This presentation is part of the Mid-Century Modern Structures: Materials and Preservation Symposium, April 14-16, 2015, St. Louis, Missouri.
The Farnsworth House
by Robert Silman and Ashley Wilson
Ashley Wilson: All right. It’s one of the most recognized and photographed buildings from the mid-century portfolio. This residential masterpiece designed by Mies van der Rohe in 1945. It’s constructed in 1950 to 1951 and it’s on the Fox River in the flood plain about an hour outside of Chicago.
The building represents some of the fully realized emergence of the glass box and has a clear span interior. Like a ship, it has a lower deck and an upper deck. The upper deck contains a living space. When you’re within that living space, you feel like you’re hovering delicately over the river.
Can I see or raise your hands of who’s actually been there? When you’re inside, you have this view and you can see how close it is to the river. It’s the delicacy of the glass and in the hovering the elevation of the land in a proximity of the river that makes it so dramatic and so special. Philip Johnson, another owner of a glass house calls this “safe danger.” It’s the proximity of the river that’s causing so much regular flooding in the building.
Mies established to finish for elevation a height to be 5 foot above grade. At the time of the construction, that was 1 foot higher than the highest recorded flood level to the time. He intended the lower deck to flood occasionally and the upper deck to ride free board exactly as you see in these photographs with the river swollen underneath it. The inset is showing the building during construction and then, the larger picture, showing it last summer. Unfortunately, he underestimated the flood mark. The highest recorded flood was in 1996. Water inundated the house by over 5 feet.
This image is from 2008 flood. In the 12 years that the National Trust and Landmark Illinois has owned the site, water has only reached the interior once, but it cost very costly repairs. The previous owner to us was Lord Peter Palumbo and he owned the property for 32 years. During his tenure, he had a similar stewardship experience that we’re having which is this constant repair, constant conservation, and constant replacement of the parts. During his tenure, the building flooded twice.
We’re also dealing with deferred maintenance. The National Trust has no endowment on this property. In between the floods, we start to see hoarding our money for a flood event, and in between the floods, we’re still dealing with the oxidation of the steel frames around the glass and then, the there’s drainage issues with the travertine pavers on the decks.
The trust and Landmark Illinois bought the Farnsworth House at auction in 2003. We bought it for a couple of reasons because it’s iconic, because we wanted to emphasize how important midcentury modern is, because the other bidder that we knew of was planning on moving the building out of the way or out of the state. They’re planning on moving it to the East Coast and we wanted this building to be open to the public.
Unfortunately, because there was another bidder, another very wealthy bidder, the bids escalated and we ended up purchasing it for more than we own or more than we had, so we took a loan out on the building. Even though we’ve paid that loan off, this building has no endowment as a consequence.
Then the other question we’re always asked is how about insurance. When it comes to insurance, after the 2008 flood, FEMA calls us a repetitive loss property. We go off, but National Trust has insurance for all of our properties together. We leave that insurance plan and federal government has a National Flood Insurance Program and it’s actually a very good insurance. We use them and then we have additional excess insurance through a private company. Our rates did go up 25% a year until we could establish an actuarial rate which we were just able to do in the recent studies.
Another way to conceptualize how much flooding and how much water is going on this site is to take a look at this Andy Goldsworthy sculpture called Flood Stones that was commissioned by Lord Peter Palumbo in 1991. This cairn was situated just on the bank of the Fox River, near the Rob Roy Creek, and they’re supposed to match the different flood lines. You can’t see them. I can only point to one.
There are 13 of these flood lines. After the 1996 epic flood, he got Goldsworthy to come back to United States. He’s a British artist. He got him to come back to United States and he had to add on to this sculpture. By having 2 scale figures standing in front of it as goofy as it looks, you can actually see the enormity of the problem, and the challenges so we asked Bob Silman to design our way out of.
That is Edith Farnsworth. I don’t think she’s thinking about flood mitigation, but we sure are. We’ve been thinking about flood mitigation for the last 3 years. The whole process started when an acquaintance of Edith Farnsworth, a friend of hers, and a big fan of the house walked up to the National Trust President, Stephanie Meeks and said very bluntly, “What are you going to do about the Farnsworth problem?” Fortunately, that woman donated some seed money and we were able to employ it to our initial studies.
The first study, we hired hydrology engineers. They were known for their international work on flood-prone cultural sites and we asked them to quantify the Fox River, and they analyzed all of the flood data. What we learned was, so the end of the report was “There’s nothing we can do locally to make a difference. There’s nothing we can do on the site. There’s nothing we can do to the river near the site that it’s a 2000 square mile water shed of the Fox River. It starts in the Milwaukee Suburbs. It comes down passed our site.”
In the 1950 when the house was built, it’s 100% farmland and now, it’s mostly suburbanized and urbanized. In the next 20 years, it’s supposed to have 30% more development.
We also had a geotechnical engineer and we learned that the water level is pretty high. It’s 4 to 5 feet below grade and that the bedrock is 17 to 20 feet below grade.
Let me orient you with the site plan here. The red star is the Farnsworth House. From here over, on this side is the original parcel. We call it parcel one. That’s what Edith Farnsworth originally purchased. Then she ended up purchasing the second whole piece of land because the highway department did the takings, a combination on her property in both that big road that’s near the Farnsworth House. They built it on that side.
Then with a survey, so the big gray swath is the Fox River. You can see the bank of the Fox River. It’s passed the gray swath, so that’s existing bank of the river, but the Fox River’s intended on any day to be and anywhere on those points so that’s why it’s showed above, it’s called the flood way. Then the 100 and the 500 year floodplain lines are also shown in here. You can see them. They’re contiguous for much of the site and then they separate. The house is steps away from where the river should be. The 500 and 100 flood line is far away from the house.
With all of this information gathered, we then hire Thornton Tomasetti engineering out of Chicago and we asked them to just do a feasibility study on everything we could think of as ways to deal with flood mitigation. We ended up with 9 different options. They put together a report about those 9 options. Then we gathered that material and we wanted to have a team to help us decide what to do with the information we’re receiving. We put together what we called a Technical Advisory Panel and it’s everybody that’s listed on here. We met at the site to talk about it.
You can see those people we’ve mention today and some people in this room are actually on this panel. Dirk Lohan who is Mies’ grandson was here. When we were discussing it and going through the different options, we were able to eliminate some and the whole reason we bought the house was to leave it here, we’re not going to move it out of Plano and to Chicago, out of the state. We were able to eliminate some of them. Then we had a pretty deep discussion about a hydraulic lift because Dirk had proposed that because he was the architect for Peter Palumbo when Peter Palumbo was dealing with it.
After we had all of these discussions about the lift, the team decided to look inward and to hire Robert Silman to pick the top 3 different options we’re looking at and to further develop them.
Robert Silman: Oh, look at what climate change has done. If you don’t believe in it, believe it, ok. We were asked to look at the 3 most promising options from the original 9 in the report. Now, the first was Option A; that would keep the house in the same location that it was originally, but raised the grade high enough, some 8 1/2 feet so that it would be above the 500-year floodplain.
Now, there were several ways to do that. You could build a little knoll and put the house on it. That wasn’t seemingly satisfactory. They would try to grade it out so that it still looked like a farm field. It meant bringing in huge quantities of filth. All the trees of course would be lost, but it was a possibility of doing this. Now, of course in order to do that, you have to temporarily move the house to a location outside of the fill area and above the floodplain presumably, demolish the existing foundation still all around it, and then build new foundations and put the house back.
This would preserve the proximity to the river which is very important. We had a technical advisory meeting in the house and I found my eye wandering to the river and you could see little pieces of bark floating in the river. You could see the bubbles in the river. It was mesmerizing. You’re that close. You’re about 95 feet away from the river at this house, but it would’ve elevated it some 8 1/2 feet and looking down that much, it would be quite a different experience. In addition of course, it would change the landscape completely.
Option B was to relocate the house up the hill to higher ground. Now, to get above the 500-year floodplain if you can see the middle picture there in Option B, you have to be some 500 feet back from the river rather than the 100 or 95 feet that we were presently. That would change things dramatically. You couldn’t see the little pieces of bark. You couldn’t see the bubbles from there. It wouldn’t change the landscape however. What was in front of you would be the same landscape, but it would dramatically change the location of the house putting it right near the road and subjecting it to road noise. The notion of being in the middle of the farm field very close to the river would be totally lost. However, of course, the house will be preserved. Obviously, here, the house gets moved also. In fact, in any option, the house has to get temporarily moved or permanently moved.
Option C is a mechanical option where some system would be employed that would lift the house up during a flood. The flood would happen, go away, and the house would come back down. When there wasn’t a flood, you would see the house in its original position. You wouldn’t know anything was there because all of the mechanism for this would be underground. When there was a flood, of course, people wouldn’t be at the house. It’s a crisis period. It loses its historicity obviously, it’s up in the air.
I think if Mies had known what climate change would do, he might’ve designed the house either in a different location or higher or whatever; but we’re stuck with this now. The notion of leaving the house as it was, 99 1/2% of the time or more, it was very attractive to the Technical Advisory Panel. They did latch on to that one option as the one to be involved more thoroughly.
Dirk Lohan, Mies’ grandson pointed out that many years ago in 1967 when they built Neue Nationalgalerie in Berlin, they lifted the roof with hydraulic jacks. The notion of using hydraulics in building technology is not unusual. The notion of using hydraulics in any technology is not unusual.
Every day you get into your car that you put your foot on the break, that’s a hydraulic system. Every time you fly in an airplane, everything that works mechanically is hydraulic. In big movable bridges, these bascule bridges or swing bridges or lift bridges; those are all operated hydraulically. They all work. We depend on them. It’s not some new technology. In fact, when you take your car to the repair shop, you drive onto the lift, somebody pushes a button, voila, your car goes up in the air on a hydraulic lift. This is not anything new. Hydraulics is something that is all around and people understand it.
The exact system that we designed first of all, involve moving the house. You have to get it off its present site temporarily up the hill. There are some places it could’ve been stored temporarily. Obviously, above the floodplain in case a flood occurs while you’re working on the house. Demolish the existing foundations. Then the notion was to dig a pit under the house. That pit would contain all of the hydraulic mechanisms. It’s not hard to dig a pit. The pit would be below groundwater; but again, many buildings with basements are located below groundwater. We know how to do that.
Instead of having like your garage, a cylinder that goes straight up in the air and lifts your car, we felt that that was not the right way to do it. Because if we put their 8 columns if we had a lift, if we put our jack under each one and push the button, they all end up in the air; each one of these things would be a separate cantilever. There’s a lot of lateral force in a flood situation. You know the currents push sideways, a lot of wind usually happens because it’s a stormy time. These jacks are not meant to take lateral force, to act as cantilevers. There’s a seal in them that holds the oil and every time you move it sideways, the seal gets stressed and the oil wants to leak out. That’s not desirable to manufacturer’s warn against using them that way.
We designed a different kind of a system, which used a set of vertical linkage trusses. In the middle picture which is called the “Down” Position here, you see 4 turquoise trusses that are lined flat. They’re lying on the top of the pit and they’re connected to these yellow things which are the hydraulic jacks. We call them actuators. There are 2 per truss, so there are 8 actuators pushing these 4 trusses.
When the warning for the flood comes, the superintendent at the house pushes the button. The jacks work by pushing these trusses up diagonally until they’re in the vertical position and I’m going to show you this later on in a video. The trusses go from being flat to being vertical, and they raise the house by 9 feet. Now, there is a platform on the top of these trusses that links all 4 of them together like a big concrete slab, and the house is resting on that slab. When you raise the slab, you raise the house as well. Of course, you have to move the house back onto it.
When the building is in the “Down” Position, you don’t see anything. The site has a dirt area around that. There’s no grass because it’s under the house. There are 2 thin steel curves that come together. When the house is raised, one goes up while the other is fixed. When it’s down, they’re just below grade and you can rake soil over them.
There’s nothing visible about what will happen when the house is down. When the house goes up, of course, it’s a different story. The notion of the house being tilted up, when it’s in the “Up” Position, all of the lateral forces are taken by the trusses, not by the jacks. Now, in the diagonal position, they can take in line forces, but not laterally because the trusses are much stiffer than they are. Also, all of the load is in the trusses. You can relax the pressure in the jacks if you want to in the actuators because all of the support is in the 4 trusses now. This is another feature.
The controls would be in a waterproof cabinet somewhere in the pit. There would be an electric emergency generator obviously in case the power failed that could be up by the road, by the house. There are many ways of handling all of the controls for this. There are many different kinds of controls. The particular one that’s most important is the one that keeps the jacks in a unified working order so one doesn’t push harder than the other. This is a system that’s again, well proved. When we move buildings, we move buildings with unified jacking so that this thing as it’s going down the road knows how to adjust for a pothole or an elevation change.
We know how to do this. There are feedback sensors that change the pressure in the jacks. Not very sophisticated as a matter of fact, but commonly used. This is a system that can be tested. You don’t have to have a flood. If you want to know if it works, you have a test. You push the button, the house goes up, the trusses go up, you see if it works. If it doesn’t work and maybe you test it 3 or 4 times a year, it was my suggestion that you increase the visitor fee that day so people can see it, ok.
You test the house and if anything is wrong, you have the opportunity to fix it. Who fixes it? A hydraulic mechanic. They’re all over the place. It’s not some guy you have to get from NASA. This is not rocket science. This is hydraulics, easy to do. It’s off-the-shelf components, hydraulic rams with them all the time. These are not huge, oversized things. The amount of force in them is not something that’s out of the ordinary. It’s designed to be single point failure proof. How’s that?
If any one item fails, the whole system does not fail. This comes from things like Disney World. All of the stuff at Disney is hydraulically powered; the rides, the bridges that move over the canals. There’s hundreds of thousands of people walking around these things and they have to be damn sure that they don’t fail. The science of doing that is reliability engineering and again, something we do all the time. Movable bridges, the same way. There is a lot of precedent for this kind of a thing to happen. The reliability of it in our opinion is not under question. It’s not something that’s never been done before, never been tried before maybe not in this application. The equipment for it is something that is commonly used.
The concrete pit. The exact configuration hasn’t been decided because we haven’t designed this thing yet other than the preliminary; but you can see in the upper picture, there’s a pit in this case, a U-shape thing. One part has to be deeper than the other, but you can walk around down there. It’s probably about 8 feet high, 9 feet high. You can see what’s going on to service it. The cover is in the bottom picture. You can see a concrete cover on it. The jacks and the trusses are under there. When it lifts, it lifts the concrete cover. What isn’t shown here is that the house is on top of that and then there’s a layer of earth. That all gets lifted together.
Is there a precedent for this? Certainly. I talked about movable bridges. Here’s a really sexy one in Greece, 1988. This is a bridge that is over a canal that has very large ships that are high. To build a road that went over the top would have meant destroying the landscape on both sides. What did they do? Where’s the bridge? What happened? It’s in the water. It goes down, not up.
People say, “These systems don’t work underwater.” They don’t? This bridge has been in operation since 1988. It works absolutely fine. This equipment is very durable, very rugged. This bridge is used daily, several times it goes down. We’re talking about something that’s seasonal. Really, it’s going to be operated for testing. This is in a halfway position where you see the 2 roadways coming up out of the water or going down in the water. These systems are used all the time. Ashley?
Ashley Wilson: All right. Bob was pretty convincing. We all thought it was a really cool idea. We loved it. Not all of us. Us inside the National Trust liked it, but we knew we needed some people to peer review it because I don’t know. So many who used to interrogate this report tells that the engineering works and if there’s any unforeseen condition, so we have this group of technical peer reviewers on the top, and then we also have these one-on-one conversations for over an hour with many of our preservation partners, and many of you are still in this room, thank you very much.
Within the Technical Advisory Panel, it was split. More people liked the hydraulic system, but there were 2 that still had reservations. Then we decided that because the Chicago Architectural Community is so engaged, we’re going to do a roadshow. We went to Chicago actually 5 times. We had 2 meetings for professionals; architects, engineers, preservationists in different places. We had 2 public meetings; one at Plano, one at IIT Grand Hall and we met with our easement holder.
While mostly there was support for the hydraulic system, there was some loud and vocal opposition against it. What we did learn from the feedback is we needed to do a couple more studies. The first one was really obvious, but we just didn’t have time to do it because of the organic growth in this project, so we had to do a preservation plan to figure out what we’re going to be doing with the building this whole time. We needed a better understanding of the landscapes, so we needed that cultural landscape plan.
This is the team that’s currently working with us right now to do these 2 things. The big thing that we learned is the significance of the NHL is pinpointed 1951 only. Peter Palumbo has this property for 32 years. Edith Farnsworth has it for 20 and she had zero management. She built the house. She drove a car on the contractor’s drive to the house, parked in front of the house between the house and a river and got out, and that was it. Peter Palumbo was a steward. He created the gardens that you see today and he created this walk and all of the screening from the highway.
One of the things that kept coming up in these conversations was why not buoyancies. Bob’s going to speak about buoyancy quickly.
Robert Silman: One of the obvious things that must’ve occurred to you is why not let nature do the work. You’ve all been out on a floating dock somewhere and if you’ve ever looked underneath it, there are either big blocks of Styrofoam or 55-gallon drums that are sealed or some kind of a tank. It’s like a boat. It displaces the water and it floats. Why not when a float happens, dig a pit? When a float happens if you have buoyancy attached to the underside of the slab and the pit, the water will raise the house up. When the float is over, it’ll go back down again. It sounds simple.
In fact, somebody made a little model at one of our presentations of the Farnsworth House in a Lego on a little piece of cardboard and in a pot. Under the house were 3 empty water bottles, what you drink from. He poured water in there and the house went up.
In my mind as an engineer, this proved that if you pour water in the pot and there are 3 empty bottles in it, they’ll rise. Sure, this is a very nice romantic notion of how to make it work and oh, that we could do that. In fact, if you know what’s going on in New Orleans now with the shotgun houses, these schemes are being proposed. Indeed, at that level, they work because the shotgun houses are rough and ready and they don’t have nearly the requirements that this has.
Here is some downsize at buoyancy. There’s no way to test it. It won’t go up in the air until there’s a flood, so how do you know it’s going to work? It needs a guidance system first of all. That other picture we showed didn’t have a guidance system, but it’s got to have something. The shotgun house is they drill 4 telephone poles and that’s sticking up in the air, and put rings around them, and attach rings to the house so it doesn’t float away. When it raises, it stays there. When it goes down, it roughly comes back to the same place. Again, you can’t test it. We have nothing that we can have up in the air. Any guidance has to go back down in the ground. You could have telescoping poles presumably that might be able to do this.
Again, when they are subject to lateral force, they move a little bit. It’s quite tricky. As I say, no mechanism for testing, and you can’t inspect it, so you don’t know if over the years, something freezes in there, some things rusts or corrodes on the guidance system. The buoyancy tanks develop a hole like a road, we don’t know.
When we have a flood, one of the issues is going to be after the flood, it isn’t so clean. It’s not just water. You all know what floods bring with them; debris, mud, gosh knows what, dead frogs, whatever. They’re going to go down into the pit and you have to clean them out. This has no mechanism for doing that because the only thing holding it up is the water. When the water goes away, the house goes back down on top of all that crud in the pit. You can’t get at it. I don’t know how you do that.
In order to do that, you’d need some sophisticated control system. The other thing is you don’t want it to rise in a minor flood. There are many floods that may be a foot, a foot and a half, that’s fine. The house is safe. It’s already 5 feet up in the air, but this would start to raise immediately. You want to have a mechanism so it doesn’t. We could do that. We could put an intake scoop for the water somewhere above the grade. As soon as you start to do that, what if the scoop clogs up, what if this is happening during a period of ice, what if debris gets in the scoop hole. There’s a lot of stuff, a lot of ifs that we worry about.
The guidance system. All these things are what’s wrong with buoyancy and there’s no precedent for this kind of thing. Rapid accelerations that might drop, buoyancy didn’t work.
Ashley Wilson: One of the options that was brought up to us, is to actually move it to the farm field and we had never looked it that way, we had looked at moving it up on the site farther away from the river.
There’s a lot of compelling arguments about this, but one of them is the house would actually be closer to the river and be farther from the road. Because it’s a farm field, there’s nothing there so you could create an appropriate setting and so we’re calling this the move it philosophy or ultimate. It does give you the ability to interpret the old site. The old site which has the Peter Palumbo layers on it, you could build basically a ghost platform where the house is at the same size, and you could stand there and understand the flood history, and understand Farnsworth House and its old location and you’d still have the preserved farm field house in a new location.
Then you can see the location that it’s in is very compromised in a sense that you have this huge highway going by this very lot. It took 2 acres of her land. She actually sold the property once that highway went it there. She fought it in legal battle. The house was situated under a black sugar maple and that’s not there. You have to look at the integrity of the original site. It doesn’t have what it used to have for the reason that it was built there.
We’re looking at that with a farm scheme and we’re going to come back to this. When it comes to the criteria, we have no solution. We have these 2 schemes that we think are very viable. We’ve come up with criteria that these schemes have to pass to get this thing built, or moved, or preserved. The first one which is thanks to the park service for helping us see this so clearly is doing nothing is not an option. We have to do a steward, something with this building. The schemes have to fit that. It has to meet the Secretary of Interior Standards for Preservation subliminally because that’s how the National Trust looks at this property’s preservation standard, but our easement just says the standards. It doesn’t say which one.
It can’t jeopardize our NHL landmark. Here’s the kicker, it has to be approved by Landmarks Illinois who’s the easement holder on the house. The unwritten criteria that I didn’t put up here is a donor has to build or buy-in to this and pay the money for it. We did have a donor who was behind us originally, but once there was local and mild support, they’ve sat back for a while. Right now, this is unresolved and we will end this.
Robert Silman: We’re done except let me show you a wonderful animation, of what the hydraulic scheme would look like. Push the button. I just did; up, up, up. Flood is coming. All right, the flood alert comes, and they push the button, and the house goes up, and it’s literally with the push of a button. You can see those turquoise trusses going up in the air.
All I’m saying is that what happens is it stays up in the air for that period of time that you need to clean out the pit. It could be days, it could be weeks, whatever it needs to be while it’s up there. You clean it all out and this is opportunity to work on it. It could be done for maintenance reasons, or it could be done after a flood; but the idea is that it’s controllable, and this is something that it’s not dependent on nature.
You can stop it anywhere you want, but you stop it when it’s all the way up and the load is in the trusses. Now, you have all the time in the world to work on it. This is the value of the scheme and that’s it, ok, very good.
Built by Mies van der Rohe in 1951, the Farnsworth House is unarguably the most iconic building of the mid-century catalog of architecture worldwide. It is revered by all who know it for its elegance, ingenuity and originality. The challenges of the floodplain site are also well known as the building has endured numerous flood events over the past six decades; once during construction and again just three years after Dr. Edith Farnsworth moved into this retreat by the Fox River. Flood waters have breached the interior of the house three times since then, and caused extensive damage—and costly repairs—to the structure, its systems and its furnishings. In 2013 alone, there were two significant flood events, and the problem is expected to become increasingly acute in the years ahead, given current flooding models.
The rising waters of the Fox River represent a potentially devastating threat to Farnsworth’s future. Given that reality, the owner, the National Trust for Historic Preservation, is analyzing the realities of the site conditions and developing solutions.
This paper will present three possible options for mitigating flood damage: (1) Placing a large volume of fill on site so that the house will be repositioned at a higher elevation in its present location (2) Moving the house back some 400 feet from its present location to higher ground, out of the flood plain (3) Constructing a mechanical system in a pit below the house, completely concealed from view, to raise the house above any future flood by temporarily lifting it on a system of hydraulic jacks and steel linkages and lowering back down after the flood subsides.
Robert Silman is President Emeritus of the 140 person firm that was founded in 1966. He is presently located in their Boston office. The firm has worked on more than 18,000 projects, with about half being new construction and the remainder renovation, adaptive reuse and historic preservation. Among the latter, the firm has consulted on more than 450 registered landmark structures. Robert also is teaching at Harvard’s Graduate School of Design. Recipient of many awards for his work in historic preservation, he is also a former Chair of the Preservation Technology and Training Board of the National Park Service.
Ashley Wilson is the Graham Gund Architect for the National Trust for Historic Preservation and oversees changes to the historic properties and landscapes owned by the National Trust. Previous to this position she was a tenured professor and founding faculty member of the Clemson University/College of Charleston Graduate Program located in Charleston, SC. She worked at a project architect for Oerhlein & Associates in Washington DC, Kapp & Robbins Architects and as the Assistant Architect for Thomas Jefferson’s Academical Village. Ashley received her architecture degrees from the University of Virginia and the University of Notre Dame and is currently serving as the 2015 Chair of the Historic Resources Committee of the American Institute of Architects.