Carbon sink enhancement is of great significance to achieving carbon peak and carbon neutrality. This study firstly estimated the carbon sink in the Beijing–Tianjin–Hebei Region using the carbon absorption coefficient method. Then, this study explored the differentiation of carbon sink enhancement potential with a carbon sink–economic carrying capacity index matrix based on carbon sink carrying capacity and economic carrying capacity under the baseline scenario and target scenario of land use. The results suggested there was a remarkable differentiation in total carbon sink in the study area, reaching 2,056,400 and 1,528,300 tons in Chengde and Zhangjiakou and being below 500,000 tons in Langfang and Hengshui, while carbon sink per unit land area reached 0.66 ton/ha in Qinhuangdao and only 0.28 t/ha in Tianjin under the baseline scenario. Increasing area and optimizing spatial distribution of arable land, garden land, and forest, which made the greatest contribution to total carbon sinks, is an important way of enhancing regional carbon sinks. A hypothetical benchmark city can be constructed according to Qinhuangdao and Beijing, in comparison with which there is potential for carbon sink enhancement by improving carbon sink capacity in Beijing, promoting economic carrying capacity in Qinhuangdao, and improving both in the other cities in the study area.

Landslides, as sudden, destructive geological events, threaten human lives, property, and infrastructure. Remote sensing and UAV imagery provide essential data for landslide identification, yet landslide targets often appear as sparse small objects, and existing lightweight models still lack sufficient accuracy under small-sample and small-target conditions. To address this, we propose LMAF-Net, a lightweight multi-attention fusion identification network employing MobileNetV3 as encoder and integrating boundary-aware coordinate attention, serial dual attention, and efficient channel attention modules to enhance feature discrimination while maintaining efficiency. Experiments on four public landslide datasets show that LMAF-Net achieves an F1-score of 96.09% and IoU of 92.47% on STTCLD, with only 3.77M parameters and 1.07G FLOPs, outperforming compared models in both accuracy and efficiency. Ablation studies confirm that boundary-aware coordinate attention reduces omission error by 1.75%, and serial dual attention improves F1-score by 1.11%. Generalization tests reveal that the model maintains F1-scores between 94.95% and 97.63% across cross-region and cross-sensor scenarios, demonstrating strong cross-domain robustness. Robustness tests indicate that under degradation conditions such as radiometric perturbation, blurring, compression, and random occlusion, the average F1-score drops by only 2.05%, verifying resistance to image degradation. This work provides an effective lightweight solution for landslide identification under challenging small-sample and small-target conditions.