How Bad Was That Storm?
A week of rain has ended. Flood crests are passing. And many are asking, “How bad was that storm?”
The short answer to the question: depending on where you live in the region, you got a 1- to 5-year rainfall.
Why is that? And how do you determine it? It’s simple. Let’s start with the rainfall.
Step One: Determine the Amount of Rainfall You Got
If you don’t have a rain gauge, go to the Harris County Flood Warning System. Elsewhere in Texas, you can go to the Texas Water Development Board’s Mesonet.
For demonstration purposes, let’s focus on the Harris County Flood Warning System.
By default, the home page shows a map with the locations of gages with rainfall in the last 24 hours. But you can also select other time periods in the sidebar. The storm dropped water during most of the week. So I selected “7 Days.” I also selected “All Gages” to see the varying amounts of rainfall across the region.
Then I clicked “Watersheds,” and “Channels.” Automagically, 7-day rainfall totals appear over a map that lets you see which watersheds the rain fell in. That determines how it will work its way downstream to the Gulf.
On the low side, values ranged from 4″ to 6″ south of Lake Houston. On the high side, they ranged from 9″ to almost 11″ north and west of Lake Conroe. Such variation is common.
Step Two: Find the Duration and Distribution of the Rainfall
When you click on any gage location, a “For more information” box pops up. Click the link to see the distribution and duration of rainfall.
The gage at the San Jacinto West Fork and US59 received 5″ during five days. If a giant peak on one day outweighed all others, you might want to investigate that particular day further. But in this case, most days were within a half inch of each other.
Now, you’re ready to find how that compares to other storms.
Step Three: Compare Recorded Totals to Precipitation Frequency Estimates
Next, compare recorded rainfall to expected rainfalls of different intensities and durations.
Most hydrologists currently use precipitation frequency estimates called “Atlas 14.” NOAA determines them.
To find the estimates for your area, enter your address here. You should see a table like the one below although your numbers may vary slightly depending on where you live.
This is where some judgment comes in.
So, I started by looking across the seven-day row and highlighted the first box. It showed 5.76 inches. The smaller numbers in parentheses indicate possible variation due to uncertainty. Almost all the lower numbers fell within this range.
At the high end of the observed rainfall totals, I highlighted the 10.4 inch box as the most representative. Again, all of the observed totals north and west of Lake Conroe fell into the range in parentheses.
Looking up to the top of the table, you can see that the highlighted boxes correspond to the volume of rain you could expect once every year to five-years. Meteorologists also refer to these as “100% and 20% annual-chance” storms.
If you live elsewhere on the map, you might find your area received a 50% annual-chance storm.
Rainfall Chances Do Not Automatically Translate into Flood Chances
For the record, the flood at the West Fork and US59 this morning peaked at 53.71 feet. That’s three feet LESS than a 10-year flood at this location.
So, in this case, the flood level was generally consistent with a 5-year rainfall upstream. But that’s not always the case.
Many people assume that a rainfall recurrence interval of 1- to 5 years automatically translates into the same probabilities for flooding. It doesn’t.
First, for large watersheds, such as the San Jacinto, rain can vary drastically. Variation upstream will determine how high the resulting water surface elevations are at various points downstream when peaks arrive.
Also understand that annual exceedance probabilities for floods incorporate many more variables than rainfall probabilities. For instance:
- Landscape/Slope – Is it flat or hilly? Will water be bottlenecked or does it have room to spread out?
- Degree of development – Are you surrounded by farms or do you live in an urban area which produces more runoff faster?
- Soil type – More water infiltrates into sandy soils than clay.
Regardless, engineers still consider rainfall probabilities.
How Engineers Use Rainfall Estimates
Why are rainfall estimates important? Engineers must design drainage and infrastructure to handle extreme rainfalls.
For instance, most storm sewers are designed to handle the type of rain you can expect every year or two. When you see water ponding in streets or underpasses, it’s because the storm sewers can’t carry the water away fast enough.
Houston’s recently upgraded infrastructure design standards aim to keep structures safe in a 100-year event. It matters not whether the rain happens in five minutes or 60-days.
When you see flooding of streets or neighborhoods, it’s generally a sign that:
- Rainfall exceeded the design standard
- Something changed, for instance, a sewer was blocked, sediment has clogged a drainage ditch, etc.
- Someone miscalculated or cut corners during design and construction
- Infrastructure was designed to old (lower) rainfall probability standards.
Experts base the probability of extreme future events on the frequency of extremely rare past events using a branch of mathematics called extreme value analysis.
It’s important to understand that rainfall probability estimates change periodically – especially after major storms, such as Tropical Storm Allison or Hurricane Harvey. Meteorologists acquire additional data on extreme storms from these events.
All rainfall probability estimates represent best guesses given knowledge at a point in time.
NOAA is already working on Atlas-15 estimates. Atlas 15 will take climate change estimates into account for the first time.
Going forward, NOAA will compile new precipitation-frequency estimates every 10-years.
Posted by Bob Rehak on 1/27/24
2342 Days since Hurricane Harvey