THE INFLUENCE OF MEASUREMENT SCALE AND UNCERTAINTY ON INTERPRETATIONS OF RIVER MIGRATION is a well-researched Physical Sciences and Mathematics Thesis/Dissertation topic, it is to be used as a guide or framework for your Academic Research.
Measuring temporal and spatial variation in river migration enables us to better understand mechanisms driving one of the most ubiquitous and effective modes of reworking Earth’s surface. Studies of river migration span multiple orders of spatial and temporal magnitude- from a single meander bend to geologic-scale evolution of rivers.
Uncertainty is inherent but often overlooked in measuring river channel evolution and few consider how spatial and temporal measurement scales bias measurements. Ignoring such uncertainties may confound measurements, obscure patterns of river behavior, and lead to false conclusions regarding processes of river change.
In three studies, we describe (1) how to quantify and account for uncertainty in measuring channel adjustments, (2) whether temporal measurement scale impact inferences about river response to agricultural management and (3) if spatial measurement scale can bias apparent mechanistic relations between meander migration and curvature.
We explore 76 years of geomorphic change along the Root River in response to shifting hydrology and land management, recorded in decadal sets of imagery. The changing conditions and extensive imagery provide an excellent natural experiment to explore our objectives.
In Chapter 2 we developed the first comprehensive framework for quantifying and accounting for uncertainty in channel erosion derived measurements from aerial imagery. We review and test best practices for quantifying uncertainty, provide context for applying each practice, and introduce new methods for handling measurements below the threshold of uncertainty.
Although this framework is developed for river platform adjustments, it is applicable to many moving boundary measurements. Chapter 3 explores how migration rate measurements from aerial images may be biased by the time interval between measurements.
Migration rates measured over longer time intervals systematically underestimate ‘true’ rates because reversals in migration direction underestimate net migration distance between images. Migration measurements must encompass short-term rate variability in order to accurately demonstrate fluvial change and estimate long-term sediment remobilization and flux for sediment budgets.
These results inform our data selection for Chapter 4, wherein we demonstrate how spatial measurement scale can influence apparent relations among factors impacting channel migration. Using measurement scales that capture longitudinal variability in shear stresses helped discern a phase lag between curvature and migration signals.
River channels are among the most dynamic landforms on Earth’s surface, sweeping laterally across valley bottoms- often in subtle and sometimes catastrophic-ways over the event-, decadal-, and millennial timescales. Measuring temporal and spatial variation in river migration enables us to better understand mechanisms driving this ubiquitous and impactful feature of Earth’s surface.
Remotely-sensed imagery is increasingly used to measure changes in river planform in response to changes in environmental drivers such as land-use, urbanization, deforestation, dam building or removal (Gurnell et al., 1994; Gaeuman et al., 2005; Constantine et al., 2014; Donovan et al., 2015, 2016),
develop a predictive understanding of channel and floodplain evolution (Lauer and Parker, 2008; Crosato, 2009; Braudrick et al., 2009; Parker et al., 2011), providing constraints for sediment budgets (Trimble, 1983; Reid and Dunne, 2005; Belmont et al., 2011; Stout et al., 2014) and improving bank erosion models (Larsen et al., 2006; Motta et al., 2012).
River meander migration also provides intriguing opportunities to test theories regarding basic principles and properties of physics (Hick in and Nanson, 1984; Furbish, 1988; Constantine et al., 2009; Crosato, 2009; Parker et al.,2011).
The complexity inherent in modeling meander migration is reflected in studies spanning multiple orders of spatial and temporal magnitude- from individual meander bends (Dietrich et al., 1979; Kasvi et al., 2017), to the evolution of floodplains and valleys (Belmont, 2011; Gran et al., 2013; Schwenk et al., 2015), to development of a stratigraphic record spanning eons (Miall, 2006).
Accurately measuring river channel change from remotely-sensed imagery is also essential for estimating risk to infrastructure (Wente, 2000; Allan, 2004), mapping flood risk (Slater et al., 2015; Call et al., 2017), quantifying sediment loading, and improving the success of stream restoration/reclamation and riparian/watershed management.
The potential accuracy and precision of meander migration analyses have improved as the result of increased availability of historical and contemporary landscape-scale data (e.g., aerial photographs and high-resolution topography, HRT) for short (<1 year) and long (>50 years) timescales.
Availability of such data has supported a new wave of Quantitative approaches that have advanced our understanding of fluvial patterns, processes, and trends (Lindsay and Ashmore, 2002; Ghoshal et al., 2010; Donovan et al., 2015; Passalacqua et al., 2015), while also illuminating new challenges and gaps in our understanding of river morphology (Allan, 2004; Lawler, 1993).
While we focus on channel migration measured from aerial images, our insights are applicable to changes measured using other platforms, such as repeat topographic surveys, lidar, digitized images, and/or orthoimages.
A critical challenge arising in quantifying fluvial change from aerial imagery is documenting and accounting for measurement uncertainty (Unwin, 1995; Edwards and Lowell, 1996; Kiiveri, 1997).
Despite an abundance of remotely-sensed data and new capabilities enabled by continually evolving software packages, studies of fluvial change based upon remote sensing lack a robust and consistent methodology for quantifying and accounting for uncertainty (Kiiveri, 1997; Schook et al., 2017; Werbylo et al., 2017).
Several studies have provided recent advances to our understanding of uncertainty in measurements of channel width and lateral migration from remotely sensed imagery (Mount et al., 2003; Mount and Louis, 2005; Hughes et al., 2006; Lea and Legleiter, 2016; Werbylo et al., 2017).
Methods for measurement of river migration rates lags considerably behind other measurements of topographic change for which rigorous, repeatable and generalizable uncertainty methods have been developed and are routinely applied by researchers (Brasington et al., 2003; Wheaton et al., 2010; Passalacqua et al., 2015; Schaffrath et al., 2015; Bangen et al., 2016; Vericat et al., 2017; Anderson, 2018).
The first goal of this dissertation is to provide a comprehensive framework for evaluating uncertainty in estimates of river migration and width change by: (1) summarizing relevant research and methods for evaluating uncertainty; (2) highlighting and testing approaches used to estimate channel migration and uncertainty;
(3) systematically evaluating how spatial autocorrelations, riparian vegetation, and geomorphic conditions influence uncertainty; and (4) evaluating and improving techniques for dealing with measurements that fall below the minimum level of detection.
Beyond planform adjustment of river channels, the guidance and results presented herein are applicable to measuring changes in other delineated boundaries, including glacier retreat or advance, erosion or deposition along coastlines and lakeshores, changes in wetland extent, expansion or contraction of vegetation (e.g., deforestation), cliff retreat, and political boundary disputes.
Ensuring effective management of the river corridor requires that we appropriately quantify and report uncertainty in river migration measurements, lest we run the risk of inappropriately prescribing costly channel and riparian management strategies, including bank stabilization and invasive restoration or rehabilitation practices.
A second challenge that pervades hydrologic, geomorphic, and other environmental science research is the issue of temporal and spatial measurement scales. (Blöschl, 1996; Kirchner et al., 2001; Sadler and Jerolmack, 2015; Donovan and Belmont, 2019).
The rates of many landscape processes are unsteady over time and non-uniform in space (Ganti et al., 2016). Thus, the time and space scales over which we measure change may have an important influence on the outcome and can bias our ability to understand and predict change (Schumm and Lichty, 1965; Harvey, 2002).
Timescale dependence occurs when measurements of process rates are directly influenced by the timescale over which they are measured, leading to biased comparisons of rates measured over different time intervals. This, in turn, confounds our ability to untangle the complexity of environmental responses to external variables (Gurnell et al.,1994; Larsen et al., 2006; Micheli and Larsen, 2011; Gallen et al., 2015; Schook et al.,2017).
Timescale dependence has been demonstrated for a multitude of unsteady processes, including sediment accumulation, aggradation, progradation, and degradation (Sadler, 1981; Gardner et al., 1987; Lindsay and Ashmore, 2002; Kessler et al., 2013; Sadler and Jerolmack, 2015), river incision (Finnegan et al., 2014; Gallen et al., 2015),mountain erosion (Kirchner et al., 2001), cliff erosion (Cambers, 1976), and slope adjustments (Penning-Rowsell and Townshend, 1978).
Process hiatuses (e.g., rapid change followed by periods of dormancy) and reversals (e.g., incision vs. aggradation) appear to be largely responsible for timescale dependence across a variety of unsteady processes (Sadler, 1981; Gardner et al., 1987; Finnegan et al., 2014; Sadler and Jerolmack, 2015).
In the case of river migration, channel reversals may lead to underestimating measured migration rates by erasing part of the migration record between sequential aerial images. Despite this intuitive connection, the potential for timescale dependence in river migration measurements
not been previously addressed. In Chapter 3, we analyze empirical and synthetic datasets to address the following questions: Does timescale dependence exist for river migration measurements? If so, how does it affect our ability to accurately measure and compare changes in migration rates over time? What mechanisms cause measurement timescale dependence, and to what degree? Can timescale dependence and actual changes in channel migration be disentangled in order to determine if/when/where real changes in migration rates have occurred?
The second component of the measurement scale-space – reflects a third challenge that has resurfaced in new ways with the use of aerial imagery and software to quantify changes in river morphology. While aerial imagery archives and new measurement platforms allow us to track detailed changes across Earth’s surface at a variety of scales, the scale at which change is documented can will impact the results and may bias our interpretation of the driving mechanisms.
When measurements of meander migration are averaged over the scale of a meander bend, rates of river migration are observed to be largest for bends with a moderate degree of curvature (Hickin, 1974; Nanson and Hickin, 1983; Hickin and Nanson, 1984). However, if change is measured at smaller, sub-zero scales, rates of erosion are observed to continuously increased with curvature (Sylvesteret al., 2019).
Additionally, the spatially continuous sub-bend measurements and geospatial analyses revealed a spatial lag of about 2 to 5 channel widths between patterns of curvature and migration rate. The role of curvature as a driver of bend migration informs our assessment of the driving mechanics and appears to depend upon measurement scale (Furbish, 1988; Hickin and Nanson, 1984; Howard and Knutson, 1984; Nanson and Hickin, 1983).
The contrast between empirical measurements obtained at bend-averaged and sub-meander scales highlight the need to better understand how spatial scale impacts curvature-migration relationships in natural river meanders.
Chapter 4 explores how the spatial scale of measurement can impact curvature-migration relationships in natural river meanders. We examine the evolution of curvature and migration at sub-meander scales using repeated aerial images spanning large temporal (76 years) and spatial scales (25 river-km). Fine-scale measurements provide an opportunity to reevaluate the contrasting forms of curvature-migration relationships.
Specifically, we ask: is there empirical evidence that migration rates peak at a critical radius of curvature that is 2 to 3 times the channel width (R/W ~2-3), or if they exhibit a direct relationship between curvature and migration? If the latter, what form does the relationship take? Is the peaked relationship between migration and curvature an artifact of using bend-averaged measurements, which fail to capture sub-meander scale variability? We also evaluate whether there is a spatial lag between curvature and migration.
A clearer relation between bend curvature and migration rate can support a better understanding of the underlying mechanisms and an improved basis for predicting meander dynamics.