Inlet design assumes that supply air from side-wall inlets mixes with house air over several metres, but whether this holds in a densely caged multi-tier house, where cages and birds obstruct the jet, is untested. The bird-zone environment of a commercial stacked-cage house was measured with 24 in-cage sensors at three distances and two tiers over ten winter days. Three of four groups showed no resolvable cooling gradient at 1.5 m; only one north-side group cooled with distance, strengthening with ventilation (0.30 K per percent, R² = 0.39), while another showed the opposite sign. Every fitted decay coefficient ($\alpha d_{\mathrm{h}}=0$ to $0.004$) lay below the free-jet design range ($0.01$–$0.03$), by five or more times over the window. Vertical stratification of $0.78$ K dominated (side × tier $p<0.001$): the upper tier was up to 2.4 K warmer on the north side, while the south-side tiers differed by under 0.5 K. Shading nets lowered north-side inlet temperature by $0.80$ K but warmed the south side. The practical consequence is that the thermal environment is organised by tier and side rather than by near-inlet distance, so monitoring, control and welfare-oriented assessment should target tiers and sides.

Accurate indoor-temperature forecasting underpins feed-forward control in broiler production, yet the horizons at which data-driven models add value are unclear. Two questions are addressed: at which horizons learned models measurably beat persistence, and what preview a feed-forward controller needs. Thirty-five days of data at 15 min resolution from a commercial stacked-cage broiler house (78,000 birds; outdoor temperature $-19$ to 16 °C) are used to benchmark 18 single-step and 10 multi-horizon models, probabilistic forecasts, and control simulation. A single time constant, $\tau =25.7$ h, organises the results: at 15 min, all learned models collapse to the persistence baseline (RMSE ≈ 0.105 K), because only 1% of the response relaxes per step; significant gains appear only at ≈$1\tau$ (LSTM $-25$% at 24 h, $-30$% at 48 h; FDR q ≤ 0.031). In control simulation, 4–8 h of MPC preview held $\sigma$ = 0.17 K with zero violations (0.74 K for 1 h control), while longer horizons added no benefit and were less robust. Quantile intervals calibrated in-sample (PICP ≈ 0.90) under-covered under distribution shift (0.61–0.72); $\tau$-anchored bands stayed calibrated (0.92–0.97). Design rules: evaluate forecasting models at horizons of at least ≈$1\tau$ and set control preview to ≈0.7–1.5${\tau }_f$, re-estimated per house and season.