Development of novel approaches to improve water resources data records, deep learning based forecasting, and participatory socio-hydrological systems modeling for integrated and adaptive water resources management

开发新方法来改进水资源数据记录、基于深度学习的预测以及用于综合和适应性水资源管理的参与式社会水文系统建模

• 批准号: RGPIN-2020-05325
• 负责人: Adamowski, Jan
• 金额: $ 2.62万
• 依托单位: McGill University
• 依托单位国家: 加拿大
• 项目类别: Discovery Grants Program - Individual
• 财政年份: 2021
• 资助国家: 加拿大
• 起止时间: 2021-01-01 至 2022-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=740511
• 关键词: Development; novel; approaches; improve; water

The long-term goal of my research program is to continue developing very original and practical methods to reduce the vulnerability, as well as enhance the resilience, adaptive capacity, and sustainability of water resources (WR) in Canada and elsewhere in the face of increasing complexity and climate change. In many parts of Canada and the world, pressure on WR, both in terms of quality and quantity, from urban areas, agriculture, industry and climate change are increasing, resulting in significant challenges for sustainable WR management.   The capacity of communities in Canada and elsewhere to mitigate against, and adapt to, such complex problems is constrained by numerous factors, including: i) a lack of, or unsuitable, WR data records in many watersheds that can help support WR modeling, planning and management; ii) challenges in providing very accurate short-term forecasts of WR variables and providing useful uncertainty estimates for decision makers; and iii) our limited understanding and modeling ability of the complex interactions and feedbacks between socioeconomic and physical processes in WR systems, combined with a frequent lack of meaningful participation of stakeholders in WR modeling, planning and management.   To address these complex challenges, my 5 year NSERC DG research program objectives consist of an interlinked three pronged/themed approach (that is a direct continuation of my current DG research) to develop, test, and implement very original and practical methods to improve: i) Water resources data records (Theme 1). We will develop new statistical approaches to: fill in records at stations with scattered missing values; extend records at short gauged stations or records with large missing gaps; and estimate records at completely ungauged sites. Some of these newly developed methods, and the improved WR data records obtained from their use, will also be used in themes 2 and 3 of my research program. ii) Short-term water resources variable forecasting (Theme 2). We will develop new `Deep Learning' based machine learning approaches that address nonlinearity, and new approaches to address non-stationarity and uncertainty estimation, in short-term (e.g., 1-3 days ahead) WR variable forecasting.   iii) Participatory coupled socio-hydrological systems modeling (Theme 3). We will develop new approaches to address the complex challenges of calibration, validation and `user friendly' software for participatory coupled socio-hydrological systems models that represent and dynamically model the interactions and feedbacks between physical (e.g., hydrological and related environmental processes) and socioeconomic processes that govern WR systems. The purpose of these coupled models will be to explore, in conjunction with stakeholders, long-term scenarios/trajectories and `what if' questions over decadal time spans (in contrast to short term `operational forecasting' such as Theme 2 models).

我的研究计划的长期目标是继续开发非常新颖且实用的方法来减少脆弱性，并增强加拿大和其他地区水资源（WR）的恢复力、适应能力和可持续性，以应对日益增加的水资源问题。在加拿大和世界许多地区，来自城市地区、农业、工业和气候变化对水资源的质量和数量的压力都在增加，给可持续水资源管理带来了重大挑战。加拿大和其他地方的社区缓解和适应的能力因此，此类复杂问题受到多种因素的制约，包括：i) 许多流域缺乏或不适合支持水电建模、规划和管理的水电数据记录；ii) 提供非常准确的短期预报面临挑战； WR 变量并为决策者提供有用的不确定性估计；以及 iii）我们对 WR 系统中土木和物理过程之间复杂的相互作用和反馈的理解和建模能力有限，加上利益相关者经常缺乏对 WR 建模的有意义的参与，规划和管理。为了应对这些复杂的挑战，我的 5 年 NSERC DG 研究计划目标包括相互关联的三管齐下/主题方法（这是我当前 DG 研究的直接延续），以开发、测试和实施非常新颖且实用的方法来改进： i) 水资源数据记录（主题 1）。其中一些新开发的方法以及从其使用中获得的改进的 WR 数据记录也将用于我的研究计划的主题 2 和 3）短期水资源变量预测（主题 2）。开发新的基于“深度学习”的机器学习方法来解决非线性问题，以及解决短期（例如，未来 1-3 天）WR 变量预测的非平稳性和不确定性估计的新方法。参与式耦合社会水文系统建模（主题 3）我们将开发新的方法来解决参与式耦合社会水文系统模型的校准、验证和“用户友好”软件的复杂挑战，这些模型代表并动态模拟之间的相互作用和反馈。这些耦合模型的目的是与利益相关者一起探索长期情景/轨迹和“假设”问题。十年时间跨度（与主题 2 模型等短期“运营预测”相反）。