Death by Average: Why “100-Year” Floods and Landslides Keep Happening, and Average Risk Models Are Failing Us

Estimated reading time: 4 minutes

If you feel like “100-year” floods are happening every few years, you’re not imagining things. The problem isn’t just the climate—it’s the math.

For decades, engineers and emergency managers have relied on recurrence intervals like the “100-year flood” to describe the likelihood of natural disasters. But as extreme weather events grow more frequent and destructive, those averages are failing to reflect reality—and they’re leaving infrastructure operators exposed.

Understanding why these estimates fall short is key to understanding the need for more dynamic, data-driven approaches to hazard modeling.

---

What Does “100-Year Flood” Actually Mean?

Despite how it sounds, a “100-year” flood does not mean a flood that happens once every 100 years. It means there’s a 1% chance of it happening in any given year.

Statistically, that’s true—but in practice, these events can (and do) happen back-to-back. For example:

• If you’ve experienced a “100-year” event recently, you’re not safer the next year. The odds reset annually.
• Certain regions and communities will encounter 100-year floods annually by the end of the century.

These probabilities are based on historical precipitation records and runoff patterns. But in a changing climate, history may no longer be a reliable guide.

---

Climate Variability Is Breaking the Models

Recurrence intervals assume that weather patterns are stationary—that past rainfall patterns are consistent and predictable over time. But that’s no longer the case.

Recent studies have shown that:

• The frequency of extreme precipitation events has increased significantly across much of the U.S., especially in the Northeast and Midwest.
  
• “Once-in-a-century” floods now occur with increasing regularity in regions that never expected them.
  
• Federal weather models (like NOAA Atlas 14, FEMA flood maps) which underpin many flood models, are decades out of date in some areas and don’t account for future climate trajectories.

As a result, many floodplain maps and recurrence-based models underestimate the true risk—particularly for infrastructure built before the last two decades of observed climate change.

---

The Limits of Averaged Risk

Beyond climate change, the way risk is averaged can obscure the very dynamics that lead to disaster. For example:

• A pipeline that crosses multiple watersheds may face vastly different flood or landslide risk profiles depending on local terrain, soil saturation, and upstream accumulation.
  
• Averaging risk across a basin may downplay localized hotspots where converging slopes, narrow channels, or recent precipitation elevate threat levels.

This is particularly dangerous when combined with static inspection schedules or field prioritization methods that don’t respond to weather conditions or terrain dynamics.

---

Why Probabilistic, Real-Time Models Are the Way Forward

Instead of relying on historical recurrence alone, modern hazard models integrate:

• Current and forecasted precipitation to model hydrologic loading in real time
  
• High-resolution terrain and soil data to determine how water will move and where it’s likely to cause failure
  
• Event-based probability distributions, which calculate not just the likelihood of occurrence, but the likelihood of consequence at a specific location

These models don’t aim to predict the impossible—they aim to quantify uncertainty better, and to update that understanding as new data comes in.

---

Terevue: Modeling Flood Risk in Real Time

One such system is Terevue, an environmental intelligence platform which uses high-resolution terrain, climate, and precipitation data to monitor rainfall-triggered landslide and flood risk in near real time. Rather than relying on recurrence intervals alone, it models how changing weather interacts with vulnerable terrain—helping operators stay ahead of the next so-called “100-year” event.

---

Final Thoughts:

The language of averages made sense in a more predictable world. But today, operators need more than probability—they need precision. Because what used to be rare is now routine, and the cost of being caught off guard is growing.