Guest Post by George Kaplan
Brazil is a major oil producing country, but in 2016 was still a net importer, though imports dropped significantly and they have been a slight exporter overall so far this year. It is one of the few countries that have consistently grown production over recent years, and possibly the only non-OPEC country that will show overall growth of conventional crude in the ten years to (say) 2022.
Brazil ANP or anp (Agência Nacional do Petróleo, Gás Natural e Biocombustíveis) publishes Excel files for monthly production on all wells. In theory it should be easy to extract field data from these, in practice not so much. The files are downloaded from a database but not always consistently, sometimes in field units sometimes SI, sometimes one month per file sometimes more, around 2010 onshore and offshore was split but naming conventions weren’t always followed, handling of wildcat wells seems a bit arbitrary, and spelling conventions can change. However after more effort than I expected I did download the data and split it by basin and field.
The total production fits Jodi data well except for three periods: 2005 when the reports stated, and doesn’t make much difference; 2010 when ANP split offshore and onshore reporting and the well files are a complete mess; and 2017, which may indicate that some of the data is revised (this should become evident as more releases are made over the next months).
Land based production is a small component at about 130 kbpd. It averages over 90% water cut (and increasing) and is in long-term decline at over 9% per year (which also appears to be increasing recently). Petrobras is trying to sell off some of their holdings.
The gas and oil rigs combined for onshore and offshore are also shown. The majority is oil rigs for both off and onshore. It is interesting that their numbers started to decline in mid-2012 when oil prices were high and rising. I don’t know the reason – could be to do with the financial and political problems in Petrobras, or associated with geology (i.e. a lack of exploration targets and development projects, which is probably the case for onshore drilling), or a move from many smaller shallow fields to fewer deep and ultra -deep rigs offshore.
The locations of the basins are shown here:
Most of the production to date has come from the Campos basin (Campos actually just means fields in Portuguese). It and Espirito Santo are older basins now in decline at about 9% per year, combined. The big recent additions have been in the Santos basin, particularly Lula (which was Tupi at one time) and Sapinhoá. Note that gas production is not considered here though there are significant gas or gas/condensate fields in these basins, such as the Jupiter field. The charts below show C&C production from individual fields within Campos, Santos and others offshore. There isn’t enough detail to show much of interest, but they are quite colourful.
Most of the offshore production comes from large FPSOs, but there are shallow water platforms in smaller fields. Each of the larger fields in the charts above would have one to three FPSOs. Historically the FPSOs seem to start declining after about a six to twelve-month ramp up period. There are few showing any long plateau periods, though this may be different for Santos. Decline rates can vary from 5% up to 20% per year – i.e. to maintain production in a basin new FPSOs need to be continually bought on line.
Water cut seems to be the biggest impact on decline rates. Campos and Espirito Santo started to decline once water cut hit 50%, and it is still rising in both. Santos fields seem to be low in water at the moment.
ANP provide reserve data following SPE guidelines, however they seem to give only 1P, 3P and 1C numbers in their bulletins – i.e. providing upper and lower limits but not an expected 2P number. They give all numbers to six significant figures, so I guess they must be right. The stacked chart below shows the history for 1P and 3P crude and condensate reserves by state. Almost all the reserves are offshore in Rio de Janeiro and Espirito Santo States – i.e. the Campos and Santos basins. Note I could only find 2016 data by basin, some basins (e.g. Campos) split across two states so I pro-rated numbers bases on 2015 figures – the totals remain as given by ANP.
There was a big hit to reserves in 2015. This may have been partly price related but Forbes reported Petrobras as saying “…decline in reserves was due to other factors, primarily revisions of well estimates at its pre-salt sites offshore.” In the context of the article the word “other” used there makes no sense, so there’s still some ambiguity.
From the 2016 reserves bulletin I estimated remaining recoverable reserves assuming 97% of 1P (which equals P1) and 50% of P2 plus P3 (which equals 3P minus 1P), as below. I didn’t use 1C numbers but include them to show there isn’t a huge upside there (these should be less than 10% probability of production). Numbers shown are mmbbls, converted from Mm³ used in the ANP report.
I fitted Verhulst equations for each of the three categories (Campos, Santos and others) based on the production data from 2005 till now, and ensuring the remaining production (out to infinity) equaled the numbers calculated above. I also added a guess at Libra production assuming 6,000 mmbbls (a bit less than the 8,000 sometimes reported, but then there hasn’t been any production yet and it looks like their other pre-salt fields might not be as good as originally thought, and some of the reserves are already included in the ANP Santos numbers). To match the profile shown they would need to be approving major FPSO and drilling budgets next year, so the profile shown may be unrealistically aggressive as the Libra extended well test project hasn’t even started yet.
(Sorry about the random changes in units.)
An alternative view of the future is to look at bottom up projection based on the projects identified by the E&P companies. Below shows such a scenario. Only the projects up to Bezios 4 have been approved and are under construction (plus possibly Peregrino, although I don’t think there has been a formal FID). After that the capacities (shown as kbpd nameplate after the project name) and start up date are guesses. Most of the projects are in Santos but some (e.g. the Marlim revamp) are in Campos. Some have been cancelled, so I’m assuming these would be revived given high enough prices.
Meeting this schedule for new projects in 2020 to 2023 is now unlikely. For Campos Petrobras use cloned FPSO designs that can be built in around three years, but these may not be suitable for the need for gas reinjection and ultra deep wells in Libra. They currently have higher priorities than development, in particular reducing debt (their interest is $6 billion per year), so other E&Ps would need to get involved. The developments also need latest generation drilling rigs and for Libra would be waiting for initial extended test results from the pilot projects. Petrobras vessels also have a history of major lost time incidents; only one or two such would knock production down considerably. While cloned designs reduce costs they can incur common mode failures; for example as all the gas injection risers on recent FPSOs seem to be seeing (pretty much the highest risk component on any offshore development) and such failures may also produce large downtime across several facilities.
Near term these projections are in-line with IEA (from OMR, June 2017): “In all, Brazilian production is forecast to increase by 190 kb/d in 2017, with gains ramping up to 260 kb/d in 2018.” The EIA STEO also has projections for Brail, but includes ethanol, which is highly seasonal and makes the numbers difficult to follow.
For longer term the new Santos pre-salt fields are the big hope. Lula is in production and Libra is being developed. Libra was discovered in in 2010 and may have up to 8 to 12 Gb recoverable, but it is expensive oil: ultra-deep, difficult wells, 40% carbon dioxide content in sour gas. The high CO2 means the gas has to be reinjected and provides pressure support, but even with this, extensive gas treatment facilities are required on board the production facilities. A consortium of companies won the rights to development, comprising Petrobras (the operator), Shell, Total, CNPC and CNOOC, They are starting with smaller developments to find the problems. To save money, I’ve seen that they are looking at combined water and gas injection, which may not be ideal for the reservoir and introduces hydrate risk.
There have been reports that there may be over 100 Gb other oil in the pre-salt play (unrisked), but there is little current exploration – in addition to the price collapse, possibly because of Petrobras finance and corruption issues, or waiting for results from the initial Libra production tests or for help from outside interests (e.g. Statoil, ExxonMobil, Shell, Total, CNOOC, Repsol etc. are invested or looking to invest there). There have been other fields discovered in Santos, e.g. Iara and Iracema in 2008/2009, but there were also a number of high profile dry wells (i.e. expensive and operated by IOCs) in the basin in the years between the Tupi and Libra discoveries, and not much positive news since Libra, with some leases being given up even in the high price years. In addition the Santos basin geology is somewhat mirrored offshore Angola in the Kwanza province, which has proved pretty barren (and expensive) for the E&Ps. BP recently took a $750 million write-off against dry wells and lease costs there and have given up all leases. Other majors, including Statoil and Total, haven’t done much better. Cobalt (Cameia) and Maersk (Azul) seem to have the only announced discoveries there, but there’s been little news on developing them; Cobalt is in litigation and Maersk has gone very quiet on Angola in general. There are differences with Brazil though, e.g. Santos is deeper and would have a better chance to seals the oil traps.
So overall does this indicate a coming Brazil peak? Probably yes near term, say around 2019/2020, but only more exploration and development will tell whether that could be exceeded again longer term. Statoil plan to be drilling wildcats later this year, the 14th bidding round has four significant offshore blocks up for lease starting in late July, but such things seem to go very slowly in Brazil: another Libra or another Kwanza – who knows, until they drill?