A Long Awaited Analysis

I hope that everyone out there is enjoying the beginnings of spring. Along with the ice thawing, the snow melting (yes, even central Texas had snow this year), and the birds singing, it is time now to turn to that even more important harbinger of spring: the February natural gas bill.
Like the swallows returning to San Juan Capistrano, the arrival of this document signals the "beginning of the end" of the heating season, if not the absolute last heating day of the year, and it hopefully foretells a spring that might last several weeks before the air conditioner must reluctantly be fired up. Its arrival also signals the renewal of that great art form both revered and reviled by nearly all sentient beings, data analysis.
It is with these words that we first take a quick look at the seasonal variability of the last 4 winters so that we can get a feel for the raw and weather-adjusted numbers. For our purposes I am defining a winter as the 4 months of November, December, January, and February, since these are the months that, in my area, cause me to have my highest natural gas usage. For reference, my home is heated by burning natural gas and blowing the heated air throughout the living area via the ductwork in the ceiling.
We will measure the seasonal variability of winter in a complementary way with the method we used for summer: instead of "cooling degree days" (CDDs) we will use "heating degree days" (HDDs). The weather data I am using is in HDD's base 65 Fahrenheit, which means that a day with an average temperature of 64 degrees Fahrenheit would add 1 HDD to the total HDDs for the month; 5 days of 62 degrees would add (5*(65-62)) = 15 HDDs to the total for the month, etc.
So, by this measure, more HDDs would mean a cooler day, week, month, or season, depending on what timeframe we are looking at. First, let's look at the data for the winter season for each of the last 4 years:

I have put the most recent winter on the left on the graph with years getting older as you look right. As you can see, this winter was roughly 30% colder than last, and probably at least 20% worse than the average in terms of Heating Degree Days. Although I don't have data on this, it certainly seemed like there were more cloudy and/or rainy days this winter than usual as well. Anecdotally as well, but perhaps worth mentioning is the fact that one neighbor has complained of heating bills roughly 30% higher this year than they are used to.
So how did my attic improvements do? Mindful readers will recall that I made improvements to address all three forms of heat flow into my attic:

1. I used increased convection to let hot air remove itself from the attic, drawing in cooler air behind it. While greatly helpful in summer, this certainly seems like it might hurt my wintertime energy situation by making the attic cooler.
2. I used a radiant barrier to reflect radiation of the sun's rays back out through the roof, before it could heat up my insulation and eventually my living space. Unfortunately, the heat of the sun's radiation is awfully nice to allow in during the winter. Fortunately, there is a potential upside to the barrier in winter: heat radiated from the top of my insulation in the attic will reflect off the barrier back down into another part of the insulation, warming it.
3. I addressed conduction from the living space to the attic by at least doubling, and in places tripling the amount of insulation present, bringing my house up to current building codes at least in the area of insulation. This improvement should have helped in the winter as much as in the summer, if not more.

So how did the attic perform? The simplest measure would be just to compare my yearly energy use (regardless of weather changes) for the last 4 years, so that is what we'll do first. I prefer to compare actual volume of gas burned rather than the energy cost in dollars because the price of natural gas is notoriously volatile, making meaningful comparison over time difficult; plus, of course, Energy Efficiency Man cares most about saving energy! So without further ado, here is the wintertime natural gas energy usage graph:

The natural gas usage is measured in "ccf" where 1 ccf = 100 cubic feet of gas.
As you can see, at first it appears that some combination of cooler weather and/or attic improvements have increased my natural gas usage slightly for this winter. But by how much? From our previous weather graph, we know that this winter was about 30% colder than last winter... but my gas usage is only up a few percent (from the numbers, about 2.5%) from the previous year. In fact, although this was by far the coldest winter of the last 4 years, the natural gas usage is below the average usage of the last 4 years by about 7%. This bodes well, and perhaps calls for another graph!
The best thing to look at is probably the heating effectiveness: that is, how much cold a given unit of natural gas can offset; or more precisely, a graph of coldness per unit energy expended. This would simply be a graph of the ratio of HDDs (coldness) to natural gas usage, with a bar on the graph for each winter season. The higher the bar on the graph, the more HDDs a given unit of natural gas was able to handle, and thus, the better heating effectiveness of the house at that particular year. Here we have it:

From the graph, it would appear that we can draw a number of interesting conclusions about the performance of the house in wintertime.

1. The attic improvements between the winter ending on 2007 and the one ending in 2008 (the rightmost two bars on the graph) caused a slight decrease in the heating effectiveness of the house. In fact, spring of 2007 is when I greatly improved attic ventilation, but didn't do anything else. That must have resulted in the roughly 3% drop in heating effectiveness by letting more cold air into the attic. That is a price, albeit a small one to pay for the roughly 15% drop in the electric usage experienced that year.
   
2. The attic improvements between the next two winters, ending 2008 and 2009, improved the heating effectiveness of the house by about 9%. That was when I installed the first half or so of the radiant barrier, as well as replaced the broken air conditioner (which also could have effected the heating system). This is the only chance we have with my house to compare winter performance without a barrier and with one; unfortunately, it's only about half of the barrier that was installed. I for one am pleasantly surprised to find out that the barrier seems to have actually helped in the wintertime, in addition to helping greatly in the summer. I say "seems to have helped" because the change to the air conditioner/blower unit could also have affected the efficiency of the heating system.
3. The final comparison, between the winter ending in 2009 and the one just about ending now in 2010, shows the true power of attacking all forms of heat transfer at once. Finally in 2010 we've got the full radiant barrier installed, we have good ventilation, and we have lots of new insulation. The result: a dramatic improvement in heating effectiveness. The house now heats itself roughly 30% more effectively than it did last winter, and 38% more effectively than it did before all the improvements began. It would appear that the mysterious winter coziness that I wrote about this winter was indeed something very, very real.
   

The verdict is in: the summertime-focused improvements, which we feared might have negative consequences in the winter season, have actually yielded wintertime savings as well... significant ones. Coming soon: a revisit to my investment percentages / payoff time calculations in light of the long-awaited wintertime analysis.

But until then, adventurers in Energy Efficiency, stay warm and stay efficient!