Weatherdem's Weblog

Bridging climate science, citizens, and policy


1 Comment

Future Emissions Scenario Requirements & Arctic Warming [With Update]

A recent research article didn’t generate anything terribly earth-shattering, but I wanted to write about some writing on it because it deals with a recurring theme on this blog.  For context, I’ll start with the news release and article (article subs. req’d).  In a nutshell,

Climate model projections show an Arctic-wide end-of-century temperature increase of +13∘ Celsius in late fall and +5∘ Celsius in late spring if the status quo continues and current emissions increase without a mitigation scenario. In contrast, the mean temperature projection would be +7∘ Celsius in late fall and +3∘ Celsius in late spring by the end of the century if a mitigation scenario to reduce emissions is followed, concludes the paper titled, “Future Arctic Climate Changes: Adaptation and Mitigation Timescales.”

Again, there is nothing terribly shocking there.  If we do nothing, the Arctic will likely warm a whole lot more than if we implement mitigation policies.

But that paragraph could use some additional context.  What do the greenhouse gas emissions scenarios look like to generate those varying warming projections?  To get a little technical (stay with me), the authors compared two out of four of the Intergovernmental Panel on Climate Change’s (IPCC) Representative Concentration Pathways (RCPs): RCP8.5 and RCP4.5.  These pathways represent an additional 8.5 W/m^2 and +4.5 W/m^2 radiative forcing at the year 2100 relative to pre-industrial values.

But even though I’ve taken a graduate level radiation course and I’m using these same pathways in my own research, I don’t really know what +8.5 W/m^2 radiative forcing is, and neither do most people.  It’s a number with units that is not intuitively obvious.  This is where climate scientists underperform in communicating with the public and where I come in.

So instead of losing ourselves in the technical details, how can we understand what these two pathways represent?  Qualitatively, RCP8.5 represents a scenario in which we do not enact GHG mitigation policies until after the year 2100.  Economic growth and GHG emissions continue to grow throughout the rest of this century due to 4x 2000′s global energy use.  The radiative forcing is induced by 1370 ppm CO2-eq (CO2 and other GHGs).

By comparison, RCP4.5 represents a scenario that stabilizes forcing at 4.5 W/m^2 without overshooting it and has 650 ppm CO2-eq by 2100 (583 ppm CO2; 2013 mean CO2 concentration: 397 ppm).  Global energy use is just over 2x 2000 levels.  RCP4.5 achieves relatively lower CO2 concentrations by steadily decreasing the amount of carbon per energy unit supplied from 2000 to 2050, then decreasing the carbon/energy ratio very rapidly between 2050 and 2075, then leveling off from 2075-2100.  It does this via wider renewable energy deployment, but predominantly fossil fuel use with carbon capture and sequestration deployment.

In other words, RCP4.5 chiefly relies on slower CO2 concentration growth by assuming widespread and rapid deployment of technologies that do not exist today.  This point is very important to understand.

In a write-up on this same research, Joe Romm concludes thusly (emphasis mine):

This study essentially writes off the possibility of humanity doing any better:

The RCP2.6 scenario requires a 70% reduction of emissions relative to present levels by 2050, a scenario that is highly unlikely in view of the current trajectory of emissions and the absence of progress toward mitigation measures. We refer to the RCP8.5 and RCP4.5 future scenarios as business-as-usual and mitigation.

But the fact is that RCP2.6 — which is about 421 ppm CO2 — is entirely feasible from both a technical and economic perspective. It is only the irrationality, myopia, and, it would seem, self-destructiveness of Homo sapiens that make it “highly unlikely.”

No, it’s not.  RCP2.6 makes many more assumptions about technological capabilities and deployment than does RCP4.5.  It does this more quickly than RCP4.5 by modeling declining carbon per energy unit between 2010 and 2025 (which hasn’t happened yet), then declining much more rapidly starting in 2025 (only 10 years away) until 2050, then slowing down in 2050 and again in 2075.  But here is the kicker: it assumes negative carbon per energy unit after 2075!  How does it do this?  By assuming more carbon will be removed from the atmosphere than emitted into it starting in 2075 and continuing thereafter.  Do we have carbon capture and sequestration (CCS) technologies ready for rapid global deployment?  No, there is to my knowledge only a couple of utility-scale projects currently operating and they haven’t achieved the level of capture and sequestration this pathway assumes.

In order for CCS to operate at the level RCP2.6 assumes, global investment in the technology would have to increase by many factors for years.  Is there any discussion of this occurring in any government?  Will we price carbon-based fuels without interference (i.e., an end to market manipulation by fossil fuel entities and governments)?  No and these things aren’t likely to begin any time soon.

Simply put, RCP2.6 is a fantasy scenario [see update below].  Absent global economic collapse that dwarfs the Great Depression, CO2 emissions and concentrations will continue to increase as economies continue to rely on relatively cheap dirty fossil fuels with manipulated prices.  At this point, I think RCP4.5 is to a lesser extent another fantasy scenario.  That’s neither irrational nor myopic, but realistic based on historical climate policy and my own reading of where international climate policy is likely to exist in the next 35 years.  We are currently on the RCP8.5 pathway.  Researchers use RCP4.5 because it is illustratively different from RCP8.5.  They think it is technically feasible simply because they understand the likely science ramifications of RCP8.5 and misunderstand the public’s desire for continued increasing quality of life that comes with fossil fuel use.  Case in point: researchers have shown the difference between “worst-case” and “best-case” climate scenarios for 30+ years.  Nobody enacted robust climate policy in response to these comparisons.  To continue to do so moving forward is a waste of resources.

[Update]

I wanted to share some updated data demonstrating my statement that RCP2.6 and RCP4.5 are “fantasy scenarios”.  Here are two plots I used in a related post in last 2012:

 photo CO2EmissionsScenarios-hist-and-RCP-2012-b.png

Figure 1. Historical (black dots) and projected (out to 2050 only) CO2 emissions from a Nature Climate Change article (subs. req’d).  Bold colored lines (red, yellow, gray, and blue) represent IPCC AR5 RCP-related emission scenarios.   Thick green dashed lines and thin green solid lines represent SRES emission scenarios used in IPCC AR4.  Light blue dashed lines represent IS92 scenarios.  Different generation scenarios are presented together for inter-report comparison purposes.

 photo CO2EmissionsScenarios-hist-and-RCP-2012.png

Figure 2. As in Figure 1 except projections shown to year 2100 and RCP scenarios highlighted.

Figures 1 and 2 show historical and projected annual CO2 emissions in Pg/year from 1980 until 2050 and 2100, respectively.  Historical data end in 2011 because the paper was published in 2012.  So there are two more year’s data available to us now.  How do you think global CO2 emissions changed since 2011?  Did they decrease, stay the same, or increase?

It’s more challenging than it should be to find similar graphics, but I found this update:

 photo CO2_emissions_Global_Carbon_Project_2013_zps7214b665.jpg

Figure 3. Historical (1990-2012; 2013 projection) global CO2 emissions in GtC/year (1 PgC = 1 GtC).

As Figure 3 shows, global CO2 emissions rose in 2012 compared to 2011, and emissions likely rose further in 2013 compared to 2012.  It further shows that emission rates increased only by 1.0%/year in the 1990s and accelerated to 2.7%/year in the 2000s.  While recent year-0ver-year increases aren’t at 2000 mean levels, they are at least twice that of 1990 levels.  In other words, there has been no stabilization of CO2 emissions, let alone a decrease, as RCP2.6 and RCP4.5 assume.

A fair counterpoint can be made that RCP2.6 assumes a decline starting in 2020, while RCP4.5′s decline starts in 2040.  Sure enough, Figure 1 and 2 demonstrate those assumptions.  To that, I say Figure 1 and 2 also shows RCP2.6′s maximum annual emissions peak at 2010 levels.  Emissions have already increased at 2%+/year since then historically.  For argument sake, let’s say emissions will peak in 2020.  Historical emissions will then be higher than RCP2.6 assumed, which would require even more CO2 removal to achieve <2C stabilization by 2100.  More CO2 removal means more efficient and widespread deployment than RCP2.6 already assumes, which makes it less likely to occur.

RCP4.5 assumes peak annual emissions in 2040 of approximately 11 PgC/year.  If annual growth rates continue near 2.1%, we’ll actually reach that level in 2018 – 22 years ahead of RCP4.5′s assumption.  What emissions growth rate is required to hit 11 PgC/year in 2040?  See the chart below:

 photo CO2Emissions-21and0475_growth_rates_zps20b1f74a.png

Figure 4. Historical (1959-2012) and projected (2013-2040) global annual CO2 emissions using mean 2000′s emissions growth (blue) and calculated emissions required to achieve 11 GtC/year in 2040 (red).  [Historical data: 2013 Global Carbon Project.]

Note that the RCP4.5 scenario has declining emissions growth rate between 2030 and 2040 while my computations uses constant growth rate assumption.  Still, this calculation sheds some light on required changes to achieve RCP4.5 scenario assumptions.  Figure 4 shows that if future emissions grow at constant rate of 2.1%/year (less than the mean 2000′s rate; more than the mean 1990′s rate), 2040 emissions will total >17 GtC/year (remember RCP4.5′s maximum of 11 GtC/year could be achieved as early as 2018).  To max out at 11 GtC/year, emissions would either have to grow at no more than 0.475%/year – less than half the 1990′s mean value – or grow more quickly in the near future, stabilize quickly, and decrease every year following 2030.

RCP2.6 and RCP4.5 demand that countries begin to change their entire energy production fleet from fossil fuels to renewables – either immediately (RCP2.6) or within the next 10-15 years (RCP4.5).  What costs are associated with this conversion?  How many people without energy access today are denied energy access in the future?  That is something that Romm doesn’t address in his talking point that “the fact is that RCP2.6 — which is about 421 ppm CO2 — is entirely feasible from both a technical and economic perspective.”  421 ppm CO2 means no higher concentration than what will occur by 2025.

A permanent emissions decline has obviously never happened historically.  What basis allows for the assumption that it will occur starting in 2030?  More sweeping and effective policies than have ever been implemented are required.  The point to this exercise is to demonstrate that we can play games with numbers all day, but the real world is quite different from economic and climate models as well as Excel spreadsheets.  Unless and until we see real world evidence that emissions stabilization occurs, I see little reason to discuss what RCP2.6 or RCP4.5 shows beyond what “could be” as a rhetorical exercise.


Leave a comment

Climate & Energy Links – Sep. 12, 2013

Here are some stories I found interesting this week:

California’s GHG emissions are already lower than the 2015 threshold established as part of California’s cap-and-trade policy.  The reasons emissions fell more than expected include the slow economy and relative widespread renewable energy deployment.  The problem with this is the lack of innovation.  We have seen what companies do with no incentive to innovate their operations: nothing that gets in the way of profit, which is the way companies should operate.  That’s why we need regulations – to incentivize companies to act in the public interest.  Should CA adjust future cap thresholds in light of this news?

No surprise here: Alter Net had a story detailing the US Department of Energy’s International Energy Outlook and the picture isn’t pretty (and I’m not talking about the stock photo they attached to the story – that’s not helpful).  Experts expect fossil fuels to dominate the world’s energy portfolio through 2040 – which I wrote about last month.  This projection will stand until people push their governments to change.

Scientific American’s latest microgrid article got to the point: “self-sufficient microgrids undermine utilities’ traditional economic model” and “utility rates for backup power [need to be] fair and equitable to microgrid customers.”  To the first point, current utility models will have to change in 21st century America.  Too much depends on reliable and safe energy systems.  The profit part of the equation will take a back seat.  Whatever form utilities take in the future, customers will demand equitable pricing schemes.  That said, there is currently widespread unfair pricing in today’s energy paradigm.  For example, utilities continue to build coal power plants that customers don’t want.  Customers go so far as to voluntarily pay extra for non-coal energy sources.  In the end, I support microgrids and distributed generation for many reasons.

A Science article (subs. req’d) shared results of an investigation into increasing amplitude of CO2 oscillations in the Northern Hemisphere in the past 50 years.  This increase is greater for higher latitudes than middle latitudes.  The increase’s reason could be longer annual times of decomposition due to a warming climate (which is occurring faster at higher latitudes).  Additional microbial decomposition generates additional CO2 and aids new plant growth at increasing latitudes (which scientists have observed).  New plant growth compounds the uptake and release of CO2 from microbes.  The biosphere is changing in ways that were not predicted, as I’ve written before.  These changes will interact and generate other changes that will impact human and ecosystems through the 21st century and beyond.

And the EPA has adjusted new power plant emissions rules: “The average U.S. natural gas plant emits 800 to 850 pounds of carbon dioxide per megawatt, and coal plants emit an average of 1,768 pounds. According to those familiar with the new EPA proposal, the agency will keep the carbon limit for large natural gas plants at 1,000 pounds but relax it slightly for smaller gas plants. The standard for coal plants will be as high as 1,300 or 1,400 pounds per megawatt-hour, the individuals said Wednesday, but that still means the utilities will have to capture some of the carbon dioxide they emit.”  This is but one climate policy that we need to revisit in the future.  This policy is good, but does not go far enough.  One way or another, we face increasing costs; some we can afford and others we can’t.  We can proactively increase regulations on fossil fuels which will result in an equitable cost comparison between energy sources.  Or we can continue to prevent an energy free market from working by keeping fossil fuel costs artificially lower than they really are and end up paying reactive climate costs, which will be orders of magnitude higher than energy costs.


Leave a comment

Car mileage evaluation criticism

The Denver Post Editorial Board took a stance on car mileage based on a recent Consumer Reports (CR) article.  Let me state at the outset that I regularly use Consumer Reports rankings as part of my purchase decision-making process.  That said, no testing is ever 100% complete and is much less regularly communicated well to non-experts.  At issue: CR performed independent tests on cars and calculated different miles per gallon values than those the EPA provided.  Should the EPA update their testing?  Perhaps, but the Editorial Board and CR didn’t provide an overwhelming case to do so.  Let’s look at what each entity said.

First, the Post:

Consumers could very well feel deceived by the numbers, but there are other issues at work.

Hybrids with just a single occupant can zip past traffic using high-occupancy-vehicle lanes in some parts of the country — including Colorado — because of their superior efficiency. The idea is to support, through public policy, efficient vehicles that generate less harmful emissions. But if they’re really not substantially more efficient, it’s neither environmentally beneficial nor fair to drivers of traditional vehicles that may, in reality, get similar gas mileage.

There are key parts to this section that I want to highlight.  What does “substantially more efficient” mean?  What value efficiency is enough to warrant public policy?  In the Denver area, suburban drivers love their SUVs and trucks.  So to start, I’ll compare an SUV, a truck, a sports car, and a 2010 Prius.  And I’ll start with the location the Board identified as a policy recipient: the highway.  The EPA’s ratings for these four vehicles are: 15, 18, 26, and 48mpg (highway).  Should Prius drivers receive plicy support for driving a vehicle that averages 2-3X as many mpg as the majority of other vehicles?

Let’s change the argument a little to better match the spirit of the article and consider the “combined” fuel efficiency.  This designation accounts for street driving and highway driving.  The same four vehicles have the following combined ratings: 13, 14, 21, and 50mpg.  The complaint that CR and the Board has is the hybrid value is too high (based on their own testing protocols, which aren’t detailed very well).  So let’s change the hybrid’s EPA combined value with CR’s “overall” value: 44mpg.  Should we direct public policy toward cars that get 2-3X as many mpg as the majority of other vehicles?  Note that the comparison, and thus my argument, didn’t change.

The argument that CR made is that their hybrid vehicle test values differed from EPA values by more than 3mpg on average.  Yes, 6mpg is a difference.  Are consumers being tricked?  I don’t think so.  And this is really where I split with CR.  Here is their take:

Overall, fuel-efficiency shortfalls have narrowed considerably over the years. When Consumer Reports conducted a similar study in 2005 that compared our gas-mileage results with the EPA estimates, we found that most cars got significantly fewer mpg than their window stickers promised. Conventional gas-powered vehicles missed their EPA estimates by an average of 9 percent, and hybrids by 18 percent.

So it isn’t as if CR is bashing hybrids; far from it.  All of the EPA sticker values differed from CR’s.  That isn’t surprising since CR tested vehicles differently.  One vital question: which test best mimics real-world driving?  Most people drive very aggressively (hard acceleration and braking), which has a big impact on gas mileage.  Are the CR values too high?  What if we consider additional factors: heat and cold, wind and rain.  Those happen in the real world.  But not in the CR tests.  Isn’t it interesting that a consumer advocacy group is challenging the EPA’s tests for being unrealistic when they themselves didn’t test vehicles in real-world conditions?  How far off are CR’s values?  We don’t know.  Should we reengineer public policy in the face of CR’s unrealistic tests?  For what purpose?  Should the EPA and CR develop a more rigorous testing protocol – perhaps one that both entities can perform and therefore directly compare against one another?

I have another problem with the CR quote.  Note the bolded words.  EPA doesn’t promise drivers that they’ll attain the sticker mileage.  In fact, the EPA goes out of its way to emphasize and explain their values are merely estimates.  That’s not a promise, not even close.  It’s annoying that people read too much into things.  In fact, CR could do what I did: check the EPA website for common misconceptions.  The ninth most common:

9. Fuel Economy Label The EPA fuel economy estimates are a government guarantee on what fuel economy each vehicle will deliver.
The primary purpose of EPA fuel economy estimates is to provide consumers with a uniform, unbiased way of comparing the relative efficiency of vehicles. Even though the EPA’s test procedures are designed to reflect real-world driving conditions, no single test can accurately model all driving styles and environments. Differing fuel blends will also affect fuel economy. The use of gasoline with 10% ethanol can decrease fuel economy by about 3% due to its lower energy density.

Like I said, the EPA goes out of its way to explain what its published values are.  CR and the Board should have done 60 seconds of checking before they called out the EPA for deceiving consumers with “promises”.

A more appropriate target are auto manufacturers, who know what the tests are and try their best to optimize the results.  The same entity that has a financial interest in optimizing estimated mileage should not test vehicles’ mileage.  Like I wrote above, this is where CR and EPA should work together to independently test vehicles.

But as far as the basic argument goes, hybrids do get better overall mileage than other vehicles.  They get the best mileage if drivers drive them where manufacturers intended them to drive: in stop-and-go city traffic.  But they still drastically outperform their competition in highway driving, enough so that I think current public policies provide nearly the correct incentive for drivers to think of another dimension when choosing which vehicle they will purchase.


Leave a comment

Choices: Fuels, Efficiency, Transit vs. Drilling

I keep writing that we as a society and a species have choices that we’re continually making today that will affect the climate of tomorrow.  Most choices involve spending some small outlay of money today in order to not have to spend much larger sums just to adapt tomorrow.  The choice I’ll write about today deals with one of the Republican Teabaggers’ favorites: gradually use less fossil fuels in our transportation sector or “Drill, bagger, drill!”  As usual, the Teabaggers are on the wrong side of the issue, as this chart from the NRDC, using data from the Energy Information Administration shows:

The black line on top would be the pathetically measly result of opening up new drilling areas to the dirty energy corporations: less than 1 million more barrels of oil per day by 2025.  Real energy independent, eh?  Aside from the fact that oil corporations will sell that oil to whomever will buy it most expensively (i.e., not in the U.S.), three of the other measures would prevent the use of the same amount of oil by themselves.   Combined with other measures, the total number of barrels of oil that wouldn’t have to be bought and used is 5x the amount made available by opening up new drilling areas.

The results of using 5 million fewer barrels of oil per day by 2025 can’t be understated: less environmental damage in all aspects of the drilling process; real steps toward energy independence; freedom to keep more money in Americans’ pockets (isn’t that what Teabaggers are supposed to be all about, anyway?); more efficient transportation system.  And on and on it goes.

Going the drilling route couldn’t be more stupid.  This choice, as is the case for others, is pretty simple.

[h/t MB @ dKos)


Leave a comment

Scientific American – Polar Meltdown & Scrubbing Carbon Articles

The June 2010 issue of Scientific American had two climate-related pieces in it that I thought were worth discussing.

Polar Meltdown

The first dealt with 12 potential events, their likelihood of occurring by 2050 and some of their effects.  The front cover lists ‘Polar Meltdown’ last among the 12.  The article has ‘Polar Meltdown” as the 8th event, despite its likelihood; I don’t really understand how they decided to organize the events.  I mention these details first because more of the general public reads Scientific American than journals or even climate-related blogs.  Given the nature of the effects – vastly more negative than positive – I would personally prefer to see this listed first both on the front cover and in the article.  Interestingly, their online version has this event first, which is good news.

Another quibble: the picture accompanying the little piece is of Antarctica instead of the Arctic.  In terms of “Polar Meltdown”, Antarctica definitely presents the larger long-term threat of the two.  But the article is supposedly about events by 2050.  I’m a climate realist – which means I recognize that Antarctica won’t completely melt by 2050.  That will likely take a century or two.  And it’s not really even the Arctic ice melting that should scare us all – it’s Greenland.  Greenland’s ice has been land-bound for thousands of years.  When that ice melts, it will raise sea levels.  Similarly for Antarctica – it’s not the sea ice melting every year that presents a threat, its 8 times as much ice as exists on Greenland melting that will raise sea levels even higher.  There are plenty of graphics demonstrating the ridiculously fast meltdown that has already occurred in the Arctic.  They should have shown some of those instead of an Antarctica-shaped piece of ice in a punchbowl.  That’s a lousy visual on multiple levels – the most important being water-borne ice melting doesn’t change the level of the water.  Sheesh.

Okay – onto the science part of this potential event.  Scientific America’s author thinks this event ranks as “Likely”, or better than “50-50″ but not “certain”.  I’m glad to see the assessment at better than 50-50, because it is.  Given the speed at which Arctic and Greenland ice has already melted, I unfortunately think that Arctic sea ice in particular is within a handful of years of disappearing every summer, as this post details.  Arctic sea ice volume is plummeting towards an effective zero point much, much faster than any expert thought possible just a couple short years ago.  If you go to their online version and click on the ‘polar meltdown’ icon, it takes you to a page describing some effects of the event occurring – none of which appear in the print version, by the way.  There is also a place to vote on how likely you think this event is of occurring.  As of this writing, the highest percentage of respondents (36%) agree with the author.  29% of respondents think it is almost certain.  14%, 11% and 8% think it is ’50-50′, ‘unlikely’ and ‘very unlikely’.  There’s freepers everywhere, I suppose.

Continue reading


Leave a comment

The American Power Act – First Reactions

The Senate’s version of climate and energy legislation was formally introduced yesterday.  Titled “The American Power Act”, the draft is 987 pages long and includes darn near everything.  Reading any substantial amount of the bill is going to take a while; understanding it will take even longer.  Of course, by the time activists read and understand it, it will probably be in the process of being modified.  Regardless, here are two links that I’m looking at.  The first is the full bill; the second is a section by section summary.

S1733- The American Power Act (pdf)

21 page Section by Section summary (pdf)

Continue reading


Leave a comment

Surprise! Electric Cars Have Plenty of Range for Daily Driving

I’ve always found people’s reticence about buying electric vehicles because they wouldn’t be able to drive far enough to be based on uninformed opinions.  With a range of 100 miles, the vast majority of Americans would be able to drive to work and back home, with short jaunts for lunch and errands in between, on a daily basis without having to rely on public charging stations.  With a recharge time of 4 to 8 hours, a majority of Americans would simply be able to plug their cars in at home every night; no other shift in driving behavior would be needed.  Drivers would spend less time on a weekly basis plugging their cars in at home, and in the worst-case scenario in public, than they do at gas stations today.  Really, the only obstacle is likely to be psychological.  Most people don’t like to change their habits.  The fact that the cost to charge an electric vehicle is less than the cost to put fossil fuels in their tank is also largely lost on the public so far.

A new study supports my gut feeling:

Studies of drivers who already have electric cars are finding that they prefer the convenience of charging at home, and despite their vehicles’ limited range, most are able to avoid public charging.  The relative lack of these recharging locations could prove less of a deterrent to electric car acceptance than was expected.

Much like the community that formed around Prius drivers supporting each others’ attempts to maximize miles per gallon, communities which have been chosen as test markets have also coalesced together.

MiniE drivers posted their locations on a Web site they shared, so if one of them found themselves far from home with a low battery, they could head to another MiniE driver’s home for some electrons to get home.  This self-organized grass-roots support network that sprung up through the use of social media is an example of how electric car test drivers have communicated with one another and with carmakers even without organized surveys like Turrentine’s.

Unsurprisingly, it seems that corporations continue to underestimate the power that social networking can provide to boost their products.  They’re largely using it in unsophisticated fashions.

Now, this isn’t to say that public charging networks won’t be needed.  Market acceptance is likely to increase as people see charging stations in places where they normally drive.  But at the end of the day, real-world use has demonstrated that the “chicken and egg” question that too many thought existed simply doesn’t.  Electric cars aren’t solely dependent on public charging stations.  Public charging stations instead are more dependent on electric cars, as I thought would be the case.

Follow

Get every new post delivered to your Inbox.

Join 247 other followers