Cynical Question 4

[longisland]

Every year, we read news about flood happening all over the world. I believed the majority of the cities or areas flooded, the drainage system were designed with storm occurance taken into consideration. I'd like to know if the storm is the most severe case ever recorded; the folks who suffered from the flood, are they entitled to claim against the insurance company under the "act of God" clause? Is the design engineer liable for neglegence?
Another scenerio is if the government decided to construct a drainage system,such as roadside drain, for the storm duration of 10 years. A vehicle broke down due to flood, say a 25 year storm, on the same road. Can the driver file a lawsuit against the designer?

 

[DPB]

In my local area we have 47 years of rainfall data avaiable, yet as engineers we design for the 100 year return period event. The 100 year return period event is a statistical extrapolation of data. The truth of the matter, we are only guessing what a 100 year return period storm is, because statistically, we haven't necessarily seen one yet.

Flooding is NOT an "Act of God". Rainfall is an "Act of God". Flooding is the result of man's encroachment upon, and disruption of natural systems without the foresight and knowledge to account for unforseen events. The "Act of God" defense is more and more being challenged by the legal systemand is becoming a "weak leg" to stand on in terms of a defensive stretegy in court.

 

[dpa]

Not all flooding is due to man's encroachment. More people have died from drowning in Death Valley than from thirst. Flash floods in desert areas with rain are a common and serious hazard. However, we are certainly taking a lot more care now to prevent runoff than we ever did before. I am working on such a project right now.

DPA

 

[LHA]

I am not a lawyer, so I will answer only from an ethical standpoint. The actual results of court cases based on your scenarios are anyone's guess.

To your first scenario:
We licensed engineers are obligated to design to not endanger health, life, property or the environment. The engineer would be liable if he or she designed without regard to some flood danger. Now, how much danger is "due diligence"? Typically, the 100-year storm. If the 101-year storm occurs and the engineer can cite and prove the design was adequate during the 100-year storm, then the engineer should not be ethically liable.

To your second (and much more common) scenario:
The principle is the same as above. If I design a storm sewer to a specified 10-year storm, and the 11-year storm floods your car, you should just have to learn to be more careful while driving in floods. If the 9.9-year storm floods your car, I am negligent and therefore, (rightfully) liable.

Just my take on this.

 

[LeoSC]

For practical purposes, the 101-year flood is equal to the 100-year flood is equal to the 99-year flood. We must not forget that the 100-year flood is the flood that has a probability of 0.01 of being equalled or exceeded in one year. However, in most cases we don't have enough records to estimate the 100-year flood within 15% of the real value.

Estimation of the n-year flood is based on fitting a probability distribution to a flood time series. Commonly used and suitable probability distributions are Gumbel, 3-parameter LogNormal, Pearson type III, and others. Naturally, the longer the flood series, the better the estimate, particularly for extreme floods. The distribution is fitted by regression analysis, with the parameters computed by one of several methods, e.g. method of moments. Because of this, we can only talk of an estimate of the n-year flood, and not "the" n-year flood.

We have to remember that the probability distribution is fitted based on a sample, not the entire population. So if we get a flood equal to or greater than our "100-year flood" estimate, our sample has changed with the new data, and we should re-fit the distribution, based on the new sample.

The determination of whether a culvert or drainage system is designed for a particular design flood return period (10-year, 25-year, 100-year) is based on a benefit-cost analysis and/or other economic analysis, including calculation of marginal benefit/cost. The government is trying to optimize the use of resources. In some cases, the 10-year flood is the correct one. In others, the 1000 year storm may be the right one. It is a question of acceptable risk. Fortunately, we engineers in private practice don't have to decide on the level of protection; it's done for us in the policy decisions of the various agencies.

In any case, our duty as engineers is to provide the standard of care required, and to make sure that we compute the flood estimates, the water levels, the flow velocities, the culvert or bridge waterway areas, the channel improvements, the erosion protection, etc using our best efforts and skills. We must make sure that if we don't have the training and experience required, we ask another engineer who does to carry out the work. If we do that, we will be alright in any of the scenarios posted above.

 

[cvg]

in addition to benefit cost, agencies do base the design flood criteria on a risk assessment, albeit a qualitative one.

For example, the selection of a 10 year storm for pavement drainage assumes that for the smaller rainfall events, people will continue to drive and expect the roadway to not be flooded. Therefore, to reduce the safety risk of driving through flooded streets, a storm drain is constructed to convey this flood. During larger events, the assumption is that most people will generally stay off the streets, therefor the safety risk is reduced without the need to upsize the storm drain.

 

[dicksewerrat]

Gov't organizations can design to the 100 year sorm. But they biuld systems to a much smaller storm. No one can construct to the 100 year event. Too costly. Property owners cannot collect from the gov't because they set a policy to construct to the 10 or 15 year event. It has been tested in courts. If the property owners want to, they can remove the policy makers by election and pay for the storm drains to take care of the 100 or 500 year event. They won't do this because the costs go up much faster from the 10 to the 20 year and up and up. Talk to any city that does their own design and cost analysis.

 

[Hookem]

LongIsland and others: I have been witness to two 100-year rainfalls 10 days apart in 1976; I have been witness to two 200-year rainfalls three days apart in 1997; I have been a witness to a 10 inch rainfall in 3.5 hours; I have been a witness to a 29" rainfall in 24 hours; I have worked in many an industrial facility that had 15 min, 30 min., 1 hr, 4 hr, and 24 hour rainfall intensities; and we designed accordingly. I have worked with one City that had only a 1"/hr design (that was why their streets and associated areas always flooded whenever anyone spit in the street). ALL of the above was in South Texas (specifics withheld to protect the guilty).

 

[Focht3]

Hmmm,

I think I live in one of those south Texas cities you described. (The 29" in 24 hours one.) Why we haven't dealt with low water crossings is beyond me. And why the Director of Public Works seems to be working against fixing the problems is beyond me, too!



Please see FAQ731-376 for great suggestions on how to make the best use of Eng-Tips Fora. See faq158-922 for recommendations regarding the question, "How Do You Evaluate Fill Settlement Beneath Structures?"

 

[queque]

From Hookem above:
have been witness to two 100-year rainfalls 10 days apart in 1976; I have been witness to two 200-year rainfalls three days apart in 1997; I have been a witness to a 10 inch rainfall in 3.5 hours; I have been a witness to a 29" rainfall in 24 hours;

The problems in Flood hydrology stem from not being able to measure rainfall and runoff concurrently. If you use a model HEC 1 or similar, the key is based on your use of known calibrated data results. That is what is lacking. If you believe that rainfall is measureable for any storm anywhere, try setting up pairs of coffee cans, each pair 25 feet apart within a one square mile. Then see if the results make sense. Then see if you can measure the resulting runoff or obtain the runoff data from someone?