Meteorology of Saturday's Colombian Flood Disaster That Killed 254

At least 254 people were killed in the in the city of Mocoa (population 40,000) in southwest Colombia near the border of Ecuador early Saturday, when torrential rains triggered a debris flow on a nearby mountain that surged into the town as a huge wall of water carrying tons of mud and debris. The disaster is the fourth deadliest weather-related disaster in Colombia’s recorded history. Reports from Colombia indicate that 130 mm (5.1”) of rain fell during a short period on Friday night and early Saturday morning, with the heaviest of the rain falling in just two hours, between 23:00 Friday, 31 March and 01:00 Saturday, 01 April, 2017. The rains fell on soils that were already wet from unusually heavy rains during March; the Mocoa region received about 50 percent more precipitation than usual during the month of March. The heavy rains of Saturday morning triggered a debris flow down the Taruca ravine on the northwest side of Mocoa, and this landslide, accompanied by floodwaters, poured into the Sangoyaco River and rampaged through the city of Mocoa. According to a USA Today interview with Jonathan Godt, coordinator of the U.S. Geological Survey’s landslide hazards program, “That mixture can move at 35-40 miles an hour, and because it’s so dense it has a lot more momentum and destructive power than water alone.”


Figure 1. Mocao, Colombia after the Saturday, April 1, 2017 flood.  A video from early Saturday morning showed cars and trucks being swept down the streets. Image credit:  Ejército Nacional de Colombia via Facebook.

Contributing causes to the disaster
The fundamental cause of the disaster was that the city of Mocoa was situated in a vulnerable location—in a valley surrounded by steep slopes, close to the Mocoa, Mulato, and Sancoyaco rivers. Deforestation on the surrounding slopes may have contributed to the landslide and flood. President Juan Manuel Santos blamed climate change for triggering the flood, and he has a point—increased evaporation from warming oceans have caused a significant rise in atmospheric water vapor and very heavy rainfall events like the Mocoa event in recent decades. The Mocoa rains were triggered by a very moist flow of air from the tropical Atlantic, where ocean temperatures were near average (see the meteorological analysis below.) The rainy season in Colombia extends from March to mid June, so additional floods and landslides can be expected the next two months.


Figure 2. Aerial view of Mocao, Colombia after the Saturday, April 1, 2017 flood.  Image credit: Cesar Carrion/Colombian Presidential Office via AP.

Colombia’s history of weather disasters
According to EM-DAT, the international disaster database, Colombia’s most expensive and second deadliest weather-related disaster occurred in 2010 - 2011, when almost non-stop heavy rains caused three separate billion-dollar flooding events, killing 418 people. EM-DAT lists one other flood that killed more than 200 people in Colombia: a December 1971 flood in Magdalena and Cauca Valsfive that killed 307. EM-DAT also lists five landslides in Colombia’s history that have killed at least 200 people:

640 killed in Villatina on Sep 27, 1987
300 killed in Quebrada Blanca on June 28, 1974
200 killed in Bogota on June 28, 1973
200 killed on June 21, 1986
200 killed in December 1971

Meteorology of the disaster
Wunderblogger Lee Grenci looked in detail at the meteorology of the disaster, and what follows is his analysis.

Below is the animation of enhanced satellite images from GOES-13 during the period, 0045 UTC to 0715 UTC April 1 (nighttime in Colombia). You can see a mesoscale convective system (MCS) . This may more properly have been called a mesoscale convective complex (MCC), given the symmetry of the cirrus canopy toward the end of the loop--but the duration of the large, circular cirrus canopy did not last very long, however, and must last at least 6 hours to be labeled a MCC. Needless to say, the MCS was slow-moving, if not stationary, for several hours. The cloud-top temperatures at the peak of deep, moist convection were as low as -90 degrees Celsius, indicative of heavy thunderstorms quite capable of producing very heavy rain.


Figure 3. The animation of enhanced infrared images from GOES-13 during the period, 0045 UTC to 0715 UTC on April 1, 2017. Courtesy of Penn State.


Figure 4. A relief map of Colombia, Courtesy of Wikipedia. The disaster occurred in the town of Mocoa, which lies in southwest Colombia in the department of Putumayo at an elevation of nearly 2000 feet.

Given the quasi-stationary nature of the MCS, it’s pretty clear that the initiation of heavy thunderstorms was orographic in nature. Check out (below) the GFS model analysis of 700-mb streamlines (about 10,000 feet) at 00 UTC on April 2. Please note that I’m using 700 mb as a proxy for the lower troposphere, given the high elevations in this region (mountains to the west of Mocoa are as high as 15,000 feet). The analysis shows a strong easterly flow from the tropical Atlantic that moved upslope over the high terrain, paving the way for heavy thunderstorms. The rapid, extreme run-off from the lofty mountains set the stage for devastating flooding and mudslides.


Figure 5. The GFS model analysis of 700-mb streamlines at 00 UTC on April 2, 2017. Courtesy of Penn State.


Figure 6. The GFS model analysis of 850-mb streamlines at 00 UTC on April 2, 2017. Courtesy of Penn State. Lower-level winds come into South America from the tropical Atlantic and Caribbean.

Moisture from the Caribbean, where ocean temperatures were up to 1°C (1.8°F) above average might have played a role in the MCS, but I believe the mountains to the north probably blocked some of this low-level moisture. To support my claim, take a look at the GFS analysis of precipitable water (PWAT, Figure 7). In my opinion, the broad east-to-west band of high PWAT, with an embedded area of 2.5+ inches, was aimed (given the prevailing easterly flow) directly at the location of the MCS.


Figure 7. The GFS model analysis of precipitable water (PWAT expressed in inches) at 00 UTC on April 2, 2017. Larger image. Courtesy of Penn State.

Final blog post in old software system
This post is scheduled to be the final one in the old software system; the new “Category 6” will use the Drupal software package, and commenting will be via the Disqus software package. The current plan is to leave the current blog archive up for two more weeks, after which time a full archive of all the old blogs will be made available, along with the first 500 comments in each post (no graphics or videos posted to the comments will be saved, though.) WU member Patrap figured out that the Internet Archive's “Wayback Machine" has backups of the blogs; for example, one can access old posts by changing the “entrynum=“ portion of the URL below:

https://web.archive.org/web/http://%20/www.wunderground.com/blog/JeffMasters/comment.htm%20l?entrynum=2835

Our FAQ page has more detailed info on how this transition will occur. Blog commenting is not going away! Check back later on Monday for the first "new" Cat 6 post. 

Jeff Masters and Lee Grenci