Large Scale Mapping of Forest Vertical Structure from DQ-1 Multi-Wavelength LiDAR: A Focus on Boreal Forests

Canopy percentile height is a key parameter for describing forest vertical structure and assessing terrestrial carbon sequestration. The DQ-1 satellite provides multi-band, single-footprint full-waveform LiDAR observations, but canopy height retrieval is affected by 1064 nm waveform saturation and cross-band differences in waveform shape and vertical resolution. To address these limitations, we developed a multi-band fusion algorithm for forest applications (MBFA-F) to retrieve ancillary canopy percentile heights from DQ-1 waveform observations and characterize their uncertainty. Validation against Finnish ALS data showed relatively small and stable errors for RH20–RH40, with MAE values of 1.99–2.84 m and RMSE values of 2.46–3.05 m, whereas errors increased toward RH100, where the MAE and RMSE reached 4.89 m and 6.30 m, respectively. In the combined point-scale comparison with GEDI and ICESat-2 observations, the MAE/RMSE values were 5.69/7.71 m for RH98, 5.74/7.66 m for RH90, and 5.53/7.24 m for RH75. These results indicate measurable consistency with independent LiDAR references, while supporting the interpretation of DQ-1 canopy percentile heights as uncertainty-aware ancillary retrievals rather than direct substitutes for established spaceborne LiDAR canopy-height products. In boreal forest regions, DQ-1 provided 0.05%–3.06% additional pixel coverage compared with the canopy percentile height products from GEDI and ICESat-2, indicating limited but useful complementarity in sparsely sampled areas. Overall, this study demonstrates the feasibility of retrieving canopy percentile heights from DQ-1 multi-wavelength full-waveform LiDAR observations and highlights their complementary value for forest vertical-structure monitoring.