the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 15 Jun 2026
Valley-bottom wetland agricultural conversion and recovery shape greenhouse gas dynamics and soil carbon sequestration in an African tropical highland system
Abstract. Tropical wetlands significantly impact greenhouse gas (GHG) budgets and carbon storage: however data from sub-Saharan African (SSA) remain limited. Highland valley-bottom wetland (HVBW) agriculture supports millions, yet its effects on GHG emissions and carbon storage remain undocumented. This study quantified soil emissions of nitrous oxide (N₂O), carbon dioxide (CO₂), and methane (CH₄) in the Taita Hills, Kenya, from 12 converted, 10 recovering, and one reference (intact) HVBWs. Agricultural conversion shifted wetland emissions from CH₄ to N₂O dominance. Converted HVBWs were N₂O sources (2.7 kg N₂O–N ha⁻¹yr⁻¹), driven by elevated soil nitrate, whereas the intact wetland was an N₂O sink (−0.3 kg N₂O–N ha⁻¹ yr⁻¹), with high soil moisture and high soil C/NO₃⁻–N ratio suggesting complete denitrification. Recovering HVBWs showed intermediate N₂O emissions (0.6 kg N₂O–N ha⁻¹ yr⁻¹). CO₂ emissions were similar between converted and recovering HVBWs (10,850 vs. 11,031 kg CO₂–C ha⁻¹ yr⁻¹) but lower in the intact (2,923 kg CO₂–C ha⁻¹ yr⁻¹). CH₄ emissions were highest in the intact HVBW (2,757 kg CH₄–C ha⁻¹ yr⁻¹), intermediate in recovering sites (879 kg CH₄–C ha⁻¹ yr⁻¹), and lowest in converted sites (37 kg CH₄–C ha⁻¹ yr⁻¹). The intact HVBW had 224 Mg C ha⁻¹, indicating carbon loss rates of 2.6 Mg C ha⁻¹ yr⁻¹ over 45 years for converted HVBWs. Restoring Taita Hills wetlands would sequester 1.1 Mg C ha⁻¹ yr⁻¹, offsetting ~0.0005 % of Kenya's annual agricultural GHG emissions, or 9.8 % when scaled nationally. These findings highlight trade-offs between GHG emissions and carbon storage in HVBWs, with wetland recovery promoting functional restoration and long-term carbon sequestration in SSA.
- Preprint
(2121 KB) - Metadata XML
-
Supplement
(1315 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2026-3028', Anonymous Referee #1, 29 Jul 2026
-
AC1: 'Reply on RC1', Sharon Gubamwoyo, 08 Aug 2026
Thank you for the positive and encouraging assessment of our manuscript and for recognizing the value of the dataset and the effort involved in its collection and analysis. We also greatly appreciate the constructive suggestions provided to improve the clarity of the manuscript, figures, and figure captions. We have carefully considered all the comments and addressed them accordingly. Below is some feedback to some of the critical questions
- L74 Please briefly explain what is meant by “recovering wetlands” here.
L74 Recovering wetlands explained in L129 to L131 “Converted HVBWs were defined as wetlands that are currently under agriculture and range in age since conversion from 1963 to 2023 (Gubamwoyo et al. 2025; Table S2). Recovering HVBWs were defined as wetlands that were used for agriculture in the past but have been abandoned for ≥ 5 years due to high soil moisture or frequent flooding and have 130 regained natural wetland vegetation (Gubamwoyo et al. 2025). Only one intact HVBW, which had never been used for agriculture, was found.”
- L126 Why there is a gap in the measurements?
L126 the gap in between was because we had to run trial field surveys at the beginning of the study. We were also short of funds to continue at that moment so we had to wait then commence in April when the funds were got. The gap in the study however, didn’t affect the results.
- L151-154 Why did you change the method for the last four sampling campaigns?
L151-L154 We changed the method to allow us have a complete year of sampling with reduced costs on gas samples taken to the laboratory. The cost of sampling into gas vials and sending them to the laboratory was becoming very expensive and would not enable us have enough data. LiCOR analyzers were also only available at the last campaigns but not before. When we started using this method, we sampled using the gas vial method and the LICOR method at the same time and compared the results to ensure reliable and comparable results. The comparison results showed a good fit with an R2 of 0.91. We also needed to run some experiments which would generate quite a number of gas samples (leading to high expenses), but by using the LICOR we were able to analyse at reduced costs. In this experiment we planted maize in one of the wetlands and varied the manure and fertilizer addition in combination with watering while measuring the GHG emissions.
- L155 How did you decide in which patches to install the collars?
Collar locations were selected to represent the main vegetation/soil/hydrology patches present in each wetland. Areas close to wetland edges, footpaths, drainage channels, or other obvious disturbances were avoided.
- L156 How many chamber bases in total?
Six chamber bases were installed in each wetland, with a set of three setups in one patch and three in another patch. In total, we had 132 chambers.
- L206 How many of CH4 fluxes were eliminated? How many data points did you have in the end for each gas?
L207: “In total, this led to the exclusion of 1% of CO₂, 6% of CH₄, and 4% of N₂O data points.” CH4 fluxes that were eliminated were 64
At the end for the data analysis, we used 1015 data points for CH4, 1069 data points for CO2 and 1036 data points for N2O
Citation: https://doi.org/10.5194/egusphere-2026-3028-AC1
-
AC1: 'Reply on RC1', Sharon Gubamwoyo, 08 Aug 2026
-
RC2: 'Comment on egusphere-2026-3028', Anonymous Referee #2, 25 Aug 2026
This study analyses greenhouse gas dynamics in the highlands of East Africa, a region that, as far as I am aware, remains underrepresented in this field of research. The manuscript presents a significant dataset that contributes to understanding biogeochemical cycling in the region, and it is clear that considerable time and effort have gone into the study. The findings related to higher CH₄ emissions and SOC stocks in intact wetlands are particularly interesting. Most of my comments are minor; however, I believe the change in methods used to analyse greenhouse gas fluxes requires further clarification. In addition, the total GWP of the combined effects of CO₂, CH₄, and N₂O should be made more explicit, perhaps in the graphical abstract.
Line 35 – The first figure might benefit from including an overall CO₂-equivalent value to provide an indicator of global warming potential.
Line 45 – Can you clarify this statement ‘with nutrient-rich soils available year-round’. I don’t understand the reference to year-round when related to how nutrient rich the soils are.
Line 69 – Can you spell out DNRA?
Line 78 – Remove full stop here? ‘The HVBWs within this and other water tower systems sustain over 95% of the livelihoods in Taita Taveta County. through their organic-rich, fertile soils, water availability, and flat topography’
Line 81 – KBA only mentioned once, so doesn’t really need to be an acronym. I would do the same with other rarely used acronyms if there are any. This can make it easier for the reader, so they don’t have to recall so many. ‘Therefore, they are recognized as Key Biodiversity Areas (KBA) and are included in international conservation initiatives (Mwacharo, 2024).’
Line 108 / Figure 1 caption – Please be consistent with m a.s.l. m.a.s.l.
Line 120 / Figure 1 – Do you have a higher quality map of the African continent? Parts of it look out of proportion from my eyes.
Line 129 – Regarding the sampling design, how can you be sure about the recovering vs intact or reference site. Did you use some historical aerial imagery? Or do plant communities determine if it hasn’t been impacted before?
Line 151 – It is a little unusual to change the method of GHG analysis between sampling runs. Were you able to calibrate the two methods to indicate the two separate methods were consistent. I understand that the LI-COR is probably much more convenient being able to measure in the field. But changing methods might add additional uncertainty to the results.
Line 191 – There are two equations here I would identify both in the text. I would also start the ‘where Gas flux’ as a new sentence.
Line 238 – Please revise this sentence, for grammar and readability.
Line 245 – How was pH determined? Soil-water mix ratio 1:5? Please elaborate.
Line 303 – eight-’folk’ ‘fold’
Line 303 – The back-to-back parentheses are slightly clunky. Perhaps you could combine them with a semicolon, changing ‘(47.4 ± 5.8 µg g⁻¹ DW) (Figure S6)’ to ‘(47.4 ± 5.8 µg g⁻¹ DW; Figure S6)’. There are a few instances of this throughout the manuscript, although if you feel strongly, please leave them unchanged.
Line 309 – Please re-word this part of the sentence ‘whereas the converted HVBWs were significantly different across seasons for all four variables’. This currently reads like each variable was significantly different from each other across all seasons for this wetland type, whereas they were not significantly different in all cases (e.g. NH4+-N in dry and wet seasons).
Line 315 – Random floating ‘A’ near the soil moisture figure near 25% label. Please remove.
Line 325 – Was there a reason for only reporting the ± SE in recovering HVBWs. Also, why didn’t the BD get reported and just a relative indication.
Line 330 – This may be a bit pedantic, but it would be worth checking consistency across the figures. For example: the y-axis labels appear bold in Figure 3 but not in Figure 4; the group labels in Figure 4, ‘Recovering’ and ‘Converted’, are capitalised; Figure 1 caption isn’t italicised, while the others are; Figure 4 caption heading isn’t bold while the other are.
Line 338 – Sentence starting ‘However’ could be reworded. The structure of ‘lower fluxes in the dry season than in the rewetting events’ then ‘higher’ is a bit confusing.
Line 348 – Suggest changing ‘statistically significant’ to ‘significantly different’ or ‘significantly lower than’.
Line 365 – Be careful with the use of en dashes and hyphens ‘N₂O–N and CH₄-C’.
Line 370 – Remove floating ‘d’
Line 388-393 – ‘CH₄-C’ This paragraph has gone back to use if a hyphen.
Line 402 – Again, the en dashes and hyphens are being used interchangeably.
Line 457 – change this ‘intact HVBWs’ to ‘intact HVBW’ given there is only one.
Line 459 to 461 – I would be more cautious with this statement, suggesting ‘likely’ or ‘potentially’ given methanogenic archaea were not measured. Similar, with the next sentence.
Line 523 – Is this correct ‘1,494,766 1 Mg C yr⁻¹’. If not please remove the extra ‘1’. Also, this might be better reported as 1.5 million Mg rather than an exact number.
Line 525 – I think you could add CO2 to the total GWP, obviously with a CO₂-eq of 1. I think that this would be of more use than giving the GWP of CH₄ + N₂O and not including CO2.
Line 526 – It could be more meaningful to readers to move from kg to Mg or t for these larger numbers.
Line 549 – This reads as part of the text. Is it associated with table 2?
Line 655 – Butterbach-Bahl, K. et al. 2011 – DOI does not work when clicked.
Citation: https://doi.org/10.5194/egusphere-2026-3028-RC2 -
AC2: 'Reply on RC2', Sharon Gubamwoyo, 30 Aug 2026
We appreciate your review, feedback and suggestions. We have carefully addressed the comments and edited the manuscript accordingly. We have also provided some feedback to the questions that were raised. Thank you once again.
Line 35 – The first figure might benefit from including an overall CO₂-equivalent value to provide an indicator of global warming potential.
The graphical abstract was updated and the global warming potential values were added.
Line 45 – Can you clarify this statement ‘with nutrient-rich soils available year-round’. I don’t understand the reference to year-round when related to how nutrient rich the soils are.
We apologize this was too ambiguous year-round in this case refers to permanent wetland saturation not to a seasonal change in soil nutrient status. We have edited this to read “with soils that remain saturated and nutrient-rich throughout the year”
Line 120 / Figure 1 – Do you have a higher quality map of the African continent? Parts of it look out of proportion from my eyes.
A clearer map was used to replace the old one
Line 129 – Regarding the sampling design, how can you be sure about the recovering vs intact or reference site. Did you use some historical aerial imagery? Or do plant communities determine if it hasn’t been impacted before?
We determined wetland condition (recovering vs. intact vs. converted) using a combination of two approaches. First, historical LULC maps (1987–2023, Section 2.1.2) allowed us to trace land-use history and identify sites that had previously been converted to agriculture and were subsequently abandoned or restored, versus sites showing no history of conversion. Second, semi-structured interviews with farmers, village heads, and county heads (Section 2.1.3; see also Gubamwoyo et al., 2025) provided local knowledge on land-use history, prior disturbance, and management interventions at each site.
Line 151 – It is a little unusual to change the method of GHG analysis between sampling runs. Were you able to calibrate the two methods to indicate the two separate methods were consistent. I understand that the LI-COR is probably much more convenient being able to measure in the field. But changing methods might add additional uncertainty to the results.
We changed the method to continue measurements with more efficiency, as measuring right at the chambers and allow to measure at reduced costs. The use of sampling into gas vials and sending them to the laboratory was becoming very expensive and would not enable us to have enough data. LiCOR analyzers were also only available at the last campaigns but not before. When we started using this method, we sampled using the gas vial method and the LICOR method at the same time and compared the results to ensure reliable and comparable results. The comparison of results showed a good fit with an R2 of 0.91. We also needed to run some experiments that would generate quite a number of gas samples (leading to high expenses), but by using the LICOR we were able to analyze right at the experiments at reduced costs. In this experiment, we planted maize in one of the wetlands and varied the manure and fertilizer addition in combination with watering while measuring the GHG emissions.
Line 245 – How was pH determined? Soil-water mix ratio 1:5? Please elaborate.
pH was determined using a 1:2.5 soil-water mix ratio
Line 303 – The back-to-back parentheses are slightly clunky. Perhaps you could combine them with a semicolon, changing ‘(47.4 ± 5.8 µg g⁻¹ DW) (Figure S6)’ to ‘(47.4 ± 5.8 µg g⁻¹ DW; Figure S6)’. There are a few instances of this throughout the manuscript, although if you feel strongly, please leave them unchanged.
This has been corrected throughout the whole manuscript
Table 2 can be moved to Supplementary.
Table 2 is one of the core results shaping the discussion. Due to the journal requirements during review, the table was converted from landscape to portrait format. The final table will be in landscape form on one page.
L287 Why does the total number of recovering HVBWs differ from the number shown in Fig. 3?Two sites (Munuka, converted, and Ngelenyi, recovering) were removed from the dataset as they were both above the 99th percentile (z-score=2.9). (Line 288-289)
Line 526 – It could be more meaningful to readers to move from kg to Mg or t for these larger numbers.
Thank you the values have been converted to Mg (Mega grams)
Line 549 – This reads as part of the text. Is it associated with table 2?
Line 549 is part of Table 2. Due to the journal requirements, the table was converted from landscape to portrait format.
Line 655 – Butterbach-Bahl, K. et al. 2011 – DOI does not work when clicked.
I have updated the reference to make it more accessible.
Citation: https://doi.org/10.5194/egusphere-2026-3028-AC2
-
AC2: 'Reply on RC2', Sharon Gubamwoyo, 30 Aug 2026
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 96 | 38 | 14 | 148 | 21 | 14 | 8 |
- HTML: 96
- PDF: 38
- XML: 14
- Total: 148
- Supplement: 21
- BibTeX: 14
- EndNote: 8
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
First of all, thank you for contributing valuable data on various wetland types in Sub-Saharan Africa. It is an impressive dataset that has required significant effort to collect and analyse. The manuscript is clearly written and well structured and the addressed scientific questions fall within the scope of BG. However, I have several suggestions to improve both the text and the figures (see the comments below). Some of the figures and figure captions should also be clarified (see the specific comments in the Results section).
Abstract:
L20 Add “global” before “greenhouse gas” so it reads “global greenhouse gas budgets”.
Graphical abstract: Nice figure! Just a suggestion: use two different colours for CO2 and CH4.
Introduction
L49 Delete “contribution to”.
L52 Change the sentence so it reads “due to population and economic growth as well as drying lowlands caused by climate change”.
L52-53 Change the sentence so it reads “Specifically, this leads to changes ..., increased nutrient loading via manure or synthetic fertilized addition application...”.
L58-60 Change the sentence so it reads “Despite the importance for
livelihoods and to climate feedbacks, the impact of land use change on SOC stocks and GHG emissions in HVBWs has not been previously studied”.
L70 Insert “, typical to intact wetlands,” after “Under carbon-rich, low-nitrate conditions” so it reads “Under carbon-rich, low-nitrate conditions, typical to intact wetlands, ...”.
L72 Add “for agricultural land use” after conversion.
L74 Please briefly explain what is meant by “recovering wetlands” here.
L77 Change from “within this and other water tower systems” to “within this and other systems”.
L79 Beggining of the sentence is missing, perhaps delete entirely “. through their organic-rich, fertile soils, water availability, and flat topography”.
L81 Delete “(KBA)” as the acronym is not further used in the text.
L82 Please specify which initiatives.
L83 Delete the sentence “Despite their economic and ecological importance, the effects of agricultural conversion have not been extensively studied”. The same is already stated in L58-60.
L86-87 Change “a significant SOC loss of approximately 50% and a corresponding increase in CO₂ emissions” to ¬“ ~ 50% loss of SOC”.
L87 Change “land-use pressure” to “agricultural pressure”.
L89 Are these forests located in the Taita Hills? Please provide more specific information.
L90-92 The sentence “Here, we ...” can be moved to the first part of the next paragraph, i.e., “This study focused on field-based GHG measurements, aiming to i) compare GHG fluxes and their potential biogeochemical controls seasonally and spatially between intact, recovering, i.e., abandoned for years to decades, and converted, i.e., currently used for agriculture, HVBWs...”.
L95 Change “planting” to “growing”.
L101 Delete “ploughing” as it is part of cultivation.
Materials and Methods
L109 Change “ideal” to “suitable”.
L110 Delete “(FAO)” as the acronym is not further used in the text.
L114 Please add the average annual precipitation values for the high- and lowlands.
L116 This should be Table S1 instead.
L120-121 Add (LULC) after the "Land use and land cover" and change "Highland Valley-Bottom Wetland sampling points" to "studied Highland Valley-Bottom Wetlands" in the figure description. In Figure 1, please use different colours for the catchment boundaries so that it is easier to distinguish the different catchments, particularly in Panel A. Under LULC, woodlands and plantation forests appear to have the same colour. Please use distinct colours for these categories. Also,please increase the line width for the stream orders so that first-order streams are clearly visible. The panels should be labelled with “a” and “b” to match the formatting of the other figures.
L126 Why there is a gap in the measurements?
L128 Change “under agriculture” to “under agricultural use”.
L131 Delete “Only one intact HVBW, which had never been used for agriculture, was found. It” and start the sentence with “The intact HVBW”.
L134 Please add genus or species of the ferns.
L152 What is meant by GC here?
L151-154 Why did you change the method for the last four sampling campaigns?
L153 Please add the manufacturer and the place where each device was produced. Please do the same for the other devices as well.
L155 How did you decide in which patches to install the collars?
L156 How many chamber bases in total?
L177 Delete (ILRI) as the abbreviation is not used further in the text and explain the abbreviation “GC-ECD/FID”.
L198 Delete “Moreover, ”.
L206 How many of CH4 fluxes were eliminated? How many data points did you have in the end for each gas?
L257 Please add reference for the nlme package.
L265 The version is mentioned twice in the same sentence.
Results
Please report the statistical test used as well as the p-value.
In Fig. 2-3, 5, 6, please use a capital letter for the wetland type, i.e., "Intact".
In Fig.3, please add the length of Rewetting season in days?
Fig. 5, I am not sure I understand the letter combinations, for example, "Cab" for the N₂O–N flux from converted wetlands during the rewetting season. Does this mean that the flux was significantly different from the fluxes of the other wetland types? What does “ab” mean here? I think the explanation could be improved to make it easier to follow. The same goes for other figures where the letter combinations are used.
In Fig.6, please change “Wetland” to “Wetland type”. Are rewetting events included here?
Table 1. The reference (Nakagawa & Schielzeth 2013) can be deleted.
Table 2 can be moved to Supplementary.
L287 Why does the total number of recovering HVBWs differ from the number shown in Fig. 3?
L296 Change “Figure” to “Figures”.
L303 Change “eight-folk” to “eight-fold”.
L358-359 Change the figure title so it reads “Greenhouse gas fluxes from intact, recovering and converted Highland Valley-Bottom Wetlands (HVBW) across dry, rewetting and wet seasons”
L372 “during the wet season” is written twice.
L373 Delete “d”.
Discussion
L402 Delete space “CH ₄-C”.
L409 For “papyrus-dominated” please change to species name.
L421 Please add reference/-s at the end of the sentence.
L426 Change “NO3-N” to “NO3--N” and check this also in the rest of the text.
L462-463 Perhaps it would be enough with two references here.
L465-466 Change “C m−2 yr−1” to ”C m−2 yr−1” and add your results for comparison, i.e., after “Taita Hills”.
L481 Is there any CO2 data available from similar studies to compare with?
L518 Change “OC” to “organic carbon”.
Supplementary
Table S2 This should be Table S1 instead. Please add the scientific name of Royal ferns.