Effects of correlated collisions and intermittency on… — Plain-Language Explanation

✨

This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The Big Problem: The "Size Gap" in Rain

Imagine you are watching a warm cloud. Inside, there are billions of tiny water droplets. Some are very small (like dust motes), and some are big enough to fall as rain.

But there is a problem: The Size Gap.

Tiny droplets (under 15 micrometers) form easily from water vapor, but they are too light to fall. They just float.
Big droplets (over 50 micrometers) fall fast because gravity pulls them down.
The Gap: How do the tiny droplets grow big enough to become rain?

Theoretically, if tiny droplets just bumped into each other randomly, it would take hours for them to grow big enough to fall. But in reality, rain starts in about 30 minutes. Something is speeding up the process.

The "Lucky Droplet" Theory

Scientists realized that you don't need every droplet to grow fast. You just need a few statistical outliers—droplets that get incredibly lucky. Let's call them "Lucky Droplets."

If even a tiny fraction of droplets can grow super-fast and cross the "Size Gap," they will start falling, drag the rest of the cloud down with them, and create rain. The question is: What makes a droplet "lucky"?

The Two New Clues: "Memory" and "Stormy Weather"

The authors of this paper investigated two specific things about turbulence (the chaotic, swirling air inside clouds) that might help these Lucky Droplets grow faster.

1. Correlated Collisions (The "Memory" Effect)

The Old View: Imagine a droplet is walking through a crowd. In the old model, every time it bumps into someone, it's a totally random, independent event. Like flipping a coin. Heads or tails, it doesn't matter what happened last time.

The New View: The authors found that in a swirling cloud, droplets have a bit of a "memory."

The Analogy: Imagine you are in a mosh pit. If you get bumped once, you are likely to get bumped again immediately because you are in a chaotic, swirling cluster of people. You aren't just bumping into random strangers; you are stuck in a "bump zone."
The Result: If a droplet has just collided, it is more likely to collide again very soon. This "memory" helps droplets grow a little faster at the very beginning.
The Catch: The authors found this is a nice bonus, but it's not the main hero. It speeds things up a little bit, but not enough to explain the 30-minute rain onset on its own.

2. Intermittency (The "Stormy Weather" Effect)

The Old View: Scientists used to assume the cloud was like a calm, uniform soup. The "turbulence" (the churning energy) was the same everywhere and every second.

The New View: Clouds are not uniform soups; they are like patchy, stormy weather.

The Analogy: Imagine the cloud is a giant room. Most of the room is calm, but there are tiny, invisible pockets where the air is spinning violently like a tornado.
The Mechanism: When a "Lucky Droplet" happens to drift into one of these high-energy pockets, the collisions happen much faster. It's like the droplet suddenly enters a high-speed conveyor belt.
The Fluctuation: These pockets come and go. A droplet might spend a few seconds in a calm zone, then get swept into a violent spin, grow huge, and then move on.

The "Toy Model" Experiment

To prove this, the scientists built a computer simulation (a "toy model").

They created thousands of virtual "cloud parcels" (tiny boxes of air).
In some boxes, the air was calm. In others, it was violently spinning (high dissipation).
They watched the droplets grow.

The Result:
When they assumed the cloud was uniform (calm everywhere), it took a long time for the "Lucky Droplets" to grow big enough to fall.
But when they allowed for the fluctuations (the stormy pockets), the "Lucky Droplets" grew 33% faster.

Why? Because the droplets that got swept into the "stormy pockets" grew so fast that they bridged the size gap in record time. The average cloud parcel looked slow, but the extreme pockets did the heavy lifting.

The Bottom Line

Rain doesn't happen because the average droplet grows steadily. It happens because of extreme luck combined with extreme weather conditions inside the cloud.

Correlated Collisions: Like being in a mosh pit; it helps a little bit at the start.
Intermittency (The Real Hero): Like getting swept up in a tornado; it creates the "Lucky Droplets" that grow fast enough to start the rain within 30 minutes.

In short: To get rain, you need a few droplets to get incredibly lucky and ride the "stormy waves" inside the cloud. Without these wild fluctuations, we might not get rain when we expect it!

1. Problem Statement

The formation of rain in warm clouds is hindered by the "size-gap problem."

The Gap: Water droplets form via condensation up to $\approx 15\,\mu\text{m}$ . Droplets larger than $50\,\mu\text{m}$ grow rapidly via gravitational settling. However, bridging the gap between $15\,\mu\text{m}$ and $50\,\mu\text{m}$ via collision-coalescence is theoretically too slow (taking hours) to explain observed rain onset (typically $\approx 30$ minutes).
The "Lucky Droplet" Hypothesis: Rain initiation may rely on statistical outliers—rare droplets ("lucky droplets") that grow significantly faster than the average due to stochastic fluctuations.
The Knowledge Gap: Previous models often assumed collision rates are constant (Markovian) and spatially uniform. However, turbulence introduces:
1. Correlated Collisions: Successive collisions may not be independent events due to particle clustering and the "sling effect."
2. Intermittency: The volume-averaged energy dissipation rate ( $\epsilon$ ) fluctuates strongly in space and time, leading to time-dependent collision rates.

The study aims to quantify how short-time correlations between collisions and spatio-temporal intermittency of dissipation affect the growth rate of these lucky droplets.

2. Methodology

The authors employ a multi-scale approach combining Direct Numerical Simulations (DNS) with stochastic modeling.

A. Direct Numerical Simulations (DNS)

Setup: They simulate homogeneous isotropic turbulence using the pseudospectral solver TurTLE.
Scale: The simulation box represents a small cloud parcel ( $\sim 1$ m scale), resolving scales down to the Kolmogorov length.
Parameters:
- Droplets are modeled as spherical Stokes particles.
- Initial size: $12.5\,\mu\text{m}$ (monodisperse).
- Number density: Up to $1000\,\text{cm}^{-3}$ (upper limit of cloud conditions) to ensure sufficient collision statistics.
- Dissipation Fluctuations: They simulate different volume-averaged dissipation rates ( $\bar{\epsilon}_r$ ), specifically the global mean ( $\bar{\epsilon}_r = 1$ ) and high-dissipation outliers ( $\bar{\epsilon}_r = 9$ ), to mimic the log-normal distribution of intermittency predicted by the refined similarity hypothesis (K62).
Collision Handling: A linear extrapolation of trajectories is used to detect collisions between timesteps, assuming a collision efficiency of 1 (perfect coalescence).

B. Non-Markovian Stochastic Framework

To analyze the DNS data, they developed a generalized master equation that accounts for memory effects.
Survival Probability: Instead of assuming a constant collision rate ( $\lambda$ ), they measure the conditional collision rate $\lambda_n(\tau)$ and the survival probability $S_n(\tau)$ (probability a droplet of size $n$ has not collided by time $\tau$ ).
Modeling Correlations: They found that $S_n(\tau)$ is not a single exponential (Poisson process) but is well-approximated by a superposition of two exponentials:
$S_n(\tau) = A_n^{\text{slow}} e^{-\lambda_n^{\text{slow}}\tau} + A_n^{\text{fast}} e^{-\lambda_n^{\text{fast}}\tau}$
This represents a mixture of a slow background process and a fast, correlated process.
Solution: They derived a solution for the droplet size distribution $P_n(t)$ that integrates over these time-dependent rates, effectively treating the non-Markovian process as a superposition of Markovian processes with different rates.

C. "Toy Model" for Intermittency

To extend the analysis to the full timescale of rain onset ( $\sim 30$ min), they constructed a simplified ensemble model.
Mechanism: They model the volume-averaged dissipation rate $\bar{\epsilon}_r(t)$ as a stochastic process (Ornstein-Uhlenbeck process applied to the logarithm of dissipation) that fluctuates over time.
Growth Law: The collision rate $\lambda_n(t)$ is parameterized as a linear function of droplet size $n$ and the instantaneous dissipation rate $\bar{\epsilon}_r(t)$ , based on DNS fits.
Ensemble: They simulate $10^5$ cloud parcels with fluctuating dissipation rates to calculate the probability of a droplet crossing the size gap.

3. Key Results

A. Effect of Correlated Collisions (Short-Time)

Acceleration: Correlated collisions (memory effects) significantly accelerate the growth of droplets on short timescales.
Quantification: For high dissipation ( $\bar{\epsilon}_r = 9$ ), the time required for the fastest $10^{-6}$ droplets to reach size $n=2$ is reduced by approximately 50% compared to a memoryless (Markovian) model.
Limitation: This effect is sub-leading for larger droplet sizes. As droplets grow, the relative speedup diminishes, and the assumption of statistically independent collisions becomes increasingly valid for the "tail" of the distribution.

B. Effect of Dissipation Intermittency (Long-Time)

Dominant Mechanism: The fluctuations in the volume-averaged dissipation rate are the primary driver for bridging the size gap.
Ensemble Results: When modeling an ensemble of cloud parcels with fluctuating dissipation (log-normal distribution):
- The ensemble mean reaches the "lucky droplet" threshold ( $10^{-6}$ probability of $>100$ collisions) significantly faster than a model with constant mean dissipation.
- Speed-up: The time to bridge the size gap is reduced by approximately 33% (from $\sim 350$ s to $\sim 230$ s in their specific parameterization).
Outlier Dominance: The ensemble average is dominated by a few parcels experiencing extreme, transient spikes in dissipation. These "burst" events drastically accelerate droplet growth, allowing them to cross the size gap much faster than the average conditions would suggest.

4. Significance and Conclusions

Resolution of the Size-Gap Problem: The study provides a quantitative mechanism explaining how rain can initiate within 30 minutes in warm clouds. It confirms that while turbulence-induced correlations help initially, the intermittency of energy dissipation is the crucial factor that creates the statistical outliers necessary for rapid rain onset.
Methodological Advance: The paper successfully bridges the gap between high-fidelity DNS (which captures short-time correlations) and large-scale stochastic modeling (which captures long-time intermittency). The derived non-Markovian framework allows for the explicit inclusion of memory effects in growth models.
Implications for Cloud Physics:
- Models assuming constant collision rates or uniform dissipation significantly underestimate the rate of rain initiation.
- The "lucky droplet" mechanism is robust, but its efficiency is heavily dependent on the statistical distribution of turbulent dissipation.
- Future cloud models should incorporate time-dependent dissipation fluctuations to accurately predict precipitation onset.

In summary, the authors demonstrate that turbulence intermittency creates rare, high-dissipation environments where lucky droplets grow explosively, effectively solving the timescale discrepancy in warm rain formation, with short-time collision correlations playing a secondary, though non-negligible, role.

Effects of correlated collisions and intermittency on the growth of lucky droplets