These questions are worth going into some detail, so let me now discuss what it means that the data is changing. Data is considered to be stationary if it is random and if the characteristics from randomness, such as mean and variance, do not change over time. You can kind of think of this like a bus schedule. Sometimes the bus comes on time. Sometimes it is late. Sometimes, it is even early. However, if the schedule says the bus is coming at 10:15, that’s still the time you should be ready to get on it.
Nonstationary data changes exhibits changes in its randomness parameters over time. Nonstationary data changes exhibits changes in its randomness parameters over time. In the context of our bus schedule, if the bus company produces a new schedule adding a stop before ours, leading to a new time of 10:18, it is no longer a “stationary process” as the mean (expected time of arrival) has changed.
In the context of flooding, we know the data is not stationary and this has important implications. It is evident that as global climate change effects weather patterns, we are getting more severe storms. The latest data (and I don’t have a link for this right now), seem to suggest we are getting storms at the same frequency, but those storms are getting stronger, which kind of makes sense from an atmospheric dynamics perspective. But those stronger storms are delivering more rain and more water on the ground. This means that places that may not have flooded as frequently are flooding more frequently, and at higher levels. Either of those factors raises the 1 percent AEP level.
If the mean water level is rising, we are getting closer to water on the ground. If the mean water level, the base water level, is changing, that affects how close we are to water on the ground. If the mean water level is rising, we are getting closer to water on the ground. Especially along coastal areas, this has important implications since we are staring down a 2-meter rise in ocean levels over the next several decades. Such a rise will erase large portions of Maryland’s Eastern Shore, and Florida. Indeed, we are starting to see the first generation of climate change refugees as a result of this.
As bad as that is for coastal areas, inland areas are not protected. The variance, mean more severe floods when they do happen, goes up. This is driving more severe weather across the United States. We are seeing more severe storms, more rain, and more powerful tornadoes across the country. Climate change is leading to a double punch for everyone.
In the case of my analysis of Ellicott City, the effects are not a change in the frequency of 100-year events. It means the 100-year events are more intense, and so are those less than 100-year events. This is why these more intense storms have suddenly become “the new normal.” Worse than normal has become worse than before.
Image by Geert Schotanus / Wikimedia Commons.