## Problem 181 — rank:166

**Title:** Ecological scaling across space and time

**Added binding assumptions (LaTeX):**

```latex
Use trailing rather than future-looking temporal windows when $Z(t)$ is available at forecast time $t$. Retained lagged summaries must be computable from observations available then. Include fluxes across the coarse-cell boundary and preserve the declared conserved quantities. Small one-step conditional-mean error does not automatically yield accurate multi-step distributions; state the horizon, stability and propagated error criterion.
```

**Direct edits:** 1. **Fresh external checks:** None; supplied sources retained as provenance.