Forest ecosystems are recognized for their efficiency in fixing atmospheric CO₂ and storing substantial amounts of carbon (Lal, 2005). They also play a pivotal role in the greenhouse gas (GHG) balance, generally acting as sources of CO₂ and N₂O while serving as sinks for CH₄ (Walkiewicz et al., 2025). In this context, while forest degradation and deforestation are major contributors to the rise in atmospheric GHG concentrations (van der Werf et al., 2009; Reygadas et al., 2023), planted forests offer a potential mitigation strategy (Waring et al., 2020). Worldwide, planted forests occupy approximately 294 million hectares across the five continents (FAO, 2020). Of this total, eucalyptus plantations occupy approximately 25 million hectares in tropical and subtropical regions (Mao et al., 2024), with 7.8 million of this area located in Brazil (IBÁ, 2024).

In recent years, there has been a growing economic interest in using forest harvesting residues as a source of renewable energy (Udali et al., 2024). In Brazil, it is projected that roughly 6.4 million tons of wood residues are generated annually in the eucalyptus and pine harvesting processes (Pincelli et al., 2017). From this perspective, several forestry companies worldwide are adopting the whole-tree harvesting system, which collects, in addition to wood, other components such as branches, bark, and leaves to facilitate the removal of these materials from the field (Nieminen et al., 2016).

However, the removal of these residues in eucalyptus areas can have adverse effects on soil quality. Possible consequences include reduced soil fertility (Menegale et al., 2016), increased susceptibility to erosion (Wichert et al., 2018), negative influence on biological activity (Chaer and Tótola, 2007) and reduced soil organic C stocks (Rocha et al., 2018). Removing eucalyptus harvesting residues can be even more impacting in sandy soils, with drastic decreases in soil organic C stocks and soil C retention rates (Epron et al., 2015; São José et al., 2023).

A further aspect that should be considered and studied is the impact of removing eucalyptus harvest residues on GHG emissions. In crop areas, recent studies have evaluated the influence of residue management on N₂O and CH₄ emissions (Pitombo et al., 2017; Vasconcelos et al., 2018; Langeroodi et al., 2019; Mirzaei et al., 2024), and such studies have generally found that maintaining crop residues contributes to reducing emissions of these two gases. In forest areas, the information is scarcer.

The major source of N₂O emissions in agriculture is the application of nitrogen fertilizers, but the N present in plant residues also contributes substantially to the emissions (Syakila and Kroeze, 2011). The magnitude of this contribution depends on the chemical composition of the residue added to the soil (Li et al., 2016). Residues with a low C/N ratio increase N₂O emissions (Chen et al., 2013). In contrast, residues with a high C/N ratio favor nitrogen immobilization, resulting in lower emissions (Muhammad et al., 2011).

Forest soils are recognized as significant CH₄ sinks due to the oxidation of this GHG by methanotrophic microorganisms (Wigley et al., 2024). However, factors such as soil temperature, moisture, fertilization, and residue management determine whether the soil will act as a source or sink of CH₄ (Vasconcelos et al., 2018). Generally, the input of organic substrates under anaerobic conditions promotes methanogenesis, resulting in high CH₄ emissions (Zhang et al., 2015). On the other hand, increased soil porosity facilitates the transport of CH₄ to methanotrophs, enhancing its oxidation and leading to lower net emissions (Prajapati and Jacinthe, 2014).

The GHG emissions can be used to calculate the global warming potential (GWP) of different eucalyptus harvest residue management. The GWP compares the warming potential of each gas to that of CO₂, which is taken as a reference (Bayer et al., 2016). Specifically, CH₄ and N₂O have a 100-year global warming potential 34 and 298 times higher than CO₂, respectively (Zhou et al., 2023). Studies in subtropical regions have demonstrated the potential of reforestation to reduce GWP values (de Godoi et al., 2016; Martins et al., 2015); however, these studies did not consider the effects of eucalyptus harvest residue management on soil carbon stocks and GHG emissions. Therefore, this study aimed to evaluate the influence of eucalyptus harvest residue and litter management on the GHG balance in sandy soil in the Brazilian subtropics.

Soil N₂O fluxes ranged from -13.90 to 11.15 µg N-N₂O m^-2 h^-1 in the different eucalyptus harvest management systems, characterizing a low intensity of fluxes in all treatments in this sandy soil ( Figure 1a ). Since no agricultural practices occurred during the sampling period, soil N₂O fluxes remained practically constant without evident influence from eucalyptus harvest residues. N₂O fluxes showed a low but significant correlation with soil NH₄ ⁺ contents ( Table 1 ).

Soil CH₄ flux ranged from -104.22 to 9.33 µg C ha^-1 h^-1, with a strong predominance of CH₄ influx into the soil. As observed in N₂O fluxes, there was no difference in CH₄ influxes between eucalyptus harvest residue management systems ( Figure 1b ). However, there was a tendency for AR and NBr management to present higher CH₄ oxidation rates compared to NRs in almost all evaluation periods.

To assess the impact of different management systems on soil C-CO₂ retention, we used the system in which both harvest residues and litter were removed (NRs) as the reference. Based on this, the contributions of the other management systems to SOC retention were calculated relative to the NRs system. This approach — using the system with minimal organic input as a baseline — has been adopted by several authors in agricultural and forest systems (de Godoi et al., 2016; dos Santos et al., 2011; Dietz et al., 2024; Souza et al., 2023).

A significant linear relationship was observed between the annual soil C-CO₂ retention rates and the amount of carbon added through eucalyptus harvest residues and litter (r² = 0.81, p = 0.03) ( Figure 2a ). The highest retention rate (–5,540 kg C-CO₂ ha⁻¹ year⁻¹) was recorded under the management system that retained both the previous crop’s residues and the current crop’s litter (AR). In contrast, the lowest rate (–1,752 kg C-CO₂ ha⁻¹ year⁻¹) occurred under the system where only the current crop’s litter was retained, while the previous crop’s residues were removed (NR). Retention values similar to those in the AR system were observed in the two systems that maintained either bark or branches (NB and NBr).

Annual soil N₂O emissions ranged from 0.11 to 0.23 kg N-N₂O ha⁻¹ year⁻¹ and did not show a direct relationship with the amount of N added by eucalyptus harvest residues and litter (r² = 0.28; p = 0.35) ( Figure 2b ). Regarding methane during the one-year evaluation period, the soil showed an annual uptake of -2.56 to -3.91 kg C-CH₄ ha⁻¹ year⁻¹, with a weak direct correlation with the amount of C added by harvest residues and litter ( Figure 2c ).

All treatments demonstrated potential for mitigating global warming. The GWP values ranged from –24,424 to –33,946 kg CO₂eq ha⁻¹ year⁻¹ ( Figure 3 ). The main contributors to this atmospheric carbon sequestration were the carbon added to the soil and the carbon stored in wood products, which offer medium- and long-term sequestration potential.

The management system that retained eucalyptus harvest residues (AR) showed the lowest GWP (–33,946 kg CO₂eq ha⁻¹ year⁻¹), representing a significant reduction in greenhouse gas emissions compared to the NRs system (p< 0.10). AR was followed by NBr (–31,879 kg CO₂eq ha⁻¹ year⁻¹), NB (–30,619 kg CO₂eq ha⁻¹ year⁻¹), and NRs (–28,416 kg CO₂eq ha⁻¹ year⁻¹), with no significant differences among these three treatments.

In the reference system (NRs), nearly all the GWP value was attributable to carbon stored in wood. In contrast, wood carbon contributed approximately 85% of the total GWP on average in the systems with partial or complete retention of eucalyptus residues. The remaining share was associated with the influence of harvest residue management on soil organic carbon retention.

The N₂O fluxes we observed are consistent with recent studies in eucalyptus areas in tropical Brazilian soils. Cuer et al. (2018) found values below 10 μg N-N₂O m⁻² h⁻¹. Silveira et al. (2022) observed that eucalyptus forests emitted up to 5.5 μg N-N₂O m⁻² h⁻¹ during the rainy spring, but acted as a N₂O sink during the dry winter periods.

In general, eucalyptus plantations established in sandy soils have low nitrogen availability and, consequently, low N₂O emissions (Livesley et al., 2009). In addition to the low levels of N-NH₄ ⁺ and N-NO₃ ^- in the soil, we also observed low WPS values, which may have contributed to the low N₂O emission ( Supplementary Figure S1 ). The WPS values were below 25%, an unfavorable condition for N₂O production by denitrification, which occurs at WPS values above 60% (Bateman and Baggs, 2005). In sandy soils, such as the one examined in this study, oxygen diffusion rates are higher than in clay soils, avoiding anaerobic conditions for a prolonged period, a key requirement for denitrification (Rochette et al., 2008). The lack of relationship between N₂O fluxes and soil N-NO₃ ^- levels reinforces this hypothesis. On the other hand, the correlation with N-NH₄ ⁺ contents suggests that the nitrification process may have contributed significantly to N₂O fluxes. During nitrification, a process favored under aerobic conditions, intermediate compounds may lead to non-obligatory N₂O production (Zhang et al., 2025).

The impact of plant residues on N₂O emissions depends on the composition of these residues, especially their C/N ratio (Li et al., 2016). In general, emissions are negatively correlated with the C/N ratio, meaning that the presence of residues with a high C/N ratio promotes N immobilization, reducing net mineralization and N₂O production (Pilegaard et al., 2006). In this context, our results support the idea that the input of forest residues with a C/N ratio greater than 30 has a low contribution to N₂O emissions (Fest et al., 2015) due to the rapid immobilization and limited availability of N for nitrification and denitrification processes (Livesley et al., 2009). The C/N ratio values in our treatments ranged from 111.7 to 162.7, accompanied by high lignin/N ratios, 58.1 to 69.1 (São José et al., 2023), which favor microbial nitrogen immobilization. Additionally, the low-organic-matter sandy soil was fertilized only during forest establishment (São José et al., 2020), contributing to low nitrogen availability and, consequently, low N₂O emissions.

However, since we did not observe differences between treatments, our results differ from other studies involving agricultural residue management (Gonzaga et al., 2018; Maris et al., 2018; Vasconcelos et al., 2018; Reeves et al., 2024). These differences are possibly related to the period between the application of residues and the collection of gas samples. In those studies, the highest N₂O emissions occurred immediately after the addition of plant residues, with a reduction after a few months and remaining constant over time. In our study, the evaluations only occurred six years after the experiment started. Thus, we believed that N₂O emissions in the management of eucalyptus harvest residues were already reduced and stabilized, and possibly the highest emissions must have occurred at the initial time of the experiment installation.

The negative CH₄ fluxes we observed are consistent with other studies conducted in eucalyptus areas (Fest et al., 2017; Fialho et al., 2018; Silva et al., 2024). In our case, the influxes were possibly favored by the sandy texture of the soil, which, regardless of the management adopted, allowed greater oxygen diffusion and, consequently, created conditions for CH₄ oxidation. This assumption is supported by the findings of Livesley et al. (2011) and Grover et al. (2012), who observed similar CH₄ dynamics in low-nutrient sandy soils of northern Australian Eucalypt savanna woodlands.

Methane influxes may also be related to low WPS values that favor CH₄ oxidation (Liu et al., 2019). In such a condition, the improved soil porosity and gas diffusivity facilitate the transport of CH₄ to methanotrophic bacteria that oxidize CH₄ to CO₂. CH₄ influxes are usually inversely related to soil moisture (Fest et al., 2017; Liu et al., 2019). Despite this, in the present study, no relationship was observed between CH₄ influxes and WPS, probably due to the reduced water retention capacity of the sandy soil. CH₄ influx is enhanced in sandy soils, where rapid drainage occurs, preventing the maintenance of high soil water contents for prolonged periods that could determine soil reduction conditions (Walkiewicz et al., 2025). Our results were similar to recent studies on agricultural residue management (Wegner et al., 2018; Langeroodi et al., 2019) and in harvesting and soil preparation operations in eucalyptus areas (Fialho et al., 2018). As in our study, these authors also attributed the low effect of plant residues on CH₄ fluxes to the small variation in soil moisture between residue managements.

Removing eucalyptus harvest residues has been widely considered a management that reduces C-CO₂ retention in the soil (Rocha et al., 2018). This practice is more relevant in soils with lower clay contents, which have a lower capacity for physical protection of soil organic matter (Dieckow et al., 2009), causing reductions in soil C compared to the maintenance of eucalyptus harvest residues (Oliveira et al., 2018). The results obtained in this study corroborate these considerations. In addition, the capacity of the soil to function as a CO₂ sink depends on the biomass input (Conceição et al., 2013). This dependence is evidenced by the correlation between C-CO₂ retention rates in the soil and the amount of C contributed by eucalyptus harvest residues and litter ( Figure 2a ), reinforcing the importance of maintaining eucalyptus harvest residues to promote C additions, mainly in sandy soils in tropical regions (Epron et al., 2015).

On the other hand, annual N₂O emissions were not correlated with the amount of N added by plant residues. Much of the N present in the residues may have already been released in the first months of implementation of the experiment, as observed by Rocha et al. (2016). The authors evaluated the decomposition and release of nutrients in different management systems of eucalyptus harvest residues, observing that the management system with the maintenance of all residues released approximately 130 kg ha^-1 of N after 300 days from the beginning of the experiment. In addition, the high C/N ratio of the bark and branches of the residues (110 and 316, respectively) and the low N content of the litter could be causing the immobilization of N by the soil microbial population.

Studies evaluating annual N₂O emissions in reforestation areas in subtropical regions have shown variable results. The annual N₂O emissions in our study were lower than those observed in areas with Acacia mearsii (0.24 ± 1.25 kg N ha^-1 year^-1) (de Godoi et al., 2016), Acacia auriculiformis (2.3 ± 3.1 kg N ha^-1 year^-1), and Eucalyptus urophylla (1.9 ± 2.1 kg N ha^-1 year^-1) (Zhang et al., 2014). However, our results were quite similar to those obtained by van Delden et al. (2018), who observed annual N₂O emissions ranging from 0.08 to 0.09 kg N ha^-1 year^-1 in eucalyptus areas grown in subtropical sandy soils in Australia. Our results demonstrate the low potential for N₂O emissions in the different management of eucalyptus harvest residues in this Brazilian sandy soil.

Despite the low linear relationship between eucalyptus harvest residue input and CH₄ influxes, we observed a trend of lower influx in the NRs and NR systems, which may be related to lower soil quality compared to the AR system (São José et al., 2022). The loss of soil quality results in a lower capacity to oxidize CH₄ (Bayer et al., 2013). Wu et al. (2019) obtained similar results. The authors observed that litter removal reduced the CH₄ oxidation capacity by approximately 30% compared to areas that maintained litter in coniferous forests in the Chinese subtropics. This reduction was attributed to the lower abundance of methanotrophic microorganisms due to the poor availability of low-molecular-weight organic compounds caused by litter removal.

Carbon sequestration is considered one of the main factors controlling GWP in agricultural systems (Schönbach et al., 2012), which was confirmed by our results. Furthermore, our results demonstrate that N₂O and CH₄ fluxes made a negligible contribution to the final GWP result, as typically observed in forests (Saggar et al., 2008; Walkiewicz et al., 2025). Our results are similar to those obtained by Zhang et al. (2015), who observed that the contributions of N₂O and CH₄ to GWP were less than 3% in forest areas in the subtropical region of China. In our study, however, the contribution of these gases was even lower, not reaching 1% ( Supplementary Table S2 ).

Few studies have related the effects of forest management impacts with GWP, and there is no standardization in the calculations of this variable, which makes it difficult to compare results. Martins et al. (2015) observed a 76% reduction in GWP in Eucalyptus saligna areas in Australia compared to pasture areas. However, these authors did not consider the soil C retention rate and included only CH₄, N₂O, and CO₂ emissions. Wang et al. (2022) also adopted this approach, excluding soil carbon retention rates and considering only CH₄, N₂O, and CO₂ emissions in their calculations. They were among the few, if not the only, researchers to evaluate the effect of harvest residue management on GHG emissions in forest ecosystems. However, their study focused on Cunninghamia lanceolata growing in a soil type markedly different from ours, with approximately eight times higher organic matter content. de Godoi et al. (2016), studying Acacia mearnsii areas in the Brazilian subtropics, found that wood contributed approximately 70% of the GWP value, while soil carbon retention accounted for about 30%.

Estimating the GWP of local agricultural and forest systems is essential for obtaining accurate and context-specific assessments of environmental impacts. To the best of our knowledge, this is the first study to evaluate the effects of eucalyptus harvest residue management on the GHG balance in a Brazilian subtropical sandy soil. In our research, the management practice that retained all eucalyptus harvest residues and litter (AR) showed significantly lower soil-associated GWP values, indicating a greater potential for carbon sequestration compared to other management strategies. These results demonstrate that maintaining eucalyptus harvest residues, besides promoting an increase in forest productivity, represents an alternative for mitigating GHG emissions in subtropical sandy soils, both due to the potential for CH₄ oxidation and the storage of soil organic carbon.