Right after "how many hurricanes will there be?", the question I'm most asked this time of year is "why can't you guys forecast rain?" It's a good question for which there really isn't a very good answer. This week is an excellent example of the problem. In order to get rain in the summertime, we need to have a) very humid air in place, b) the right conditions in the upper atmosphere and c) a trigger to allow warm, humid air to rise rapidly and produce thunderstorms. It's generally very humid in Houston in the summer, but the other two elements may be lacking.
Early on, we saw indications that a weak upper-level low pressure system would move our way from west Texas. Upper-level lows are basically small swirls in the atmosphere that are good for lifting the surface humid air. On Wednesday, this feature drifted just west of Dallas and was forecast to move toward Houston Thursday and Friday. The key word is "forecast". These systems are very difficult to forecast because they are meandering in the upper atmosphere and are usually cut off from faster jet stream winds. So, if the forecast of their location is off, the forecast of where storms will form will also be off.
Additionally, we don't have really good information about the location of moisture and wind throughout the entire atmosphere over the Houston area. Twice daily, weather balloons are launched in Corpus Christ and Lake Charles. Houston sits roughly in the middle of those two locations so computer models estimate the upper air conditions based on the other two sites. This means that estimates of the atmosphere's ability to allow storms to form in Houston isn't based on real data. Its a wonder we ever get the forecast for late-day storms in the summer right at all.
Usually, what I do is look at several stability parameters that assess the ability for air to rise when heated. Sometimes, though, even if the parameters indicate that storms will form, nothing happens. Other times, none of the conditions that produce storms are present and there's a big eruption. Even then, storms form in localized areas so it might pour in one part of town, but not the other. You've probably even seen where it's pouring rain on one side of the street and dry on the other. There's no model detailed enough to depict that scenario.
The bottom line is that we based a forecast for the next day's storm chance on the information we have at the time. The better the information, the better the forecast.