Thursday, September 12, 2013

Front Range Flooding - CoCoRaHS Deja Vu

500 millibar map for 6:00 a.m. MDT September 12
An upper level low over the southern Great Basin has been funneling moisture northward along the
Rocky Mountains, producing a conveyor belt of thunderstorms that have dropped up to a foot of rain on some locations in the past few days . The most recent event is the widespread and destructive flash flooding that occurred in the foothills from Boulder to Fort Collins, CO Wednesday night and Thursday.   24-hour rainfall amounts in Boulder County this morning exceeded eight inches, with a relatively narrow band of 7.00 to more than 8.00 inches of rain through the city of Boulder.

24 hour rainfall ending at 0700 MDT September 12

The rain gauge at Centennial Middle School in Boulder, CO (CO-BO-337) this morning with 8.43 inches of rain.

The radar-estimated precipitation map for the period from September 9 through 10:00 a.m. MDT September 12 (just a little less than three days) shows the two bands of heavy rainfall, one northwest of Denver and the other just east.

Radar estimated precipitation from the Denver NWS radar
Normally trickling rivers at this time of year became raging torrents, and the flash flooding washed out many roads, flooding many communities and isolating others. At least three people have lost their lives in the flooding.  Emergency personnel conducted numerous water rescues. The main business district in Estes Park, CO, the gateway to Rocky Mountain National Park, was wall to wall water this morning. More rain fell during the day Thursday and is expected to continue into Friday.

The flooding the last two days is much like the heavy rainfall event that is behind the formation of the CoCoRaHS program. on July 28, 1997 a major flash flood hit Fort Collins causing devastation in excess of $140 million and five fatalities.

It was the attempt to document and explain this event that is behind the origins of CoCoRaHS. Rainfall data was available from only a few "official' rain gauges. In this situation rainfall varied from 2 inches to more than ten inches over a distance of about three miles, and none of the official gauges captured the rainfall maximum.  You can read more about the rainfall for this event at this link:  An Analysis of Rainfall for the July 28, 1997 Flood in FortCollins, Colorado.

An excellent source of information on rainfall-induced floods that have caused damage in the Front Rain since 1953, including the 1997 Fort Collins flood, can be found on the Colorado Front Range Historical Flood Summaries web site provided by the National Center for Atmospheric Research. There is little doubt that this week's flooding will make its way on to the list.


  1. Based on rainfall frequency stats from NOAA Atlas 14, the three day rainfall through yesterday morning of 10+ inches reported by some CoCoRaHS observers was already more rare than a 1000 year storm. And it was still raining at the time. Unbelievable!

  2. Pam, can you let me know how that was calculated? CoCoRAHS records aren't that long, but can you still extrapolate to 1000-year-storm frequency confidently?

  3. Pretty incredible. The photo of the rain gage amazes me... having one of those gages in Utah (Salt Lake City area), I have yet to see the water level just in the INNER cylinder get that high. Hope you guys can soon get a break!

  4. I'm CoCoRAHS CO-BO-33 and in four days, I've recorded 18.22" rain. I've topped the inner cylinder four times now - much too much. We need it to stop and get some sun! I don't ever want to record more rain than the numbers I've gotten.

  5. This comment has been removed by the author.

  6. Greg, I used to work for the NOAA office that produces the extreme statistics. There are specialized statistical formulas that help determine extreme values when you have short record lengths, such as the Weibull and Gumbel distributions. However, there is always going to be some uncertainty in the values, so now if you look at the NOAA Atlas 14 publication online, it will give a range of values for the 1000 year storm. All of these assume the climate is stationary, however, and that is not necessarily a good assumption now. But you have to start somewhere.

  7. I should add that the NOAA Atlas 14 values are not based on CoCoRaHS data but NWS cooperative data plus whatever additional rainfall records that they can find from mesonets, etc. that they have a good degree of confidence in. Using just five or ten years of CoCoRaHS data is not going to give you a good estimate for a 1000 year storm, but using 100 years of coop data will do much better.