Since Kyoto 1997 the greenhouse gases subject to emission reduction commitments under the UN Climate Convention include the fluorinated compounds SF6, PFCs and HFCs (hydrofluorocarbons). The latter are by far the most important ones among these so-called "F-gases". The 5.2% (industrial countries) or 21% (German) reduction target refers not to each individual one of the basket of gases (before Kyoto only CO2, methane, and nitrous oxide had been under consideration), but to their total expressed in CO2 equivalent emissions.

For this reason the HFC 134a, which represents three quarters of the worldwide HFC production, got under international pressure just seven years after its introduction into the market.

The Kyoto decision confronted environmental policymakers with a number of problems: What are the applications of fluorinated gases like? What quantities are they used for? How much of the gases in use are released to the atmosphere in the short, middle or long run? What can be said about their present and future consumption and emission levels, what can be done to reduce their emissions?

In Germany our consultancy was asked to address these questions. The result was the 1999 report on behalf of the German Umweltbundesamt (Öko-Recherche 1999), of which numerous findings were adopted in the recently passed Climate Protection Programme 2000 of the German Government.

Car air-conditioning second largest HFC emitter: ca. 3.500 t/a

Refrigerants (only R-134a) from car air-conditioning (A/C) will contribute about a quarter of the overall HFC emissions, namely roughly 4.5 million tonnes CO2 equivalent. Quantitatively, car air-conditioning will be the second largest source of HFC emissions after the sector of refrigeration and stationary air-conditioning. The index of the global warming potential (GWP) of R-134a is 1,300, i.e. the 1,300-fold of CO2. The 4.5 million tonnes CO2 equivalent are calculated from roughly 3,500 metric tonnes emissions which in our estimations are most likely to be the emission levels by 2010/11.

Climate effect: 2/3 from additional fuel consumption, 1/3 from direct refrigerant emissions

The direct emissions of R-134a depend on the emission rate. The problem is that nowhere in the world empirical studies of the long-term emissions from car A/C systems have been carried out. That is not so surprising in the case of R-134a systems, since the first of them started running just ten years ago, in 1990. But concerning R-12 systems, there are no empirical investigations available, either. We get reassured: "The emissions of 134a are by far not so high as in the days of R-12 when three charges of refrigerant during car life time had to be calculated." The engineers point out the more careful handling while servicing (intermediate suction extracting and storing of the refrigerant), the little losses during car life time (more impermeable hoses and improved shaft sealing), the possible suction extraction in end-of-life car scrapping instead of releasing the refrigerant into the atmosphere.

But real measurements actually proving the lowered refrigerant emission rates are not available. Up to now they have not been carried out at all. Thus we also have to rely on experts' estimations for the time beeing.

Diagram 1. Average global warming emisssions per car air-conditioning system in kg CO2 and kg CO2 equivalents, respectively, within 12 years. Two thirds result from additional fuel consumption connected with the A/C system, one third results from direct refrigerant emissions (R-134a).

Most experts assess the emission rate of the average car A/C at "roughly 10%", taking together service losses, running losses (including such in case of frontal car accidents), and disposal losses. To us, this value seems to be rather low under current technological conditions.

Applying a 10% as the lower limit of annual refrigerant loss and assuming a refrigerant charge of 0.84 kg (the latter was the average charge of new A/C units installed in new registered cars in Germany 1997), the 134a emissions total within 12 years to 1 kg.

In terms of climate impact the one kilogramm of refrigerant equals 1.3 tons of CO2. To recall: CO2 from additional fuel consumption per average car is in the range of 2.6 tons.

Old CFC stock emissions not suitable as orientation for forecasting HFC emissions

Alternatives to HFC refrigerants face the following objection again and again: HFCs are a very good solution. Without them the CFC replacement would not have taken place. The CFC 12, which is being substituted by HFCs, contributes not only to damaging the ozone layer, but also to damaging the climate due to its high GWP of 8000 which is six times more than the 1,300 GWP of R-134a.

No doubt that's right. However, firstly CFCs are not subject to the Kyoto climate protocol, since they are already covered by the Montreal Protocol. Secondly, it is no accident that this argumentation (lower global warming contribution) is rather used in the sector of refrigeration and stationary air conditioning, but hardly in the sector of mobile air conditioning. The reason is simple. Regarding their refrigerant consumption level stationary systems could be called stagnant, because neither today nor in the future they will need more HFCs than formerly CFCs/HCFCs. Therefore, forecasting the HFC stock in the year 2010 we could take the former CFC stock as an orientation.

That cannot be done in case of car A/C systems. The former CFC stock is too small for that. In car A/C systems, R-134a stock had already exceeded by early 1996 the historic peak of R-12 in 1993 (2000 metric tons). Even in terms of direct global warming impact, the contribution of R-12 at that time, namely 4.25 million t CO2 equivalent, will be exceeded by R-134a emissions by the year 2010, even though the GWP of R-134a is far lower than that of R-12.

Different to USA or Japan, in Germany the upswing in A/C penetration (A/C ratio) of annual new car registrations set up only after 1990. In 1992, the last full year with CFC refrigerants, the A/C ratio was only 9%. Since then, it hat risen steeply, first in German car brands, then also in the imported brands. For both together, the A/C ratio had reached 24% by 1995. In 1997, at 52%, it exceeded half of all new car registrations for the first time. The rise is continuing. The A/C ratio for 1998 was 68%. All forecasted saturation values of the A/C ratio up to now are regularly surpasseed. In our study is assumed that the A/C ratio of new registrations reaches 90% in German makes in the year 2001, and then remains constant. In foreign makes, the correspondig plateau is at 80% and applies from 2005 onwards. Meanwhile we are not sure about this long term average A/C ratio of 87%. Probably even this assumption will be overtaken by the reality.

In other words, the historic CFC emissions from mobile air conditioning systems are not suitable to serve as a model to assess the future global warming HFC emissions. Forecasting the R-134a emissions from car A/C up to the year 2010 and afterwards requires a model with dynamic parameters.

Methodological approach and assumptions in emission forecasting

What are the most important assumptions made in our study? Forecasting always needs simplifying assumptions which will never completely be proven by reality but just more ore less. Our study proceeded as follows. First of all we asked all seven domestic car manufacturers and the 21 importers selling more than 8,000 car units annually for all their 202 models sold in Germany from 1995 to 1998: How many cars were sold on the German market, how many of them were equipped with A/C systems, and how much refrigerant did these systems contain each. The collected answers provide a rather good survey as well as a solid data base for estimating the real domestic car air conditioning trend.

In foreign makes the A/C ratio of new registrations rose 1995-1998 from 12% to 42%, in German makes the A/C ratio moved on a generally higher level from 30% to 74%. The specific refrigerant charge dropped in the same time frame in foreign makes from 0,77 to 0,73 kg. In German makes the average charge also dropped, again on a higher level, from 0,86 to 0,83 kg.

The data up to now presented were empirically based. They all end in 1998. Beyond that the emission calculation needs assumptions for the future. These are the following ones.

1. Annual new car registrations average 3.6 million till 2020. (That is the mean of the years 1995-1999).
2. Two thirds of new registrations are manufactured in Germany, one third abroad. (That is also the mean of the years 1995-1999). This point is important in so far as the foreign and the German makes differ considerably concering the charges.
3. The A/C ratio of new registrations reaches 90% in German makes in the year 2001, and then remains constant. In foreign makes, the corresponding plateau is at 80% and applies from 2005 onwards. The average A/C ratio of new registrations will then be 87%. The total car fleet reaches this value 12 years later in 2016.
4. The service life of a new A/C unit figures 12 years.
5. The average charge of new A/C units drops as in the past by 1% anually. The overall average charge of all new A/C units drops till 2010 from 0,86 to 0,73 kg. From 2010 onwards no further reduction is expected.
6. The operating emission rate is steady at 10% of the refrigerant stock. More exactly: 10% of the sum of the half stock of the previous year plus the half stock of the current year, according to the equation "10% x (0.5 x stock n-1 + 0.5 x stock n)".
7. The rate of refrigerant recovery losses in end-of-life car scrapping is steady at 30%.

Certainly, the most important assumptions are the 6th and 7th dealing with the emission behavior of the car A/C systems.

Constant emission rates of 10% and 12,5%, respectively

For lack of empirical measurements in our model the assumed operating emissions rate is steady at 10% of the accumulated refrigerant stock. Over a system life of 12 years, 10% loss per year corresponds to 120% refrigerant loss, 100% being one system charge. This 120% figure is made up of two half-refills to top up emission losses, and two times 10% maintenance loss. In our eyes, that is the lower limit. (Further estimates of the emission rate: AFCE 1998; IPCC 2000; UNEP-RTOC 1998; Baker 1999; Preisegger 1999.)

Additionally we have to consider disposal emissions arising 12 years after the registration. Our study assumes a one-off recovery loss upon scrapping of 30% of the total charge. Surely, that is a crucial point. For this evacuation of the refrigerant out of the A/C system before shredding the car body according to the German Statutory Ordinance on end-of-life vehicles does not really take place. We could not find one single dismantling plant at least owning a suction extraction device to empty the refrigerant. Certainly among the about 1000 dismantlers there are some having such a device. So we made the optimistic assumption that from 2005 onwards everything will be much better, at the time when the first annual cohorts containing R-134a will be scrapped and the EU end-of-life vehicle directive will show its effects.

On the other hand we are not so naive as to assume that an old car will enter a dismantling plant with a new-filled A/C system. (That's why we assume 30% refrigerant loss at disposal and not 70% recovery. That's a difference.) Moreover it is well known, that numerous old cars (second-hand cars) evade the legally prescribed way of end-of-life treatment by being exported to foreign countries (mostly to the east), where as yet little attention is given to the refrigerant in scrapping. Nevertheless, 30% disposal loss from 2005 onwards is not exaggerated. This is, by the way, the recovery rate from old vehicles with 134a systems aimed by Japan, an island state, by the year 2010 (Tujibayashi 1999).

To sum up, the total annual emission rate of an average A/C system is 12.5%. The altogether emissions of 150% (120% operating emissions and 30% disposal emissions) divided by 12 years are 12.5.

Business-as-Usual emissions forecast 1995 - 2020

Diagram 2 grafically illustrates the calculated total R-134a emissions from car A/C systems over the 1995-2020 period. Until 2010 these emissions will steeply rise. From 2010 onwards to 2020 and after 2020 they will in principle remain at that high level.

The operating emissions (10% of the refrigerant stock) will peak in 2012, reaching 2,800 t. The refrigerant stock of the total car fleet will then amount to some 28,000 t. From then on, the operating emissions sink till 2020 slightly due to the continually dropping A/C system charges.

Compared with that the disposal emissions (30% of the charge of the annually scrapped A/C systems) still keep rising until 2016. They peak in accordance with the model assumptions 12 years after the maximum annual R-134a filling in new A/C systems (2004). The peak in the year 2016 figures about 700 t emissions. This emission level remains constant until 2020.

The total emissions (12.5% rate) reach their maximum between 2012 and 2016. They peak at 3,530 t in the year 2013. Then the global warming potential of these emissions comes to 4.6 million t CO2 equivalent (cf. Diagram 4).


Diagram 2. R-134a operating and disposal emissions from car A/C systems in 1995-2020 in t (Business-as-Usual Scenario), in Germany. Operating emissions (upper area) reach 2,800 t in 2012 (peak). Disposal emissions (lower area) commence in 2005 and rise to more than 700 t in 2016. The overall emissions peak in 2013 figuring over 3.500 t.

Emission reduction scenario: CO2-systems from 2007 onwards

The operating conditions of mobile air conditioning (MAC) systems (vibrations, open compressors, flexible piping) limit the possibilities to reduce emissions of their refrigerants. Dispensing entirely with air-conditioning may be the most straightforward solution from an environmental perspective, but has been unrealistic since A/C systems became mass-consumed goods in Europe, too (see Holdack-Janssen 1998). The low-GWP alternative refrigerant with a good chance of succeeding is above all carbon dioxide.

This is not the place to go into technical details of that alternative. We just present the emission reduction potential resulting from the possible series introduction of CO2 A/C systems replacing R-134a systems from 2007 onwards. Our model assumes more or less arbitrarily that within two years new cars exclusively will have been equipped with new A/C systems using CO2 instead of R-134a. That means that 2009 is the first complete annual cohort without R-134a systems. Diagram 3 shows the reduction potential from 2007 onwards compared with the R-134a model presented before.



Diagram 3. R-134a operating and disposal emissions from car A/C systems in 1995-2021 according to Reduction Scenario. The series introduction of CO2 systems from 2007 onwards makes the R-134a operating emissions (middle area) sink. They drop from 2009 onwards continuously, reaching zero in 2021. Disposal emissions disappear in the same year either. Since 2009, the overall emissions drop, this providing a reduction potential (white area) compared to the total emissions of BaU-Scenario. This potential grows to almost 600 t by 2010, to 1,700 t by 2015, and to almost 3,500 t by 2021.

If series introduction of CO2 A/C systems replacing R-134a systems started in 2007, from 2021 onwards practically no HFC emissions occurred any more from car A/C systems. Within 12 years annual refrigerant emissions amounting to 3,500 t could cease entirely.

Global warming contribution: 4.5 million t CO2 equivalent or nought

Diagram 4 shows separately for the BaU Scenario and the Reduction Scenario how global warming contribution of R-134a emissions develops over the 1995-2020 period.


Diagram 4. R-134a emissions from car air-conditiong in million t CO2 equivalent - BaU Scenario and Reduction Scenario. The total HFC global warming contributions of the two Scenarios begin to differ in 2007, when car A/C systems are assumed to be converted to carbon dioxide refrigerant. In the BaU Scenario the R-134a emissions remain between 2010 and 2020 relatively constant at ca. 4.5 million t CO2 equivalent. In the Reduction Scenario they drop from 2009 onwards within 12 years to zero.

As a result of the series application of CO2 A/C systems to new cars greenhouse gas emissions amounting to 4.5 million t CO2 equivalent annually could be saved from the year 2021 onwards. This reduction equals a half per cent of the overall German greenhouse gas emissions - a quite considerable quantity considering the fact, that the total of greenhouse gas emissions is made up of a great array of individual sources, each of which is generally small in itself, but all of which need to be scrutinized for abatement options as a part of the crucial effort to reduce greenhouse gas emissions.

Quoted Literature

• AFCE (Alliance Froid Climatisation Environment), Projection à 15 ans des émissions de HFC, Rapport d`etude par L. Palandre, D. Clodic, A.M. Pougin, mai 1998;
  
• Baker, James A. (Delphi Automotive Systems, Inc): Mobile Air Conditioning: HFC-134a Emissions and Emission Reduction Strategies, in: Joint IPCC/TEAP Expert Meeting on Options for the Limitation of Emissions of HFCs and PFCs, Petten, 26-28 May 1999;
  
• Behr GmbH & Co.: Motorkühlung und Klimatechnik: Technische Innovation '97. Darin: Brigitte Taxis-Reischl, Energieverbrauch von Klimaanlagen und Wege zur Verbrauchsreduzierung. Sonderdruck aus ATZ Automobiltechnische Zeitschrift, 99. Jg., Heft 9/97, S. VII-X;
  
• Holdack-Janssen, Hinrich: Entwicklung der Pkw-Klimatisierung, in: Ki Luft- und Kältetechnik 1/1998, 8 ff.;
  
• IPCC 2000. Good Practice Guidance and Uncertainty Management in National Greenhouse Gas Inventories, accepted by the IPCC Plenary at its 16th session held in Montreal, 1-8 May, 2000. Chapter 3.
  (http://www.ipcc-nggip.iges.or.jp/public/gp/pdf/3_Industry.pdf)
  
• Öko-Recherche (Schwarz, W./Leisewitz, A.): Emissions and Reduction Potentials of Hydrofluorocarbons, Perfluorocarbons and Sulphur Hexafluoride in Germany, on behalf of the German Umweltbundesamt, Berlin, October 1999. See: http://www.oekorecherche.de;
  
• Preisegger, Ewald (Solvay Fluor und Derivate GmbH): Automotive Air Conditioning Impact of Refrigerant on Global Warming, in: Joint IPCC/TEAP Expert Meeting on Options for the Limitation of Emissions of HFCs and PFCs, Petten, 26-28 May 1999;
  
• Tujibayashi, Yoshiyuki (Nissan Motor Co., Ltd.), Action plan related to emission control of HFC134a refrigerant for automobile air-conditioning systems, in: Joint IPCC/TEAP Expert Meeting on Options for the Limitation of Emissions of HFCs and PFCs, Petten, 26-28 May 1999;
  
• UNEP-RTOC 1998. UNEP 1998 Report of the Refrigeration, Air Conditioning and Heat Pumps Technical Options Committee, 1998 Assessment, Nairobi 1998.