The Massachusetts Global Warming Solutions Act (“GWSA”) requires that Massachusetts must limit its emissions in four important sectors to less than 80% of emissions in the year 1990 beginning in 2050, and its emissions must decrease year after year beginning from when the law went into effect in 2009. The figure below, based upon just released data from the U.S. Energy Information Administration through December 2014, shows why increasing use of natural gas is a problem for compliance with the GWSA, even if it provides temporary relief by weaning Massachusetts off coal and oil:

(Click on graph to see a bigger picture, and use browser Back Button to return to blog.)
The data are shown in dots, and a smoothing spline is used to interpolate and then extrapolate the trends. The spar (or regularizing coefficient) for the smoothing splines used was always 0.7.

Point update, 2016-11-05, 10:39 EDT
[Due to a question from Paul Lauenstein, I am adding that, apparently, the emissions above include methane emissions from leaks, as the cited page states “These leaks were the source of about 29% of total U.S. methane emissions, but only about 2% of total U.S. greenhouse gas emissions in 2013”. The latter is a little disingenuous, since in 2013 natural gas was a relative small part of the energy mix and long term it will be larger. But, nevertheless, the chart apparently includes these leaks in its trending for natural gas. Note that if these leaks were to be fixed, to the extent they can be, then the natural gas trendline will be shallower.]

Massachusetts seems to be doing well on its targets, as long as it can keep reductions going, and there isn’t too much uncertainty (“variability behind the scenes”) in the CO₂ emissions. Unfortunately, natural gas use is increasing, and at some point, emissions from burning natural gas will predominate in the Commonwealth’s emissions. At that point, natural gas infrastructure will need to be retired if the trend is to continue. None of the existing natural gas infrastructure built in the last decade has built into its price early retirements or accelerated depreciation. I made that point in my testimony from yesterday. Some time around 2030, natural gas is going to have to begin to rapidly go away, to be replaced by zero Carbon sources.

We might as well transition rapidly to zero Carbon source for energy now. Even allocating just a significant fraction of the investments proposed for natural gas infrastructure would yield a great deal of efficiency measures, energy storage, and zero Carbon generation, allowing the Northeast United States and Massachusetts in particular to continue as a champion of clean energy.

Update, 2016-11-04

The choice of spar in the above was somewhat arbitrary, so I re-did the calculation using an evidence-based metric, and penalized spline regression via the R package pspline. The message turns out the same:

(Click on graph to see a bigger picture, and use browser Back Button to return to blog.)

The code, data, and the above figure are available.

Update, 2016-11-06

(Click on graph to see a bigger picture, and use browser Back Button to return to blog.)

Confidence and prediction intervals derived from the U.S. EIA data for Massachusetts overall emissions and natural gas emissions. Intervals were obtained by 3-fold validation, sometimes called 3-fold jackknifing. Essentially all combinations of the 35 data points in each set taken 35-3 or 32 at a time were taken, and splines constructed in each case. These then were each plotted. The draws from the natural gas set were independent of those from the MA emissions set.

A surprise is that the data itself contains suggestions that both MA emissions could increase, and natural gas emissions decrease. Also, the prediction interval densities are strikingly multimodal.