Today’s post is part 3 of a series on Operational Context (View the other posts in the series here). In this third series, we will be looking at the same questions that we’ve been looking at throughout this series. Today’s post will be looking at rainfall amounts, and is tied very closely to the last post on drought. Essentially, many of the same datasets for extreme rainfall and drought can be used with one another, it’s just that the rainfall data is looking for an absence of rainfall when you’re looking at drought. This will also tie into the next post on Operational Context – Flooding which will be looking at the extreme rainfall events and their impact as the water runs off. So, let’s go ahead and look at today’s theme – extreme rainfall.
There are several ways to measure or estimate rainfall. Most people know about rain gages and using them to measure rainfall (see examples in post on Innovative Ways to Teach the 3 M’s – Math, Maps, Measurement). Rain gages are great for measuring what fell in a specific location, but there will always be gaps in coverage. Rainfall can be enhanced or reduced by terrain or other geographic features, so point data by itself isn’t enough. There are rain gages at major airports, and there are companies with portable weather stations where rainfall can be measured. Additionally, did you know that you can participate in rainfall measurement directly by participating in CoCoRaHS (stands for Community Collaborative Rain, Hail, & Snow Network)? CoCoRaHS takes the individual reports from trained observers who measure rainfall once a day. These measurements help to verify rainfall amounts throughout the country.
We also have access to NWS Weather Radar and several derived products. One of these products is radar indicated precipitation. Radar indicated rainfall amounts are helpful because they can estimate rainfall even in places where direct gage measurements aren’t available. Additionally, when compared against measured rainfall from gages, the level of accuracy can be assessed.
Each radar site can only estimate rainfall out to ~100 miles. There are more than 100 radar sites in the United States, and these rainfall estimates can be merged together to estimate rainfall amounts throughout the entire country. Looking at the next graphic, this shows a view of the data from all of the radar stations merged across the entire country, and then compared to a “normal” distribution of rainfall – normal being an average over a specific period of time.
In this example below from http://water.weather.gov/precip/, you can clearly see the areas where we have seen excessive rainfall as well as drought. Much of the Ohio River Valley is well above normal and many of the areas hard-hit by Irene and Lee flooding are well above normal for the year. On the flip side, many parts of Texas and the Deep South are well below normal.
Finally, I want to share with you about the NOAA Precipitation Atlas available from http://hdsc.nws.noaa.gov/hdsc/pfds/. This atlas provides an opportunity to identify approximate “return period” for extreme rainfall amounts throughout the country. There are GIS files available for these maps, but there is also a very easy to use Google Maps interface where you can point and click on a part of the country, and the resulting table will give you rainfall periods from 5 minutes to 60 days and recurrence intervals from 1 year to 1000 years.
How do we pull all this information together for operational context? You can go look at the National Weather Service Office from Sterling, Virginia and some of their products from the remnants of Tropical Storm Lee this past September (see full files from http://www.erh.noaa.gov/lwx/irene_lee_rainfall.php)
ON WEDNESDAY...SEPTEMBER 7TH...THE IFLOWS GAUGE IN BOWIE RECORDED 4.57 INCHES IN THREE HOURS...WHICH HAS APPROXIMATELY A 0.5 PERCENT CHANCE OF OCCURRING IN ANY GIVEN YEAR /SOMETIMES CALLED A 200-YEAR RAINFALL/. FOR ELLICOTT CITY...WHERE SEVERE FLASH FLOODING OCCURRED...THE RAINS WERE ACTUALLY LESS RARE...ROUGHLY A 1 TO 2 PERCENT CHANCE OF OCCURRING IN ANY GIVEN YEAR. THE HEAVY RAINS ON THURSDAY...SEPTEMBER 8TH...WERE EVEN MORE RARE. THE KINGSTOWNE IFLOWS GAUGE NEAR FRANCONIA IN FAIRFAX COUNTY RECORDED 5.47 INCHES IN JUST THREE HOURS...WHICH HAS APPROXIMATELY A 0.2 PERCENT CHANCE OF OCCURRING IN ANY GIVEN YEAR /SOMETIMES CALLED A 500-YEAR RAINFALL/. THE SAME IS TRUE FOR RAINS IN RESTON...WHERE 6.57 INCHES WAS RECORDED IN SIX HOURS. AT FORT BELVOIR...THE AUTOMATED WEATHER STATION THERE RECORDED AN INCREDIBLE 7.03 INCHES IN JUST THREE HOURS. THIS HAS LESS THAN A 0.1 PERCENT CHANCE OF OCCURRING IN ANY GIVEN YEAR /SOMETIMES CALLED A 1000-YEAR RAINFALL/.
- The rainfall amounts measured at the gages had a starting and ending time.
- Those values could be compared to the NOAA Precipitation Atlas for those specific locations
- Approximate recurrence interval values can be created to help provide operational context for how much rain fell.
This type of context helps to really get your arms around how significant of an event this really was, but it does rely upon quality rain gage information as well as connecting the dots together to answer the questions of context.
Next Post: Operational Context – Flooding