Raising the Molly Brown--lessons from an unsinkable boathouse
President, Life Is A Project, LLC
Sometimes projects appear to be snake-bit, seemingly attracting more than their fair share of uncertain events or conditions. Over the past 10 years an unassuming boathouse on a lake in Tennessee has served as the perfect object lesson, experiencing:
Flood – The absolute excess of things that must be dealt with on most projects can overwhelm the project manager and divert our attention from what is important, staying moored to the project's objectives. When a “100 year event” set the Molly Brown drifting away, only consistent monitoring of forecasts and quick response prevented disaster.
Fire – Putting out fires is an inevitable part of our job, no matter how well planned our efforts. When historical lessons learned were not effectively communicated, the Molly Brown was at risk of burning to the waterline. Fortunately, the impact of this event resulted in the opportunity to accelerate another project.
Wind – Freakish, improbable, high impact events, Black Swans, can strike any project and potentially destroy any likelihood of success. When an unexpected tornado almost generated a Wizard of Oz outcome and left the Molly Brown half-sunk, a return to basic, cost effective planning and engineering succeeded in raising her.
Origins of an Anomalous Structure
The Molly Brown is a 16'x24' floating boathouse on Watts Bar Lake in 10 Mile, Tennessee. A one-of-a-kind structure that is “grandfathered in”, no more like it can ever be built. Exactly how old it is no one knows, but the previous owner constructed it on four Army Corps of Engineers' aluminum pontoons that must have fallen off the barge since they are the only ones ever seen in private use.
The boathouse and associated dock were initially quite humble in appearance (Exhibit 1). Several risk events over the last decade have threatened their very existence. Responding to these in conjunction with an engineer's incessant tinkering over time, has transformed and saved it (so far) and delivered important lessons in project risk management.
Exhibit 1 – Humble Beginnings
Origins of Risk Management
The origins of the word “risk” go back to Greek roots and its first use by Homer in describing the threat that Odysseus faced from the cliffs of Scylla. “English borrowed the term from Spanish, German from Italian and both were backed up by the French “risqué” of the 18th century. Dictionaries confirm that it was a metaphor for “difficulty to avoid at sea”… The various forms of systematic use of risk analysis and risk management strategies are normally traced back to the late fifties or early sixties. Their use spread largely driven by statutory frameworks for the regulation of safety, health, and environmental risks.” (DNV, 2012)
Dealing with uncertainty has been a recognized part of formal project management from the beginning. The PMBOK® Guide has been remarkably consistent in its proportion of risk management content, close to 10% from the 1987 draft right through the Fourth Edition.
The Purpose of the Tennessee Valley Authority
“To improve the navigability and to provide for the flood control of the Tennessee River; to provide for reforestation and the proper use of marginal lands in the Tennessee Valley; to provide for the agricultural and industrial development of said valley; to provide for the national defense by the creation of a corporation for the operation of Government properties…”
Tennessee Valley Authority Act, 1933
Originally chartered in 1933 for the mission above, by 1942 the nation was desperate for electricity to help feed the WWII military manufacturing behemoth. Harry Truman famously declared “I want aluminum. I don't care if I get it from Alcoa or Al Capone.” Today TVA is one of the largest electricity generators in the U.S. and most of the objectives are now in a mature maintenance mode. The management of 47 lakes and reservoirs that TVA performs is dedicated to balancing energy production, navigation, flood control, water supply, and recreation. From a purely personal point of view recreation is not only last on this list, but last in priority.
This presents special challenges to landowners with structures on the water and homes near the TVA high water mark. A floating boathouse has both advantages and disadvantages compared to its fixed counterparts.
|Floating Boathouse (Molly Brown)||Fixed Boathouse (Everyone Else)|
|Easy mooring/access to boats no matter the water level||Able to lift boats for access and to prevent hull fouling|
|Cool, floaty feeling inside||Stable footing in any condition|
|More engineering and structural concerns||No worries on weight loading of house and contents|
|Can float off pilings at high water||Boats and/or house can be destroyed at high water|
|Can land on bottom at low water||Can strand boats at low water|
The Molly Brown floats in its own small cove that is dry when the reservoir is lowered, roughly October to April. At any time of year TVA may need to make exceptional alterations to lake levels in response to weather and flood potential (Exhibit 2).
During the shoulder seasons the boathouse can be yo-yoed many times, landing on the mud each time the water reaches about 736’.
Normally the lake cannot not rise above 745’.
Exhibit 2 – Attempting to Manage the Uncertainty
An Unlikely Intersection of Events
“On September 27, 2002, at approximately 8:30 a.m., the Rhea County, Tennessee 9-1-1 Center began receiving calls about a fire at the Watts Bar Hydroelectric Plant. The plant, built between1939-42 and operated by the U.S. Tennessee Valley Authority (TVA), is constructed of steel and concrete, and sits on the Tennessee River midway between Knoxville and Chattanooga. The dam supplying water to the hydroelectric plant is 112 feet high, and approximately one-half mile long. The plant has a generating capacity of 175,000 kilowatts. It supplies power for TVA, and provides back-up power for the Watts Bar Nuclear Power Station located directly south of dam.
The fire began in the vertical cable shaft and spread so rapidly to the control building and burned so intensely that fire suppression personnel were unable to make entry into the building until 9:10 a.m. The hydroelectric plant has remained closed since the fire, at a business loss of approximately $100,000 per day. Although many people perceive a steel and concrete structure as not being a fire risk, this fire illustrated, once again, that a fire in an electrical system can quickly be life-threatening regardless of the structure type.” (USFA, 2002)
Rebuilding the control room took over a year, and during that time TVA had limits on its ability to affect the flow of water through the dam. One of the rainiest springs in history came in 2003. On May 13th the flow down the river system from East Tennessee overcame Watts Bar Dam's ability to release water, resulting in record high water and the flooding of Interstate 75 in Chattanooga.
Impacts in Action
In addition to the Operating Guides, TVA's website also has current and three day predictions of water levels. Fortunately, monitoring this became a habit when not at the lake. One afternoon I looked at the numbers and a shirt-cuff calculation revealed that the Molly Brown was going to float off the pilings. I arrived at the lake after midnight and the next morning awoke to a potential disaster (Exhibit 3).
Exhibit 3 – All that Floats Can “Turn Turtle”
The dock and its poorly set support posts were trying to rise from the lake bottom. The boathouse was held only by a safety cable to shore, had done a 360°, and was now coming down on a wood and PVC post as fast as TVA could let the water out. Worse yet, the ski boat I had inherited from my father was still inside! Although effective risk monitoring of the water level trigger had caught the event, no risk response planning had been performed. Only a workaround involving a neighbor, a chainsaw, and a ramming speed maneuver saved the Molly Brown.
Calculating the Odds
Whenever possible we should attempt to quantify risks in the analysis process or post-event. Watts Bar Lake was completely filled in 1942; between that time and 2003 water levels have exceeded 746’ only seven times. The odds of that occurring on any given day are about 1 out of 3,180 or .000315.
The Apparent Problem
The dock/walkway out to the Molly Brown is over 50’ long and was supported by 4”x6” pressure treated wooden posts sunk into the mud with a minimal amount of surrounding concrete. The main horizontal members were the same size and the surface was 2x12s four feet long with no center support. The structure appeared to be 25-30 years old (minimum) and showed no signs of any annual maintenance. Over time the sun, weather, and repeated submerge/dry cycles had turned much of the structure into a soft, spongy hazard.
The Unintended Consequences of Applying Band-Aids®
Serious work on the dock was always delayed due to the mass of other pressing projects at the lake: grading, building a garage, landscaping, swimming, boating, fishing, etc. Instead, over the years a handrail was added, larger boards were tacked onto the outside, new wood replaced the old surface, and large concrete footers were poured around the posts at low water during the winter. This sufficed to keep the structure standing, but only served to hide the old, punky wood inside a deceptively new cloak.
In 2007 some friends going home to Virginia from the PMI Global Congress in Atlanta stopped at the lake. After a day of boating, tubing, and barbeque we set off an impressive number of fireworks on the dock; a good time was had by all. Sitting in the house several hours later their son said “It looks like there is something on fire down by the water.” Even though he was a budding Eagle Scout I found that hard to believe until I looked out the window. The four of us ran pell-mell down to the shore and started pouring lake water on the flames.
A spark from the fireworks had fallen down into the old dock structure, smoldered there, and finally burst into flame right before bedtime. If it had caught 30 minutes later the dock, and very possibly the boathouse, would have been a total loss.
Poor Communications and the Tribal Memory
I told everyone that goes to the lake about the incident and made sure that they knew how important it was to be aware of the hazard that existed after pyrotechnics. Despite the fire we continued to set off fireworks several times a year from the dock. The communication was not iterated enough to become part of the tribal memory and was quickly forgotten. In November of 2010 some of my in-laws went up to the lake without me. Fireworks again topped off a long day. Late that night my mother-in-law was heard to exclaim “It looks like there is something on fire down by the water.” This time the spark had much longer to work its damage, but was finally put out. A six foot section of the entire dock was burned out; again the Molly Brown was unscathed.
The Normalization of Deviance
If a threat risk is present, yet the event does not cause project failure, we have escaped a deviant condition. Often this behavior is accepted over and over until it becomes “normalized.” This concept was first defined after the investigation into the NASA Challenger disaster (Exhibit 4).
“The attention paid to managers and rule violations after the disaster deflected attention from the compelling fact that, in the years preceding the Challenger launch, engineers and managers together developed a definition of the situation that allowed them to carry on as if nothing was wrong when they continually faced evidence that something was wrong. This is the problem of the normalization of deviance.” (Vaughan, 1996)
Exhibit 4 – The Challenger Disaster
Even more troubling is that it appears to be human nature to consider a lucky outcome as good as a success resulting from executing to a well-designed plan. This “near-miss blind spot” effect is supported by several studies.
“We believe one factor that can lower inappropriately (from a normative perspective) people's perception of the risk of a hazard is information about prior near-miss events. A near-miss occurs when an event (such as a hurricane), which had some nontrivial probability of ending in disaster (loss of life, property damage), does not because good fortune intervenes. People appear to mistake such good fortune as an indicator of resiliency… Our research thus shows how people who have experienced a similar situation but escape damage because of chance will make decisions consistent with a perception that the situation is less risky than those without the past experience.”
(Dillon, 2011, p.1)
Another factor in the acceptance or outright denial of risk is the generally optimistic tendency that seems to be a basic part of human nature. “It will come as no surprise to the average project manager that most people tend to be overly optimistic in their assessments. Whether reporting status, providing an estimate, or forecasting an outcome, the majority will err on the positive side. What may surprise most of us is that this is a hard-wired condition.”
(Kinser, 2011, p.2) This is similar to the cultural bias about uncertainty that organizations experience causing them to adopt a risk avoidance or risk taking mindset.
Calculating the Odds
Over 10 years fireworks were launched about 50 times; twice the dock caught fire. The odds of that occurring are 1 out of 25 or .04.
Leveraging Risks into Opportunities
Sometimes it takes a significant risk event to provide the impetus to initiate a long overdue project. Instead of continuing to put Band-Aids® on the dock we decided to do it over right (Exhibit 5). Any construction on TVA property (technically they own the land up to the 750’ mark) requires going through a long involved permitting process. Fortunately we had started this almost a year before the fire in anticipation of stabilizing the shoreline on “the point”, a finger of land that protects the cove which had been eroding for decades. While I was at it anyway, I had submitted drawings for the dream dock that would provide actual seating space to enjoy the lake and eventually capture the boathouse in its own slip. Only by expediting the process with the bureaucracy for several months were we ready to go early in 2011.
Tearing out the old dock and utilizing the new footers, we rebuilt the walkway wider and better. Watching the forecasts of weather and water levels was critical in the timing of setting new posts for the widening of the dock beyond the boathouse on both sides.
In some cases the posts had to be set in holes full of water that would seep in; we would count on the weight of the concrete to displace the puddle. Air temperatures in the 20-30 degree range and water of 50 degrees was the norm.
Exhibit 5 – Leveraging a Risk Event
Sometimes the dock structure would have to be built while perched on a 4'x10' raft, the Kon-Tiki, since the water would often rise in the interim. Doing solo installations of 12 foot, 2x10s while balancing on a floating object brought along risks of its own. By spring of 2011 we were in a position to run the dock down both sides of the Molly Brown the next winter and have more control over its movement than the existing two pilings with cable loops provided.
The grinding of the thin walled pontoons on the ground during transitions between floating and landing had always been a concern. In addition, the lake bottom is not level, so the structure experiences significant racking or twisting during low water. Over the years I had added extra hardware, straps, plywood, and flotation to help compensate for these issues, but creating a slip offered much more potential for the future survival of the boathouse.
Like many of the reservoirs in East Tennessee, Watts Bar Lake lies at the edge of the Cumberland Plateau, an escarpment of about 400’ that rises from the foothills west of the Appalachian Mountains. As a result it experiences katabatic winds that result from the cooler, denser air on the plateau rushing downhill. Add that to being in a high frequency tornado area (Exhibit 6) of the United States and one might expect some unusual events.
Impacts in Action
Although 72 miles long, Watts Bar is a fairly narrow inland lake. The summers can be so calm that sailboats struggle to move. Other seasons are much more variable; I was forced onto a lee shore in a 13’ sailboat one March and lost my only paddle. After walking the boat around the point I raised the lateen sail about two feet out of the water and flew home.
Even in summer the wind occasionally plays hob with objects at the lake. Freestanding lawn swings with fabric sunshades are repeatedly blown over; anchoring them to the ground just results in the sunshades being torn off. While on the water, sudden windstorms, usually preceding rain, can come out of nowhere.
Exhibit 6 – On the Edge of Tornado Alley
This sometimes results in winds and waves so severe that even 20’ powerboats are wise to take cover. Once a neighbor was blown to shore in a pontoon boat and escaped unscathed, although the boat was pounded to pieces. In the last decade several towns 30-40 miles away have experienced severe tornados, but as far as we knew 10 Mile, Tennessee (Pop. 581) had never been subject to any. This all changed on June 28, 2011.
An Eye Witness Account
Our neighbor's newly built dock is over 80 feet from shore and another 100’ from their house with two slips, a sizable roof, and a small closet in the center. A couple visiting their place were enjoying a sunny, summer day when the “Black Swan” occurred. Afterward this is what they related.
The day had been picture perfect until about 2:00 PM when it started getting a little dark and windy. They decided to wait until till it started raining to take cover since there are often “false alarms” at the lake. Suddenly the wind increased dramatically and a cone of water started hurtling from the lake just beyond the power lines - about a quarter mile away. It touched the land this side of Sam's Boat Dock and a huge cloud of debris flew into the sky. The tornado was moving towards them so fast they knew we could never make it to the house; instead they hunkered down behind the closet and prayed.
Next thing they knew the noise, wind, and water were a solid mass around them and they saw our entire boathouse (which they were facing) literally rise up into the air. The rear end tipped higher and something blew out from under … it seemed to hesitate … and fell back into the water. The tornado struck the far shore of the cove, toppling some trees, cleared the hill and disappeared. The entire time from start to finish was under a minute.
Some 50 trees north of us were blown down. We lost one tree by the lake and two even larger ones in the front yard over 200 feet away. The boathouse two doors away ended up in the owner's backyard; the one that shielded the witnesses was essentially untouched. The Molly Brown lost the two rear pontoons and was half-sunk in 40” of water (Exhibit 7) with my twenty foot ski boat, the TANSTAAFL, still in the slip without a scratch on it. The trees the next cove over did not even lose a leaf.
The options at this point were limited.
- Wait on an insurance settlement
- Raise the Molly Brown
- Demolish and rebuild
- Wait on low water to repair
These were not mutually exclusive, for instance the insurance adjuster might declare it totaled, pay, and we could raise it. Unfortunately, the adjustor could not come for 10 days, and the longer it was in the water the worse it was for the structure. Initial estimates from commercial firms for any approach to raising it started at $10,000 and went up. We decided to photograph and document everything and try to raise it ourselves.
Exhibit 7 – Happy 4th of July
The top priority was to get the boat out, even though I was out of town when the event occurred and could not return for three more days. The on-site project manager (my 81 year old mother) managed to do this with the assistance of friends and complete strangers. In the tradition of small-town country folk everywhere, when they saw troubles they stopped and volunteered to help. The next task was to estimate the weight of the boathouse so that solutions could be realistically assessed. The techniques available were:
- Analogous - relate the unknown to the known by comparing to previous similar items and adjusting
- Expert Judgment - use SMEs with experience in similar efforts to provide numbers
- Modified Delphi - synthesize expertise from a group who perform rounds of estimating and discussion
- Parametric - apply a mathematical model using previous factors with known relationships
- PERT - gather Optimistic, Most likely, and Pessimistic numbers and calculate using (O + 4M + P)/6
- Bottom Up - estimate each of the smallest project elements and add or “roll” the numbers up
Initially it seemed the Bottom Up process was the logical choice, but in the end the parametric technique provided a simple, quick, and reliable number. The Molly Brown weighed 6,350 pounds ±5%.
Next, a variety of solutions were considered to raise the boathouse ASAP and at minimum cost. In the end it was achieved by two men in a day and a half for a materials cost of $281. This paper does not document all the steps involved, but the associated presentation does.
Calculating the Odds (with a historical perspective)
I figured the Black Swan tornado was surely something that was a once in a lifetime event in our area. Talking to my cousins (they lost 28 of the 50 trees north of us) whose family has had a cabin there since the 1940's revealed otherwise. A similar event occurred in the 50's and again in the 70's. Only by engaging someone with a long term historical perspective could the real odds be determined.
Watts Bar Lake was completely filled in 1942; between that time and 2011 three extreme wind events (tornados or microbursts) have been experienced near our cove. The odds of that occurring on any given day are about 1 out of 8,395 or .000119.
Risk Relationships and Continuing Work
Every risk event at the lake has initiated more changes and improvements. The work done after one incident usually impacted the next either positively or negatively.
- The variability of water levels observed at Watts Bar led to a habit of continual monitoring of metrics. If not for that, the flood would probably have destroyed the boat and boathouse.
- The flood led to the reinforcement of the dock to keep it from floating loose. This resulted in a secondary risk, hiding the defects that led to the fire.
- The fire led to building a better, stronger dock and extending it past the full width of the boathouse. Large gouges in the dock indicate where the Molly Brown hit the new structure, hesitated, and fell back. If not for that, it and the boat would probably have ended up in the back yard.
- The chains that held the pontoons in place had rusted in two; I never knew their original purpose until the tornado. They have since been replaced to prevent the pontoons from blowing away again.
- The dock on the right side now extends to the end. One rotted and leaning piling has been replaced by a higher dock post and trolley system. This is the next step in mitigating the risk of racking.
- Extra flotation has been added between the pontoons to mitigate the chance of sinking if a pontoon fails.
Other projects are in the queue to extend the life and utility of this snake-bit, but remarkable boathouse.
“What are the odds?” is a question often asked when an unlikely risk occurs. This paper explains how to actually answer this (usually rhetorical) question, anticipate uncertain events, and use common sense and common folk efforts to apply effective workarounds when a crisis strikes.
The odds of all three of these risk events impacting this one boathouse are found by multiplying them together.
No team can ever anticipate or control all uncertainty, but using a risk management process is a vital part of any project. Despite our best efforts, some “unknown unknowns” will always exist. Applying ingenuity in developing responses to these unanticipated risks is an important part of our jobs as project managers.
Exhibit 8 – It's a Tough Life, but Someone's Gotta Live It
- Examine your project for strengths and weaknesses to identify relevant risks
- When an ongoing risk is detected, determine an effective trigger and monitor continuously
- Be aware of other risk events' potential impacts on your project's uncertainty
- Applying too many Band-Aids® to your project can mask risky conditions
- Risk communications need to be repeated to become part of the tribal memory
- Don't fall prey to the normalization of deviance
- Understand that one bullet dodged does not decrease the risk of the next
- Leverage risk events into improvements for your project and its risk exposure
- Seek out historical statistics to help anticipate Black Swans
- Let the situation guide the selection of techniques for risk or estimating
- Engage in flexible thinking when planning workarounds to risk events
Dillon, Tinsley & Cronin (2011) Why Near-Miss Events Can Decrease an Individual's Protective Response to Hurricanes. Risk Analysis, 31: 440–449
DNV (2012) Risk - a word from ancient Greece Retrieved on 12 August, 2012 from http://www.dnv.com/focus/risk_management/more_information/risk_origin/
Kinser, J. (2011, October) Nurturing Human Nature in Projects – It's All Psychobabble Rap to Me. PMI Global Congress 2011, Dallas, TX
U.S. Fire Administration/Technical Report Series (2002) Fire at Watts Bar Hydroelectric Plant Retrieved on 12 August, 2012 from http://www.usfa.fema.gov/downloads/pdf/publications/tr-147.pdf
Vaughan, D. (1996) The Challenger Launch Decision: Risky Technology, Culture, and Deviance at NASA. Chicago: The University of Chicago Press
© 2012, John Kinser, PMP
Originally published as a part of 2012 PMI Global Congress Proceedings – Vancouver, Canada