You can call Isaias a “two-year flood." What does that even mean?

We were told El Nino would keep the storms at bay this year. Which was true, right up until last week, when Hurricane Isaias made history as the latest first-hurricane-of-the-season in over a century. Isaias quickly gathered strength over the warm Gulf waters, becoming a Category 3 cyclone before El Nino-induced wind shear weakened it just before landfall in the Florida Panhandle. Here’s an excellent summary of the atmospheric dynamics that both caused the hurricane drought and allowed Isaias to form.

In the hours before the arrival of Isaias’ remnants in WNC, we were whipsawed by rapidly changing forecasts. First, we were told to expect 3-5 inches of rain – more than we’ve had at once in over two years, but not terribly concerning. Then, Friday night, models began showing the likelihood of much higher totals: 8”-10”, even 12” in the highest southeast-facing mountains of the Blue Ridge. Since the South Toe drains peaks that are notorious for catching the brunt of tropical storm remnants, folks grew nervous. A completely natural response, considering the trauma still reeling through the community since Helene. 

In the end, Isaias delivered a middling amount – more than the initial forecasted rainfall, but less than the most dire predictions. Many people were awakened at 6 am Sunday by the jarring sound of emergency flood warning alerts coming out of their phones. The final belt of moisture was raking the mountains, and rainfall rates reached the dangerous threshold of one inch per hour. Fortunately, it only rained at this rate for about an hour. The storm passed. An hour and half later, the river crested at the streamgage at a vigorous-but-not-violent 4,800 cubic feet/second. There had been an earlier, slightly higher, crest of 5,700 cfs around midnight. To put those flow rates in perspective, the river reached 28,000 in 2004, 33,000 in 1977, and 71,000 cfs during Helene.  

Now, let’s talk about recurrence intervals. We’ll look at two different but related variables: river flow and rainfall. As for river flow, USGS uses the term Annual Exceedance Probability or AEP. What people call a 100-year flood has a 1% chance of occurring in any given year: a 1% AEP. South Toe has had three of these in the past 50 years (1977, 2004, 2024). Bad luck? Perhaps. More likely, the AEP is off due to the challenges of estimating a river that is flashy, uniquely vulnerable to severe storms, and doesn’t have a very deep dataset. Isaias’ peak flow of 5,700 cfs has a 50% AEP, meaning we might expect an inundation like Isaias every other year, so you could call it a 2-year flood, aka a nothingburger. Isaias fell far short of the threshold for “minor flooding,” when water begins to cover sections of low-lying primary roads. South Toe’s threshold for minor flooding is around 13,000 cfs – this is considered a 10-year flood, or one that has a 10% chance of occurring in any given year. In reality, the South Toe has exceeded the minor flood stage at least 14 times in the past 70 years. The highest of these are represented on our flood tree.  

As for rainfall, National Weather Service & NOAA use the term Precipitation Frequency Estimates: essentially how often should we expect x amount of rain to fall in y amount of time. For us, a good drenching is 4” in a day – according to NOAA’s precipitation frequency estimator (Atlas 14), we can expect a storm like this about every other year. You could call it a 2-year deluge: an event that, on average, recurs every two years. Or you can think of it this way: an event that has a 50% chance of occurring in any given year. During the record-breaking year of 2018, we got four of these 4”-in-24-hrs storms. In 2025, we didn’t get any. 

Isaias was a 24-hr event. My NWS COOP station, Celo 2S, received just a hair under 8 inches of rain between Saturday morning and Sunday morning. That’s a lot of rain in 24 hours. Eight inches in a day is estimated to be a 25-year event, or something that we might have a 4% chance of seeing in any given year. Interestingly, we have had four such “25-year” events in the past 25 years. Not a statistical impossibility, but enough to make you go, hmmmm.

Which brings us to the climate change wildcard. Given the limitations of our historical data, and given the complicated nature of atmospheric dynamics, recurrence intervals are difficult to estimate – even before accounting for climate change. NOAA has long been aware of the issues with their currently in-use Atlas 14 model. A new model, Atlas 15, has been in the works for several years, and is due for a preliminary release any time now. According to NOAA, “Atlas 15 represents a shift from a stationary assumption (i.e., extreme precipitation events do not change significantly over time) to a nonstationary assumption (i.e., extreme precipitation events change over time).” This means that climate change will finally be factored into the precipitation frequency estimates. The laws of physics dictate that a warmer atmosphere can carry more moisture, which is why we’ve seen a dramatic increase in severe storms lately. When Atlas 15 becomes available, we should have a better sense of what to expect in terms of the frequency of big storms.

Eight inches of rain is still a lot of rain! If 8” in 24 hours is a 25-year event, why didn’t it cause a 25-year, or even a 10-year flood? As I’ve written before, the South Toe is a headwaters river, which means every flood is a flash flood. The watershed is small enough that the river is more sensitive to rainfall rate than total rainfall. It could rain ½” per hour for days on end and the river might not even reach minor flood stage. If, however, it were to rain 1 ½” per hour for a few hours, we could suffer a catastrophic flood. In addition to the Atlas 15 effort, NOAA is also working on updating Probable Maximum Precipitation Estimates. These are absolute worst-case scenarios for any given watershed. It is likely that Helene was pretty close to the “probable maximum” for a 24-48 hour storm. But since we are a modest-sized watershed, we could have just as severe (or worse) flooding from a six-hour storm.  

It’ll take a while for all of us to recover from the traumatic events of September, 2024. Friday night, on the eve of Isaias’ predicted arrival, a worried friend asked for my take on the coming storm. I responded that the forecast was in line with what TS Fred gave us in August, 2021. I reminded her that the river crested just above minor flood level during Fred. She responded “It’s good to remember that there are normal floods too.” How true. The South Toe has a long history of frequent “normal” floods, as well as the occasional catastrophic one. 

The river was a foot higher on Sunday morning than it was reading on the USGS website. I messaged USGS with this picture and they immediately corrected the measurements. They sent someone out Sunday morning to check on it; the pipe running from the transducer into the river had broken loose during the crest in the middle of the night. By Wednesday, USGS had a crew out to set a new pipe.