Ma bamboo (Dendrocalamus latiflorus Munro) is a fast-growing woody grass that offers significant economic benefits, including materials for construction, furniture, biofuel, food, and handicrafts. It also provides ecological benefits like soil conservation, wildlife habitats, and carbon sequestration. However, its species distribution patterns are influenced by various factors, including climate (mainly temperature and precipitation), soil attributes, and landscape characteristics such as topography, land use, and vegetation. Understanding these impacts is essential for the sustainable management of D. latiflorus resources and fostering related economic activities. To address these challenges, we developed a comprehensive habitat suitability (CHS) model that integrates climate, soil, and landscape variables to simulate the distribution dynamics of D. latiflorus under different shared socio-economic pathway (SSP) scenarios. An ensemble model (EM) strategy was applied to each variable set to ensure robust predictions. The results show that the current potential distribution of D. latiflorus spans 28.95 × 10⁴ km², primarily located in South China and the Sichuan Basin. Its distribution is most influenced by the annual mean temperature (Bio1), the cation exchange capacity of soil clay particles in the 20–40 cm soil layer (CECc 20–40 cm), vegetation, and elevation. Under future climate scenarios, these habitats are projected to initially expand slightly and then contract, with a northward shift in latitude and migration to higher elevations. Additionally, the Sichuan Basin (Sichuan–Chongqing border) is identified as a climatically stable area suitable for germplasm development and conservation. To conclude, our findings shed light on how climate change impacts the geographic distribution of D. latiflorus, providing key theoretical foundations for its sustainable cultivation and conservation strategies.

China’s rare and endangered bamboo species belonging to the Indocalamus genus, specifically Indocalamus decorus Q. H. Dai, Indocalamus hirsutissimus Z. P. Wang & P. X. Zhang, and Indocalamus pedalis (Keng) P. C. Keng, possess notable value in biodiversity conservation and have significant potential for use in landscape design. Using an enhanced MaxEnt model, this study forecasted shifts in the species’ potential range under four separate climate scenarios (SSP1-RCP2.6, SSP2-RCP4.5, SSP3-RCP7.0, and SSP5-RCP8.5), considering both the historical period (1970–2000, referred to as “the current period”) and upcoming timeframes (2021–2040, 2041–2060, 2061–2080, and 2081–2100). The analysis disclosed that the present total potential habitat area for these species is approximately 251.79 × 10⁴ km², with high, medium, and low suitability areas occupying 0.15 × 10⁴ km², 125.39 × 10⁴ km², and 126.26 × 10⁴ km², respectively. The minimum temperature of the coldest month emerged as the critical determinant of their potential habitat distribution. Expected climate changes are likely to increase the suitable habitat for these species, although areas with low suitability might slightly diminish, with Guizhou and Chongqing showing the least impact. The distribution centers of suitable habitats for the three plant species consistently exhibit a westward shift under various climate scenarios. These results contribute valuable insights for the spatial distribution, continuous monitoring, sustainable management, and ex situ conservation in response to climate change.

High-resolution image segmentation for landscape applications has garnered significant attention, particularly in the context of ultra-high-resolution (UHR) imagery. Current segmentation methodologies partition UHR images into standard patches for multiscale local segmentation and hierarchical reasoning. This creates a pressing dilemma, where the trade-off between memory efficiency and segmentation quality becomes increasingly evident. This paper introduces the Multilevel Contexts Weighted Coupling Transformer (WCTNet) for UHR segmentation. This framework comprises the Mult-level Feature Weighting (MFW) module and Token-based Transformer (TT) designed to weigh and couple multilevel semantic contexts. First, we analyze the multilevel semantics within a local patch without image-level contextual reasoning. It avoids complex image-level contextual associations and eliminates the misleading information carried. Second, MFW is developed to weigh shallow and deep features for enhancing object-related attention at different grain sizes from multilevel semantics. Third, the TT module is introduced to couple multilevel semantic contexts and transform them into semantic tokens using spatial attention. Then, we can capture token interactions and obtain clearer local representations. The suggested contextual weighting and coupling of single-scale patches empower WCTNet to maintain a well-balanced relationship between accuracy and computational overhead. Experimental results show that WCTNet achieves state-of-the-art performance on two UHR datasets of DeepGlobe and Inria Aerial.

Multi-class land-cover change detection from bi temporal remote sensing imagery has become a critical task for large-scale Earth observation. Nevertheless, prevailing segmentation-based change detectors remain prone to false alarms and missed detections, especially along fine-grained boundaries and in complex backgrounds where unchanged regions are easily confused with particular change classes. To mitigate these issues, we propose a decoder-enhanced UNet style segmentation framework (USNet) to improve change-class discrimination by jointly reinforcing semantic representation, boundary/detail reconstruction, and computational efficiency. Specifically, USNet strengthens the bottleneck to enrich high level contextual semantics while adopting lightweight convo lutional blocks to constrain decoding complexity. Moreover, spatial-and-channel squeeze¨Cexcitation (SCSE) recalibration is incorporated into the up-sampling stages to suppress background interference and facilitate accurate boundary recovery. Extensive experiments on the SECOND dataset demonstrate that USNet consistently outperforms representative segmentation baselines. In particular, it achieves 72.69% mIoU, 49.35% F1-score, and 88.97% accuracy, surpassing strong competitors such as FCN, UNet, PSPNet, and DeepLabV3+ across all evaluation metrics. These results confirm the effectiveness of USNet for robust multi class land-cover change detection in challenging remote sensing scenarios.