What are the data showing that Natural Gas is worse than coal in terms of greenhouse gas emissions? [Alan]

The first point to acknowledge, as Ray pointed out, is that when burned, natural gas is much more efficient at generation energy than coal. Because of this, the carbon dioxide emitted per unit of energy produced is much lower for natural gas than coal.  Indeed, coal results in some 216 pounds of carbon dioxide per million British Thermal Units of energy generated.  Meanwhile, natural gas results in 117 pounds per million BTUs, and diesel/gasoline produces a little under 160 pounds of CO₂. So, in terms of the emissions at combustion (the power plant or home furnace), there is no doubt that natural gas is better than coal (and oil products) (https://www.eia.gov/tools/faqs/faq.php?id=73&t=11). But that is far from the whole story…

Recall that natural gas is between 85 and 95% methane (https://www.uniongas.com/about-us/about-natural-gas/Chemical-Composition-of-Natural-Gas) and its Global Warming Potential is much greater than carbon dioxide.  Because of it shorter longevity in the atmosphere than carbon dioxide (before it breaks down into carbon dioxide and water) methane has a carbon dioxide equivalent of 34 times CO₂ on a 100-year basis and 86 times CO₂ on a 20-year basis. That would not be a problem unless the methane leaks. As one can readily see, I suspect, not much methane leakage into the atmosphere is necessary before that leakage negates the combustion benefit emissions of carbon dioxide by natural gas compared to coal as described above. In fact, the leakage rate for natural gas at which the benefit is negated (assuming the 20-year equivalence) is 2.8% (Howarth 2014, http://www.eeb.cornell.edu/howarth/publications/Howarth_2014_ESE_methane_emissions.pdf).   Incidentally, there is no reason why a 100-year basis should be used for this comparison rather than a 20-year basis. Indeed, since the timeline we have for addressing this problem is measured in decades, the 20-year basis seems much more reasonable.

So, the question arises as to what the leakage rate for natural gas really is.  For many years, everyone relied on the EPA for these assessments, and they were reporting leakage rates down in the 1 – 2% range. If that were accurate, we might still be able to make a case for natural gas as an improvement over coal.  However, more recent analyses have generated two concerns: (a) the equipment that EPA was using was systematically underestimating the leakage (Howard et al. 2015 – http://www.tandfonline.com/doi/full/10.1080/10962247.2015.1025925), and (b) the leakage rate is grossly understated (Howarth 2014, http://www.eeb.cornell.edu/howarth/publications/Howarth_2014_ESE_methane_emissions.pdf; Howarth 2015, http://www.eeb.cornell.edu/howarth/publications/f_EECT-61539-perspectives-on-air-emissions-of-methane-and-climatic-warmin_100815_27470.pdf).  These studies, and the reports reviewed therein, suggest that the actual rate if leakage is as follows: 1.7 – 6.0% (mean 3.8%) for conventional natural gas (non-fracked sources), and 3.6 – 7.9% (mean 5.8%) for shale fracked natural gas.  This leakage comprises two elements: downstream transmission leakage at 1.3 – 3.3% (average 2.5%), similar for both conventional and fracked natural gas, with the remainder from upstream (extraction and processing) activities.  Additionally, Schneissing et al. (2014, https://www.deepdyve.com/lp/wiley/remote-sensing-of-fugitive-methane-emissions-from-oil-and-gas-V1JYhQ7H0S) reported the mean leakage for shale-fracked natural gas at 9.5%.

The evidence suggests that natural gas extracted from the ground is, on average, worse than coal regardless of its source.  Since the advantage of natural gas over oil during combustion is nowhere near as great as its advantage over coal, the inference has to be drawn that this leakage (fugitive emissions) natural gas is also never better than oil.