Keywords
Statistical modelling, Process-based modelling, environmental flows, land cover change, land management, water management, climate change, trend, Mann-Kendall, Generalized Additive Modelling
Start Date
16-9-2020 3:00 PM
End Date
16-9-2020 3:20 PM
Abstract
Quantification of uncertainty in hydrological modeling is imperative and meant to improve reliability of modeling result. Both process-driven and data-driven modeling approaches could be implemented to drive cause-and-effect relationships between influencing drivers (e.g. direct or indirect) and eventual impacts on study areas. Land cover changes and management practices are most likely to influence and/or alter the hydrological functioning in the catchments, while it could either exacerbate or mitigate climate change consequences especially in drought-sensitive regions. This uncertainty might lead to inaccurate definition of underlying processes including initial and boundary conditions leading to unknown errors in new, untested environmental circumstances. Statistical modelling may clear up the difficulties faced while numerically solving process-based models through the quantification of correlations and confidence intervals. The study aims at linking environmentally critical stream flows (environmental flows) and the integrated impacts of land use, river regulation and climate change on the example of two Central German catchments (Eine; 180 km2 and Goetsche; 50 km2). The catchments are subject to increasing water demand from different water users (e.g. agricultural, industrial and domestic) as well as water quality issues. The driver-impact relationship needed to be analyzed and understood in order to develop future adaptive measures. Trend-Free-Pre-Whitening-Mann-Kendall (TFPW-MK) trend analysis has been implemented to assess the significance of temporal patterns, while Generalized-Additive-Modelling has been used to attribute causal links of associations between explanatory variables (e.g. climatic and anthropogenic) and the response variables (e.g. water quantity and quality). A SWAT+ (Soil and Water Assessment Tool) process-based model is being set up with a detailed description in land and water management. Comparing the outputs between statistical and physical modellings might pinpoint key areas where further research may be most useful thereby permitting better future model refinements.
Statistical processing and modeling for performance tuning of integrated river basin physical modelling
Quantification of uncertainty in hydrological modeling is imperative and meant to improve reliability of modeling result. Both process-driven and data-driven modeling approaches could be implemented to drive cause-and-effect relationships between influencing drivers (e.g. direct or indirect) and eventual impacts on study areas. Land cover changes and management practices are most likely to influence and/or alter the hydrological functioning in the catchments, while it could either exacerbate or mitigate climate change consequences especially in drought-sensitive regions. This uncertainty might lead to inaccurate definition of underlying processes including initial and boundary conditions leading to unknown errors in new, untested environmental circumstances. Statistical modelling may clear up the difficulties faced while numerically solving process-based models through the quantification of correlations and confidence intervals. The study aims at linking environmentally critical stream flows (environmental flows) and the integrated impacts of land use, river regulation and climate change on the example of two Central German catchments (Eine; 180 km2 and Goetsche; 50 km2). The catchments are subject to increasing water demand from different water users (e.g. agricultural, industrial and domestic) as well as water quality issues. The driver-impact relationship needed to be analyzed and understood in order to develop future adaptive measures. Trend-Free-Pre-Whitening-Mann-Kendall (TFPW-MK) trend analysis has been implemented to assess the significance of temporal patterns, while Generalized-Additive-Modelling has been used to attribute causal links of associations between explanatory variables (e.g. climatic and anthropogenic) and the response variables (e.g. water quantity and quality). A SWAT+ (Soil and Water Assessment Tool) process-based model is being set up with a detailed description in land and water management. Comparing the outputs between statistical and physical modellings might pinpoint key areas where further research may be most useful thereby permitting better future model refinements.
Stream and Session
false