Integrated Modeling of Precipitation-Recharge-Runoff at the River Basin Scale: The Susquehanna

尽管流域是陆地水循环研究的组织原则，但直到最近，将这种规模的气候、地形、生态和地质的复杂性与水资源预测的特殊需求相结合的模型还没有被认为是实用的。本研究提出在复杂地形和水文地质条件下研究降水-补给-径流的多尺度动态，以及水和能量收支的分配。该办法在每个感兴趣的尺度上保持了地表和地下过程之间的自然耦合，但认识到地表水流域和地下水流域可能有不同的界线。该建模方法基于有限体积表示法，其中守恒方程和本构方程在指定的支撑尺度上被平均。该模型是多尺度的，即以保留特定水资源预报(洪水动态、河流-含水层对干旱的反应等)所需的时空尺度的方式来解决景观中的气候、植被、地形和水文地质因素。这项研究解决了在计算规模和需要在水资源预测中包括精细材料特性之间的权衡问题。对SRB的长期流量预报必须考虑到控制盆地对气候或土地利用强迫的反应的地表和地下过程的相互竞争的时间尺度。我们的策略是使用多目标进化算法，根据快速地表响应和较慢的地下响应来进行预报。这一多目标框架将调整模型参数和分辨率，以考虑地表和地下水流不同的时间尺度和物理复杂性。这项研究最初将专注于为SRB内的组成部分流域开发区域概念性地表水-地下水模型。下一阶段的研究将制定一项综合区域概念模型的战略，将地理空间数据作为整个SRB地表地下水动态大型模型的输入。研究的最后阶段将开发一个决策支持系统，该系统将把新的气候数据同化为长期径流预测。这些长期预测将支持为整个SRB制定更好的水管理政策。这项研究涉及四个基本问题：1)水文地质在长期和短期径流中扮演什么角色，与气候和土地利用动态有什么关系？2)小尺度土壤和地下变异性何时控制径流，模型如何“适应”外部(气候)和内部条件(土地利用)的变化？3)流域支流动态耦合的时空尺度是什么？4)如何将进化计算策略和“定性”概念模型结合起来，以更好地解决流域尺度上的模型维度、参数化和预测？

0310122DuffyAlthough the river basin is the organizing principal of the terrestrial water cycle research, until recently, models which couple the complexity of climate, terrain, ecology, and geology at this scale to the particular needs of a water resource forecast, have not been considered practical. The present research proposes to investigate the multi-scale dynamics of precipitation-recharge-runoff, and the partitioning of water and energy budgets over complex terrain and hydrogeological conditions. The approach maintains the natural coupling between surface and subsurface processes at each scale of interest but recognizes that surface water basins and groundwater basins may have distinct delineations. The modeling approach is based on a finite volume representation, where conservation and constitutive equations are averaged over a specified support scale. The model is multi-scale in the sense that climatic, vegetative, topographic and hydrogeologic elements of the landscape are resolved in such a way as to preserve the necessary space-time scales for a particular water resource forecast (flood dynamics, stream-aquifer response to drought, etc.). The research addresses the tradeoff between the scale of computing and the need to include fine-scale material properties in water resource predictions. Long-term flow forecasts for the SRB must account for the competing time scales of the surface and subsurface processes governing the basin's response to climatologic or landuse forcing. Our strategy is to condition forecasts on both rapid surface responses as well as slower subsurface responses using multiobjective evolutionary algorithms. This multiobjective framework will adapt model parameters and resolution to account for the disparate time-scales and physical complexities of surface and subsurface flow regimes.This research will initially focus on developing regional conceptual surface-groundwater models for component watersheds within the SRB. The next phase of the research will develop a strategy for synthesis of the regional conceptual models with geospatial data as input to a large-scale model for the surface-groundwater dynamics of the entire SRB. The final phase of the research will develop a decision-support system that will assimilate new climatologic data into long-range runoff predictions. These long-range predictions will support the development of improved water management policies for the entire SRB. The research addresses four fundamental questions: 1) What role does hydrogeology play in long-term and short-term runoff and what is the relation to climate and landuse dynamics? 2) When does small-scale soil and subsurface variability control runoff and how can models "adapt" to changing external (climate) and internal conditions (landuse)? 3) What are the space-time scales at which tributaries of the river basin are dynamically coupled? 4) How can evolutionary computing strategies and "qualitative" conceptual models be incorporated to better resolve model dimensionality, parameterization, and prediction at the river basin scale?