Power law scaling of streamflow in mountainous basins
Shenavaei Abbasi, Mohammad (2023)
Shenavaei Abbasi, Mohammad
2023
Ympäristö- ja energiatekniikan DI-ohjelma - Programme in Environmental and Energy Engineering
Tekniikan ja luonnontieteiden tiedekunta - Faculty of Engineering and Natural Sciences
This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Hyväksymispäivämäärä
2023-06-19
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:tuni-202306066564
https://urn.fi/URN:NBN:fi:tuni-202306066564
Tiivistelmä
Floods and droughts are significant threats to communities worldwide. Streamflow is a key predictor of riverine floods and watershed-scale drought events. River discharge is a response of various underlying and interacting climatic, land cover, and hydrological processes in the draining watersheds. Spatiotemporal scaling of streamflow regimes (low through high discharge events) may uncover emerging hydrologic patterns, leading to the development of robust prediction models. The main objective of this thesis is to investigate the scaling property of streamflow using statistical modeling and data analytics with a focus on mountainous areas. The goal is to identify the dominant watershed hydroclimatic and land cover drivers controlling various streamflow regimes, as well as to develop power law-based scaling models to robustly predict streamflow at different time and space scales. The dominant controls of streamflow regimes were identified by using Pearson correlation analysis and principal component analysis. Robust scaling models of various streamflow regimes were then developed by using univariate to multivariate power law functions iteratively.
The study area of this thesis is the Monongahela River Basin of USA. It is a large watershed located mostly in mountainous Appalachian regions of USA with variable land cover characteristics. The watershed passes through three states, and the region's varied topography and landcover make it a unique place to study. Eleven hydro-climatic variables are collected including different land cover characteristics (six categories) drainage density, imperviousness, area, slope, and precipitation percentiles. A cumulative approach is then implemented on all variables to capture scaling at multiple spatial scales. The spatial correlation of variables is then broken using a bootstrap resampling approach to prevent pseudoreplications. After performing regression analyses, the efficiency and accuracy of the model are assessed by using, respectively, the Nash-Sutcliffe efficiency (NSE) and the root-mean-square error to the standard deviation of observations (RSR) ratio.
The results revealed that drainage area and mean precipitation are the most dominant variables that influence all flow regimes in largely forested mountainous basins. Other hydroclimatic and land cover variables exhibited significant mutual correlations. For example, the most common land cover types, vegetated and agricultural land cover, showed a strong correlation with land area and built-up lands, respectively. The principal component analysis results provide information about model collinearity correlation and significant variables. Previous analyses that provide a reliable scaling model are validated. The power-law model also shows promise in predicting streamflow in an ungauged basin, with an average NSE of 0.90 and RSR of 0.27. As a result, studying streamflow scaling not only reveals details about the mechanisms governing streamflow and how it interacts with other variables, but it also provides useful insights into how to predict using simple power law equations. This study provides information about the mechanisms of streamflow in mountainous watersheds, which can help water re-source scientists, engineers, and managers predict streamflow regimes and mitigate floods in similar regions, as well as regulate flood management schemes in different watersheds.
The study area of this thesis is the Monongahela River Basin of USA. It is a large watershed located mostly in mountainous Appalachian regions of USA with variable land cover characteristics. The watershed passes through three states, and the region's varied topography and landcover make it a unique place to study. Eleven hydro-climatic variables are collected including different land cover characteristics (six categories) drainage density, imperviousness, area, slope, and precipitation percentiles. A cumulative approach is then implemented on all variables to capture scaling at multiple spatial scales. The spatial correlation of variables is then broken using a bootstrap resampling approach to prevent pseudoreplications. After performing regression analyses, the efficiency and accuracy of the model are assessed by using, respectively, the Nash-Sutcliffe efficiency (NSE) and the root-mean-square error to the standard deviation of observations (RSR) ratio.
The results revealed that drainage area and mean precipitation are the most dominant variables that influence all flow regimes in largely forested mountainous basins. Other hydroclimatic and land cover variables exhibited significant mutual correlations. For example, the most common land cover types, vegetated and agricultural land cover, showed a strong correlation with land area and built-up lands, respectively. The principal component analysis results provide information about model collinearity correlation and significant variables. Previous analyses that provide a reliable scaling model are validated. The power-law model also shows promise in predicting streamflow in an ungauged basin, with an average NSE of 0.90 and RSR of 0.27. As a result, studying streamflow scaling not only reveals details about the mechanisms governing streamflow and how it interacts with other variables, but it also provides useful insights into how to predict using simple power law equations. This study provides information about the mechanisms of streamflow in mountainous watersheds, which can help water re-source scientists, engineers, and managers predict streamflow regimes and mitigate floods in similar regions, as well as regulate flood management schemes in different watersheds.