GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter Instagram YouTube Icon Navigation Search Icon Main Search Icon Video Play Icon Plus Icon Minus Icon Picture icon Hamburger Icon Close Icon Sorted
Musings of an Energy Nerd

Rethinking Durability

Lessons learned from a visit to a Roman aqueduct

The Pond du Gard in Languedoc-Roussillon, France, was built to convey water to the ancient city of Nemausus (modern Nîmes). The stone bridge with an aqueduct on top is 2,000 years old.
Image Credit: Unless otherwise noted, photos are by Martin Holladay and Karyn Patno

Does durability matter? Most green building advocates seem to think that green builders should always aim to build durable structures. My own opinion differs; in fact, as I explained in a 2009 article on the topic, it’s hard to see any correlation between durability and “greenness.”

I recently had an opportunity to reconsider the advantages and disadvantages of durability when my wife and I visited the Pont du Gard in Languedoc-Roussillon, France.

The 2,000-year-old Pont du Gard is a remaining section of what used to be a 31-mile-long Roman aqueduct that conveyed water from a spring in Uzès to the city of Nîmes. Most of the aqueduct was buried; the buried sections of aqueduct resembled a masonry culvert lined with waterproof plaster. But Roman engineers knew that they couldn’t bury the section of the aqueduct that crossed the Gardon River. The technical solution to the river crossing was a 160-foot-tall stone bridge — the Pont du Gard — with an aqueduct on top.

The Pont du Gard is built of dressed limestone blocks, assembled without mortar (see Image #2 at the bottom of the page).

Look at the people walking across the lowest span of the bridge to get an idea of the bridge’s scale. The bridge is 160 feet high.

In ancient Roman times, the aqueduct conveyed 44 million gallons of water per day to the public baths, fountains, water spouts, and cisterns of Nîmes. (That amount of water would adequately meet the needs of 138,000 modern American homes.)

Without any maintenance, the aqueduct continued to supply fresh water to Nîmes for 300 to 400 years. Unfortunately, mineral buildup gradually reduced the size of the aqueduct’s channels; eventually, in the 4th or 5th century A.D., the aqueduct ceased functioning.

For centuries after water stopped flowing through the aqueduct, the…

GBA Prime

This article is only available to GBA Prime Members

Sign up for a free trial and get instant access to this article as well as GBA’s complete library of premium articles and construction details.

Start Free Trial


  1. Expert Member
    Dana Dorsett | | #1

    Regarding Satsop & WPPS
    The notion that the Satsop reactor was obsolete is a bit oblique. Satsop the handful of other nukes under construction under WPPS at that time were simply not needed on the regional power grid. With escalating costs due to construction delays, it became apparent that it would not be able to compete economically with the incumbent large-scale hydro power driving that regional grid (both then and now.)

    With no markets for higher-priced electricity on the horizon and rising costs, the WPPS was unable to make good on their bond payments, leading to the largest bond failure of it's kind in US history, a failure that has rendered nuclear power un-financeable in the US without governmental guarantees ever since. (Often impossible to finance even WITH guarantees.)

    Similar designs have been completed & commissioned since then in markets where the was needed and could compete on price. It's perhaps inaccurate to call it "obsolete". "Redundant", or "uneconomic" or "an ill timed investment" would be closer to the mark.

    It's probably a good thing that the WPPS fleet was never commissioned, and large scale reactors of that ilk have a horrible track record of hitting the financial benchmarks promised, even with loan guarantees. And it's not just a US thing- the 3 gigawatt Hinkley Point C reactor currently under construction in the UK is only able to attract financing with a guaranteed (and inflation adjusted over time) 15 US cents/kwh for it's output for 35 years, which is more than the cost of baseload off-shore wind power in the UK, and WAY over the cost of onshore wind & PV (even in the not so sunny foggy-dew British Isles.) Just the cost of the generation that has to be built just to back up that reactor for when it has suddenly go off line is daunting. It's the opposite of "energy security", and about as bad an investment in energy infrastructure that the UK taxpayers & ratepayers have ever bankrolled, given where the ~$40,000,000,000 USD of recently adjusted estimate (and not guaranteed limit) total capital cost MIGHT have been spent.

    But is uneconomic and unnecessary the same thing as obsolete? Maybe, if that's the last of it's kind ever built. In 2015 it's arguable that the existing power grid model of large-scale centralized generation has become obsolete, large nuclear being just part of that model. We'll see.

  2. GBA Editor
    Martin Holladay | | #2

    "Obsolete before it was commissioned"
    My "obsolete before it was commissioned" comment was a quip intended to highlight the stupidity of the developers and regulators who approved the massively expensive project in the first place.

    "Ill-timed" may indeed be more accurate -- but "ill-timed" is overly charitable, I think. How about "unbelievably stupid"?

  3. Expert Member
    Dana Dorsett | | #3

    Unbelievably stupid works!
    Unbelievably stupid surely applies to Hinkley Point C, and probably the Vogtle 3 & 4 in GA too.

    Some of these projects would be cheaper & better off to take the write-down now, before they load the fuel. As expensive as the WPPS fiasco was, it's cheap compared to the total cost of having to decommission a nuke that didn't run anywhere near the lifecycle built into the financial models. Given that the lifecycle cost of utility scale PV in 2015 in the US is cheaper than just the fuel costs of a US nuke, with onshore wind being even cheaper than that, it's not clear how you're going to get another 35-50 years of service out of big-iron generators (any type), new or existing. Heavy hydro won't be going away, but nukes & big fossil burners are too expensive to fuel & maintain in a smarter-grid environment heavily supplied with distributed zero-marginal cost renewables.

    "Ill timed" also works, even for Satsop. If they had waited for electricity markets to expand to where it was actually needed before breaking ground they would still be waiting, but that's what timing your investment is all about, eh? Even the most aggressive assumptions about growth in power sales didn't really support the scale of construction WPPS was going for. They seemed to be going at it with a "if you build it, they will come" mentality, a concept almost worked in the power industry the 1930s & 1940s, but by the late 1950s even finishing some of the Roosevelt era dam projects didn't quite make sense, but they did it anyway, and all but forced it into the market with "all electric home" incentives and discounts, advertising resistance electricity space-heating as modernity.

    By the 1970s when the WPPS nuke plans were being hatched there was ample reason to question who would be buying up all that excess power, and at what price. But at the time there were plenty of people in state government (including governor Dixy Lee Ray, formerly the head of the Atomic Energy Commission) who were steeped in the heady cooling water outflow, with the conviction that expanding energy use would continue unabated forever, despite clear signals in the marketplace (from the oil price shocks, and elsewhere), that energy would have to get cheaper to keep growing at prior decades' rates.

    Nuclear power went from "Too cheap to meter", to "Too expensive to matter" in about one generation, two at most. But some haven't gotten the memo yet, ergo Hinkley Point C. It remains to be seen if the Little Nuke that Could designs such as scalable molten salt reactors can beat those economics. I'm fairly pessimistic about their prospects, but could see TransAtomic's design becoming economically rational as a means of processing the existing stockpile of spent fuel rods into power + dramatically reduced half-life of the re-spent fuel. It's worth building a few to find out. Designing and building containment for radioactive waste that would be reliable for 100,000 years without much maintenance isn't something humans have any experience with. But containment for the ~500 years it would take for the molten-salt reactor's waste to cool off after extracting most of the rest of the energy in those spent fuel rods is something we can probably handle.

  4. Expert Member

    A very timely blog. Durability plays a very small role in which residential buildings in North America are abandoned or torn down. Demographics, regional economics, and fashion are much larger factors.
    There is also a good argument to be made that the older building in cities like New York or Montreal (like much of Europe), benefit more from their place in a well planned urban environment than their inherent durability. When something is a functioning part of a whole you tend to value it more.

    My own approach is to try and distinguish which elements of a building will need periodic replacement and detail them to make that easier. The way windows are installed, or even simple things like whether step flashing is nailed on to the roof or to the wall can mean a great deal to a future owner facing these tasks.

  5. GBA Editor
    Martin Holladay | | #5

    Response to Malcolm Taylor
    The chambre d'hôtes (bed and breakfast) that my wife and I stayed at near the Pont du Gard was an old stone farmhouse. It was lovely. Many people might say, "It's a good thing that the old farmer who built this house 300 years ago made it durable."

    But the house had undergone a recent renovation. I examined the work and talked to our host about it. The building had basically been gutted. It had new flooring, new wall finishes, and mostly new ceilings. It had new plumbing and electrical work. It had a brand new heating system. The bathrooms and kitchen were built from scratch. The windows and doors were all new. In some of the rooms, they left the old walls exposed -- and those sections of wall were uninsulated stone.

    So it's fair to say that the cost of renovating this farmhouse was basically the same as building a new house, except that the work may have cost a little more, because there were stone walls in the way of the work. Oh, and except for the fact that the house they ended up with didn't perform very well from an energy perspective, because of those uninsulated stone walls.

  6. Reid Baldwin | | #6

    How to measure durability
    For any structure, the probability of the structure surviving X years declines as X increases. For a house, it is not important that the median lifespan be >100 years. The 90% lifespan is a better way of measuring house durability.

  7. GBA Editor
    Martin Holladay | | #7

    Response to Reid Baldwin
    I'm not sure I understand your point. Are you talking about durability, or durability estimates?

    A researcher looking into durability statistics would probably gather data on houses that have either been abandoned or demolished. If you know the demolition date and the construction date, you know how long the house lasted.

    Estimates of the lifespan of existing buildings are, of course, uncertain.

    If you are looking at a set of houses -- for example, the set of houses built in Salem, Massachusetts in 1850 -- you could probably say that 90% of the houses lasted at least x years. But you can only determine that number if at least 10% of the houses built in Salem, Massachusetts in 1850 have already been abandoned or demolished.

  8. Reid Baldwin | | #8

    Durability Metrics
    Measuring durability probably wasn't a good title for my point. As you point out, you cannot measure lifespan until the end of the lifespan. However, we can design for a life expectancy probability distribution. That life expectancy distribution can be characterized by several points, such as how long before there is a 10% chance that the building will fail, how long before there is a 50% the building will fail, and how long before there is a 90% chance that the building will fail. The steps we take to increase the 10% number may not be the same as the steps that we would take to increase the 50% number or the 90% number. My point was that taking the steps to increase the 10% number is probably worthwhile. Taking the steps to increase the 90% number is probably not justified.

  9. GBA Editor
    Martin Holladay | | #9

    Response to Reid Baldwin
    I'm still lost. You wrote, "We can design for a life expectancy probability distribution."

    Really? Can we do that?

    All kinds of assumptions are built into that sentence. It sounds as if you know (a) Which neighborhoods will be desirable in 50 years, (b) Which style of home will be so cherished that it will be taken care of, and (c) Which type of homeowner will have a high enough income in 50 years to maintain their homes well, and (d) Which neighborhoods will be hotbeds of future teardown activity, because developers in the future can make more money bulldozing houses than fixing them up, and (e), (f), and (g) ...

    I'm not as confident as you are. I have a long list of reasons why buildings are abandoned or demolished -- but the quality of the roofing or the existence of a rainscreen gap aren't on my list.

  10. xhammerhurtsthumbx | | #10

    Enduring service is one of the proper goals of workmanship in construction. We all know that the future will change the purpose for a building and the way it works. But the rule of thumb is whatever you had a hand in putting together, should still be going strong when the your work is deconstructed.

  11. GBA Editor
    Martin Holladay | | #11

    Response to Timothy James Robinson
    I can't figure out whether your statement is a meaningless tautology or an impossibility.

    You wrote, "Whatever you had a hand in putting together, should still be going strong when your work is deconstructed." Are you talking about when the components of the building are deconstructed or when the entire building is deconstructed?

    If you are saying that the components of the building need to last until the day that a maintenance worker performs deconstruction of the (failed) components, that is a tautology. The components have to last until they fail; when the components fail, they are deconstructed and repaired.

    If you are saying that the components of the building need to last until the building is deconstructed, that's a high bar. So the only roofing permitted is roofing that last for the life of the building? All roofing has to be slate or 16-ounce copper? And when acid rain eats through the copper, what then? The building is bulldozed?

    Many building components need to be replaced before the life of the building is over. So what?

  12. JIM BAERG | | #12

    An influential book on this topic is Stewart Brand's "How Buildings Learn" It's been 30 years since I read it, but I remember Brand advocating that we build assuming that buildings will be re-purposed. He advocating thinking about buildings having 3 systems; the first was a simple structure that supported and protected the space It should be built to last for centuries. The 2nd system, if I remember correctly, was the inner walls, which would get moved around on occasion as the building's use changed. Finally, interior finishes and mechanical systems would get replaced fairly regularly so we so the designer/builder should make them easily replaced.

    Worth a re-read, I think.

  13. GBA Editor
    Martin Holladay | | #13

    Response to Jim Baerg
    I certainly agree with your suggestion that How Buildings Learn is worth a re-read.

    Many GBA blogs and articles have made the same suggestion. Among the GBA articles that discuss How Buildings Learn is this one: Low-Road Buildings Are Homeowner-Friendly.

    Inspired by Steward Brand, builder Tedd Benson developed the Open-Built Platform, a system that attempts to "disentangle" building components that need to be remodeled regularly from the more durable building envelope. For more on Tedd Benson's Open-Built Platform approach, see these two GBA articles:

    Unity Homes: Pushing the Boundaries of Home Building

    Service Cavities for Wiring and Plumbing

  14. kim_shanahan | | #14

    Does attention to detail create durability?
    My understanding of the green/durability question comes from the belief that increased attention to detail required to make a house tight, which then triggers the need for controlled air exchanges, will necessarily make a building better built and therefore more durable. As it is only a belief, and not a fact proven by time, since these notions are relatively new, we can only assume them to be true. And even if time proves them to be untrue, we do know the occupants are likelier to live in a healthier environment with relatively less expense.

  15. GBA Editor
    Martin Holladay | | #15

    Response to Kim Shanahan
    You propose an interesting theory: that "increased attention to detail required to make a house tight, which then triggers the need for controlled air exchanges, will necessarily make a building better built and therefore more durable."

    Unfortunately, recent decades provide plenty of counter-examples to undermine your theory. Many enthusiastic builders have jumped into the field of energy-efficient construction, and have managed simultaneously to make a house that is tight and in need of controlled air exchanges -- and have also ensured that the house has moisture problems and rots quickly. The cluster of EIFS failures is North Carolina is just one example of the type of failure I'm talking about.

    Of course, builders who pay attention try to learn from each failure cluster, and strive to do things differently in the future. But not all builders pay close attention, and we're all learning something new every year. (For example, look at all the conditioned attics insulated with open-cell spray foam that now have damp roof sheathing.)

    So, sadly, your formula for success -- make a house tight, and make sure the house has controlled air exchanges -- is insufficient to ensure durability.

  16. kim_shanahan | | #16

    Passive Houses too?

    Of course you are correct, but this statement gives me hope: " who pay attention try to learn from each failure cluster, and strive to do things differently in the future." I also know you have documented spectacular failures of certain passive houses, specifically in Europe, that had underground ventilation systems for air exchanges.

    But have we seen evidence that the notion of attention to detail and conditioned air exchanges that is taken to extremes in certified Passive House is creating disastrous unintended consequences?

  17. GBA Editor
    Martin Holladay | | #17

    Response to Kim Shanahan
    Like you, I'm a big believer in reducing air leakage through building envelopes. I also believe that a tight home needs a mechanical ventilation system.

    Most builders who believe in these principles are building good buildings. Yet we still have plenty of problems to keep us all busy, repairing and learning.

    Lots of builders have trouble with wall flashing details and WRBs, so wall rot problems will be with us for many decades to come. To see some fun photos, check out All About Wall Rot.

  18. morganparis | | #18

    Neglect tolerance
    Tolerance of extended periods of neglect is a characteristic of both the aqueduct and the old stone farmhouse but not so much of the airtight mechanically dependent structures that (for many good reasons) we find ourselves building today. Maybe in time we can learn to create homes that can remain sound even through the times we can't afford to actively maintain them.

  19. GBA Editor
    Martin Holladay | | #19

    Response to James Morgan
    You're right that (at least in centuries past) an old stone farmhouse could survive decades of neglect -- especially if it had durable roofing like clay tiles. But decades of neglect aren't kind to the interiors of any kind of home that people want to live in in the 21st century, because our interiors include kitchen cabinets, insulation, and electrical wiring.

    If a house includes these features, as did the stone-walled bed-and-breakfast where we stayed in France, then decades of neglect would result in the need for a total gut rehab job. The stone walls don't protect the expensive stuff.

Log in or become a member to post a comment.



Recent Questions and Replies

  • |
  • |
  • |
  • |