Why flooding is so difficult for meteorologists to predict

Flooding during Hurricane Harvey in August 2017.
Flooding during Hurricane Harvey in August 2017.

HOUSTON – Meteorologists use a great number of tools to determine the timing and severity of storms. Every severe weather situation is different, too, so forecasters will customize each forecast analysis based on the situation at hand.

Breaking a storm forecast down into its simplest terms, there are two atmospheric ingredients we look for when forecasting severe weather:

  • Ample atmospheric moisture.
  • Strong instability, which causes rising air, or uplift in the atmosphere.

Without these two elements, storms don't form. Meteorologists need to determine the extent of moisture and instability in order to generate a forecast.

Common parameters we look at are: 

1. How much moisture is in the air? This is known as Precipitable Water and helps evaluate the potential for how much water can fall out of that air. Obviously, the more moisture that is there, the higher the predicted inch amounts can go.

2. What will lift that moist air up into the sky where it will cool, condense, form clouds and precipitate? The common methods are a front, whereby cooler air or a different wind direction causes the lift, or good old-fashioned convection -- our warm atmosphere supplies enough heat to lift the air mass high enough to begin cooling and raining.

3. Given that we live in a tropical atmosphere, the 'inch' amounts forecasted by the computer model can often be doubled, or even tripled, as to what actually falls. Simply put, warm air expands and the farther apart the warm, air molecules go, the more room there is to put in water molecules. So with more water in the air mass, when it does unload, we can see huge amounts of rain over what was predicted. 

Here are some of the common tools we use:

  • Satellite and radar imagery.
  • Numerical weather prediction models.
  • Skew-T diagrams, which provide a graphical vertical analysis atmospheric moisture, temperature and wind.
  • Current conditions weather analysis.
  • Analysis and insight from the National Weather Service, Storm Prediction Center, Weather Prediction Center, National Hurricane Center and other NOAA agencies.
  • Historical data, including a forecaster’s own experience dealing with previous heavy rain or extreme weather events.
  • Intuition.

We glean different information from different sources, but it all goes to determine just how juicy and just how unstable the atmosphere will become during the forecast period.

There is a degree of uncertainty in every forecast we make. In some cases, all the data we study points to the same solution. In these cases, there is a high degree of confidence in our forecast. In some cases, the data all point in different directions. That's when we have to make a best judgement call on a forecast. In situations like that, a forecaster's experience and intuition play a key role.

In the end, we come up with a forecast that predicts timing, placement and severity of storms. But the job doesn't end there. New data is always coming in, making forecasting a continual process. Satellite and radar imagery updates constantly. There is always a new forecast model output to review. New current conditions are published. New forecast discussions are issued. And each piece of new information can change a forecaster's projection -- a little, or a lot.

Tuesday's heavy rain was a good example -- plenty of moisture in the very warm air. We've been in the 80s every afternoon and only around 70 at night. Upper-level winds moved air so that lower level winds could replace it. Think of how you can sweep the water away at the top of your pool, but the water below quickly comes up to replace it. That replacement was the lifting mechanism. But the overall surface winds were light and not enough to move the storms quickly. As a result, they built up, collapsed and by collapsing actually caused a mini-cold front behind them. 

This mini-cold front, or gust front, became ANOTHER lifting mechanism causing MORE lifting of the moist air and, therefore, more rain. This is a training effect and while it's a small train, we still had boxcar after boxcar over the same area. At that point, we can only look at how much rain is falling and predict how much more will fall in the immediate future. 

Around these parts, when it will rain is easier than how much it will rain. To that end, predicting how much flooding will occur because of those rains becomes very difficult due to the urban structures, highways, bayous, creeks and overall flat geography.

Forecasting at KPRC2 is truly a collaborative effort. We collaborate with each other continually during the forecasting process, weighing all the information at our fingertips, to come up with the best forecast for our viewers on a daily, even hourly, basis.

About the Authors: