Handles huge datasets (n)

Option	Description
Yes	Scales smoothly to arbitrarily large n without fundamental changes.
Most do	Works for large n with most method declinations but may need optimization.
Most don't	Only a few variants or implementations of the method can handle large n.
No	Breaks down or becomes computationally infeasible when n grows.

Definition

The method’s scalability to very large datasets, indicating if it remains computationally and statistically tractable as sample size n grows.

Explanation

Big‑data settings present both opportunity and challenge for detection and attribution: scalable algorithms can leverage high resolution or massive sensor networks, but some methods become infeasible or lose interpretability at scale. Knowing which methods handle large n informs decisions about algorithmic implementations and resource needs.

Tools/rationale for helping assessment

Depending on the number of samples available in the case study, the user estimates if this aspect is important to consider when filtering methods.

Example

A biogeographer wants to process 3 million 10 m Sentinel-2 pixels; only a distributed causal random-forest implementation would complete in hours, whereas a full Bayesian SEM fails beyond ~10 000 pixels, so Yes is selected.