Rainfall Event

[wettpaint]

I have been asked to find the rain fall event year for a recent storm. I have the following information to go off of...

Time of Day total rainfall
858 0.02
901 0.15
905 0.34
918 1.10
928 1.25
947 1.41
954 1.44

The rainfall is a running total of the entire day.

The area is for North Mississippi.

Thanks

 

[psmart]

If you take the total 24-hour rainfall depth, you can compare this to the standard 24-hour events for your location and determine the approximate return period for that depth. For example, the depth may fall between the 10 and 25 year events.

Note that the rainfall distribution doesn't enter into this process. You're just using the total depth. Every storm has a different distribution which is not normally considered when you evaluate the return period (frequency).

On the other hand, if you want to simulate the effects of this storm on a given drainage system, you could use the actual rainfall observations to define a mass curve for use with your runoff model. The fine detail of a specific storm can cause different results than would be predicted by the return period alone, and these effects will be different for every watershed and drainage system.

So you need to take a careful look at the question being posed to you: What are they really asking and what is the intended use of this information?

Peter Smart
HydroCAD Software

http://www.hydrocad.net

 

[cvg]

24 hour depth may or may not be the best approach. if the storm was a shorter duration such as 2 or 6 hour than you should compare to 2 or 6 hour rainfall durations. check the precipication frequency data server

http://hdsc.nws.noaa.gov/hdsc/pfds/index.html

 

[beej67]

It's an hour long storm, so 24 hour data doesn't apply at all to his case.

You have to find an IDF curve or IDF table for your region. Without that, you can't proceed with the analysis. Search the web, and talk with local county engineers.

You've got 1.44 inches in one hour, so that's an intensity of 1.44 in/hr for a 1 hour storm. Go across the bottom of your IDF graph, find the line for 60 minutes, read up to 1.44 in/hr, and then see what curve you fall on. That would be right around a "1 year 60 minute" storm in Atlanta, but you need to find an IDF curve or table for your region and check on it.

Next, look at smaller segments and do the same analysis. For instance, from 9:05 to 9:18, you had (1.10-0.34=) 0.76 inches of rain fall, over a (9:18 - 9:05) 13 minute time frame. That's a rainfall intensity over that 13 minute segment of 3.5 inches per hour, which is quite a lot more intense, but would still only crack about a "1 year 13 minute" storm in Atlanta. Again, you need to go by local IDF data.

Repeat the analysis for any number of segments of the storm, or any combination of contiguous segments. It's completely possible that a storm could be a "1 year 1 hour" storm while also being a "25 year ten minute" storm when you look at a portion of the storm, if the storm is flashy. So few engineers seem to realize this. Doubt this storm is such a storm though.

If you can't find better local data, CVG's link is a good one, but watch your units. I believe they're giving the table in inches of total rainfall, not in intensity. Based on a quick glance, looks like your storm is "around the 1 year storm" for your area.



Hydrology, Drainage Analysis, Flood Studies, and Complex Stormwater Litigation for Atlanta and the South East -

http://www.campbellcivil.com

 

[psmart]

Although the data given is for about one hour, the OP also said "The rainfall is a running total of the entire day". So the actual storm duration was unclear.

In any case, a single event can certainly be viewed at shorter durations, each of which can turn out to have a different return period. Hence my question: We really need to know exactly why the question is being posed and how the information will be used. It's not a simple question at all.

Peter Smart
HydroCAD Software

http://www.hydrocad.net

 

[Drew08]

Questions like this are usually asked in litigation when a structure fails, or a construction site is flooded delaying work. Parties involved want to know if the failure/delay was an act of God, or should it have been expected and planned for. They don't care about the details of the duration. If you have cumulative data, convert it to incremental, find the largest 5-min, 10-min, 30-min, 1-hr, 2-hr, etc. etc. 24-hr, 2-day, etc. etc. cumulative depths and compare to the NOAA Atlas 14 point precipitation frequency estimates. The duration with the largest recurrence interval rules the day – or the negotiations. This could be all done simi-automatically in excel. Also, the largest cumulative depth for a given duration won’t necessarily be from the beginning of rainfall, but likely somewhere encompassing the peak of the storm.

Longer durations (>24-hours) would be important indications of antecedent conditions. A retaining wall failure might be caused by longer-term saturation, not necessarily instantaneous peak intensity. NOAA Atlas 14 has return frequency depths up to 60 days.

 

[beej67]

^^ Yep.

I did this exact analysis for a construction site BMP failure here in Atlanta a couple months back. It's not that hard to automate a field of rainfall data in excel to flag certain events if it's recorded on regular intervals.



Hydrology, Drainage Analysis, Flood Studies, and Complex Stormwater Litigation for Atlanta and the South East -

http://www.campbellcivil.com

 

[wettpaint]

Thanks for the input.

This is all the rain that fell on this particular day. There was an inlet issue and after further inspection after the event it appears the inlet was 50%-75% full of construction debris including 2X4s. There has not been an issue at this location in the past, this inlet has been in operation for about 6 years.

 

[goeggel]

Two information sources you might like:

Excellent IDF data:

http://hdsc.nws.noaa.gov/hdsc/pfds/

All available weather data (cost about $ 12/wk access) :

http://weather-warehouse.com/index.html

 

[NJHighway]

We had a similar issue. Our problem was finding accurate enough data to analyze. Our closest rain gauge to the site is 7 miles away in Philly. Recently we have been having smaller intense localized storms making that rain data useless. We had inlets and manholes surcharging and the gauge in Philly recorded only 0.20 inches. Unless your site is close to the station that you're getting your precipitation amounts from you may be analyzing useless data.

 

[elev8848]

Not an exact answer to the original question.
But I have put this website together for our particular IDF data for routine storm analysis.

http://www.hydrocalc.com/beta/rain_analysis/analyzer.html

It can be modified relatively easily for other IDF curve data.
Thanks

 

[Spartan5]

We had a similar issue. Our problem was finding accurate enough data to analyze. Our closest rain gauge to the site is 7 miles away in Philly. Recently we have been having smaller intense localized storms making that rain data useless. We had inlets and manholes surcharging and the gauge in Philly recorded only 0.20 inches. Unless your site is close to the station that you're getting your precipitation amounts from you may be analyzing useless data.

A great source for local precipitation data are Personal Weather Stations. I have one that I maintain with the data dumped to Weather Underground. Granted, you have to have things pass the smell test because of how people might have sited their equipment, as well as the quality of their equipment (Davis for instance is particularly good). But it's certainly another source of information. I also look at cumulative radar returns for storm totals.

Personal Weather Stations here:

http://www.wunderground.com/wundermap/

Note that not all have rain gauges. Click a station on the map and then click the station ID for detailed daily logs. Again, take the data with a grain of salt. But if you can find one you can trust, it can be very valuable.

 

[SMIAH]

As NJHighway pointed out, watch out using only 1 Precipitation gage.