1 Introduction

The Former Soviet Union (FSU) regionFootnote 1 is a major contributor to the world’s fossil fuel production. The region accounts for over 7% (coal), 15% (oil) and 21% (gas) of the world’s production in 2018 BP (2019). The large contribution of the FSU is matched by its resources which are over 18% (coal), 12% (oil) and 28% (gas) of the world’s total BGR (2016). The fate of the FSU’s fossil fuel future production therefore will have a major influence on the world.

Despite the importance of the FSU region, the literature has limited detailed projections for this region compared to comparable regions such as China and USA. For example, Mohr et al. (2015b) projected both China and USA by province/state for fossil fuels and Höök and Aleklett (2009) examined USA coal production by state. A literature review highlights the limited current fossil fuel production modelling for the FSU region. The literature can be divided into three categories:

The first is to model the world fossil fuel production as a whole and differences of regions are excluded in these analyses. For example, Cavallo (2004) modelled the whole world oil production. Brecha (2008) analysed the whole world fossil fuel production in different scenarios. Kharecha and Hansen (2008) analysed the whole fossil fuels production for the world and their impacts on CO2 and climate. Nel and Cooper (2009) forecast the whole world fossil fuels production and their implications on economic growth and global warming. Maggio and Cacciola (2009) projected the world oil production as a whole by using a variant of the Hubbert curve. Wang et al. (2011) analysed the whole world conventional oil production by using two different multi-cycle curve-fitting models. Maggio and Cacciola (2012) modelled the peak of world oil, gas and coal by using the multi-cycle Hubert method. Similarly Nehring (2009) projected fossil fuels for the world. Ward et al. (2012) presented a high estimate for the whole world fossil fuels production. In these studies, the contribution of FSU is unknown.

The second category includes world fossil fuel production estimates by geographic/political regions. For example, Al-Fattah and Startzman (2000) and Imam et al. (2004) forecast the gas production of Eastern Europe and FSU as a whole in their world natural gas production modelling. Mohr and Evans (2011, 2009) have projected natural gas and coal production at the FSU region level. Mohr et al. (2015b) projected fossil fuel scenarios at the country level for most countries, however the FSU region was mostly projected as a whole. Höök et al. (2010) analysed Russian coal production and total Euroasian coal in their forecast of global coal production. Nashawi et al. (2010) analysed the crude oil production of Russia and Kazakhstan when they forecast world crude oil production. Rutledge (2011) analysed the coal production of Russia when they estimated long-term coal production. Reynolds and Kolodziej (2008) forecast FSU oil production as a whole by using a modified multi-cycle Hubbert model. Wang and Bentley (2020) modelled CIS gas production as a whole when they forecast world natural gas production. In these analyses, FSU is primarily treated as a whole.

The third category is to model the fossil fuel production for specific countries in FSU. Henderson (2019) projected Russian oil production in high detail to 2030, and Kapustin and Grushevenko (2019) projected Russian oil production to 2040. In terms of gas projections, Anon (2020) modelled Russian gas production by region to 2030.

Based on the above analysis, we note that the number of studies for FSU fossil fuels production is limited, despite the importance of the FSU region. Furthermore, several studies on FSU fossil fuels generally treated the region as a whole in their modelling. This appears to be due to the paucity of disaggregated production data during the Soviet Union years. The importance of the region necessitates the need for more detailed and disaggregated projections of this region.

The purpose of this paper is to examine by region the Former Soviet Union fossil fuel production in an attempt to reduce the uncertainty in global fossil fuel projection models and the associated greenhouse gas emissions. This study will continue to use the three URR scenarios of Mohr et al. (2015b) for all other regions of the world. The GeRs-DeMo approach assumes no global action to reduce global greenhouse gas emissions and no significant breakthroughs in alternative (non fossil fuel) energy technologies. The resultant models are therefore not intended as a prediction of future fossil fuel energy use, but instead estimate an informative, geographical and mineralogical picture of the upper limits to business as usual growth in fossil fuel use and its associated greenhouse gas emissions (Mohr et al. 2015b).

Due to the border disputes in what was until recently Eastern Ukraine, the Donetsk, Luhansk and Crimea regions have been modelled individually. This has been done to ensure that data is as granular as possible and to remain as neutral as possible to the politics surrounding these regions. The GeRS-DeMo model has the term ‘country’ and these regions will be modelled as such. This labelling by the authors is for modelling purposes only and is not an indication of support for or against any separatist movements in these regions or for any particular nations claims to these regions.

2 Modelling methodology

The model used to create the projections is the Geologic Resources Supply-Demand Model (GeRS-DeMo). GeRS-DeMo incorporates a supply and demand components with interact, so that if demand is high, supply is increased and vice versa. The model has been used to model a wide variety of resources such as fossil fuels, lithium, copper, lead, zinc, and iron ore (Mohr et al. 2012, 2015a; Northey et al. 2014; Mohr et al. 2018). The model was selected due to its ease of use and capability to model supply and demand interaction and handle supply disruptions (e.g. global conflicts). The model was developed previously (Mohr 2010), and has been briefly described elsewhereFootnote 2. The model has two methods of supplying resources either from mines or from oil/gas fields as indicated in Fig. 1.

Fig. 1
figure 1

Idealised production from fields and mines

Full size image

2.1 Supply–Oil and gas fields

The production for a region is determined from the production of all idealised fields. The production of an individual idealised field has a one year ramp up to a plateau period, followed by an exponential decline in production, as shown in Fig. 1. Two key variables to calculate are the number of fields on-line over time, and the URR of the individual fields. The number of fields on-line n(t) is determined by Eq. 1.

$$\begin{aligned} n(t)= & {} \left\lceil r_Fn_{T}\frac{Q(t)}{Q_T} \right\rceil \end{aligned}$$
(1)

where, \(n_T\) is the total number of fields to be placed on-line, \(r_F\) is a rate constant, \(Q_T\) is the URR of the region, and Q(t) is the cumulative production. The URR of the individual field, is calculated through the exploitable URR. The exploitable URR, is the sum of the URR in fields (or mines) that have already been brought on-line. The exploitable URR \(Q_e(t)\) is estimated via Eq. (2).

$$\begin{aligned} Q_e(t)= & {} Q_T\left( \frac{n(t)}{n_T}\right) ^{r_Q} \end{aligned}$$
(2)

where, \(r_Q\) is a rate constant. The URR of an individual field brought on-line in year t, \(Q_F(t)\) is determined as:

$$\begin{aligned} Q_F(t)= & {} \frac{Q_e(t)-Q_e(t-1)}{N(t)-N(t-1)} \end{aligned}$$
(3)

2.2 Supply–Coal, natural bitumen, extra heavy and kerogen mines

The production from mines is determined from the sum of the individual idealised mines’ production. The idealised mines have a four year ramp up and ramp down period, with a steady production rate in between, as shown in Fig. 1.

The life of an individual mine and its production rate is dependent on the year the mine is brought on-line as described in Eqs. (4) and (5).

$$\begin{aligned} M_P(t)= & {} \frac{M_H+M_L}{2}+\frac{M_H-M_L}{2}\tanh (r_t(t-t_t)) \end{aligned}$$
(4)
$$\begin{aligned} L_M(t)= & {} \left\{ \begin{array}{ll} L_H +(L_L - L_H)\frac{\log _{10}\left( M_P(t)/M_H\right) }{\log _{10}\left( M_L/M_H\right) } &{} \text{; } \text{ if } M_L \ne M_H\\ \frac{(L_L + L_H)}{2} &{} \text{; } \text{ otherwise } \end{array}\right. \end{aligned}$$
(5)

where, \(r_t\) and \(t_t\) are rate and time constants, \(M_L\), \(M_H\) is the minimum and maximum mine production rates, and \(L_L\), \(L_H\) are the minimum and maximum mine lives. The rate and time constants used are the same as those from Mohr (2010). Finally, the number of mines brought on-line in year t is calculated via the estimated exploitable URR \(Q_E(t)\) as:

$$\begin{aligned} Q_E(t)= & {} \frac{Q_T - Q_{T1}e^{-r_T}}{1-e^{-r_T}} - \frac{Q_T-Q_{T1}}{1-e^{-r}}e^{-r_T\frac{Q(t)}{Q_T}} \end{aligned}$$
(6)

where, \(Q_{T1}\) is the URR of the first mine brought on-line in the region and \(r_T\) is a rate constant. The number of mines brought on-line is determined by increasing the number of mines on-line until the actual exploitable URR is larger than the estimated exploitable URR.

2.3 Demand

The demand used is identical to Mohr et al. (2015b). Specifically, the global population p(t) (in billions) is estimated to level off at 11 billion U.N. (2013) based on the following equation:

$$\begin{aligned} p(t)= & {} \frac{11-0.82}{\left[ 1+1.5\exp (-0.023\times 2(t-2014))\right] ^{1/2}}+0.82 \end{aligned}$$
(7)

The per-capita demand, D(t) is calculated as:

$$\begin{aligned} D(t)= & {} \left\{ \begin{array}{ll} 60\exp (0.025(t-1973)) &{} \text{; } \text{ if } t<1973\\ 60 &{} \text{; } \text{ if } t\ge 1973 \end{array} \right. \end{aligned}$$
(8)

3 Data source

Historic production for the FSU needed to be split into the individual countries. Russia’s production was split into regions (oblast’s/krai’s etc) where possible due to Russia’s importance to world fossil fuel supply. Where this was not possible the production was reported at the Federal Districts level. The regions of the Former Soviet Union are shown in Fig. 2. In general the word krai, oblast or republic is dropped with the exception to distinguish between Altai Republic and Altai Krai. The region Tyumen denotes the Tyumen oblast excluding Khanty-Mansi AO and Yamalo-Nenets AO which are modelled separately. In addition the Donetsk, Luhansk and Crimea regions' production was split out into individual regions. Acquiring production data at this granular level proved to be difficult. To the best of the authors’ knowledge a comprehensive, publicly available dataset does not exist covering the full time period and region, which our current paper seeks to address.

Fig. 2
figure 2

Regions of the Former Soviet Union. A – Armenia, B – Azerbaijan, C – Belarus, D – Crimea, E – Donetsk, F – Estonia, G – Georgia, H – Kazakhstan, I – Kyrgyzstan, J – Lativa, K – Lithuania, L – Luhansk, M – Moldova, N – Russia, O – Tajikistan, P – Turkmenistan, Q – Ukraine, R – Uzbekistan, I – East Kazakhstan, II – Karaganda, III – Kostanay, IV – Pavlodar, a – Central, b – Far Eastern, c – North Caucasian, d – Northwestern, e – Siberian, f – Southern, g – Ural, h – Volga, 1 – Yaroslavl, 2 – Amur, 3 – Buryatia, 4 – Chukotka AO, 5 – Jewish AO, 6 – Kamchatka, 7 – Khabarovsk, 8 – Magadan, 9 – Primorsky, 10 – Sakhalin, 11 – Yakutia, 12 – Zabaykalsky, 13 – Chechnya, 14 – Dagestan, 15 – Ingushetia, 16 – Kabardino-Balkaria, 17 – Karachay-Cherkessia, 18 – North Ossetia-Alania, 19 – Stavropol, 20 – Kaliningrad, 21 – Komi, 22 – Murmansk, 23 – Nenets AO, 24 – Novgorod, 25 – Altai Krai, 26 – Altai Rep, 27 – Irkutsk, 28 – Kemerovo, 29 – Khakassia, 30 – Krasnoyarsk, 31 – Novosibirsk, 32 – Omsk, 33 – Tomsk, 34 – Tuva, 35 – Adygea, 36 – Astrakhan, 37 – Kalmykia, 38 – Krasnodar, 39 – Rostov, 40 – Volgograd, 41 – Chelyabinsk, 42 – Khanty-Mansi AO, 43 – Sverdlovsk, 44 – Tyumen, 45 – Yamalo-Nenets AO, 46 – Bashkortostan, 47 – Kirov, 48 – Orenburg, 49 – Penza, 50 – Perm, 51 – Samara, 52 – Saratov, 53 – Tatarstan, 54 – Udmurtia, 55 – Ulyanovsk

Full size image

Recent production data after the end of the Soviet Union is readily available through the various statistical agencies and yearbooks e.g. (Ukrstat 2017; Rosstat 2018) and usual sources such as the BP (2019) and BGS (2017). Declassified documents from the US Central Intelligence Agency contain a wealth of data on Soviet fossil fuel production from both before and during the Cold War. Production data between 1955 and 1980 in particular was challenging to acquire and typically was only reported every 5 years. As a result, production data in between these 5 year intervals had to be estimated. The historical production dataset was constructed by combining the data from the following literatureFootnote 3. The historical production data for the FSU is shown in Fig. 3.

The dominance of the Kuznetsk basin (in Kemerovo Oblast), Khanty-Mansi Autonomous Oblast and Yamalo-Nenets Autonomous Oblast to Russia’s coal, oil and gas production respectively is readily observed. Coal production in regions closer to Moscow have historical peaked and declined, such as Central, Northwestern, Ural and Volga regions. To assist future researchers the collated production dataset is available in the electronic supplement.

Fig. 3
figure 3

Historic fossil fuel production of the FSU

Full size image

4 Fossil fuel URR

The Ultimately Recoverable Resources (URR) are the total amount of the fossil fuels that can be recovered from the resource in the ground before production starts ASPO (2014). Due to the uncertainty surrounding the URR, three URR values have been used, specifically a Low Estimate, a High estimate and a Best Guess (BG) estimate. The URR estimates for the FSU region have been collated from a wide range of sources (see Table 8). The Low estimate was determined primarily through Hubbert Linearisation, and the High estimate was primarily from BGR (2016). The new URR values for the FSU are compared to Mohr et al. (2015b) results in Table 1 and detailed URR values for FSU are shown in Tables 2–4. As shown the High URR is higher than the previous estimate across each fuel source. Similarly the Low URR is slightly lower than the previous estimate. The main difference is in the BG estimate, with the current URR substantially higher in this study, particularly for coal and gas.

The mass to energy conversions are the same as Mohr et al. (2015b). A small number of regions the coal quality is not known for these regions, the energy density assumed is half way between brown and black coal energy densities (19.5 EJ/Gt). The conversion to greenhouse gas emissions, carbon dioxide equivalents (CO2e), assumes the bituminous values for these regions.

Table 1 URR in EJ used in this study; Mohr et al. (2015b) (in brackets) for comparison
Full size table
Table 2 Coal URR values used in this study by country and type
Full size table
Table 3 Oil URR values used in this study by country and type
Full size table
Table 4 Gas URR values used in this study by country and type
Full size table

5 Results and discussion

The results and discussion will examine first detailed projections of Russia’s fossil fuels and Kazakhstan’s coal production. Following this the results for the entire FSU region will be examined. All results shown are the dynamic model where the new FSU model was combined with projections from the rest of the world from Mohr et al. (2015b). The electronic Supplement contains the complete results of the projections.

5.1 Regional results

Fig. 4
figure 4

Russian coal projection

Full size image

The projections of Russian coal are shown in Fig. 4. Coal production for Russia is likely to increase for several more decades with the earliest peak estimated at 2042 in the Low projection. In all projections of Russian coal production we can see the dominance of the Kuznetsk basin (in Kemerovo Oblast) will continue into the future, with the earliest peak estimated 2 decades away in 2042 (Low estimate triggering Russia’s coal peak). The projection in this study is slightly higher than the Russian Government’s estimate for 2035 (This Study 465–734 Mt, Russian Government 429–588 Mt) (Mishustin 2020). More generally the dominance of Siberian and Far Eastern regions is evident. The sharp decline evidenced in the projections is due to the dynamic interactions in the model attempting to keep coal production for the world increasing. Note that this model assumes continuing underlying demand for coal to explore the character of peak estimates arising due to constrained supply. In practice, reduced future demand for coal could alter estimates of peak production to be earlier or later.

Fig. 5
figure 5

Russian oil projection

Full size image

Russian oil production is rather disjointed as indicated in Fig. 5. The collapse of the Soviet Union caused oil production to sharply decline, and while it has managed to approximately reach its pre collapse heights there is cause for concern. An important factor is that the dominant Khanty-Mansi AO oil production has been in declining since 2007. All projections indicate that there will be a short term decline in Russian oil production in the near future as a result. The conventional oil decline is in line with other literature projections, however the projections presented here are on the more optimistic end of the literature (Table 5) (Henderson 2019; Kapustin and Grushevenko 2019). These projected declines are partially offset in the short term by Yamalo-Nenets AO production and in the longer term by unconventional oil sources.

Table 5 Russia conventional oil production comparison to literature (EJ/yr)
Full size table
Fig. 6
figure 6

Russian gas projection

Full size image

Russian gas production is driven almost entirely by Yamalo-Nenets AO production (Fig. 6) and this region has been producing a steady production level for decades. It is difficult to predict what will happen to Russian gas production in the future, but the BG and High scenarios indicate that substantial growth is possible. In contrast, the Low scenario with a substantially smaller URR indicates that Russian gas production would peak in 2022 before sharply declining.

Fig. 7
figure 7

Kazakhstan coal projection

Full size image

Kazakhstan coal production projection is highlighted in Fig. 7. Coal production in Kazakhstan is currently declining due to stagnant production in Karaganda and declining production in Pavlodar. For the Low scenario this declining production is expected to continue. In the BG and High scenarios however production is projected to start increasing again in the near future, and decline after 2100.

5.2 FSU total results

Fig. 8
figure 8

Comparison between this study and Mohr et al. (2015b) for FSU

Full size image
Table 6 Peak year comparison between this study (Mohr et al. 2015b in brackets)
Full size table

The FSU projections are compared to Mohr et al. (2015b) in Fig. 8 and Table 6. FSU coal production in the High scenario is projected to increase faster than Mohr et al. (2015b) and ultimately peak at over 100 EJ/year compared to under 50 EJ/year in Mohr et al. (2015b). The substantial increase in the FSU BG coal URR in this study is evident as the projection shows BG FSU coal production peaking after 2100 instead of choppily continuing to decline. In terms of oil, the current projection is more optimistic than Reynolds and Kolodziej (2008) with a peak year estimate of 2017–2038 at 28.4–40.1 EJ compared to a peak at 26 EJ in 2009. For the fossil fuels overall, compared to Mohr et al. (2015b), there is little difference in the Low scenarios; the High scenario peak year is almost identical (2082–3), however the peak rate is notably higher (222 EJ/yr compared to 162 EJ/yr).

The results shown in Fig. 8 highlight that the specific URR value used has a large impact on the projections. It could be argued that detailed modelling of the FSU region was not necessary, and efforts instead could be restricted to towards more detailed and accurate URR information. Modelling at a granular level does however result in a more nuanced understanding that would otherwise have been missed. For example the rapid increase gas production in the Far Eastern and Siberian regionsFootnote 4. Similarly the depletion of coal closer to Russia’s population such as the Central lignite and the increases in more remote locations such as the Kuznetsk basin.

6 Global implications

The impact of the new FSU projection for the world fossil fuel production is shown in Fig. 9 and the peak year and rates are shown in Table 7.

Fig. 9
figure 9

Comparison between this study and Mohr et al. (2015b) for the world

Full size image

The comparisons for the world between the two FSU models shows little difference to world oil production, with the slight change in the BG scenario of a longer slower decline compared to Mohr et al. (2015b). For gas the new FSU projection causes world production to increase slightly higher and faster in the BG and High cases, with the Low scenario mostly unchanged. World fossil fuel production from the new FSU projection is anticipated to be virtually unchanged in the Low scenario, decline more gradually in the BG scenario and peak at a higher rate in the High scenario. The comparison to selected IPCC projections (Nakicenovic et al. 2001; IPCC 2013; Meinhausen et al. 2011) is shown in Fig. 10. The high scenario now very closely aligns with the A1 Aim, and the BG scenario declines more slowly than the A1Fl or RCP4.5 scenarios. The potential decline in near future could have significant implications on responses to climate change, and accelerate the use of renewable energy.

Table 7 Peak year comparison between this study (Mohr et al. 2015b in brackets)
Full size table
Fig. 10
figure 10

World Emission projections compared to IPCC scenarios (Nakicenovic et al. 2001; IPCC 2013; Meinhausen et al. 2011)

Full size image

7 Conclusions

This paper utilises comprehensive data from the FSU to establish scenarios for future projections of fossil fuel supply from known FSU resources, with comprehensive geographical and mineralogical detail. This additional detail is added to the work of Mohr et al. (2015b) to produce updated global projections of fossil fuel supply from known resources assuming an increasing global demand arising from population growth (with demand per person assumed constant). Comparisons of emissions from the scenarios presented in the paper with IPCC projections representing significant climate change are also given. The most striking finding is the substantial increase in FSU ultimately recoverable resources, particularly for coal but also for gas and oil. At the aggregate global level, the Best Guess and High supply projections increase somewhat whilst Low scenario is broadly similar to the 2015 study. The value of geographically resolved projections for future work, is to more readily be able to visualise both upper bound scenarios – were fossil fuel demand to continue at current per capita rates – as well as the contribution to meeting climate change goals which might be achieved through reducing demand and in turn supply from various regions, or the impact of supply interruptions from various regions. Given that fossil fuel demand has declined in 2020 due to the global impact of the coronavirus, the assumption of constant per capita supply must be qualified. Rather than likely projections of demand, the projections presented in this paper illustrate a time-dependent supply landscape from different countries under low, high and best-guess estimates of ultimately recoverable resources.