The study area for this research was Polesgo landfill, ranging from 12˚25'08" to 12˚25'53"N and from 1˚30'41" to 1˚31'12"W. The Polesgo landfill was built and commissioned in 2005. It has a capacity of 6.1 million cubic meters of waste and offers an operating capacity of twenty (20) years. It covers an area of 70 hectares and is located about ten kilometers north of the city center. The Polesgo landfill has two missions: firstly, solid waste burial and secondly, solid waste valorization (i.e. composting, plastic recovery). Twenty-four wells are integrated into landfill for future in situ measurement. Figure 1 gives an aerial view of Polesgo landfill, which was named Polesgo Waste Treatment and Recovery Centers, according to this second mission.

LFG generated in a landfill will have different pathways, as illustrated in Equation (1). If CH₄ generation is significant (in the LFG), LFG can be collected and utilized for power and/or heat production, upgraded to biogas or flared for thermal conversion to CO₂ (Stegmann, 1996). Part of the LFG will migrate to the top cover, where it can be oxidized to CO₂ by methanotrophic bacteria in the cover soil. However, some of the migrating LFG may also escape into the atmosphere without any oxidation and add to anthropogenic CH₄ emission and accumulation in the atmosphere. LFG can also migrate laterally to surrounding areas, and finally some will be temporarily stored inside the landfill. This can be summarized in the following CH₄ balance for a landfill [28] (with all units in mass time⁻¹)

$\begin{array}{l}{\text{CH}}_4\text{generated}={\text{CH}}_4\text{emitted}+{\text{CH}}_4\text{oxidised}+{\text{CH}}_4\text{recovered}\left(\text{flared or utilised}\right)\\ {\text{+CH}}_4\text{migrated+}\Delta {\text{CH}}_4\text{storage}\end{array}$ (1)

Recorded CH₄ is based on flow and CH₄ concentration measurements, generated CH₄ is often modeled based on landfilled waste amounts and waste compositions, using different models for LFG generation [29] . The emission of CH₄ can be quantified by using either remote methods, such as tracers gas dispersion, DIAL (Differential Absorption LiDAR) or radial plume mapping [30] [31] , or surface-based point measurements (e.g. flux chamber measurements) that integrate total emissions [28] [32] .

Most available global models which predict biogas generations from landfills are among the ones developed based on first-order decay models. These models consider quality of waste (i.e. moisture content, carbon content, age of waste and ability of waste to be digested), waste quantity and condition of the landfill (i.e. climate, temperature, precipitation) implicitly. In order words, the effect of depletion of carbon in the waste through time is accounted for in a first-order model.

1) US EPA’s LandGEM version 3.02

The LandGEM was developed by the Control Technology Center (CTC) of the United States Environmental Protection Agency (US EPA) [33] . It is an automated tool on MS Excel based interface to determine release rates of total LFG, methane and other contributing gases in LFG. However, it doesn’t include the categorization of wastes. LandGEM predicts the LFGs emissions based on a first order decay equation, which assumes that the CH₄ generation rate reaches its peak shortly after the initial waste is placed and decreases exponentially after that. The LandGEM model also assumes that the volume emission rate of CO₂ and CH₄ emissions are the same, with trace amounts of non-methane organic compounds and other air pollutants. In 2005, the current version 3.02 has been released, and it works in Windows Excel Environment. Equation (2) shows the first-order decay equation used to estimate CH₄ generation rate (Q, in m³/year) [34] [35] :

$Q_{{\text{CH}}_{\text{4}}}={\displaystyle \underset{i=1}{\overset{n}{\sum }}{\displaystyle \underset{j=0.1}{\overset{1}{\sum }}k{L}_0\left(\frac{M_i}{10}\right)\times \left(e^{-k{t}_{ij}}\right)}}$ (2)

where $Q_{{\text{CH}}_{\text{4}}}$is the annual methane generation in the year of calculation (m³/year), k is the methane generation constant (year⁻¹), L₀ is the methane generation potential (m³MgMSW⁻¹), M_i is the mass of waste in the i^th year (Mg) and t_ij is the age of j^th section of waste mass M_i accepted in the i^th year.

According to Equation (2), LandGEM model exhibits two input constants: k, the methane generation constant and, L₀, the potential methane generation capacity. Default or site-specific values of these constants may be used and are distinguished as “CAA” and “inventory” [36] . The “CAA” defaults are in accordance with the federal Clean Air Act (CAA) regulations for landfills receiving solid household garbage while the “Inventory” is based on emission factors from the “US EPA”.

2) First Order Multi-phase SWANA Model

This model is described by Van Zanten and Scheepers (1995). It assumes that the methane generation initially may be low. The generation then rises to a peak before declining in what is essentially an exponential fashion. Equation (3) shows the first-order multi-phase decay equation used to estimate CH₄ generation rate (Q, in m³/year):

$Q_{{\text{CH}}_{\text{4}}}=M_i\times L_0\times \left[F_{\left(r\right)}\times k_{\left(r\right)}\times e^{-k_{\left(r\right)}\left(t-t_1\right)}+F_{\left(s\right)}\times k_{\left(s\right)}\times e^{-k_{\left(s\right)}\left(t-t_1\right)}\right]$ (3)

where $Q_{{\text{CH}}_{\text{4}}}$is methane generation in m³/year; M_i is waste in place (tons) in the year i; Lo is methane yield potential in (m³MgMSW⁻¹); t is time after waste placement in years; t₁ is lag time (between placement and start of generations) in the year; k_(r) is the first order decay rate constant for rapidly decomposable waste; k_(s) is the first order decay rate constant for slowly decomposable waste; F_(r) is the fraction of rapidly decomposable waste and F_(s) is the fraction of slowly decomposable waste.

3) Models inputs

The value of the degradation constant of rapidly degradable waste k_(r) or slowly degradable k_(s) is estimated from Equation (4) [36] - [38] :

$k_{\left(r\right)/\left(s\right)}={\displaystyle \underset{i=1}{\overset{N}{\sum }}{k}_i\times }{F}_i$ (4)

where k_(r)/(s) is the first order decay rate constant for rapidly/slowly decomposable waste; k_i is the first order decay constant degradation of each waste category (rapidly/slowly degradable waste); F_i is the fraction of rapidly or slowly decomposable waste.

The quantity of Municipal Solid Waste (M_i) buried in the Polesgo landfill from 2005 to 2023 was provided by the municipal authorities of Ouagadougou. That from 2024 to 2025 was calculated according to Equation (5) [39] :

$M_i=t_c\times t_e\times n_{ja}\times w_c\times P_0\times {\left(1+r\right)}^t$ (5)

where t_c (%) is MSW collection rate (41%); t_e (%) is the rate of landfill of waste at the Polesgo landfill (90%); n_ja is the number of days in a year (365 days); w_c is the specific daily production of waste in Ouagadougou (0.62 kg/inhabitant/day reported by [13] ); P_o (3, 000,000 inhabitants) is the population of Ouagadougou for the year 2020 taken as a reference (in years); r (%) is the population growth rate of Ouagadougou (4.1%) and t is the number of extrapolation years (in year). Table 1 shows the waste annual tonnage, Q_waste, at Polesgo landfill from 2005 to 2025.

^*Projected with Equation (5).

According to Machado et al., 2009, if waste composition is known, methane yield potential, L₀ can be calculated from Equation (6) [40] :

$L_0=\frac{B{F}_w\times C_m}{1+w}$ (6)

where BF_w is waste biodegradable fraction, C_m the methane generation and, w the water content. The values of BF_w and C_m were calculated from Equations (7) and (8). Where BFw is waste biodegradable fraction, C_m the methane generation and, w the water content.

$B{F}_w={\displaystyle \underset{i=1}{\overset{n}{\sum }}B{F}_i\times F{R}_i}$ (7)

where BF_i is biodegradable fraction of each waste component and FR_i fraction of each component of dry-based waste.

$C_m=\frac{{\displaystyle \underset{i=1}{\overset{n}{\sum }}B{F}_i\times F{R}_i\times C_{m_i}}}{B{F}_w}$ (8)

where C_mi is the methane generation of each waste component. Table 2 shows MSW characteristics related to Polesgo’s landfill.

Concerning the methane generation constant, k, it can be obtained by an empirical expression given by Equation (9) if landfill meteorological data are available (Alexander et al., 2005):

$k=3.2\times {10}^{-5}\times \left(x\right)+0.01$ (9)

where x is the average annual rainfall (millimeters). The rainfall data available for Ouagadougou city since 1902 gives an annual average of 784.93 millimeters.

In the second approach, the quantities of generated and collected LFG are related by the collection efficiency as shown in Equation (10) [7] :

${\text{Q}}_{{\text{CH}}_{\text{4c}}}{\text{= Q}}_{{\text{CH}}_{\text{4g}}}\times {\eta }_{{\text{CH}}_4}$ (10)

where $\eta \left(\%\right)$ is methane extraction efficiency, ${\text{Q}}_{{\text{CH}}_{\text{4g}}}$is methane generated and ${\text{Q}}_{{\text{CH}}_{\text{4c}}}$is methane collected. The value of ${\eta }_{{\text{CH}}_4}\left(\%\right)$ of the Polesgo landfill was assessed in our previous studies. It has been estimated at 27% and 23% in 2017 and 2018 respectively [7] . In this study, an average value of the two determined in our previous study was applied to the other years.

The models used for modeling of the CH₄ production were corroborated by using two statistical parameters; namely (i) coefficient of determination (R²) and (ii) Root Mean Square Error (RMSE). These parameters are comprehensive enough to quantify the accuracy of the models used. R² was calculated by Equation (11) [8] . It represents the association between experimental CH₄ (Q_exp) and modeled CH₄ (Q_mod) at the time i for n number of years. The model with a greater value of R² establishes the better prediction.

$R^2=1-\frac{{\displaystyle {\sum }_{i=1}^n{(Q_{\exp ,i}-Q_{\mathrm{mod},i})}^2}}{{\displaystyle {\sum }_{i=1}^n{(Q_{\exp ,i}-\overline{Q_{\mathrm{mod},i}})}^2}}$ (11)

Table 3 shows the values of the input parameters of the models used. The value of methane yield potential obtained is close to that of conventional landfills or those located in arid areas and bioreactor type landfills (the value obtained is 0.983 times the conventional value for these types of landfills). Also, the waste degradation constant obtained from rainfall and Equation (9) is close to that given in the inventory for bioreactor type landfills (the value obtained is 0.88 times the conventional value for these types of landfills). This is close to reality insofar as the city of Ouagadougou can be considered an arid zone with respect to the average rainfall recorded (784.93 millimeters and spread over 3 to 4 months out of 12 months).

Two predictive models (SWANA and LandGEM) were used to estimate methane emissions from the Polesgo landfill. The methane emission results are shown in Figure 2 . Note that the model assumes that there is no biogas production in the first year of waste deposition, 2005.

According to models, from 2005 ≤ t ≤ 2026, the amount of methane increases linearly with the years, reaching a maximum value in 2026. And for t > 2026, the amount of methane produced will decrease exponentially after landfill closure, in parallel with the decrease in the amount of decomposable material in the landfill.

SWANA has predicted maximum methane emission rates of 4,426,559, 4,491,083 and 4,538,419 m³/year for the years 2023, 2024 and 2025 respectively, compared with 8,243,196, 8,733,855 and 9,213,521 m³/year for the same years according to the LandGEM model. The experimental data obtained over this period are 4,191,845 and 4,295,454 m³ in 2022 and 2023 respectively. Data for 2024 are not yet available at the time of writing. Trends in these models show that maximum methane emissions are expected in 2026, one year after the planned closure of the landfill.

Methane production data obtained from the two above-mentioned models and experimental data showed that the SWANA model better characterizes methane production at the Polesgo landfill. This is because the LandGEM model considers municipal waste to consist of a single category (source). According to this concept, all waste categories degrade in the same way and at the same rate, with the same average methane production potential. This is not realistic in practice [47] . However, when considering municipal waste by category, with the specific characteristics of each category (rapidly degradable, moderately degradable and slowly degradable), as considered in the SWANA model, methane production is overestimated but less exaggerated than with the LandGEM model. The discrepancies observed with the SWANA model (see Figure 2 ) can be explained by the difficulty of obtaining experimental data. As a reminder, the Polesgo landfill does not have a technical platform for monitoring landfill gas emissions. In situ measurements only began in 2016. The consistency of the experimental data remains rather low for a good correlation.

It should be noted that the data on the mass of waste buried in the landfill is approximate, given the weighing methods used in situ. Also, the characteristic data used to come from the 2017 characterization campaign, after the waste was already being degraded in the cells. As a result, there may be discrepancies between the data used and that for waste already landfilled.

In addition, the rate of landfill gas production decreases exponentially after the predicted closure of the landfill (2025), as the amount of decomposable material in the landfill decreases. This is because the model assumes that maximum production normally occurs in the year of closure (2025) or the following year (2026), and that landfill gas production decreases exponentially as the organic fraction of waste is consumed. Biogas concentration decreases with the age of the final landfill site; this fact is included in the model with parameters k and L₀.

The results of methane prediction and experimentation at the Polesgo landfill show that there is an untapped energy potential and thus a huge contribution to global warming. It is important to note that this potential can be used for energy purposes and broaden the country’s energy mix, as [48] pointed out. Mor & Ravindra, (2023) [49] go further, pointing out that landfill gas-to-energy is an essential component of an integrated municipal solid waste management strategy.

Table 4 shows the values of statistical parameters of the two models.

Considering methane production rate simulations, SWANA model showed better R² of 0.59 compared with the LandGEM model which had a value of 0.006 as depicted in Figure 1 and Figure 2 . Estimating the rate of methane production in landfills depends largely on the basic and secondary parameters that form the basis for developing equations or computer models. Fewer parameters and/or imprecise data would lead to more inaccurate results than field surveys. More complex and detailed parameters would provide a more accurate estimate [48] . A different mathematical model can be used to evaluate the various parameters, but due to poor analysis or lack of data, the model results are subject to considerable uncertainty. One of the major problems in simulating CH₄ emissions from landfills is the quality and abundance of inputs to the models used. This is the case with Polesgo’s landfill.

Comparing the measured values against those simulated via SWANA (model data/experimental data), we can see that the ratio evolves from 1.03 to 1.5 between 2017 and 2023. In a similar work, Plocoste et al., 2016 had obtained a ratio of 1.94 for the Gabarre landfill [36] , less interesting than those of the present work. In view of all these aspects, it should be noted that the SWANA model better reproduces CH₄ generation at the Polesgo landfill.

Figure 3 shows the potential electrical energy generated and carbon dioxide equivalent avoided by the Polesgo landfill in 27 years of operation (between 2024 and 2052).

Knowledge of the biogas generation potential of the Polesgo landfill is essential before implementing a biogas recovery plant for energy production. It should be noted that biogas generation at the Polesgo landfill is continuous due to the anaerobic decomposition of the organic fraction of solid waste. Consequently, if the biogas recovery facility is not operational, there will be an increase in pressure that will cause biogas to be released into the atmosphere. This would pose a real environmental threat, as methane from biogas is a greenhouse gas.

As with methane production, the estimated energy potential increases from 2024 to 2026. From 2024 to 2026, it rises from 3,829 to 3,896 MWh, with a maximum in 2026. After 2026, the energy potential decreases exponentially. Assuming that the biogas plant starts operating in 2025, by the 27th year of operation (in 2052), the recovered biogas will still be sufficient to generate 1038 MWh, representing a significant energy opportunity for Burkina Faso.

These high values of estimated energy potential could be explained by high values of biogas volume. The increase in energy potential from 2024 to 2026 can also be explained by the increase in methane volume over the same period. The exponential decrease in energy potential observed from 2026 onwards is explained by the exponential decrease in methane volume linked to the gradual reduction in the organic matter contained in landfilled waste.

Producing energy from waste at the Polesgo landfill can help Burkina Faso achieve its energy transition by gradually replacing fossil fuels. Converting waste into energy can become a green, renewable and sustainable energy source. What’s more, converting biogas into energy prevents the release of CH₄, a greenhouse gas, into the atmosphere. It’s important to note that these energy potential values can increase with biogas capture rates, which are currently very low (25%).

Over the 27 years of operation of the biogas plant (from 2025 to 2052), energy recovery from biogas at the Polesgo landfill would generate more than 61GWh, avoiding 1,340 GgCO₂e emissions into the atmosphere. Also, the evaluation of revenues from electricity and carbon credits derived from methane production yielded US$27.38 million who reported an average cost of US$10.5 CO₂e /tonne. This income represents a significant added value for the economy of the municipality hosting the landfill.

It was noted that at the Polesgo landfill, in 2026, 1 year after its planned closure, an electrical potential of 3,896 MWh was estimated, with a total of 4,547.040 Gg of waste landfilled since its opening in 2005. This electrical potential corresponds to 857 Wh recovered per tonne of landfilled waste, compared with 84,158 Wh/tonne of landfilled waste at the Akouedo landfill. This difference is explained, on the one hand, by the low biogas collection rate considered at Polesgo (25%) versus (66%) at Akouedo, and, on the other hand, by the difference in waste characteristics, notably the organic loads in the waste from the two landfills (55% - 60% at Polesgo versus 80.09% at Akouedo) and the CH₄ production potential (107.56 m³CH₄/kg at Akouedo [43] versus 98.03 m³CH₄/kg at Polesgo [39] ). In addition to the above, the difference in climates between Côte d’Ivoire (more humid) and Burkina Faso (less humid) should be noted. Humidity is a factor favoring the degradation of waste, and therefore higher CH₄ emissions [18] [36] [50] [51] .