Abstract: To meet the need for early leakage identification and sensor layout for buried hydrogen-blended natural gas pipelines in refining and petrochemical sites, an open-top soil-bin system measuring 6.00 m × 0.80 m × 1.25 m was constructed. Methane (CH₄) and hydrogen (H₂) sensors were arranged on the soil surface. Controlled leakage experiments followed by continuous monitoring were conducted under different absolute pressures, hydrogen blending ratios, and leakage locations. The hydrogen blending ratio was defined as the volume percentage of H₂ in the H₂/CH₄ mixture. The first gas-detection time (FGT) and first dangerous time (FDT) were used to characterize the surface response, hazard formation, and warning window. The warning window was calculated as the earliest FDT minus the earliest FGT. The results showed that FGT, FDT, and the warning window did not vary monotonically with pressure or hydrogen blending ratio. Changes in leakage location shifted the earliest response zone and the earliest hazardous zone, and the two zones did not always coincide. Under test T1, the earliest FGT was 2 min, the earliest FDT was 22 min, and the warning window was 20 min. At all four monitoring groups under T1, the H₂ channel responded earlier than the corresponding CH₄ channel. This result supports the use of H₂ sensors for early leakage identification. Combined CH₄/H₂ measurements and the FDT criterion can be used for hazard verification. Under the present experimental conditions, FGT can serve as the primary indicator for sensor layout, while FDT can be used to verify the warning margin. Sensor density should be increased near elbows, valves, inspection-well walls, and structural boundaries. Final sensor locations and spacing should also be verified against site-specific soil, underground structures, and surface environmental conditions.