Integrating complexity, GIS and soft computing for multi-scale spatial modeling and decision-making

人类活动对自然环境的影响日益增加，导致许多无计划的土地利用和土地覆盖变化。城市化、毁林和工业发展等活动现在与城市无计划扩展密切相关，而森林火灾和虫害等自然灾害普遍降低了自然景观的质量。LULC变化过程难以预测，因为它们受到人类和生物物理系统变量之间许多复杂和非线性相互作用的影响。该研究计划的长期目标是促进对LULC变化过程的结构和机制的科学理解，以改善空间预测和决策。以复杂性科学和地理信息系统科学为理论基础，建立动态、现实的土地利用变化模型。主要目标是：（1）在地理信息系统（GIS）和Web-GIS框架内开发新的基于代理的建模（ABM）和CA方法;（2）在多个空间和时间尺度上处理和处理复杂地理空间数据集的软计算方法;（3）大型地理空间数据集的鲁棒地理模拟的并行处理方法和算法;以及（4）校准和验证方法，以生成用于规划和管理背景下的空间决策的可靠和有用的模型。将开发的方法和工具与利益相关者协作方法联系起来，将缩小研究级建模与有意义的现实世界应用之间的差距。这项研究有助于扩大LULC变化过程的科学知识，并推进GIS和GIScience领域，包括使用创新的空间建模方法的动态功能。基于LULC模型的空间决策支持系统和工具的实施将进一步帮助协作环境规划和管理。林业、农业和城市土地利用方面的科学家、规划者和决策者将从这项研究中受益，因为他们有客观和科学的基础来做出有效的决策。

The increasing impact of human activities on the physical and natural environment has resulted in many unplanned land-use and land-cover (LULC) changes. Activities such as urbanization, deforestation and industrial development are now strongly associated with urban sprawl, while natural hazards such as forest fires and insect infestations have generally reduced the quality of the natural landscape. LULC change processes are difficult to predict as they are influenced by many complex and non-linear interactions between variables from human and biophysical systems. The long-term goal of this research program is to advance scientific understanding about the structure and mechanisms of LULC change processes for improving spatial prediction and decision-making. Complexity science and GIScience will be used as the theoretical foundations to develop dynamic and realistic land transformation models. The main objectives are to develop: (1) novel agent-based modeling (ABM) and CA approaches within geographic information systems (GIS) and Web-GIS frameworks; (2) soft computing approaches for handling and processing complex geospatial data sets at multiple spatial and temporal scales; (3) parallel processing methods and algorithms for robust geosimulations of large geospatial datasets; and (4) calibration and validation methods to generate reliable and useful models for spatial decision-making in planning and management contexts. Linking the developed approaches and tools to stakeholder collaboration methods will narrow the gap between modeling at a research level and meaningful real-world applications. This research contributes to expanding the scientific knowledge of LULC change processes and to advancing the field of GIS and GIScience to include dynamic capabilities using innovative spatial modeling approaches. The implementation of spatial decision support systems and tools based on the LULC models will further aid collaborative environmental planning and management. Scientists, planners and policy-makers in forestry, agriculture, and urban land-use will benefit from this research by having objective and scientific foundations on which to make effective decisions.