Overdependence on and exploitation of forest resources have significantly transformed the natural reserve forest of Sundarban, which shares the largest mangrove territory in the world, into a great degradation status. By observing these, a most pressing concern is how much degradation occurred in the past, and what will be the scenarios in the future if they continue? To confirm the degradation status in the past decades and reveal the future trend, we took Sundarban Reserve Forest (SRF) as an example, and used satellite Earth observation historical Landsat imagery between 1989 and 2019 as existing data and primary data. Moreover, a geographic information system model was considered to estimate land cover (LC) change and spatial health quality of the SRF from 1989 to 2029 based on the large and small tree categories. The maximum likelihood classifier (MLC) technique was employed to classify the historical images with five different LC types, which were further considered for future projection (2029) including trends based on 2019 simulation results from 1989 and 2019 LC maps using the Markov-cellular automata model. The overall accuracy achieved was 82.30%~90.49% with a kappa value of 0.75~0.87. The historical result showed forest degradation in the past (1989–2019) of 4773.02 ha yr⁻¹, considered as great forest degradation (GFD) and showed a declining status when moving with the projection (2019–2029) of 1508.53 ha yr⁻¹ and overall there was a decline of 3956.90 ha yr⁻¹ in the 1989–2029 time period. Moreover, the study also observed that dense forest was gradually degraded (good to bad) but, conversely, light forest was enhanced, which will continue in the future even to 2029 if no effective management is carried out. Therefore, by observing the GFD, through spatial forest health quality and forest degradation mapping and assessment, the study suggests a few policies that require the immediate attention of forest policy-makers to implement them immediately and ensure sustainable development in the SRF.

中文翻译：

---

应用多时态卫星卫星数据和马尔可夫元胞自动机预测孟加拉国桑达尔班保护区森林覆盖率变化和森林退化

对森林资源的过度依赖和开发已使拥有世界上最大的红树林地区的Sundarban自然保护区森林大大地退化。通过观察这些，最紧迫的问题是过去发生了多少退化，如果持续下去，将来会发生什么情况？为了确认过去几十年的退化状况并揭示未来趋势，我们以桑达尔班保护森林（SRF）为例，并使用1989年至2019年之间的卫星地球观测历史Landsat影像作为现有数据和主要数据。此外，还考虑了一个地理信息系统模型，该模型基于大树和小树类别来估计1989年至2029年SRF的土地覆盖（LC）变化和空间健康质量。采用最大似然分类器（MLC）技术对具有五种不同LC类型的历史图像进行分类，这些图像将进一步考虑用于未来的投影（2029），包括基于1989年和2019年LC映射的2019年模拟结果的趋势，以及使用马尔可夫细胞自动机的趋势模型。总体准确率为82.30％〜90.49％，kappa值为0.75〜0.87。历史结果表明，过去（1989-2019年）森林退化为4773.02公顷^-1，被认为是严重的森林退化（GFD），并且按照预测（2019-2029）的1508.53公顷^-1迁移，并且呈下降状态，而在1989-2029年间总体下降了3956.90公顷^-1。期。此外，研究还发现，茂密的森林逐渐退化（好到坏），但相反地，轻度森林得到了增强，如果不进行有效的管理，这种情况将持续到2029年。因此，该研究通过空间森林健康质量和森林退化的制图和评估来观察全球森林论坛，提出了一些政策，需要森林决策者立即注意，以立即执行这些政策，并确保可持续发展框架中的可持续发展。