the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 06 Jul 2026
Exchange fluxes of CO2, CH4, N2O and NO between soil and atmosphere along an ecoclimatic gradient in West African savannas: a missing piece for regional budgets
Abstract. This study investigated greenhouse gas (CO₂, CH₄, N₂O) and reactive nitrogen (NO) fluxes at three West African savanna sites: the international research reserve of Lamto (Taabo district, Côte d’Ivoire), the Observatoire de Recherche en Environnement de Nambekaha (OREN) (Korhogo, Côte d’Ivoire), and the Centre de Recherches Zootechniques (Dahra, Senegal). Measurements were carried out during intensive field campaigns conducted in 2024 and 2025, during the wet seasons at the three sites, across tree areas and grassy areas in savannas, and cropland ecosystems subjected to different treatments from March 2023 to September 2025.
Overall, soil moisture, vegetation type (grassy areas, trees areas, crops) and site location (Lamto, Dahra, Nambekaha) were the main factors controlling gas fluxes (CO2, NO and CH4), whereas treatments containing different ratio of nitrates and ammonium had no significant effect according to the statistical analysis (ANCOVA).
CO₂ fluxes ranged from 8.21 ± 2.5 to 91.35 ± 73.2 µg C m⁻² s⁻¹ and were controlled by soil moisture, with a decrease in soil respiration as water content increased (β = −1.105 ± 0.236 µg C m⁻² s⁻¹; p <0.001), due to a limitation of oxygen diffusion in the soil, highlighting the key role of soil moisture in regulating both heterotrophic microbial respiration and autotrophic plant respiration, in relation to soil aeration conditions. NO emissions, ranging from 0.01 ± 0.0 to 497.39 ± 146.3 ng N m⁻² s⁻¹, showed a significant correlation with vegetation type. The highest values were observed in the cropland plots of Nambekaha (β = +76.779 ± 15.82 ng N m⁻² s⁻¹; p < 0.001) compared with natural savannas, reflecting intensified nitrification processes linked to background fertilization inputs (150 kg NPK ha⁻¹ yr⁻¹).
CH₄ fluxes were primarily determined by vegetation type: grassy areas within savannas behaved as net sources (β = +3.836 ± 0.62; p < 0.0001), whereas croplands acted as sinks, suggesting methanotrophic activity capable of oxidizing atmospheric methane in the soil. In contrast, N₂O fluxes were mostly low or even negative across all ecosystems and treatments, with no significant relationship to soil moisture, vegetation type, or treatments. The results indicate that soils could occasionally function as net N₂O sinks: indeed, N2O uptake may occur in nitrogen-poor soils under oxic conditionswhere the limited availability of mineral nitrogen restricts N₂O production and where atmospheric N2O diffuses easily into the soil.
These findings highlight the microbial and environmental coupling of carbon and nitrogen dynamics in tropical savanna soils and provide critical insight for predicting greenhouse gas and reactive gas emissions under changing land-use conditions.
- Preprint
(6173 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 25 Sep 2026)
-
RC1: 'Comment on egusphere-2026-3592', Anonymous Referee #1, 24 Aug 2026
reply
-
AC1: 'Reply on RC1', Moussa Zoure, 10 Sep 2026
reply
We thank you for your careful evaluation of our manuscript and for your constructive and encouraging comments. We appreciate your recognition of the scientific significance of this work, the experimental design, and the collaborative effort involved in conducting this study.
We have carefully considered your suggestions and revised the manuscript accordingly. In particular, we have:
1. checked and corrected the terminology and English translation, particularly the expression “forage water”;
2. revised the description of the soil texture at the Nambékaha site and clarified that the soil samples used for pH analysis were collected to a depth of 20 cm;
3. improved the presentation and organization of the results to better reflect the multifactorial design of the experiment and to ensure greater consistency among the different gas fluxes;
4. revised the statistical tables and figures to improve their clarity and readability, particularly by removing the lines connecting the data points to their associated variability;
5. revised the discussion, where necessary, to account for the revised presentation of the results;
6. expanded the literature review and discussion to include relevant studies conducted in Southern Africa.We also suggest changing the title by removing the last part, “a missing piece for regional budgets”
Best regardsCitation: https://doi.org/10.5194/egusphere-2026-3592-AC1
-
AC1: 'Reply on RC1', Moussa Zoure, 10 Sep 2026
reply
-
RC2: 'Comment on egusphere-2026-3592', Anonymous Referee #2, 04 Sep 2026
reply
This manuscript describes an experiment conducted at three sites in West Africa with different rainfall levels, where different amounts of nitrate and ammonium were added to microplots. Four trace gases (CO₂, CH₄, N₂O and NO) were measured over a three- to four-day period in June (the wet season) of both 2024 and 2025. At each site, the microplots were located in areas dominated by grass, trees and crops. Soil water content was measured as an independent variable.
While the sentences in the manuscript are understandable, the organisation, design and description of the experiment and results are poor and sometimes chaotic.
In the introduction, the main argument of the experiment appears to be that very few studies have been conducted in this region (lines 89, 96, 121 and 139), a point that is also made in the discussion (line 720). Given the importance of this argument, it is surprising that several studies of which I am aware are not considered in the manuscript, e.g.:
Anderson et al 1998 https://doi.org/10.2134/jeq1998.00472425002700050017x ; Anderson et al 2003 https://doi.org/10.1029/2002JD003345 ; Brümmer et al 2008 https://doi.org/10.1007/s10021-008-9144-1 ; Cofer III et al 1996 J. Geophys Res. 101; Delmas et al 1991b J. Geophys Res. 96; Dick et al., 2006 https://doi.org/10.1111/j.1365-2699.2005.01421.x ; Levine at al 1996 J. Geophys Res 101; Priemé & Christensen 1999 J. Geophys Res. 1004; van Straaten et al 2019 https://doi.org/10.1016/j.jaridenv.2019.02.013 ...and this list is incomplete, especially if you include papers on trace gas emissions from southern African savannas. While there is no doubt that many of these studies have shortcomings (e.g. limited temporal resolution), it seems to me that the current study is also suffering from some of these problems.
The introduction goes on to discuss: vegetation depending on precipitation, and the importance of studying N and C cycles; a brief overview of N2O and CH4 uptake and emissions, without going into detail; an increase in N deposition, and how important reactive N fluxes can be. Then, rather abruptly, the NitroAfrica project is introduced, along with the fact that several questions will be raised. This is not how you write an introduction. No hypotheses are formulated, which is not surprising given that the paragraphs in the introduction do not clarify what the manuscript will be about apart from adding a study in an area where little has been done.
Materials and methods: The site description is inadequate. Rather than describing the sites themselves, we learn about the research being conducted at the various stations. No soil classification is provided, which might be important if you want to comment on the representativeness of the sites. It is unclear whether there are differences in the soil between the microsites. We only learn about the climate and vegetation to a limited extent, and it remains unclear how important tree- and grass-dominated areas are. Are there no trees in grass-dominated areas, and vice versa? If there are trees in grass-dominated areas and vice versa, what is the difference between the two? It is only later in the manuscript that we learn that at least one cropped area received a lot of fertiliser (150 kg), which makes it unsurprising that no difference was found between treatments.
The experimental design is very confusing. Apparently, the treatments mostly focus on simulating N deposition. However, I have not seen the actual deposition rate (in kg N/ha or g N/m²) nor is it clear in what form the nitrogen is deposited (NH₄, NO₃?). It seems that the authors did not use demineralised water to dissolve the nitrogen salts, but rainwater, which they refer to as 'forage water'. This makes everything even more confusing, since sometimes the forage water contained more nitrogen than was added. While the adjustment of the N concentration in the treatment water is explained in detail, the amount of N added per microplot or per m² is never specified. The plot size is very small (2.5 x 2.5 m), yet possible edge effects are never mentioned anywhere in the manuscript.
If I understand correctly, only one chamber per microplot was measured. This means that each chamber measurement was considered a replicate, which is risky given the spatial variation in trace gas fluxes. The chambers were inserted 5 cm into the soil, much deeper than I would normally do. In savannah climates, where most of the roots are in the topsoil, cutting a substantial part of these roots would have affected the results. Were the chamber bases installed each time measurements were taken? If so, how might this have affected the results? Trace gases were measured in very short campaigns, so it is not surprising that measurements taken over three or four days are not going to tell you much. Studies have shown that seasonality can be important in these regions, especially the onset of the rainy season, but this was not measured at all. Some papers have suggested that burning is important and may affect fluxes in these ecosystems, but this has been completely ignored.
The authors measured soil moisture with a probe and used the standard calibration, which they adjusted per site based on texture. It is unclear whether the moisture content is reported as volumetric or gravimetric water content, and why it was never converted into water-filled pore space (WFPS). In trace gas studies, it is generally accepted that water should be expressed as WFPS, because it is the only way to compare fluxes of sites with differences in bulk density and pore systems, as is the case here.
What is obviously missing from the supporting measurements is any indication of extractable mineral N or N mineralisation, nitrification. Without these measures, it is virtually impossible to explain the observed trace gas measurements, especially N₂O and NO, but also CH₄. These omissions mean that there is almost nothing to explain or discuss. Although there are claims that the vegetation type was important for CH₄ (line 516) and NO (line 577), the discussion becomes pure speculation due to these missing measurements. We do not learn anything about the mechanisms responsible for the few things that are measured. In the case of CH₄, I would at least expect termites to be discussed as a potential source, but they are not mentioned anywhere in the manuscript. All Tables do not contain a lot of useful information, nor does Figure 2, 4, 6 and 9, which do not help the clarity of the results.
In summary, this manuscript describes a confusing study that does not even come close to explain the few measurements conducted. Apart from the confusing treatments, which I still do not understand why they were done, trace gases were only measured as a snapshot of 3 to 4 days in the rainy season. You cannot expect to find any meaningful correlation with soil water content if that is your data basis. Furthermore critical measurements that might be able to explain fluxes were not conducted (e.g. extractable mineral N) and the only measurement that may helped to explain observed trace gas flux values (soil moisture) was not processed in such a way that can be useful. I agree that there is not a lot of studies done in this region of the tropics, but from the few studies that exist, several were apparently never read and unfortunately, the current manuscript does not help to fill the lack of data from this important region, it only makes the confusion bigger. My advice is to reject this study, it does not meet the standards of publication of Biogeosciences. Given the poor design and the lack of critical data to explain fluxes, I can also cannot see how this manuscript can become publishable after revision.
Citation: https://doi.org/10.5194/egusphere-2026-3592-RC2 -
AC2: 'Reply on RC2', Moussa Zoure, 10 Sep 2026
reply
We thank you for the careful evaluation of our manuscript. We agree that several aspects of the study could be improved, particularly the description of the experiment, the presentation of the study sites, the formulation of the scientific questions, and the discussion of the study limitations. Necessary revisions can be made to clarify these different points.
1. Description of the experiment and literature review
We acknowledge that the description of the experimental design could be further clarified to avoid any ambiguity. Studies conducted in the 1990s were not systematically included because we initially prioritized more recent references. However, we recognize that this approach was too restrictive, particularly in the context of South African savannas. We were also already familiar with the articles mentioned by the reviewer.
2. Introduction and formulation of hypotheses
We acknowledge that the hypotheses could be formulated more explicitly and better linked to the observations and scientific questions raised in the introduction. Although the main scientific questions are already presented in the manuscript, we can improve their formulation to make the hypotheses and the associated expectations of the experiment more explicit.
3. Description of the study sites and watering protocol
We also acknowledge that some aspects could be described in greater detail. In particular, we will provide further information on the characteristics of the study sites and the application procedures for the watering solutions. The information regarding the nitrogen forms used in the experimental treatments, particularly the proportions of nitrate and ammonium, is already provided in the manuscript, but we will make these details more explicit to avoid any confusion regarding the experimental design.
4. Gas flux measurement protocol using chambers
The measurement protocol is based on a methodology that has already been widely used and published in several previous studies, and the main conditions required to ensure reliable measurements were followed in our experiment. Nevertheless, we acknowledge that the protocol description could be improved in the manuscript, particularly to allow readers to better understand the measurement conditions and procedures applied.
5. Study objectives and representativeness of the dataset
We would also like to clarify that the main objective of this study is not to characterize seasonal variability in gas fluxes or specifically assess the effects of fires, but rather to investigate if there was an effect of experimental water addition combined with different nitrogen forms on greenhouse gas and nitrogen compound fluxes during the wet season, after 18 and 30 months of experimentation. From our results, this effect is non significant compared to soil moisture, site or type of ecosystems, at local scale.
The establishment and monitoring of such an experimental design across three tropical savanna sites with contrasting climatic and environmental conditions is relatively complex. In this context, the study is based on a total of 144 flux measurements, representing a substantial dataset for documenting the responses of these ecosystems to the experimental treatments.
We therefore consider that these results provide original and relevant information on soil–atmosphere exchanges in tropical savannas, which remain relatively understudied, while acknowledging the need to better clarify certain limitations and methodological aspects in the revised version of the manuscript.
6. Absence of measurements of soil mineral nitrogen
Finally, we acknowledge that the absence of measurements of soil mineral nitrogen concentrations constitutes a limitation of the study. Such information would have helped us better link variations in gas fluxes to nitrogen transformations in the soil.
However, the present article primarily focuses on the characterization of soil–atmosphere fluxes and their responses to the experimental treatments. A more detailed mechanistic interpretation incorporating soil nitrogen and carbon pools and forms, nitrogen-cycle functional genes, as well as nitrification and denitrification potentials, would go beyond the scope of the present study. These aspects will be investigated further and addressed in a subsequent study.We also suggest changing the title by removing the last part, “a missing piece for regional budgets”
Best regardsCitation: https://doi.org/10.5194/egusphere-2026-3592-AC2
-
AC2: 'Reply on RC2', Moussa Zoure, 10 Sep 2026
reply
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 153 | 51 | 28 | 232 | 26 | 27 |
- HTML: 153
- PDF: 51
- XML: 28
- Total: 232
- BibTeX: 26
- EndNote: 27
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
I would like to congratulate the authors on a very nice piece of work that addresses an important gap in the understanding of biogeochemical processes in a significantly understudied region.
Scientific Significance
This paucity of research on the African continent has impacts on the global understanding of these processes and the global understanding of soil atmosphere exchange on a global basis.
The composition of the research team, comprising researchers from many countries and institutions in West Africa, should be lauded as an example of regional collaboration and scientific exchange.
I believe that this submission is work of excellent scientific significance. it highlights an in-situ study of the emissions of CO2, CH4, N2O and NO from three different sites with contrasting bioclimatic characteristics representing some of the major systems of West Africa and being indicative of tropical and subtropical regions in Africa and globally. Specifically, the sites studied include Lamto and Nambekaha in the humid savanna systems with approximately 1100mm MAP and the Dahra site in a Sahelian dry savanna biome MAP 380mm. These sites are subjected to agricultural activity and therefore provide an outstanding gradient of land uses and agro-ecological contexts.
Scientific Quality
The experimental design followed a multifactorial approach, and sampling occurred over multiple years, under differing land use and vegetation cover, and with differing levels of nitrogen addition. This design required significant efforts to deploy and to maintain, and the authors and their support teams should be commended on the dedication required to deploy and maintain this experiment in what are likely trying conditions.
The experimental design is appropriately replicated with 4 replicates per fertilisation treatment (4 treatments) per vegetation cover (3 cover types), resulting in 48 plots per site. Therefore, a total of 144 plots. I'd like to reiterate that this is a significant amount of work to set up and maintain this experimental design, and the results emanating from this are invaluable. At this point I'd like to question the use of the term "forage water", which I couldn't find defined in the text, I assume this may refer to something like irrigation water. I do, however, suggest the Authors check the translation to English.
The flux measurements seem to have been well conducted, utilising high-precision sensors and a well-described chamber system; this approach seems appropriate and is likely to yield defensible results. Analyses applied for the flux calculations seem appropriate. Measurements of ancillary data such as soil moisture and soil pH seem appropriate. I do recommend the authors rephrase the description of the soil texture at the Nambekaha site (line 393) and the depth to which the soil was sampled for pH analysis (Line 400) (rather "collected to a depth of 20 cm" ).
Presentation Quality
Presentation of the results is generally appropriate by separating the presentation of the results into the gas species of interest, i.e CO2, CH4, N2O, NO.
In the presentation of the fluxes of each of the gaseous species Id like to suggest that the authors consider reorganising the presentation of the results to account for the multifactorial design of the experiment, first by site, then by vegetation cover, then by the N addition treatments. This is done to some extend (eg fig, fig 6 and fig 9 ) but it would be better to have the treatment effects (Site, vegetation, moisture) represented consistently for each of the gas fluxes. As it stands I find the presentation of the results somewhat confusing and hard to follow and the statistical tables are not clear.
Please remove the lines of connection between the point and variance figures (4b, 6a, 6b, 9a, 9b); these should not be included.
The discussion section is generally well written and clear although I believe there would be value in following the logic of the multifactorial design in the presentation of the results (Site effects, vegetation effects, fertilisation effects, soil moisture effects), if the authors restructure the presentation of the results the presentation of the results the discussion will need to be updated to reflect those changes.
While the introduction and discussion appropriately reference existing research in West Africa, due to the paucity of studies on this (and related) topics in general, I would encourage the authors to expand their literature assessment to include work undertaken in Southern Africa.