研究印度次喜马拉雅跨界河流流域的平面动力学及土地利用类型变化，并提供预测性见解

《Physics and Chemistry of the Earth, Parts A/B/C》：Investigating the Planform Dynamics and LULC Changes with Predictive Insights in a Sub-Himalayan Transboundary River Basin, India

土地利用与河流平面形态动态关联性研究——以印度Raidak河流域为例。通过CA-Markov模型与遥感技术分析1930-2024年河道迁移、侵蚀沉积演变及土地利用变化规律，发现左岸平均年迁移量达429米，右岸372米，侵蚀面积累计114.91平方公里，植被减少显著，农业和建设用地扩张率达7.39%和11.55%。预测至2034年，湿地减少23.56%，城市扩张驱动LULC持续向人类活动主导型转变。

　　
Raidak River Basin: Decades-Long Planform Dynamics and Land-Use Changes Analysis

1. **Geomorphic Context and Study Objectives**
The Raidak River system, spanning parts of India, Bangladesh, and Bhutan within the Himalayan foothills, exhibits unique fluvial behaviors shaped by steep gradients, high sediment loads, and complex interactions between natural processes and human activities. This study focuses on three interconnected dimensions:
- **River planform dynamics** (1930–2024): Analyzing lateral bank migration, erosion-accretion patterns, and sinuosity changes.
- **Land-use dynamics** (1973–2024): Mapping transitions between seven LULC classes through supervised classification of multi-temporal satellite data.
- **Future projections** (2034): Predicting LULC patterns using the CA-Markov model to guide sustainable resource management.

2. **Methodological Framework**
Geospatial technologies were integrated to capture spatiotemporal changes:
- **River bank mapping**:NDWI and MNDWI indices distinguished water bodies from land, with manual digitization of banklines from historical topographic sheets (1930, 1955) and satellite imagery (1973–2024).
- **Erosion-accretion quantification**: Cross-sectional monitoring (31 for Raidak I, 20 for Raidak II) calculated lateral migration distances and land area shifts.
- **LULC classification**: Maximum likelihood classification (MLC) with ground-truthed polygons identified seven classes (dense forest, open forest, cropland, built-up, swamp, sandbar, water body).
- **Future projection**: CA-Markov model with MLP integration used elevation, distance to roads/rivers, and historical transitions to forecast 2034 LULC.

3. **Key Findings**
**3.1 River Planform Evolution**
- **Raidak I**: Characterized by progressive meandering with lateral bank migration averaging **429 m/year** (left bank) and **372 m/year** (right bank). Over 100 years, total erosion was **59.19 km2**, while accretion occurred in **44.96 km2**. Notable bank shifts included 1,822 m in Cross-Section 19 (left bank) and 1,493 m in Cross-Section 6 (right bank).
- **Raidak II**:展示 braided behavior with avulsive events, averaging **460 m/year** (left bank) and **474 m/year** (right bank). Total erosion was **64.73 km2**. The most significant migration occurred at Cross-Section 16 (left bank: 2,520 m; right bank: 2,548 m).

**3.2 Spatiotemporal LULC Shifts**
- **Vegetation decline**: Open forests decreased from **395 km2** (1973) to **224 km2** (2024), while dense forests shrank from **298 km2** to **116 km2**. Total vegetation loss: **40%** over 51 years.
- **Agricultural expansion**: Cropland increased by **251 km2** (from 117 km2 in 1973 to 369 km2 in 2024), driven by food demand and population pressure.
- **Urbanization**: Built-up areas expanded from **98 km2** (1973) to **262 km2** (2024), a **168% rise**.
- **Wetland degradation**: Swamp land decreased from **50 km2** (1973) to **10 km2** (2024), with water bodies shrinking by **57%**.
- **Sandbar dynamics**: Sandbody areas fluctuated between **45–55 km2**, reflecting deposition-erosion balance.

**3.3 Future Projections (2034)**
- **Cropland**: Projected to grow by **0.14%** (from 369 km2 to 369.32 km2).
- **Built-up areas**: Increase by **7.39%** (261.68 km2 → 281.01 km2).
- **Vegetation**: Open forests may decline by **5.8%** (223.64 km2 → 207.82 km2), while dense forests drop **5.8%** (116.38 km2 → 109.64 km2).
- **Wetland loss**: Swamp land and water bodies are projected to shrink by **23.56%** and **4.91%**, respectively.

**3.4 Model Validation**
- **AUC value**: 0.84 for 2024 LULC prediction, indicating high accuracy.
- **Kappa coefficients**: Range from **0.76–0.87**, confirming strong agreement between predicted and observed data.

4. **Drivers and Mechanisms**
- **Natural factors**: Steep gradients (up to **801 m bank migration/year** for Raidak I) and sediment load variability.
- **Human activities**:
- **Agriculture**: Conversion of abandoned channels and swamps into croplands, accelerated by irrigation infrastructure.
- **Urbanization**: Proximity to roads and riverbanks facilitated unplanned settlements.
- **Engineering**: Dam construction (e.g., Bhutan Ghat) altered sediment supply, exacerbating channel avulsion in Raidak II.
- **Hydro-morphological feedback**: Channel migration created new habitats (ox-bow lakes, sandbars) that were later converted to agriculture and settlements.

5. **Policy Implications**
- **Floodplain management**: Restrict urban encroachment in vulnerable zones (e.g., abandoned channels, swamps).
- **Sustainable agriculture**: Implement agroforestry and reforestation to mitigate vegetation loss.
- **Hydrological regulation**: Monitor sediment transport post-dam construction to prevent extreme bank shifts.
- **Adaptive planning**: Integrate CA-Markov projections with climate scenarios to prioritize conservation areas.

6. **Research Gaps and Future Directions**
- **Data limitations**: Lack of hydrological (e.g., discharge, sediment load) and socio-economic data in models.
- **Model refinement**: Incorporate dynamic variables like infrastructure projects and climate extremes.
- **Transboundary governance**: Address shared risks in the Brahmaputra–Ganges–Multipass systems where Raidak merges.

This study provides a comprehensive blueprint for balancing development and ecological preservation in Himalayan river basins, emphasizing the need for interdisciplinary approaches integrating geomorphology, remote sensing, and socio-economic data.