The new normal?

Louisiana isn’t the hilliest of the 50 states in the Union, nor is it particularly dry. In August this year its topography helped to make a bad situation worse when the skies opened and the rain fell.

From August 12 to August 14 an estimated 61cm of rain fell onto parts of Louisiana, overwhelming flood defences, creeping into homes and businesses, washing away cars and damaging roads, culverts and bridges. Five people died and 30,000 inhabitants had to be evacuated.

Governor John Bel Edwards of Louisiana claimed in a letter to President Barack Obama that the total economic cost of the floods could be as high as $15 billion, but this figure is still an initial estimate. The total amount of insured damage is likely to be a lot lower, due to the relatively low level of insurance take-up in parts of the State.

Why did this event take an area that can hardly be described as arid and unfamiliar with floods by surprise? What made it so different from other storms?

“We’ve been seeing very heavy rainfall events every now and then, but this event was unique,” says Hemant Chowdhary, principal scientist at AIR Worldwide. “It’s evident that such heavy rainfall was not observed in the past, in the historic records of under 200 years. At several locations that rainfall amounted to a 1,000-year return period event—that’s a 0.1 percent probability of its occurring. For everyone living in those areas that’s unprecedented.

“Second, the extent of the area affected by this rainfall was quite big—an area of 80km by 160km, oriented in a south-west to north-east direction, a relatively large area to be hit by such heavy rainfall.”

According to Chowdhary, as a consequence of all this rainfall, several relatively small rivers in the area had a huge amount of run-off from their watersheds and high flows in the rivers at the same time. As a result there were many unfavourable processes happening concurrently. In a sense this was not very uncommon, but all the possible unfavourable conditions that could happen were happening at the same time during this event in a small watershed.

For example, riverine on-plain flooding was happening, the back water was pushing the water from the smaller tributaries because of the high flood stages in the main rivers and at the same time a significant inundation was happening because of high local flooding, which is called off-plain flooding, caused by intense rainfall.

“This area has relatively flat terrain, and the event was also quite lengthy, spanning two to three days. As a result the soil was saturated, there was limited drainage capacity after that much rainfall and there was the back water effect that I mentioned earlier. So all the possible unfavourable conditions that could have happened, did happen at the same time, which made it worse,” says Chowdhary.

Unusual effect

Jeff Waters, product manager and meteorologist at RMS, agrees that the flood was not a normal event. “A few things contributed to this event being as unique as it was,” he says.

“First, from a meteorological perspective, a slow-moving low pressure system moved uncommonly westward before stalling over the Southern Gulf of Mexico region, including Louisiana. During that time, it pulled record amounts of moisture onshore from the warm Gulf of Mexico waters, creating favourable conditions for excessive and prolonged rainfall.

“The way that the storm system moved was also unique. Generally speaking, most storm systems in the mid-latitudes tend to move in an eastward manner due to the prevailing winds in the upper atmosphere. The fact that this storm system moved westward before stalling over Louisiana was rather unusual.”

According to Waters the disturbance essentially created a conveyor belt of torrential rain that dumped more than 25cm of precipitation throughout much of Louisiana—in some extreme cases more than 75cm—resulting in widespread major and minor riverine flooding.

The local conditions also contributed to the uniqueness of these recent floods, he says. Louisiana was experiencing one of the top 20 wettest years on average leading up to this event. The already heavily saturated soil conditions inhibited water from being absorbed effectively, which exacerbated the amount and severity of flooding that took place.

The art of modelling

The uniqueness of the event raises an important question. Meteorology—not to mention the science of creating flood modelling maps and other tools that can be used by insurers—is about predicting future flooding events. If something is unique then it’s obviously not going to fall in the usual parameters for future predictions. That does not mean however that it can’t be predicted.

Chowdhary says that low probability events like the Louisiana floods will always occur at some point in the lifetime of the market, but that the low likelihood of them occurring means that people will always be alarmed when they actually happen.

“The frequency of these events these days, such as in the last 10 to 15 years, is relatively higher. For example there have been three significant flooding events in Texas and Louisiana in the last six months, each of them a billion dollar event. Considering these types of big events with this kind of frequency it’s not hard to imagine that there will be more such events happening relatively frequently,” he says.

“We hear about climate change and global warming more when we see these frequencies and intensities, but it is very difficult to label these phenomena as possible reasons behind any single occurrence.”

According to Chowdhary, as far as modelling is concerned, it’s not very difficult to model these events in the forecasting scenarios. For example in this event, the Louisiana floods, there was enough information two to three days in advance from the National Weather Service that such a big event was likely to happen, for institutions to have a good idea of the likelihood of what was going to come in the near future.

“These types of events have a huge impact on flood insurance, on the take-up rates and greater coverage of the insurance,” says Chowdhary.

“Such events are very grim—and frequent—reminders these days that damaging floods can occur outside of the usual areas that see floods, the one-in-a-100-year event zones, where the Federal Emergency Management Agency (FEMA) requires you to have flood insurance. It makes it clear that local or off-plain flooding in moderate or low-flood risk areas can also cause disastrous floods, so a lot of individuals or companies will now look at flood insurance very seriously.”

There is another issue involved, a more problematic one. Unusual as this event was, it hasn’t been the only one of its kind—in September 2013 a slow moving storm stalled over parts of Colorado and caused widespread flooding. There have been other events as well. Are they connected to a wider possible cause such as global climate change?

“A lot of discussion has arisen with regard to climate change and to what extent global warming is contributing to these types of flood events,” says Waters.

“It’s difficult to tie any one event to climate change. However, the Louisiana floods and similar flood events that have taken place elsewhere over the past few years are an example of what some scientists believe could happen as a result of increased greenhouse gases—an increase in frequency and intensity of extreme precipitation events. It’s something to keep in mind as we look ahead to the future.”

There is another possible reason that recent floods have been so bad in the US. According to Waters, exposure trends are also contributing to the severity of recent flood events throughout the country. A combination of population growth and urbanisation, particularly along the coast, is leading to an increase in water-impermeable surfaces, such as streets, parking lots, structures themselves. As a result, more water tends to stay on or above ground during precipitation or storm surge events, subsequently amplifying the severity of the flooding and inundation.

The challenge for flood modellers now is to deal with such factors. Are these changes predictable? More flood events might need to happen to make the full ramifications visible for reinsurers.