Abstract

Hydro-climatic (HC) models have complex environments due to the integration of hydrological processes and climate indices for the assessment of historical and future scenarios. The approximation of HC models leads to a major uncertainty in the selection of optimal methods for processing, enhancement, and assessment. The present work developed a User-Friendly Interface (UI) in the R programming platform to enhance the geospatial HC models using machine learning concepts. Here, UI complies with various technologies together to perform consistently with input control, processing, and visualization. To validate this interface, a snow-dominated alpine watershed was selected. The results showed that, (a) UI assisted to downscale of the future climatic data into finer resolution, (b) boosted the efficiency of the geospatial model by adaptive random forest regression with NSE = 0.92 and 0.84, respectively. Moreover, UI designed to apply for different geospatial optimization problems which assist academicians, scientists, decision-makers, planners, and stakeholders, etc.

摘要

由于水文过程和气候指数的整合用于评估历史和未来情景，水文气候 (HC) 模型具有复杂的环境。HC 模型的近似导致在选择用于处理、增强和评估的最佳方法时存在重大不确定性。目前的工作在 R 编程平台中开发了一个用户友好界面 (UI)，以使用机器学习概念增强地理空间 HC 模型。在这里，UI 符合各种技术，以与输入控制、处理和可视化保持一致。为了验证这个界面，选择了一个以积雪为主的高山流域。结果表明，(a) UI 有助于将未来的气候数据缩小为更精细的分辨率，(b) 通过 NSE = 0.92 和 0.84 的自适应随机森林回归提高了地理空间模型的效率。此外，UI 旨在应用不同的地理空间优化问题，帮助院士、科学家、决策者、规划者和利益相关者等。