Integrating hydrological, hydro-geological, soil-physical and hydrodynamic processes by means of particle based simulations

由于不同的操作尺度，如空间和时间尺度以及过程的复杂性，将水文过程整合到模型中是具有挑战性的。尽管如此，水文系统建模中的一些任务需要具有可靠的科学描述和水与连续或离散介质相互作用的相关模拟概念的整体方法。例如，这包括关于山洪暴发的风险评估，在这种情况下，至少水文和水动力过程的耦合是必不可少的。该项目的目的是在一个模型中全面和充分地综合考虑水文、水文地质、土壤物理和水动力过程。该模型将采用无网格数值插值方法SPH （smooth particle hydrodynamics）与智能标度方法相结合，发展成为一种新型的水系统模拟工具。与普通欧拉方法的主要区别在于，时间积分是在粒子上而不是在网格上进行的。尽管SPH方法近年来在水动力应用中得到广泛应用，但尚未考虑与相关水文或土壤水力过程的耦合；尽管该方法由于其数值稳定性和一般可扩展性而具有这样做的潜力。该实现将通过使用Nvidia显卡的GPU-CUDA框架来实现。动态标度被认为是实现水运动过程在现实环境中不可避免的整合的创新关键。将通过遵循源自土壤物理学的相似概念来实现标度。土壤和植被的信息将存储在属性域中，属性域中包含特定的参数作为分布函数。多概率粒子的参数分配反映了自然系统的多样性。粒子的密度和速度由属性场传递的信息控制。粒子的密度和速度控制着动态缩放。里程碑1是一个功能齐全的建模系统，具有地块尺度上的详细过程描述，同时已经考虑了上述学科的整体情况，例如水与植被的相互作用，与非饱和和饱和土壤带的相互作用。里程碑2侧重于使用上述缩放方法将地块尺度建模系统转移到更大的尺度。里程碑3是该模型在集水区规模的应用，包括对观测数据（鸡溪）进行测试和验证。验证后，我们将集成的建模系统用于不同工艺规模之间高度交互的应用。其目的是为诸如山洪暴发或灌溉等跨学科领域的决策者提供可靠和现实的基础，以便评估潜在的损害或灌溉的需要。

The integration of hydrological processes into models is challenging due to diverse operation scales, such as spatial and temporal scales as well as the complexity levels of the processes. Nonetheless, several tasks in hydrosystem modelling require holistic approaches with a reliable scientific description and its associated simulation concepts of the interaction of water with continuous or discrete media. This includes for instance risk assessment with regard to flash floods, where at least a coupling of hydrological and hydrodynamic processes is essential.The aim of the project is the holistic and fully integrated consideration of hydrological, hydro-geological, soil physical and hydrodynamic processes within one model. The model will be developed as a new type of hydrosystem simulation tool by using the grid-free numerical interpolation method SPH (smoothed particle hydrodynamics) in connection with smart scaling methods. The main difference to common Eulerian methods is that the integration in time is performed over the particles and not over the grid. Even though the SPH method is recently wide spread in hydrodynamic applications, a coupling with related hydrological or soil hydraulic processes is not yet considered; though the method has the potential to do so due to its numerical stability and general expandability. The implementation will be achieved by using the GPU-CUDA framework for Nvidia graphic cards.The dynamic scaling is considered to be an innovative key to perform the inevitable integration of water movement processes in real-world environments. The scaling will be achieved by following similarity concepts originated from soil-physics. Information of the soil and the vegetation will be stored in property fields, which contain the specific parameters as distribution functions. The assignment of parameters to the particles with multiple probabilities reflects the variety of natural systems. The density and velocity of the particles are controlled by information transferred via property fields. The density and velocity of the particles control the dynamical scaling.Milestone 1 is a fully functioning modelling system with detailed process descriptions at the plot scale, while already considering the holistic picture of the mentioned disciplines such as water interaction with the vegetation, with the unsaturated and saturated soil zone. Milestone 2 focuses on the transfer of the plot scale modelling system to larger scales with the mentioned scaling methods. Milestone 3 is the application of the model at the catchment scale including the test and validation with observed data (chicken creek). After the validation, we use the integrated modelling system for applications with high interactions between different process scales. The aim is to achieve a reliable and realistic basis for decision makers in interdisciplinary fields such as flash floods or irrigation in order to assess damage potential or need for irrigation.