Ah, the scientific method. We all remember it from our school days. Observations are made. Observations lead scientists to develop hypotheses intended to explain the observed phenomenon. Those hypotheses are subsequently tested against experimental data. A hypothesis that is found to be in contradiction of the data, no matter how compelling that hypothesis may be, must be discarded and new hypotheses generated that better reflect the data at hand.

You may ask why I am starting this post with a simplified explanation of the scientific method? The answer is because last week we had a classic example of a compelling hypothesis being confronted by a wealth of data from a well-conducted empirical studies and the hypothesis being found to be lacking. In this case we are talking about Cornell Biologist Robert Howarth‘s latest paper: Ideas and perspectives: is shale gas a major driver of recent increase in global atmospheric methane?

As anyone who has read this blog knows, I follow the BC liquid natural gas (LNG) debate pretty closely and Dr. Howarth’s is a name easily recognized in this debate. I can’t count the number of times I have heard the expression “natural gas is a bridge to nowhere“ which I have debunked in the Canadian context. I have repeatedly demonstrated that Dr. Howarth’s research is simply not applicable to BC LNG. My opinion is consistent with the peer-reviewed academic literature on this topic.

This week I was directed to Dr. Howarth’s his latest work. This new article looks at the global increases in methane concentrations and concludes that:

shale-gas production in North America over the past decade may have contributed more than half of all of the increased emissions from fossil fuels globally and approximately one-third of the total increased emissions from all sources globally over the past decade.

Now let’s admit, that is a pretty interesting hypothesis. We know global methane concentrations have been increasing and we know that the US is extracting a lot more natural gas, these two must be related mustn’t they? As Dr. Howarth puts it:

Shale gas makes up 2/3rds of all new natural gas development over last decade. An emissions rate of 3.5% for full lifecycle (which should surprise no one) gives you about 10 Tg per year of new atmospheric methane. My paper is entirely consistent with this, giving me confidence.

In his paper Dr Howarth presents his hypothesis and supports this hypothesis with some very interesting isotope analyses. Now when I say “interesting” I don’t mean “convincing” because it really isn’t all that convincing to those familiar with this type of analysis. A friend (who happens to be a geochemist) I spoke with on the topic pointed out that isotopic fingerprinting of shale gas is simply not as clear cut as Dr. Howarth suggests it should be. That being said this blog post is not about whether Dr. Howarth got the isotope argument right, it is about a more fundamental issue: whether Dr. Howarth’s general hypothesis can stand up when confronted with actual data about methane concentrations in our atmosphere.

The thing Dr. Howarth glosses over in his paper is that methane is a highly studied gas globally. We have satellites and ground stations that can detect methane signatures and our governments monitor methane concentrations in the atmosphere on an ongoing basis. We know a LOT about what is happening on this topic.

In the US, the National Oceanic and Atmospheric Administration maintains a Global Greenhouse Gas Reference Network that includes the Earth System Research Laboratory methane tracker. If releases from North American shale gas extraction were increasing, that should be evident in all the methane sampling and observations. Especially, if as suggested by Dr. Howarth, the North American shale gas extraction were responsible for one-third of the global increase in methane over the last decade.

As a scientist, I decided to go look at the data from the methane tracker and was surprised to see that there is no signal of massive increases in methane emissions that would be necessary to tip the global balance. The spike is simply not there.

For those of you who don’t want to wade through the data, a peer-reviewed article was recently published in Geophysical Research Letters, by Lan et. al., (2019) titled: long‐term measurements show little evidence for large increases in total U.S. methane emissions over the past decade that summarizes the results. The “Plain Language Summary” for this paper says:

In the past decade, natural gas production in the United States has increased by ~46%. Methane emissions associated with oil and natural gas productions have raised concerns since methane is a potent greenhouse gas with the second largest influence on global warming. Recent studies show conflicting results regarding whether methane emissions from oil and gas operations have been increased in the United States. Based on long‐term and well‐calibrated measurements, we find that (i) there is no large increase of total methane emissions in the United States in the past decade; [my emphasis] (ii) there is a modest increase in oil and gas methane emissions, but this increase is much lower than some previous studies suggest; and (iii) the assumption of a time‐constant relationship between methane and ethane emissions has resulted in major overestimation of an oil and gas emissions trend in some previous studies

So we have Dr. Howarth’s hypothesis that says that US fossil fuel emissions of methane are responsible for “one-third of the total increased emissions from all sources globally” while the people responsible for looking for this spike says that fossil fuel emissions have seen “no large increase” in US emissions in the last decade. These are two mutually exclusive statements. Sounds like time for more data.

In his paper Dr. Howarth refers back to an earlier paper A large increase in U.S. methane emissions over the past decade inferred from satellite data and surface observations by Turner et al., (2016) from which Dr. Howarth draws this information:

Since virtually all shale-gas development globally through 2015 occurred in North America (mostly in the United States but also western Canada), we conclude that at least 33 % of the increase in methane fluxes came from North America. This is consistent with the work of Turner et al. (2016), who used satellite data to conclude that 30 % to 60 % of the global increase in methane emissions between 2002 and 2014 came from the United States.

The Lan et al., (2019) paper notes the existence of the Turner et al., (2016) paper and questions some of its conclusions. That being said, the Turner et. al., (2016) paper includes an important qualifier that Dr. Howarth seems to have missed. Turner et al., (2016) write:

The U.S. has seen a 20% increase in oil and gas production [US EIA, 2015] and a ninefold increase in shale gas production from 2002 to 2014 (Figure 1, bottom), but the spatial pattern of the methane increase seen by GOSAT [Greenhouse Gases Observing Satellite] does not clearly point to these sources [my emphasis]. 

The Turner paper literally says that the satellite data they used did not see the methane increases in the vicinity of where the shale gas was being extracted. This is pretty darned important. If the increase in methane, is as large as Dr. Howarth indicates it should be (i.e. enough to significantly effect global numbers) this should be visible in the satellite data…and yet it isn’t.

To summarize: Dr. Howarth’s hypothesis is that shale gas extraction has resulted in a massive increase in North American methane emissions. According to Dr. Howarth’s hypothesis, that massive increase is so intense that it could be responsible for “one-third of the total increased emissions from all sources globally over the past decade”.

But the Global Greenhouse Gas Reference Network sees no large increase in US methane emissions and the GOSAT satellite did not see increases in methane in the areas where “a nine-fold increase in shale gas production” was taking place.

The empirical data does not support Dr. Howarth’s hypothesis. Rather the empirical data makes it clear that shale gas is not the source of the increased North American emissions. From a Canadian perspective this makes a lot of sense. In Canada shale gas is being extracted under tight regulatory regimes that do not allow the practices that were the norm a decade ago. Most natural gas wells are “green completed” which avoids the venting and flaring that were the norm in the past. Newer infrastructure is just that – newer – which means that it meets the current generation’s best practices and typically includes in-line monitoring and leak detection. Finally, our gas is being extracted from extremely deep formations which provide less opportunity for seal failures or for gas to migrate to the surface.

To conclude; in this story we have a compelling hypothesis that has been confronted by a whole lot of very solid empirical data and that empirical data says that the hypothesis is wrong. As we know from our understanding of the scientific method, when confronted by data that contradicts the hypothesis, the hypothesis, no matter how compelling, must be discarded. Dr. Howarth’s suggestion that shale gas is responsible for the increase in global methane emissions simply isn’t supported by the empirical data available at this time.