Carbon stock of Agoro Agu Central Forest reserve, in Lamwo district, Northern Uganda

Abstract

Terrestrial ecosystems such as forest landscapes provide critical ecosystem services such as carbon sequestration, fundamental to people, society and the global climate change discourse. Just like moist forest, dry Afromontane forests too present a high carbon sequestration potential. However, Uganda has since not undertaken carbon stock inventories for these conservation areas, especially with dry Afromontane forests like Agoro-agu central forest reserve (CFR). So, their potential to capture and store carbon is yet to be understood in Uganda. This study was carried out to estimate carbon stock of Agoro-agu CFR, for its potential in climate change mitigation. A stratified sampling design was used, where 65 sample plots were established. Nested, fixed area circular sample plots with sub-plots of varying radii for tree height, diameter measurements and soil sampling, were used. The mean total carbon stock of Agoro-agu CFR was estimated at 606.7 Mg C ha⁻¹, for which 409, 72, 124 and 0.24 Mg C ha⁻¹ was stored as above ground carbon, below ground carbon, soil organic carbon and carbon in litter herbs and grass respectively. The study illustrates the carbon sequestration potential of the forest reserve for any results-based payment projects for climate change mitigation. This is mainly due to the interventions through collaborative forest management arrangements and the pro-poor reducing emissions from Deforestation and forest Degradation (REDD+) pilot project in the landscape. This however, calls for more multi-stakeholders’ collaboration from direct resource users to national level to enhance forest conservation and reduce forest degradation for sustainable resource benefits and other ecosystem services.

1. Introduction

Forest ecosystems cover about a third of the total land mass on the earth's surface [1]. These terrestrial biomes sequester and store carbon (C) through the process of photosynthesis thus contributing to the stabilisation of atmospheric Carbon dioxide (CO₂) concentrations [2,3] Forests store C in five major pools which include; above-ground and below-ground biomass, deadwood, litter and soil carbon [4]. Intergovernmental Panel on Climate Change (IPCC) estimates a net uptake by terrestrial ecosystems from less than 1.0 to as much as 2.6 Pg C year⁻¹ for the 1990s [4]. Globally, forests store about 289 G t of carbon (1 G t = 10¹⁵ g) in biomass alone, representing about 1060 G t CO₂-equivalents (CO₂-eq) [5], thereby playing a crucial role in global climate regulation [6].

International negotiations, for example the Kyoto Protocol and Paris agreement [7,8], to limit greenhouse gases also require an understanding of the current and potential future role of forest carbon emissions and sequestration in both managed and unmanaged forests. The 2007 Bali Action Plan acknowledges the importance of carbon accounting through reducing emissions from Deforestation and forest Degradation (REDD+), and the nationally determined contributions (NDCs) providing avenues of results-based payments for carbon offset projects [8]. Certainly, developing countries have taken commitment to meet these ambitions which require an explicit understanding of the carbon sequestration potential of forest landscapes. It is clear that inventorying carbon of a forest is key for estimating carbon offsetting and its potential as a sink for carbon dioxide.

In Uganda, forest cover was estimated at 24% of total land area in 1990, which has declined to 2.5 million hectares (12%) in 2015 [9,10]. Forest area loss to other land uses such as agriculture and settlements, contribute to loss of forest carbon stocks. This poses a great threat to the future of forests potential of providing ecosystem services and sector contribution to the country's gross domestic product (GDP). To this effect, Uganda reported its forest reference emissions level (FREL) in 2017 to the UNFCC [9] which predicted that Tropical high forests stored up to a range of 61–240 t ha⁻¹ woodlands encompassing tropical shrubland and dry forests sequestered up to a range 9–94 t ha⁻¹ and plantations up to a range of 9–71 t ha⁻¹ [9].

Carbon stock assessments in tropical forest landscapes is through either direct or indirect approaches. Direct approaches involve the use of destructive sampling techniques to develop plot level or species-specific models to quantity biomass that is later converted into Carbon or tons of carbon dioxide equivalents (t CO_2eq), while indirect approaches employ developed allometric models or equations, tested form factors, area expansion factors and wood basic density [10]. Each method has its pros and cons for example direct approach yields results with a high precision and reliability, it is costly in terms of time, with restrictions where a forest or a species is threatened with degradation and extinction respectively. Allometric equations are less time consuming and do not harm the subject under study [11]. Depending on the concentration of study, models can be developed from species level, to habitat or biome level, to regional level to regional level [12]. This study uses allometric equations for dry pan tropic forests by Ref. [11].

Dry forests represent about half of tropical and subtropical vegetation stretching from expanses of Africa, South America and the Asian Pacific [13]. Dry forests occupy areas receiving a tropical climate with precipitation ranging from 500 to 1500 mm per annum. Africa's dry forest landscapes account 70–80% of forested areas, and are mostly in highland areas between 1000 and 2700 m above sea level [14]. In East Africa, dry Afromontane vegetation extend from the southern part of Arabian Peninsula to the eastern arm of the rift valley in the Drakensberg Mountains. These ecosystems have been reported to be rich in biological diversity and as centres of endemism with various range-restricted species [13]. Agoro-Agu CFR is one of the dry Afromontane Forest in Uganda at the border of Uganda and South Sudan, in Agoro hills. Recent scholars have reported that Afromontane landscapes are biodiversity hotspots of both flora and fauna [15,16]. The vegetation of Agoro-Agu CFR is largely classified as dry Combretum savanna, wooded-savanna mosaics and dry highland forest dominated by Combretum molle R. Br ex G. Don, Vachellia hockii De Willd, Hagenia abyssinica Steud. ex. A. Rich and Terminalia glaucescens Benth species [13,16].

Dry land forests also act as sink and sources of carbon given their expansive nature and are reported to store about one third of global carbon budget [10]. There is a wealthy body of literature on carbon stocks and storage potential of dry forests at both international and regional level but not to a commune level of Agoro-Agu CFR, considering the heterogenous vegetation communities of the area. Further, previous studies in Uganda have quantified forest carbon stocks but in moist montane forests and tropical rainforests. In the recent past, forest disturbances have been reported in the Agoro-Agu CFR landscape which is anticipated to affect carbon storage and greenhouse gas emissions of the forest. Therefore, this study was carried out to understand the carbon stock potential of Agoro-Agu CFR in Northern Uganda, whose result could provide baseline information to resource users and managers for any carbon-based projects in the Landscape.

2. Materials and methods

2.1. Study area

Agoro-Agu CFR is found between 3°40-3° 53′ N and 32°42′-33°04E, in Lamwo district, Northern Uganda which is about 460 km (km) from Kampala, the capital city of Uganda [15], and at an altitude of 2600 m above sea level (Fig. 1). Gazetted in 1937, it occupies 26,508 ha (Ha) of land with enclaves of Barley plantations [16].

Fig. 1.

Location Agoro-Agu CFR in Northern Uganda.

The forest reserve is drained by Okura stream in the East and Aringo stream in the West of the reserve. The CFR receives a bimodal rainfall pattern from late March or early April and late November, with highest in April and August, ranging from 800 mm–1000 mm (Fig. 2). December to mid-March are the hottest months with temperature ranging between 20 °C and 28 °C [15,16].

Fig. 2.

Average rainfall and temperature of Agoro-Agu CFR (1990–2021).

Gray-brown sands superimposing red clay or brown sandy loam soils describe the area. The vegetation of the landscape is largely Afromontane undifferentiated forest majorly of Combretum wooded landscapes, assemblages of low land bamboo, and Acacia thickets at the foothills of Agoro Agu hills [16]. The CFR is also said to have endemic species of the Eastern Afro Montane hot spot, and among the 15 forest reserves in Northern Uganda that were reported to have endemic species than any other forest reserve in Uganda [13].

2.2. Sampling design and procedure

Agora-Agu landscape was purposively selected because of the its participation in Reducing Emissions from Deforestation and Forest Degradation (REDD+) pilot project, a high conservation value area and long history of deforestation and forest encroachment by the local communities [17]. A reconnaissance field tour of Agora-Agu CFR was conducted in August 2021 with the help of the forest rangers and support staff under the instruction of the NFA administration. From this reconnaissance survey, land cover variations were identified based on the [18] guidelines and the Uganda Biomass land cover map [19,19]. It is from this reconnaissance survey that vegetation measuring and soil sampling were taken, since the vegetation was nearly homogenous (tropical highland woodland forest). A total of 65 sample plots were established randomly across the forest reserve.

Stratified random sampling design based on land cover was employed to achieve precise biomass and carbon stock estimates. Stratification ensures that measurements are more alike within each stratum than in the whole sample frame [12]. For this study, temporary sample plots (TSPs) were laid. Fifteen parallel transects with 1 Km distance were laid throughout the forest at every 500 m distance apart. Nested, fixed area circular sample plots with sub-plots of varying radii (Fig. 3), were laid out in the identified LCTs. Nested circular plots are efficient in inventorying individual trees that grow at different rates accounting uneven size distribution [12]. For each of these nested plots, three concentric plots of radii 20 m, 14 m and 4 m were laid out at each of the sampling plot.

Fig. 3.

Size and shape of nested, fixed circular sample plots.

So at every 400 m along the transect, sample plots of 1257 m², (20 m radius) and 50 m² (4 m radius) (within the main plot) size were established for trees and shrub assessment, and grass, herb and litter sampling, respectively [20]. Five quadrats of (1 m × 1 m) were established along and the middle of the 4 m radius circular plot for grass, herb and litter sampling from which a composite was prepared. Overall, a total of 64 sample plots were set up for C stock estimation and vegetation records. Number of plots were calculated using the following formula (Eq. 1) adopted from Ref. [12]:

$$n=\frac{{\left(\sum _{i=0}^{n}Ni*Si\right)}^2}{\frac{N⌃2*E⌃2}{t⌃2}+\left(\sum _{i=0}^{n}Ni*Si⌃2\right)}$$ Equation 1

where: E = allowable error or the desired half-width of the confidence interval. Calculated by multiplying the mean carbon stock by the desired precision (that is, mean carbon stock × 0.1, for 10% precision, or 0.2 for 20% precision), t = the sample statistic from the t-distribution for the 95% confidence level. t is usually set at 2 as the sample size is unknown at this stage, N_i = number of sampling units for stratum i (area of stratum in hectares or area of the plot in hectares), n = number of sampling units in the population and Si = standard deviation of stratum i.

Plot data were extrapolated to full-hectare area to generate carbon-stock estimates (Mg ha⁻¹). Extrapolation uses expansion factors by calculating the proportion of a hectare (ha) (1 ha = 10,000 m²), that is occupied by a given plot and so, this study adopted expansion factors outlined by Ref. [12]. That is, for a series of nested circular plots measuring 4, 14, and 20 m in radius being used, their areas equal 50, 616 and 1257 m² respectively and so, their expansion factors for converting the plot data to a hectare basis are 198.9, 16.2, and 8.0 respectively.

2.3. Biomass inventory and carbon estimation

2.3.1. Biomass inventory

Tree species were identified and measured for diameter at breast height (DBH) (1.3 m aboveground level) and height (H) [21]. Whereas for the aboveground biomass (AGB) estimation, field transects of 200 m distance of circular sampling method were employed [12]. In the plot, local names of trees were recorded and identified with their English and scientific names based on [22] book. Species not identified in field were collected (specimens), pressed and taken to Makerere University, Kampala, Uganda herbarium for further identification.

The biomass of herbs and litter were collected during the end of the rainy season, since it is the expected peak growth period. All the herbaceous vegetation emerging within the quadrat areas (1 m × 1 m) were cut at the ground level, weighed with a 100 g precision balance to get a composite sample that was placed in a marked bag for oven-dry mass determination in the laboratory [12]. For all samples collected, location data (x and y) coordinates were captured using Garmin GPS.

2.3.2. Soil sampling

Samples were collected for bulk density and soil organic carbon (SOC) analysis. Soil samples were taken from quadrants (1 m²) found in the five directions (north, south, centre, east and west) of the inner circular sample plot. Samples for the determination of SOC were collected from 30 cm soil depth within 1 m × 1 m quadrat where five (5) samples were picked by using auger, and the five soil samples were mixed to form a composite mixture [12]. The collected soil samples were air-dried, then titrimetric method was used to determine SOC based on [23].

2.4. Carbon stock estimation

2.4.1. Above ground biomass and carbon

A combination of forest inventories with allometric tree biomass regression models were used to determine the aboveground biomass (AGB) of the study area [4,24]. These included; 1) The selection and application of an allometric biomass function for the estimation of individual tree biomass, 2) The summation of individual tree AGB to estimate plot AGB, and 3) The calculation of an across-plot average to hectare based.

In this study, the most recent pan-tropical tree AGB equation [11] based on D, H and wood density (ρ, g cm⁻³) was adopted (Eq. (2)). Therefore, the following allometric equation was used to calculate the AGB:

$$AGB\left(Mg/ha\right)=0.0673{\left(\rho D^{2}H\right)}^{0.976}$$ Equation 2

where, where AGB is aboveground biomass of trees (Mg ha⁻¹), ρ is the specific wood density (g cm⁻³), D is the diameter at breast height (cm), and H is the average height of trees (m). Above ground carbon was estimated at 50% of AGB of the study area [4].

Wood densities (ρ) were sourced from the online databases [25,26]. However, where species-wise value was not found, genus-wise mean was used, and if genus-wise mean could not be calculated, the mean values for species in different regions were used [27].

The Chave [11] dataset is limited to only old-growth or secondary woody vegetation and deficient of other species and forestry systems. Therefore, a separate equation was used for Acacia sp, which are common in dry Afromontane forests in Uganda. The equation (Eq. 3) for Acacia sp AGB estimation was;

$$AGB\left(Mg/ha\right)=\mathit{exp}\left(-1.59+2.19x\mathit{ln}\left(D\right)x1.05\right)$$ Equation 3

2.4.2. Below-ground biomass and carbon

A regression model for tropical forests was adopted from Ref. [28] to estimate below-ground biomass (Eq. (4)) (living and dead, fine and coarse roots) as a function of above-ground biomass.

$$BelowGroundBiomass\left(BGB\right)\left(Mg/ha\right)=\mathit{exp}\left(-1.0587+0.8836x\mathit{ln}AGB\right)$$ Equation 4

where BGB is below-ground biomass in Megagrams per hectare (Mg/ha) and AGB is above-ground biomass (Mg/ha).

2.4.3. Herb and litter biomass estimation

All litter and herbs within the frame and samples from the subplots were pooled and weighed of 100 g (g). A well-mixed subsample was used to determine the oven-dry-to-wet mass ratios to convert the total wet mass to oven-dry mass (Eq. (5)). The collected samples from herbs and litter were taken to laboratory, oven-dried at 70 °C till constant weight was reached. For the forest floor (understory wood vegetation, herbs, grass, and litter), the amount of biomass per unit area was given by Ref. [12]:

$$Oven-drymass=\frac{SubsampleDrymass\left(kg\right)}{SubsampleFreshmass\left(kg\right)}x\left(FieldMass(kg\right)$$ Equation 5

The carbon content in herbaceous biomass was calculated by multiplying herbaceous biomass by 0.5 [4].

2.4.4. Soil organic carbon

Soil carbon analysis based on soil samples collected from the field temporary sample plots. Soil samples of 100 g were air-dried for 48 h, well mixed and sieved through a 2 mm mesh size sieve. Then SOC was analysed following [23], method with laboratory work conducted at Makerere University Soil Testing Laboratory. Bulk density (Eq. (6)) was determined after drying the core samples of soil at 105 °C where the weight of the soil was divided by the volume of the core sampler. The weight of the gravel above 2 mm diameter was subtracted to determine the bulk density (Eq. (6)) of the soil samples.

$$\rho =\frac{ODW}{CV-\left(\frac{RF}{PD}\right)}$$ Equation 6

Where, $\rho$ = Bulk density of the < 2 mm fraction, (gm⁻³), ODW= Oven-dry mass of fine fraction (<2 mm) in g, CV= Core volume (cm3), RF = Mass of coarse fragments (>2 mm) in g, PD = Density of rock fragments (g/cm³) (Estimated at 2.65 g m⁻³).

The SOC was estimated using (Eq. (7)) [12]:

$$SoilOrganicCarbonStock\left(SOC\right)=\left[\rho xdx\%C\right]100$$ Equation 7

Where, $\rho$ = Bulk density (g m⁻³), d = depth of soil sample (cm) and %C= Percent carbon concentration

2.4.5. Total carbon stock estimation

The total carbon stock was calculated as a summation of all carbon pool stock using the following formula (Eq. 8) according to Ref. [12]:

$$\mathrm{T}{\mathrm{C}}_{\mathrm{S}\mathrm{t}\mathrm{o}\mathrm{c}\mathrm{k}}={\mathrm{C}}_{\mathrm{A}\mathrm{G}\mathrm{B}}+{\mathrm{C}}_{\mathrm{L}\mathrm{H}\mathrm{G}}+{\mathrm{C}}_{\mathrm{B}\mathrm{G}\mathrm{B}}+\mathrm{S}\mathrm{O}\mathrm{C}$$ Equation 8

where, T C _stock = carbon Stock [Mg C ha⁻¹], C_AGB = carbon in above-ground biomass [Mg C ha⁻¹], C_BGB = carbon in below-ground biomass [Mg C ha⁻¹], C_LHG = carbon in litter, herb & grass [Mg C ha⁻¹], and SOC = Soil organic carbon [Mg C ha⁻¹].

The total carbon stock was converted to tons of CO₂ equivalent by multiplying it by 44/12, or 3.67 [12].

Further, a Pearson correlation analysis was performed to determine the relationship between SOC and vegetation parameters. All analyses were performed in MINITAB version 14 at 95% confidence level. Tukey's honestly significant difference test was executed to separate means.

3. Results

3.1. Biomass and carbon stock of different carbon pools

The total number of trees enumerated in the forest reserve were 677. Eighty-five species (85) were identified from 54 genera belonging to 28 families. The mean, maximum and minimum DBH of sampled trees was 21.04 cm, 42.84 cm and 2.87 cm respectively. Ficus ingens and Oxytenanthera abyssinica exhibited the maximum and minimum DBH respectively. The mean, maximum and minimum height (H) of trees were 9.2 m, 24.8 m and 1.7 m respectively. Prosopis africana and Pterygita mildbraedii recorded the minimum and maximum tree heights respectively.

The mean AGC stock of Agoro-agu CFR was 409 ± 32.5 Mg C ha⁻¹. The first top five species which stored the highest above ground carbon stock of the reserve were Vangueria apiculata, Ficus glumosa, Vitellaria paradoxa, Ficus saussureana, and Hymenocadia acida with values of 784.87, 704.54, 525.47, 498.20 and 497.99 Mg C ha⁻¹ respectively. While the lowest above ground carbon stock was reported in Stereospermum kunthianum, Steganotaenia araliacea, Bridelia scleroneura, Vernonia amygadalina, and Bridelia scleroneuroides at 8.10, 8.07, 6.35, 4.43 and 3.54 Mg C ha⁻¹ respectively.

The mean BGC, SOC and carbon stock of litter, herbs and grass was 72.32 ± 6.13, 124 ± 3.51, and 0.24 ± 0.011 Mg C ha⁻¹ respectively.

The mean soil bulk density was 0.60 ± 0.012 g cm⁻³ with a range of 0.32–0.88 g cm⁻³. Plots located in the higher elevation, where human disturbances and encroachment are hardly reported recorded high soil bulk density values depicting the high mineral soils in the hills of Agoro-agu CFR. There was no much variation in soil organic matter, where 17.4% and 13.72% were the highest and lowest soil organic values of the study area. The total carbon stock values of the forest reserve ranged from 242.4 in the area of plot 55 to a maximum of 804.17 Mg C ha⁻¹ in the area of plot 17 (Fig. 4), respectively.

Fig. 4.

Total carbon stock (TC) and CO_2eq per plot.

The mean total carbon stock of all carbon pools in Agoro-agu CFR was 606.7 ± 37.9 Mg C ha⁻¹, with the total carbon values ranging from 168.6 to 1284.4 Mg C ha⁻¹. The mean carbon stock across the carbon pool was statistically significant (P = 0.000), with a 58% variation. There was a significant difference between soil organic carbon and DBH (P = 0.03), while other pools remained insignificant.

3.2. Relationship between carbon pools and dendrometric parameters

3.2.1. Contribution and pearson correlation between carbon pools and vegetation parameters

The analysis of individual carbon pool contribution to the total carbon stock sequestered by the forest reserve showed significant variations. Much of the carbon in the forest is stored in above ground biomass (67.55%) followed by soil organic matter (20.49%), below ground carbon pool (11.92%), and litter, herbs and grass (0.04%) respectively (Fig. 5). From this study result, litter, herbs and grass contribution to the carbon stock of the forest was insignificant.

Fig. 5.

Percent Carbon Contribution of different Carbon Pools in the study area.

A Correlation between carbon pools and tree parameters was tested using Pearson's correlation coefficient (Table 1), which showed a significant positive correlation between height and DBH, AGC and DBH, and BGC and AGC.

Table 1.

Pearson's correlation coefficients and P-values between carbon pools and dendrometric parameters.

	DBH	H	AGC	BGC	SOC	LHG
H	0.906*	1
	0.000
AGC	0.239*	0.108	1
	0.057	0.394
BCG	0.134	0.024	0.748*	1
	0.291	0.851	0.000
SOC	0.367	0.235	0.132	0.051	1
	0.33	0.62	0.297	0.689
LHG	0.092	0.039	0.010	0.171	0.137	1
	0.467	0.762	0.938	0.176	0.281

Notes: Pearson's correlation coefficients (Upper cell values), P-values (lower cell values); *P < 0.05, H-height, BDH-Diameter at breast height, AGC-Above ground carbon, BCG-Below ground Carbon, SOC-Soil Organic Carbon, and LHG-Carbon in Litter, herbs and Grass.

3.3. Limitations of the study

The current study faced a limitation of absence of species-specific allometric equations. This was overcame by adopting pan-tropical allometric models and genus models applied on Acacia sp respectively. The other challenge was absence of reference biomass data for the study area. Uganda has reference biomass data from 1990, but previous assessments never covered the northern part of Uganda due to political insurgency between 1986 and 2007. This present study provides baseline carbon stock survey of the forest reserve for future studies and national monitoring.

4. Discussion

4.1. Carbon stock of Agoro-agu central forest reserve

International efforts such as the United Nations Framework Convention on Climate Change (UNFCCC) Bali conference in 2007 [8] and the Paris agreement of 2015, made resolutions to combat climate change [7]. These among others included; developed countries adopting national emission reduction targets and providing least developed countries with climate financing and capacity building; while developing countries majorly undertook mitigation actions [29]. By 2020, member states were to submit their nationally determined contributions (NDCs) for climate action.

Uganda committed to reduce GHGs emissions by 24.7% by 2030 relative to the business as usual scenario of 148.8 Mt CO_2-eq in 2030 and 235.7 Mt CO_2-eq by 2050 under the Business-As-Usual (BAU) Scenario [9]. The total emission reduction potential for the country is about 112.1 Mt CO_2-eq. The remaining 37 Mt CO_2-eq is secured for results-based carbon payments such as sale of carbon credits. 82.7% of the mitigation impact was estimated to come from the agriculture, forestry and other land uses (AFOLU) sector, while 7.56%, 6.36%, 3%, and 0.4% were to come from the transport, energy, waste, and Industrial Processes and Product Use (IPPU) sectors, respectively. This underscores the need for reliable estimates of carbon sequestration potential of forest reserves in Uganda. In this study, Agoro-agu CFR alone has the potential to sequester about 9.7 M t CO_2-eq without consideration of carbon emission through land use change and degradation. To successfully participate in carbon market, reliable estimation of total carbon storage is essential [10]. The robust estimate is also precarious for sustainable forest management, for monitoring status of the forest and reporting carbon stock dynamics as required by Reducing Emissions from Deforestation and Forest Degradation (REDD+) mechanism [8,30].

From the current study, species that reported high carbon stock had high mean DBH. The result showed that DBH has a significant influence in carbon capture and storage of a forest over time [31]. Most of these tree species were encountered in the higher elevation areas of the forest that are in undisturbed forest zone [13]. One of the impeding factors for forest growth and trees attaining a bigger diameter is rampant forest degradation and subsistence farming triggering deforestation, especially in the lower elevation of the forest reserve [32]. The AGC and BGC in Agoro-agu CFR were higher than those estimated by Ref. [4] for Afromontane dry forests in the tropics, especially on the African continent. The higher average carbon stock in above ground biomass in the study area is also largely attributed to DBH [14] and elevation that greatly discourages human activities in the reserve.

The mean C stock in LHG in Agoro-agu CFR was generally low. This could be attributed to high rates of decomposition in the tropical areas but also because of the steep slopes, vegetation is under growth is limited upon litter being washed off to valleys [33]. Further, a large section of the forest reserve is under a closed canopy with less light under growth. This later affects the soil properties to support grass and herb growth. The mean SOC of Agoro-agu CFR was higher than mean SOC of Tropical & Subtropical Moist Broadleaf forests (57 t ha⁻¹) [34]. The differences in SOC between different biomes could be due to the diversity in tree, shrub species, variation in the physical and chemical properties of soil and differing soil profiles, climate, elevation, anthropogenic disturbances, and process use to determine organic soil carbon content [34]. Whereas the current study reported more carbon in above ground biomass, it has been reported that the soil carbon pool stores vast amounts of carbon but under situations of deep rooting systems by the overlaying vegetation that recharges the soil bank through decomposition and transportation, minimum to no disturbances of the soils especially by human activities. Dedicated efforts to maintain the forest reserve with minimal disturbances will increase amount of carbon sequestered over time [34,35].

The mean carbon stock of Agoro-agu CFR was higher than that reported of any Afromontane dry forest in African region. Recent estimates by Ref. [4], show that sub-Saharan Africa tropical forests, tropical seasonal forests and tropical dry forest stored 212.9, 190 and 69.6 t ha⁻¹ respectively. This study is in contrast with previous studies in the region that have reported low carbon estimates in dry forests, for example [27], reported a range of a mean of 340.92 Mg C ha⁻¹ from five forests in Northern Ethiopia, while [14] reported 555.03 Mg C ha⁻¹ from Wujig Mahgo Waren forest, Northern Ethiopia. The differences in carbon stock across different biomes could be explained by inaccurate determination of tree dendrometric parameters of height and DBH, absence of site-specific or species specific allometric models, the presence relics with huge diameters accounting for high basal area, a higher population of woody species especially in secondary forests following natural (floods, natural fires, hailstorms, landslides) and anthropogenic modifications [32,36].

Statistically significant positive correlation between DBH and H as well as AGC stock and BGC explains the fact that vegetation parameters, and diversity of plant communities affect carbon dynamics of a forest ecosystem [14,37,38], as they have been reported to facilitate decomposition of plant material for soil carbon formation. The high carbon stock of the forest reserve maybe largely attributed to the large diameter, heights and density of trees. The presence of high carbon stock in the forest demonstrates its potential for climate change mitigation. Agoro-agu CFR should therefore be sustainably managed through reduction of deforestation and land degradation, promotion of sustainable landscape management approaches such as collaborative forest management to enhance the reserve carbon storage and carbon sequestration potential to mitigate effects of climate change and ensure continued provision of other ecosystem services.

5. Conclusion

The current study quantified carbon stock across carbon pools in Agoro-agu central forest reserve and the result show on average, a high carbon stock storage compared to the average value previously reported of tropical forests. Across the carbon pools, AGC was the highest followed by SOC in the forest reserve. Nevertheless, the influence of litter, herbs and grass carbon pool was statistically insignificant. All carbon pools were positively correlated with dendrometric parameters (DBH and H). AGC pool was statistically significantly correlated with BGC, and so does DBH and H. The result of a high carbon stock in Agoro-agu forest reserve illustrates the carbon sequestration potential of the forest for any results-based payment projects like REDD + for climate change mitigation. It has also been understood that the forest landscape was used as a pilot area for the pro-poor REDD + pilot project by IUCN. This certainly had implications on forest coverage and subsequent carbon storage of the landscape, where communities were supported through demonstration of a payment for environmental services (PES) scheme that rewarded for tree planting and carbon sequestration, as well as formation of Community Environment Conservation Fund (CECF) as the financial mechanism to support the deployment of livelihood options. This reduced encroachment for wood products and promoted sustainable use of the forest reserve. Therefore, dedicated efforts to enhance forest conservation and reduce forest degradation will require multi-stakeholders’ collaboration from direct resource users to national level to enhance the provision of ecosystem services.

Author contribution statement

Vicent Birungi: Conceived and designed the experiments; Performed the experiments; Analysed and interpreted the data; Wrote the paper. Sintayehu Workeneh Dejene, Michael S Mbogga: Conceived and designed the experiments; Analysed and interpreted the data; Wrote the paper. Marc Dumas-Johansen: Analysed and interpreted the data.

Funding statement

This work was supported by Africa Centre of Excellence for Climate Smart Agriculture and and Biodiversity Conservation.

Data availability statement

Data will be made available on request.

Declaration of interest's statement

The authors declare no conflict of interest.

Biographies

Vicent Birungi is a Graduate Associate at the Africa Centre of Excellence for Climate Smart Agriculture and Biodiversity Conservation, Haramaya University. His research interests include climate change, Ecological Modelling, GIS and Remote sensing, and Landscape ecology.

Sintayehu Workeneh Dejene is an Associate Professor at the Africa Centre of Excellence for Climate Smart Agriculture and Biodiversity Conservation, Haramaya University. His research interests include Biodiversity and Ecosystem Services, Disease Ecology and Climate Change

Michael S Mbogga (PhD) is Senior Lecturer at the Department of Forestry, Biodiversity and Tourism, in the School of Forestry, Environmental and Geographical Sciences, Makerere University. His research interests include Climate change, bioclimate envelope modelling and conservation genetics

Marc Dumas-Johansen is an agriculture and food security specialist with the Division of Mitigation and Adaptation at the Green Climate Fund (GCF), based in Incheon, Republic of Korea. His research interests include Climate change, food security and natural resources management.

Appendix A. Supplementary data

The following is the Supplementary data to this article.