitrous oxide fluxes, Soil Sci. Soc. Am. J., 45, 311–316, 10.2136/sssaj1981.03615995004500020017x, 1981.IPCC: Climate Change 2021: The Physical Science Basis, Contribution of Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change, edited by: Masson-Delmotte, V., Zhai, P., Pirani, A., et al., Cambridge University Press, 10.1017/9781009157896, 2021. Jeffrey, L. C., Maher, D. T., Johnston, S. G., Kelaher, B. P., Steven, A., and Tait, D. R.: Wetland methane emissions dominated by plant-mediated fluxes: Contrasting emissions pathways and seasons within a shallow freshwater subtropical wetland, Limnol. Oceanogr., 64, 1895–1912, 2019. Krauss, K. W., Whitbeck, J. L., and Howard, R. J.: On the relative roles of hydrology, salinity, temperature, and root productivity in controlling soil respiration from coastal swamps (freshwater), Plant Soil, 358, 265–274, 2012.Krull, C. R., Choquenot, D., Burns, B. R., and Stanley, M. C.: Feral pigs in a temperate rainforest ecosystem: Disturbance and ecological impacts, Biol. Invasions, 15, 2193–2204, 10.1007/s10530-013-0444-9, 2013.Liu, Y., Liu, X., Yang, Z., Li, G., and Liu, S.: Wild boar grubbing causes organic carbon loss from both top- and sub-soil in an oak forest in central China, Forest Ecol. Manag., 464, 118059, 10.1016/j.foreco.2020.118059, 2020.Matthews, H. D., Zickfeld, K., Dickau, M., MacIsaac, A. J., Mathesius, S., Nzotungicimpaye, C.-M., and Luers, A.: Temporary nature-based carbon removal can lower peak warming in a well-below 2 °C scenario, Communications Earth & Environment, 3, 1–8, 10.1038/s43247-022-00391-z, 2022. O'Bryan, C. J., Patton, N. R., Hone, J., Lewis, J. S., Berdejo-Espinola, V., Risch, D. R., Holden, M. H., and McDonald-Madden, E.: Unrecognized threat to global soil carbon by a widespread invasive species, Glob. Change Biol., 28, 877–882, 2021.O'Bryan, C. J., Patton, N. R., Hone, J., Lewis, J. S., Berdejo-Espinola, V., Risch, D. R., Holden, M. H., and McDonald-Madden, E.: Invasive wild pigs (Sus scrofa) as a human-mediated source of soil carbon emissions: Uncertainties and future directions, Glob. Change Biol., 28, e1–e3, 10.1111/gcb.15992, 2022.Ocko, I. B., Sun, T., Shindell, D., Oppenheimer, M., Hristov, A. N., Pacala, S. W., Mauzerall, D. L., Xu, Y., and Hamburg, S. P.: Acting rapidly to deploy readily available methane mitigation measures by sector can immediately slow global warming, Environ. Res. Lett., 16, 054042, 10.1088/1748-9326/abf9c8, 2021.Pettit, N. E., Bayliss, P., and Bartolo, R.: Dynamics of plant communities and the impact of saltwater intrusion on the floodplains of Kakadu National Park, Mar. Freshwater Res., 69, 1124–1133, 10.1071/MF16148, 2018.Reddy, K. R. and DeLaune, R. D.: Biogeochemistry of Wetlands, CRC Press, Taylor and Francis Group, 10.1201/9780203491454, 2008.Risch, D. R., Ringma, J., and Price, M. R.: The global impact of wild pigs (Sus scrofa) on terrestrial biodiversity, Sci. Rep., 11, 13256, 10.1038/s41598-021-92691-1, 2021.Robinson, C. J., Macdonald, J. M., Douglas, M., Perry, J., Setterfield, S., Cooper, D., Lee, M., Nadji, J., Nadji, S., Nayinggul, A., Nayinggul, A., Mangiru, K., Hunter, F., Coleman, B., Barrowei, R., Markham, J., Alderson, J., Moyle, F., May, K., and Bangalang, N.-g.: Using knowledge to care for country: Indigenous-led evaluations of research to adaptively co-manage Kakadu National Park, Australia, Sustain. Sci., 17, 377–390, 10.1007/s11625-021-01015-9, 2022.Rodriguez, K., Ricci, F., Ni, G., Iram, N., Palfreyman, R., Gonzalez-Garcia, R., Heffernan, J., Greening, C., Adame, M., and Marcellin, E.: Abundant and active acetogens enhance the carbon dioxide sink of Blue Carbon ecosystems, Microbiome, 13, 10.1186/s40168-025-02209-4, 2025.Rosentreter, J. A., Borges, A. V., Deemer, B. R., Holgerson, M. A., Liu, S., Song, C., Melack, J., Raymond, P. A., Duarte, C. M., Allen, G. H., Olefeldt, D., Poulter, B., Battin, T. I., and Eyre, B. D.: Half of global methane emissions come from highly variable aquatic ecosystem sources, Nat. Geosci., 14, 225–230, 10.1038/s41561-021-00715-2, 2021.Rowland, P. I. and Lovelock, C. E.: Global impacts of introduced ungulates on wetland carbon and biodiversity: A review, Biol. Conserv., 290, 110432, 10.1016/j.biocon.2023.110432, 2024.Saunois, M., Stavert, A. R., Poulter, B., Bousquet, P., Canadell, J. G., Jackson, R. B., Raymond, P. A., Dlugokencky, E. J., Houweling, S., Patra, P. K., Ciais, P., Arora, V. K., Bastviken, D., Bergamaschi, P., Blake, D. R., Brailsford, G., Bruhwiler, L., Carlson, K. M., Carrol, M., Castaldi, S., Chandra, N., Crevoisier, C., Crill, P. M., Covey, K., Curry, C. L., Etiope, G., Frankenberg, C., Gedney, N., Hegglin, M. I., Höglund-Isaksson, L., Hugelius, G., Ishizawa, M., Ito, A., Janssens-Maenhout, G., Jensen, K. M., Joos, F., Kleinen, T., Krummel, P. B., Langenfelds, R. L., Laruelle, G. G., Liu, L., Machida, T., Maksyutov, S., McDonald, K. C., McNorton, J., Miller, P. A., Melton, J. R., Morino, I., Müller, J., Murguia-Flores, F., Naik, V., Niwa, Y., Noce, S., O'Doherty, S., Parker, R. J., Peng, C., Peng, S., Peters, G. P., Prigent, C., Prinn, R., Ramonet, M., Regnier, P., Riley, W. J., Rosentreter, J. A., Segers, A., Simpson, I. J., Shi, H., Smith, S. J., Steele, L. P., Thornton, B. F., Tian, H., Tohjima, Y., Tubiello, F. N., Tsuruta, A., Viovy, N., Voulgarakis, A., Weber, T. S., van Weele, M., van der Werf, G. R., Weiss, R. F., Worthy, D., Wunch, D., Yin, Y., Yoshida, Y., Zhang, W., Zhang, Z., Zhao, Y., Zheng, B., Zhu, Q., Zhu, Q., and Zhuang, Q.: The Global Methane Budget 2000–2017, Earth Syst. Sci. Data, 12, 1561–1623, 10.5194/essd-12-1561-2020, 2020.Schmitz, O. J., Raymond, P. A., Estes, J. A., Kurz, W. A., Holtgrieve, G. W., Ritchie, M. E., Schindler, D. E., Spivak, A. C., Wilson, R. W., Bradford, M. A., Christensen, V., Deegan, L., Smetacek, V., Vanni, M. J., and Wilmers, C. C.: Animating the carbon cycle, Ecosystems, 17, 344–359, 10.1007/s10021-013-9715-7, 2014.Treby, S. and Grover, S. P.: Carbon emissions from Australian Sphagnum peatlands increase with feral horses (Equus caballus) presence, J. Environ. Manage., 347, 119034, 10.1016/j.jenvman.2023.119034, 2023.Trumbore, S., Da Costa, S., Nepstad, D. C., De Camargo, P. B., Martinelli, L. A., Ray, D., Restom, T., and Silver, W.: Dynamcs of fine root carbon in Amazonian tropical ecosytesm and the conribution of roots to soil respiration, Glob. Change Biol., 12, 217–229, 10.1111/j.1365-2486.2005.001063.x, 2006. Waltham, N. and Schaffer, J.: Thermal asphyxia exposure risk to freshwater fish in feral-pig damaged tropical wetlands, J. Fish Biol., 93, 723–728, 2018. Ward, D. P., Petty, A., Setterfield, S. A., Douglas, M. M., Ferdinands, K., Hamilton, S. K., and Phinn, S.: Floodplain inundation and vegetation dynamics in the Alligator Rivers region (Kakadu) of northern Australia assessed using optical and radar remote sensing, Remote Sens. Environ., 147, 43–55, 2014.Ward, D. P., Pettit, N. E., Adame, M. F., Douglas, M. M., Setterfield, S. A., and Bunn, S. E.: Seasonal spatial dynamics of floodplain macrophyte and periphyton abundance in the Alligator Rivers region (Kakadu) of northern Australia, Ecohydrology, 9, 1675–1686, 10.1002/eco.1757, 2016. Yvon-Durocher, G., Allen, A. P., Bastviken, D., Conrad, R., Gudasz, C., St-Pierre, A., Thanh-Duc, N., and Del Giorgio, P. A.: Methane fluxes show consistent temperature dependence across microbial to ecosystem scales, Nature, 507, 488–491, 10.1038/nature13164, 2014.