In recent years, Transformer-based methods have demonstrated proficiency in capturing complex patterns for time series forecasting. However, their quadratic complexity relative to input sequence length poses a significant bottleneck for scalability and real-world deployment. Recently, the Mamba architecture has emerged as a compelling alternative by mitigating the prohibitive computational overhead and latency inherent in Transformers. Nevertheless, a vanilla Mamba backbone often struggles to adequately characterize intricate temporal dynamics, particularly long-term trend shifts and non-stationary behaviors. To bridge the gap between Mamba’s global scanning and local dependency modeling, we propose C-T-Mamba, a hybrid framework that synergistically integrates a Mamba block, channel attention, and a temporal convolution block. Specifically, the Mamba block is leveraged to capture long-range temporal dependencies with linear scaling, the channel attention mechanism filters redundant information, and the temporal convolution block extracts multi-scale local and global features. Extensive experiments on five public benchmarks demonstrate that C-T-Mamba consistently outperforms state-of-the-art (SOTA) baselines (e.g., PatchTST and iTransformer), achieving average reductions of 4.3–18.5% in MSE and 3.9–16.2% in MAE compared to representative Transformer-based and CNN-based models. Inference scaling analysis reveals that C-T-Mamba effectively breaks the computational bottleneck; at a horizon of 1536, it achieves an $8.8\times$ reduction in GPU memory and over 10× speedup compared to standard Transformers. At 2048 steps, its latency remains as low as 8.9 ms, demonstrating superior linear scaling. These results underscore that C-T-Mamba achieves SOTA accuracy while maintaining a minimal computational footprint, making it highly effective for long-term multivariate time series forecasting.

This study focuses on Huangshi Town, Longchuan County, Guangdong Province, as the research area. By utilizing existing data and field surveys, the study identifies geological hazards and risks in the area, deeply analyzes the formation mechanisms and disaster-causing patterns, and systematically summarizes the developmental characteristics and spatial distribution of these hazards. Methodologically, the research combines regular grids and slope units, selecting ten evaluation factors for correlation analysis. The information theory model is used to assess disaster susceptibility. The study further evaluates the impact of three rainfall scenarios—100 mm of rainfall over 24 h, 250 mm of rainfall over 24 h, and 240 mm of effective rainfall over 72 h—on geological disasters in Huangshi Town. As a result, a comprehensive hazard assessment under multiple rainfall scenarios is produced. The findings show that the Receiver Operating Characteristic (ROC) accuracy of the disaster-susceptibility evaluation reaches 0.8739, ensuring high data quality for the geological hazard assessment in Huangshi Town. The zoning results align closely with field survey observations. In conclusion, incorporating rainfall as a triggering factor enhances the accuracy of the susceptibility analysis by better capturing the temporal and spatial patterns of landslide occurrences, thereby offering a more comprehensive understanding of the geological hazard development in the study area.