The Application of Rate Transient Analysis for the Production Performance Evaluation of the Temane Gas Field–Mozambique: The Use of the Per-Well Basis Approach

Abstract

The Temane gas field, the first producing natural gas field in Mozambique, remains a key supplier to southern Mozambique and the South African market. In recent years, however, the field has experienced an accelerated production decline, raising concerns regarding its long-term supply sustainability. Between 2020 and 2024, gas production decreased by approximately 25%, motivating a comprehensive reserve assessment to quantify the remaining potential and support informed reservoir management. This study applied three modern rate transient analysis (RTA) methods (Blasingame, normalized rate–cumulative, and flowing material balance) to twenty years of daily production data from thirteen producing wells across three reservoirs (G-9A, G-9B, and TEast) on a per-well basis. The RTA methods yielded consistent estimates, indicating an original gas-in-place value of 1576.38 Bscf, a remaining gas-in-place value of 503.37 Bscf, an estimated ultimate recovery of 1405.25 Bscf, and a field-average recovery factor of 76.35%. Reservoir-level recovery factors are estimated at 79% for G-9A, 74.92% for G-9B, and 58.01% for TEast. Despite the high depletion level, the magnitude of the observed production decline is not fully explained by reservoir exhaustion alone, suggesting that the field retains significant remaining recovery potential.

1. Introduction

The Temane gas field is an onshore, wet natural gas field located in northern Inhambane Province, Mozambique. It is the first natural-gas-producing field in the country, with production officially beginning in 2004, followed by the adjacent Pande gas field in 2009. The two fields share a central processing facility (CPF) and, together, constitute the principal source of natural gas for both Mozambique and South Africa.

Approximately 95% of the total gas processed at the CPF is exported to South Africa via the 863 km pipeline operated by the Republic of Mozambique Pipeline Company (ROMPCO) [1]. Within Mozambique, the ROMPCO pipeline has five offtake points at Temane, Chigubu, Magude, Marracuene, and Ressano Garcia for local gas distribution [2]. The gas exported to South Africa accounts for approximately 85% of that country’s total gas demand [3], highlighting the critical role these two fields play in maintaining the energy supply for both countries. However, in recent years, production from these fields has declined sharply, raising concerns about the long-term gas supply. In the fiscal year 2025 (FY2025), gas production declined by 4.5% relative to that in fiscal year 2024, primarily due to reserve depletion [4,5]. As the older and more mature gas field, the Temane field is likely a significant contributor to the observed production decline across the complex (Pande–Temane gas complex).

This study focuses on the Temane gas field and aims to evaluate its depletion status by estimating the original gas-in-place (OGIP) value, remaining gas-in-place (GIP) value, estimated ultimate recovery (EUR), and recovery factor (RF). The analysis employed three rate transient analysis (RTA) methods (Blasingame, normalized rate–cumulative (NRC), and flowing material balance (FMB)) on a per-well basis. This approach provides both field-level reserve estimates and detailed individual well characterization, including the flow regimes, flow capacity, and skin factor, which are often obscured by the conventional field-wide material balance or decline curve analysis.

To the best of the authors’ knowledge, no previous study has integrated three RTA methods to conduct a comprehensive reserve assessment of the Temane gas field using 20 years of daily production data.

Macie [6] applied two RTA methods (Blasingame and NRC) to analyze 10-year production data from the Temane gas field and estimated an ultimate recovery (EUR) of approximately 921.7 Bscf. However, cumulative production has since reached 1073.0 Bscf as of 1 December 2024, exceeding Macie’s EUR estimate by 151.3 Bscf (14.1%). This substantial discrepancy, combined with the field’s extended production history (now spanning 20 years), demonstrates the need for an updated quantitative assessment to support the sustainable management of the remaining reserves and to facilitate the timely identification of alternative sources.

The results of this study, besides providing a quantitative basis for the optimization of the remaining reserves, are expected to support future CO₂ and H₂ storage initiatives in the Temane gas field. The OGIP is a fundamental parameter for assessing the CO₂ and H₂ storage capacities [7].

This paper is organized as follows. Section 2 reviews the theoretical background of RTA techniques, with an emphasis on the Blasingame, NRC, and FMB methods. Section 3 describes the methodology, including the field description, data preparation, identification of the flow regime and spurious data, and the estimation of OGIP and the recovery factor. Section 4 presents the results and discussion, and Section 5 concludes the paper.

2. Rate Transient Analysis Methods: Blasingame, Normalized Rate–Cumulative, and Flowing Material Balance

Rate transient analysis (RTA) is the science of analyzing production data (time–rate–pressure data) to evaluate well and/or reservoir performance, estimate in-place volumes and formation properties, and forecast future production. It also serves as a diagnostic tool used to identify the cause of production decline (reservoir depletion or productivity loss); changes in the wellbore skin, liquid loading, and external pressure support; and the occurrence of well interferences [8,9].

Initially, oil and gas well/reservoir performance analysis was based on time–rate data, and the first systematic and empirical models for production rate analysis were introduced by Arps in 1945 [10]. Based on the work of various researchers [11,12,13], Arps presented a set of exponential, hyperbolic, and harmonic equations to analyze the rate–time data of wells produced at constant bottomhole flowing pressures and under pseudosteady-state conditions. Despite their empirical nature and inability to capture early time-production behavior, Arps’s equations gained widespread acceptance in the oil and gas industry and remain embedded in many commercial RTA software tools, including the Topaze module of KAPPA Engineering Software (Version 5.60.04) used in this work, due to their simplicity and practicality.

It is worth noting that despite their simplicity, application of Arps’s equations requires caution, particularly for gas wells where the decline exponent (the b-exponent in Arps’s equations) varies significantly throughout production, as this can lead to the erroneous estimation of reserves. Blasingame and Rushing [14] and Ilk et al. [15] recommend using Arps’s hyperbolic equations for diagnostic purposes and to supplement more rigorous analysis.

In 1980, Fetkovich [16] presented a set of analytical solutions for the transient flow regime by solving the diffusivity equation under the assumptions of a slightly compressible fluid and production at a constant pressure. A key aspect of these solutions is that irrespective of the reservoir size, they asymptotically converge to an exponential decline trend, consistent with Arps’s exponential model. This convergence provided theoretical support for the empirical basis of Arps’s exponential decline. Fetkovich then combined these analytical solutions with Arps’s equations and developed a complete production-type curve, covering both transient and boundary-dominated flow regimes. The Fetkovich-type curve, despite its limited applicability, specifically for uses in variable-rate and variable-pressure production and for gaseous fluids, served as the primary tool for production data analysis and played a pivotal role in advancing modern decline curve analysis.

Following the work of Fetkovich, several efforts [17,18,19,20,21,22] were made to extend the analysis to oil and gas wells produced under both constant and variable operating conditions. For gas production analysis, a critical challenge was accounting for the pressure dependence of fluid properties, particularly the gas viscosity–compressibility (${\mu }_g{c}_g$) and viscosity–gas deviation factor (${\mu }_{g}Z$) products, which vary significantly with pressure changes.

Fraim and Wattenbarger [19] demonstrated that when normalized pseudotime and pseudopressure functions are applied to linearize the gas diffusivity equation, the resulting rate solutions are analogous to those for slightly compressible liquids, developed by Fetkovich [16]. This indicated that Fetkovich’s decline-type curves could be applied to gas production analysis when pseudotime and pseudopressure functions are used and production occurs at a constant bottomhole flowing pressure.

2.1. Blasingame Method

The Blasingame method is the culmination of a long search for a complete tool to analyze variable rate and variable pressure production scenarios of oil and gas wells. The cornerstone of this method is the application of the material balance time, a concept initially introduced by Blasingame and Lee [18] and later derived rigorously by Palacio and Blasingame [22]. When the material balance time is used as a substitute for conventional time, the constant rate ($1/p_D$) and constant pressure ($q_D$) solutions are analogous, and in boundary-dominated flow regimes, all decline following a harmonic stem on log–log coordinates.

To account for the pressure dependence of gas properties, specifically for ${\mu }_{g}Z$ and ${\mu }_g{c}_t$ products, Palacio and Blasingame [22] modified the Al-Hussainy et al. [23] pseudopressure (for ${\mu }_{g}Z$) and the Fraim and Wattenbarger [19] normalized pseudotime (for ${\mu }_g{c}_t$) functions to the normalized pseudopressure and material-balance-normalized pseudotime, given by Equations (1) and (2), respectively.

$$m_n\left(p\right)={\displaystyle \frac{{\mu }_{g_i}{Z}_i}{p_i}}\underset{p_i}{\overset{p_{wf}}{\int }}{\displaystyle \frac{p}{{\mu }_{g}Z}}dp$$

(1)

$$t_{ca}={\displaystyle \frac{{\mu }_{g_i}{c}_{t_i}}{{q_g}_i}}\underset{0}{\overset{t}{\int }}{\displaystyle \frac{q_g}{{\mu }_g\left(\overline{p}\right)c_t\left(\overline{p}\right)}}dt$$

(2)

where $m_n\left(p\right)$ is the normalized pseudopressure (in psia), p is the pressure (in psia), $\overline{p}$ is the reservoir’s average pressure (in psia), ${\mu }_g$ is the gas viscosity (in cp), ${\mu }_{g_i}$ is the gas viscosity at the initial pressure (in cp), $c_{t_i}$ is the system’s total compressibility at the initial pressure (in 1/psia), $c_t\left(\overline{p}\right)$ is the system’s total compressibility at the reservoir’s average pressure (in 1/psia), $q_g$ is the production gas rate (in Mscf/D), $q_{g_i}$ is the production gas rate at the initial pressure (in Mscf/D), Z is the gas deviation factor (dimensionless), $Z_i$ is the gas deviation factor at the initial pressure (dimensionless), and $t_{ca}$ is the material-balance-normalized pseudotime (in days).

Palacio and Blasingame [22] demonstrated that when material-balance-normalized pseudotime (Equation (2)) and normalized pseudopressure (Equation (1)) functions are used, all the transient solutions converge to a single trend (a harmonic decline trend) in the boundary-dominated flow regime, regardless of flowrate and flowing pressure variations, as shown in Equation (3). This development provided a robust technical basis for production data analysis by enabling the analysis of variable production scenarios and eliminating ambiguity in the decline trend selection for type curves.

$${\displaystyle \frac{q_g}{m_n\left(p_i\right)-m_n\left(p_{wf}\right)}}{b}_{apss}={\displaystyle \frac{1}{1+\left({\displaystyle \frac{m_a}{b_{apss}}}\right)t_{ca}}}$$

(3)

where

$$b_{apss}={\displaystyle \frac{141.2B_{g_i}{\mu }_{g_i}}{kh}}\left[{\displaystyle \frac{1}{2}}ln\left({\displaystyle \frac{4A}{e^{\gamma }{C}_A{r}_w^2}}\right)\right]$$

(4)

$$m_a={\displaystyle \frac{1}{G{c}_t}}$$

(5)

In Equations (3)–(5), $b_{apss}$ is the pseudosteady-state constant (in psia/Mscf/D), $m_n\left(p_i\right)$ is the normalized pseudopressure at the initial pressure (in psia), $m_n\left(p_{wf}\right)$ is the normalized pseudopressure at the bottomhole flowing pressure (in psia), $B_{gi}$ is the gas formation volume factor at the initial pressure (in RB/Mscf), k is the effective permeability (in md), $\gamma$ is Euler’s constant ($\gamma \approx 0.577216$), A is the drainage area in (ft²), $C_A$ is the Dietz shape factor, and G is the OGIP (in MMscf or Bscf).

To perform the analysis using the Blasingame method, three rate functions, the normalized rate ($q_g/\Delta m_n\left(p\right)$), normalized rate integral (${\left(q_g/\Delta m_n\right)}_i$), and normalized rate integral derivative (${\left(q_g/\Delta m_n\right)}_{id}$), are plotted against the material-balance-normalized pseudotime on a log–log scale and overlaid on Blasingame-type curves. This allows the estimation of the OGIP, GIP, permeability (k), drainage area (A), skin factor (s), and either the external radius ($r_e$) or the fracture half-length ($x_f$), depending on the perforation. The latter two rate functions are specifically used to reduce noise in the production data and improve the accuracy of type-curve matching. They are defined by Equations (6) and (7), respectively [22].

$${\left({\displaystyle \frac{q_g}{\Delta m_n}}\right)}_i={\displaystyle \frac{1}{t_{ca}}}\underset{0}{\overset{t_{ca}}{\int }}{\displaystyle \frac{q_g}{\Delta m_n\left(p\right)}}dt$$

(6)

$${\left({\displaystyle \frac{q_g}{\Delta m_n}}\right)}_{id}=-{\displaystyle \frac{d\left[{\left(q_g/\Delta m_n\left(p\right)\right)}_i\right]}{dln\left(t_{ca}\right)}}$$

(7)

2.2. Normalized Rate–Cumulative (NRC) Method

In 1995, Callard and Schenewerk [24] demonstrated that for a gas well producing at a pseudosteady-state flow, there is a linear relationship between the dimensionless rate and the viscosity–compressibility-normalized cumulative production. Subsequently, Agarwal et al. [25] demonstrated that when the dimensionless rate ($q_D$) is plotted against the area-based dimensionless cumulative production ($Q_{DA}$), regardless of different reservoir sizes ($r_e$), in the pseudosteady-state flow, all the plots decline linearly and converge at a single value ($1/2\pi$) of $Q_{DA}$. The relationship between $q_D$ and $Q_{DA}$ is expressed by [25] as follows:

$$q_D={\displaystyle \frac{1}{2\pi }}-Q_{DA}$$

(8)

where the dimensionless rate and dimensionless cumulative production are defined by Equations (9) and (10), respectively

$$q_D={\displaystyle \frac{1422T}{kh}}{\displaystyle \frac{q_g}{\left[m\left(p_i\right)-m\left(p_{wf}\right)\right]}}$$

(9)

$$Q_{DA}={\displaystyle \frac{4.5T{Z}_i}{\varphi hA{p}_i}}{\displaystyle \frac{G\left[m\left(p_i\right)-m\left(\overline{p}\right)\right]}{\left[m\left(p_i\right)-m\left(p_{wf}\right)\right]}}$$

(10)

In Equations (9) and (10), $m\left(p\right)$ is the pseudopressure and is given by [23]

$$m\left(p\right)=2\underset{p_b}{\overset{p}{\int }}{\displaystyle \frac{p}{{\mu }_{g}Z}}dp$$

(11)

Therefore, the analysis based on the normalized rate–cumulative method relies on a Cartesian plot of $q_g/\left[m\left(p_i\right)-m\left(p_{wf}\right)\right]$ versus $\left(G\left[m\left(p_i\right)-m\left(\overline{p}\right)\right]/\left[m\left(p_i\right)-m\left(p_{wf}\right)\right]\right)$. During boundary-dominated flow, the data exhibit a linear trend which x-intercept corresponds to the OGIP.

2.3. Flowing Material Balance (FMB) Method

The flowing material balance (FMB) method was introduced by by Mattar and McNeil [26] as an alternative approach to estimate fluid-in-place (gas and oil) values using production rate and flowing pressure data. Similar to the classical material balance method, FMB also honors the linearity of the p/Z plot; however, it has the advantage of not requiring multiple build-up tests. The average shut-in reservoir pressure is inferred directly from either the flowing pressure or initial reservoir pressure (for known OGIP values), based on the principle that under pseudosteady-state flow conditions, the rate of change of the average reservoir pressure equals the rate of change of any pressure within the reservoir.

The constant rate solution for a well producing at a pseudosteady-state flow and the material balance equation used to infer the average reservoir pressure are given by Equations (12) and (13), respectively [27].

$$m\left(\overline{p}\right)=m\left(p_{wf}\right)+q_{g}bapss$$

(12)

$$m\left(\overline{p}\right)=m\left(p_i\right)-{\displaystyle \frac{1}{G{c}_{ti}}}{G}_p$$

(13)

where $G_p$ is the cumulative gas production (in MMscf or Bscf).

In Equations (12) and (13), neither the pseudosteady-state constant ($b_{apss}$) nor the original gas-in-place value (G) is known a priori. Therefore, when a Cartesian plot of $\left[m\left(p_i\right)-m\left(p_{wf}\right)\right]/q_g$ versus the material balance time ($G_p/q_g$) is constructed, the value of G is iteratively adjusted until convergence is achieved. The converged value of G maximizes the linearity of the plotted data, and the value of $b_{apss}$ is then obtained from the y-intercept of the straight line. Once the value of $b_{apss}$ is obtained, the average reservoir pressure values are calculated using Equation (12) and plotted as $\overline{p}/\overline{Z}$ versus the cumulative production ($G_p$). The resulting plot is parallel to the $p_{wf}/Z$ vs. $G_p$ plot, shifted by the initial pressure term ($p_i/Z_i$).

A fundamental constraint in FMB analysis is that the reciprocal of $b_{apss}$, which represents pressure loss due to the steady-state gas inflow, must remain constant over time [28].

3. Methodology

3.1. Field Description and Gas Production History

The Temane gas field (Figure 1) comprises thirteen production wells distributed across three Lower Grudja (Cretaceous) formation reservoirs: G-9A, G-9B, and TEast. The field formation consists of unconsolidated glauconitic sandstones interbedded within shale layers [29]. The average porosity is approximately 0.31, and the permeability ranges up to 5000 md. The G-9A reservoir is vertically separated from the underlying G-9B reservoir by a succession of silty claystones, which likely act as an effective communication barrier. The TEast reservoir is stratigraphically distinct from both the G-9A and G-9B reservoirs.

The net pay thicknesses are 26.25, 16.4, and 22.97 ft. for G-9A, G-9B, and T-East, respectively. The producing wells are distributed as follows: T-03, T-04, T-05, T-12, T-13, and T-15 in G-9A; T-06, T-07, T-10, T-11, and T-16 in G-9B; and T-09 and T-14 in TEast. The condensate–gas ratio (CGR) ranges from 6.25 to 14.25 STB/MMscf. The G-9A and G-9B reservoirs have the same datum depth of approximately 4185.04 ft, while TEast has a datum depth of about 4199.48 ft. The initial reservoir pressure varies between 1930 and 1940 psia, and the reservoir temperature ranges from 131 to 135 °F. The wellhead temperature is 85 °F, and the gas’s specific gravity (${\gamma }_g$) ranges from 0.597 to 0.794 [30].

The wellbore radius ($r_w$) and tubing internal diameter are 3.0 and 5.0 in., respectively.

The gas rate and cumulative production profile for the Temane gas field are shown in Figure 2.

3.2. Data Preparation

The production data (time–rate–wellhead pressure data) used in this study are from 13 producing wells over a 20-year period (2004–2024). Missing values in flow rate and wellhead pressure data account for less than 5%. To maintain temporal continuity and preserve the integrity of the production history, missing values were handled differently, depending on the duration and operational context of the gap. Short gaps (<3 days) were linearly interpolated because such interruptions typically result from errors in data logging or recording (whether due to human error or equipment malfunction). During such brief intervals, reservoir and surface conditions are not expected to change significantly. In contrast, longer gaps correspond to periods during which various operational activities (workovers or well tests) occur. For longer gaps associated with confirmed shut-in periods, production rates were set at zero. These zero-rate values were later excluded when computing material balance times while preserving the chronological integrity of the available production data.

Wellhead pressure data were converted to bottomhole flowing pressures using the Cullender and Smith correlation (modified for the gas gravity) [31]. The selection of this correlation was based on its relatively lower mean absolute percentage error (MAPE) of approximately 4.1% when compared to those of three other correlations (Gray [32], Duns and Ros [33], and Reinicke et al. [34]) incorporated into the Topaze module. The baseline for the comparison was the measured bottomhole flowing pressure data from well T-03 (the only well with available measured bottomhole pressure data).

Pressure–volume–temperature (PVT) properties were estimated using the following correlations: (1) Dranchuk and Abou-Kassem [35] for the gas deviation factor (Z-factor), gas isothermal compressibility ($c_g$), and gas formation volume factor ($B_g$); (2) Lee et al. [36] for the gas viscosity (${\mu }_g$). The subroutines of these correlations, along with those used for the bottomhole pressure conversion, are integrated into the Topaze module of KAPPA Engineering Software (version 5.60.04, Academic License), which was used in this study.

3.3. Flow Regime and Spurious Data Identification

To identify correlations between gas flow rate and pressure data, detect spurious data points, and determine flow regimes, the methodology proposed by Ilk et al. [37] was applied using four plots: (1) a Cartesian plot of $q_g$ and $p_{wh}$ vs. time, (2) a semi-log plot of $q_g$ and a Cartesian plot of $p_{wh}$ vs. $G_p$, (3) a Cartesian plot of $p_{wh}$ vs. $q_g$, and (4) a log–log plot of $q_g/\Delta p_p$ vs. $G_p/q_g$. Plots (1)–(3) are used to identify the correlation between the flowing pressure and gas flow rate and to identify potential data inconsistencies. Plot (3) also aids in identifying flow regimes, where the transient regime is characterized by a negative slope and the boundary-dominated flow regime by a positive slope [38]. Plot (4) is, likewise, used for flow regime identification; however, unlike plot (3), the boundary-dominated flow regime is characterized by a negative unit-slope. All these plots are collectively referred to as diagnostic plots and are used to verify data quality and assess suitability for RTA.

3.4. Original Gas-in-Place and Recovery Factor Estimations

The OGIP and GIP were estimated using the Blasingame, NRC, and FMB methods. All these methods were conducted using the Topaze module referred to in Section 3.2.

The total field OGIP was obtained by summing the OGIP values of all thirteen wells in the analysis. However, because the analysis was conducted on a per-well basis, summing individual well contributions assumes no well interference. To test this assumption and evaluate potential interference effects, two diagnostic approaches were applied:

• Rate-trend deviation analysis: identifying deviations in the normalized rate function from the characteristic negative-unit-slope declining trend in the Blasingame plot, following the methodology proposed by Anderson and Mattar [8];
• Reserve trend analysis: the evaluation of OGIP as a function of the well completion date, as proposed by Medina Tarrazzi [39].

The recovery factor is calculated by dividing the cumulative gas production by the estimated ultimate recovery ($RF=G_p/EUR$). The EUR is determined from the intersection of the cumulative gas production with the depletion endpoint line in both the NRC and FMB plots.

4. Results and Discussion

The results presented in this section are based on thirteen producing wells. Well T-06 is selected as a representative example to illustrate the analysis and interpretation methodology in detail. The production history of this well is shown in Figure 3. In a manner similar to that for other wells in the Temane gas field, pressure data were not recorded from the start of production. For well T-06, pressure recording began 444 days after production commenced.

Figure 4 shows plots of the produced condensate–gas ratio (CGR), water–gas ratio (WGR), and liquid–gas ratio (LGR) as functions of time. These plots are used to guide the selection of the RTA approach (single phase or multiphase) for this analysis. The CGR is defined as the ratio of the cumulative condensate production to the cumulative gas production, the WGR as the ratio of the cumulative water production to the cumulative gas production, and the LGR as the sum of the cumulative condensate and water production divided by the cumulative gas production.

As shown in Figure 4, all three ratios (CGR, WGR, and LGR) remain essentially constant (5.71, 0.27, and 5.98 STB/MMscf, respectively) over the production interval between 2000 and 5000 days. A slight change is observed after 5000 days of production, characterized by small decreases in CGR and LGR (approximately 0.71 STB/MMscf for CGR and 0.58 STB/MMscf for LGR) and an increase in WGR of approximately 0.13 STB/MMscf. Despite this slight increase, the maximum WGR (approximately 0.4 STB/MMscf) remains well below the 100 STB/MMscf threshold [40], above which a two-phase approach would be necessary. Similarly, the CGR remains below the 20 STB/MMscf threshold [41] throughout the production period.

Therefore, since CGR, WGR, and LGR remain stable and below their respective thresholds throughout the production period, a single-phase gas approach is selected for this analysis. However, it is worth noting that if the observed CGR, WGR, and LGR values indicated condensate build-up or significant water influx, the single-phase approach would require modifications to account for saturation changes and relative permeabilities. Various researchers [42,43,44,45] have adapted the single-phase approach for two-phase (gas–water and gas–condensate) flows; however, these modifications are beyond the scope of this work.

4.1. Data Quality and Diagnosis

Figure 5 presents four diagnostic plots used to evaluate the quality and characteristics of the production data for well T-06. As shown in Figure 5a–c, the pressure and flow rate data exhibit erratic behaviors, characterized by noticeable mismatches. The origins of these mismatches are unclear; they cannot be attributed to liquid loading, as water production is relatively stable (see Figure 3 and Figure 4). These mismatches are resolved in the log–log plot of the productivity index versus the material balance time (Figure 5d), which reveals a clear negative-unit-slope trend (a characteristic of the boundary-dominated flow regime). The plot also includes the transient flow region (corresponding to the material balance time below 10,000 days). Therefore, having all the flow regions present, the data are suitable for further analysis and interpretation.

4.2. Original Gas-in-Place (OGIP) and Remaining Gas-in-Place (GIP) Values

Figure 6 shows the match between the production flow rate ($q_g$), calculated bottomhole flowing pressure (BHFP), and cumulative gas production ($G_p$) with the corresponding theoretical models. To obtain this match, a dataset starting at approximately 14,356 h was selected. All the parameters ($q_g$, BHFP, and $G_p$) show good agreement with the models, except for the last few flow rate data points, which exhibit minor deviations.

Figure 7 presents the specialized plots used in this study. The log–log plot (Figure 7a) reveals a clear boundary-dominated flow regime, characterized by a positive-unit-slope trend and a well-defined radial flow in the derivative curve (log–log derivative (Model)). The plot also indicates that the well produced in the transient regime for no more than approximately five weeks (MBT < 1000 h), which is indicative of the presence of a highly permeable formation. The Blasingame plot (Figure 7b) shows good agreement between the model (Model PI) and the observed rate data. In this plot, a few data points deviate from the negative-unit-slope line (the depletion stem), but without clear indication of well interference or external pressure support. Similar behavior is observed in Figure 7c,d, where nearly all the data points align with their respective linear models (the model-normalized rate and FMB model), indicating that the well produced under volumetric depletion.

The analysis methodology demonstrated for well T-06 was applied to twelve additional wells in the field. All the wells except T-11 exhibited boundary-dominated flow regimes characterized by well-defined positive-unit-slope trends in the diagnostic plot (Figure 7a). Well T-11 operated for approximately eighteen months and produced 1.82 Bscf of cumulative gas. Its production data show a transition from transient to boundary-dominated flow regimes, but the limited data available preclude a detailed analysis. Consequently, the estimated OGIP for this well (approximately 1.82 Bscf) is less reliable. However, due to its relatively low magnitude, its inclusion has no significant impact on the overall field results.

The three RTA methods (Blasingame, NRC, and FMB) yielded consistent estimates of the original gas in place (OGIP) and gas in place (GIP). Although slight differences were observed for wells T-05 and T-09 (Figure 8), these deviations do not indicate the consistent superiority of any single method. As noted by Mattar and Anderson [46], no single method consistently provides the most reliable results.

Figure 9 shows the correlation between the cumulative gas production estimated using the three RTA methods and the cumulative production from the dataset. The two cumulative gas estimates show good agreement, with a mean absolute percentage error (MAPE) of 2.3%. The maximum deviation is approximately $-8.41$ Bscf and occurs in well T-05.

The values of OGIP, GIP, and Gp for each well are summarized in Figure 10. These values represent the averages of the results obtained from the three RTA methods. The majority of the gas production originated from reservoir G-9A wells (T-03, T-04, T-05, T-12, T-13, and T-15).

Figure 11 shows the estimated ultimate recovery for the Temane gas field, including the ultimate recovery factor for each reservoir (G-9A, G-9B, and TEast). The recoverability of all three reservoirs is high, with approximately 90% of the gas estimated to be producible. A recovery factor of 90% is typically considered as the upper limit for volumetric gas reservoirs [47], which is consistent with the production rate trends obtained from the three RTA methods, as shown in Figure 7.

Table 1. Estimated gas volume and recovery factor per reservoir.

Reservoir	OGIP (Bscf)	Gp (Bscf)	GIP (Bscf)	EUR (Bscf) 1	RF (%) 2
G9A	1099.88	763.80	336.08	966.79	79.00
G9B	341.80	243.02	98.78	324.37	74.92
TEast	134.70	66.19	68.50	114.09	58.02
∑	1576.38	1073.01	503.37	1405.25	......

¹ The EUR for each reservoir is calculated using the the corresponding correlations in Figure 11. ² RF is calculated by dividing the cumulative production by the estimated ultimate recovery ($RF=(G_p/EUR)\times 100\%$).

presents the gas volume and recovery factor for each reservoir.

Based on the EUR (1405.25 Bscf) and cumulative production (1073.01 Bscf) reported in Table 1, the current recovery factor of the Temane gas field is approximately 76.36%. Although this depletion level is high, the observed production decline (either the 25% production rate drop that occurred from 2020 to 2024 or the 4.5% FY2025 [4] production drop) cannot be fully attributed to reservoir exhaustion, as the field still contains approximately 331.64 Bscf of remaining recoverable gas. A part of this production drop is attributable to the temporary shut-in (from 2015 to the present) of the production in the TEast reservoir. The two wells in the TEast reservoir were shut in due to integrity problems, and a replacement well was drilled but has not yet been brought to production.

As discussed in Section 3.4, in addition to the rate trend deviation procedure, reserve trend analysis was also employed to diagnose well interference, as shown in Figure 12. The objective of this plot is to assess whether newer wells tend to drain reserves already contacted by earlier wells, which would indicate well interference. Across all three reservoirs, there is no clear evidence of declining OGIP with longer times since the completion date. Well T-11 was produced for a short period (approximately 18 months) before being shut in due to poor performance; therefore, it may not be representative of a declining OGIP in the G-9B reservoir.

The absence of interference signatures in the RTA results and reserve trend analysis across all thirteen wells raises the question of whether the drainage areas are truly non-overlapping. Based on well spacing in the Temane field block (Figure 1), neighboring wells, for example, T-06 and T-16, are separated by approximately 7877.3 ft., with estimated drainage radii of 6648.65 ft. and 7123.12 ft., respectively. Assuming circular drainage areas, the estimated radii relative to the inter-well spacing would suggest overlap in drainage areas. However, circular geometry, like other regular shapes (represented by Dietz shape factors), represents an idealized configuration used in the development of analytical models. In practice, while the magnitude of the drainage area may be realistic, its actual shape is rarely regular. Therefore, apparent overlap based on circular drainage assumptions is not a strong indicator of actual inter-well interference. Inter-well interference typically manifests as a data trend falling below the closed-system model trend (the normalized rate model) in the Blasingame plot. In Figure 7b, the data points that fall below the model trend do not indicate interference, as the overall match remains strong; these deviations are more likely attributable to data noise.

To further verify the consistency of this analysis, we compare the calculated EUR with the operator-reported mid (P50) reserves. The operator’s mid reserves for the Temane gas field are approximately 1334 Bscf. Compared to the EUR estimated in this study (1405.25 Bscf), the difference is approximately 71.25 Bscf (5.34%). This close agreement indicates that both the assumption of substantially independent drainage areas and the per-well-based analysis methodology employed in this work are appropriate for analyzing this field. However, conclusive confirmation of communication barriers within the formations across all three reservoirs would require additional tests, such as interference testing or the use of tracers (e.g., chemical tracers or radioactive proppant tracers). Such tests were not available during the course of this work.

5. Conclusions

In this study, three RTA methods (Blasingame, NRC, and FMB) were applied to quantitatively evaluate the depletion level of the Temane gas field. The analysis was conducted on a per-well basis using twenty years (2004–2024) of production data from thirteen wells across three reservoirs (G-9A, G-9B, and TEast).

The key findings of this study are summarized as follows:

• All three methods yielded comparable estimates for the OGIP, cumulative production (Gp), and remaining gas in place (GIP), with average values of 1576.38 Bscf, 1073.01 Bscf, and 503.37 Bscf, respectively;
• The estimated ultimate recovery (EUR) of the Temane gas field is approximately 1405.25 Bscf, corresponding to an ultimate recovery factor of approximately 89.14%. As of 1 December 2024, approximately 76.36% of the EUR has been produced;
• Reservoirs G-9A and G-9B are the most depleted, followed by the TEast reservoir, with recovery factors of approximately 79%, 74.92%, and 58.01%, respectively. The lower recovery factor in the TEast reservoir is attributed to the temporary shut-in of both production wells due to well integrity issues.
• The estimated EUR of 1405.25 Bscf compares well with the operator’s reported mid reserves of approximately 1334.00 Bscf, showing a difference of only 5.34%. This close agreement supports the assumption of substantially independent drainage areas. However, because this analysis was conducted assuming the non-occurrence of inter-well interference based on rate trends observed in the Blasingame method and reserve trend analysis, additional studies, such as interference tests or tracer studies, would be required for the definitive confirmation of flow barriers between wells across all three reservoirs in the Temane gas field.