Search

Many studies of Earth surface processes and landscape evolution rely on having accurate and extensive data sets of surficial geologic units and landforms. Automated extraction of geomorphic features using deep learning provides an objective way to consistently map landforms over large spatial extents. However, there is no consensus on the optimal input feature space for such analyses. We explore the impact of input feature space for extracting geomorphic features from land surface parameters (LSPs) derived from digital terrain models (DTMs) using convolutional neural network (CNN)‐based semantic segmentation deep learning. We compare four input feature space configurations: (a) a three‐layer composite consisting of a topographic position index (TPI) calculated using a 50 m radius circular window, square root of topographic slope, and TPI calculated using an annulus with a 2 m inner radius and 10 m outer radius, (b) a single illuminating position hillshade, (c) a multidirectional hillshade, and (d) a slopeshade. We test each feature space input using three deep learning algorithms and four use cases: two with natural features and two with anthropogenic features. The three‐layer composite generally provided lower overall losses for the training samples, a higher F1‐score for the withheld validation data, and better performance for generalizing to withheld testing data from a new geographic extent. Results suggest that CNN‐based deep learning for mapping geomorphic features or landforms from LSPs is sensitive to input feature space. Given the large number of LSPs that can be derived from DTM data and the variety of geomorphic mapping tasks that can be undertaken using CNN‐based methods, we argue that additional research focused on feature space considerations is needed and suggest future research directions. We also suggest that the three‐layer composite implemented here can offer better performance in comparison to using hillshades or other common terrain visualization surfaces and is, thus, worth considering for different mapping and feature extraction tasks. Plain Language Summary: Characteristics of the land surface, such as steepness, relative slope position (e.g., ridge vs. valley), and roughness, can be digitally represented using a variety of methods. These digital representations, along with human‐annotated labels, can be provided to artificial intelligence algorithms to generate maps of landforms, such as those associated with river systems, glaciers, or human‐induced changes to the landscape. Given the large number of terrain derivatives that can be generated from digital elevation data, it is unclear how the chosen inputs impact the utility of the resulting maps. This study suggests that artificial intelligence mapping algorithms are sensitive to representations provided to them since different inputs resulted in varying map accuracies. We suggest a combination of features, which collectively describe relative slope position, steepness, and local terrain texture, as a means to represent the landscape and to serve as input to artificial intelligence algorithms. Our results can make it easier to train artificial intelligence algorithms to consistently and objectively find surficial features of interest across large swaths of Earth's surface. Key Points: We test the sensitivity of convolutional neural network geomorphic feature extraction approaches to the input feature spaceA three‐layer composite incorporating slope and topographic position index outperformed common inputs like hillshades and slopeshadesResults held true across four use cases including both natural and anthropogenic landforms [ABSTRACT FROM AUTHOR]

Passive margin stratigraphy contains time‐integrated records of landscapes that have long since vanished. Quantitatively reading the stratigraphic record using coupled landscape evolution and stratigraphic forward models (SFMs) is a promising approach to extracting information about landscape history. However, there is no consensus about the optimal form of simple SFMs because there has been a lack of direct tests against observed stratigraphy in well‐constrained test cases. Specifically, the extent to which SFM behaviour over geologic space and timescales should be governed by local (downslope sediment flux depends only on local slope) versus nonlocal (sediment flux depends on factors other than local slope, such as the history of slopes experienced along a transport pathway) processes is currently unclear. Here, we develop a nonlocal, nonlinear SFM that incorporates slope bypass and long‐distance sediment transport, both of which have been previously identified as important model components but not thoroughly tested. Our model collapses to the local, linear model under certain parameterizations such that best‐fit parameter values can indicate optimal model structure. Comparing 2‐D implementations of both models against seven detailed seismic sections from the Southeast Atlantic Margin, we invert the stratigraphic data for best‐fit model parameter values and demonstrate that best‐fit parameterizations are not compatible with the local, linear diffusion model. Fitting observed stratigraphy requires parameter values consistent with important contributions from slope bypass and long‐distance transport processes. The nonlocal, nonlinear model yields improved fits to the data regardless of whether the model is compared against only the modern bathymetric surface or the full set of seismic reflectors identified in the data. Results suggest that processes of sediment bypass and long‐distance transport are required to model realistic passive margin stratigraphy and are therefore important to consider when inverting the stratigraphic record to infer past perturbations to source regions. [ABSTRACT FROM AUTHOR]

Bedrock landslides shape topography and mobilize large volumes of sediment. Yet, interactions between landslide‐produced sediment and fluvial systems that together govern large‐scale landscape evolution are not well understood. To explain morphological patterns observed in steep, landslide‐prone terrain, we explicitly model stochastic landsliding and associated sediment dynamics. The model accounts for several common landscape features such as slope frequency distributions, which include values in excess of regional stability limits, quasi‐planar hillslopes decorated with straight, closely spaced channel‐like features, and accumulation of sediment in valley networks rather than on hillslopes. Stochastic landsliding strongly affects the magnitude and timing of sediment supply to the fluvial system. We show that intermittent sediment supply is ultimately reflected in topography. At dynamic equilibrium, landslide‐derived sediment pulses generate persistent landscape dynamism through the formation and breaching of landslide dams and epigenetic gorges as landslides force shifts in channel positions. Our work highlights the importance of interactions between landslides and sediment dynamics that ultimately control landscape‐scale response to environmental change. Plain Language Summary: Landslides are natural hazards that predominantly occur where geologic forces have created steep, rugged terrain. Yet from a geologic perspective, landsliding is also one of the forces that contributes to shaping the landscape. Gaining insight into the role of landslides in shaping topography is critical to better understand the interplay between the factors driving long‐term landscape evolution, and ultimately controlling landslide hazard. Here, we use a computer simulation model to study how landslides help shape terrain. Simulating landslides enables us to identify the origins of several common features in mountainous topography. Model results illustrate how repeated landsliding can cause rivers and ridgelines to gradually but continually shift position over time, such that some mountain drainages grow while others shrink. This type of computer model, which describes how topography and landsliding influence one another, can help the scientific community understand the response of steep terrain to environmental changes, such as those related to climate and earthquakes. Key Points: Over geological timescales, landslides produce diagnostic topographic signatures that can be detected in real‐world topographic dataStochastic landsliding results in slopes exceeding stability angles and quasi‐planar hillslopes decorated with closely spaced channelsThe combination of landslide stochasticity and channel‐hillslope feedback mechanisms causes persistent landscape dynamism [ABSTRACT FROM AUTHOR]

Long‐term erosion can threaten infrastructure and buried waste, with consequences for management of natural systems. We develop erosion projections over 10 ky for a 5 km2 watershed in New York, USA. Because there is no single landscape evolution model appropriate for the study site, we assess uncertainty in projections associated with model structure by considering a set of alternative models, each with a slightly different governing equation. In addition to model structure uncertainty, we consider the following uncertainty sources: selection of a final model set; each model's parameter values estimated through calibration; simulation boundary conditions such as the future incision of downstream rivers and future climate; and initial conditions (e.g., site topography which may undergo near‐term anthropogenic modification). We use an analysis‐of‐variance approach to assess and partition uncertainty in projected erosion into the variance attributable to each source. Our results suggest one sixth of the watershed will experience erosion exceeding 5 m in the next 10 ky. Uncertainty in projected erosion increases with time, and the projection uncertainty attributable to each source manifests in a distinct spatial pattern. Model structure uncertainty is relatively low, which reflects our ability to constrain parameter values and reduce the model set through calibration to the recent geologic past. Beyond site‐specific findings, our work demonstrates what information prediction‐under‐uncertainty studies can provide about geomorphic systems. Our results represent the first application of a comprehensive multi‐model uncertainty analysis for long‐term erosion forecasting. Plain Language Summary: Erosion of ground material is a hazard to buildings and other infrastructure, and can pose an environmental risk when it occurs in areas such as radioactive waste repositories and post‐industrial sites. We make projections for erosion over the next 10,000 years and assess uncertainty sources at a 5 km2 watershed in New York state. Natural systems, like the study site, are not as well understood as engineered systems. However, information from studies like this one can provide useful insights to guide management decisions. Parts of the watershed experienced up to 50 m of erosion over the past 13,000 years. The type of model used to simulate land surface evolution over thousands of years is a Landscape Evolution Model (LEM). We use a set of alternative LEMs identified by prior work as successful in simulating this watershed from 13,000 years ago to the present. We consider uncertainty in future climate, incision of downstream rivers, the models used and their parameters, and small changes to the initial topography. We find that just under 20% of the study site will experience erosion exceeding 5 m at +10,000 years. The most important uncertainty source varies across the watershed and increases with time. Key Points: Projections of erosion to 10 ky and uncertainty assessment demonstrate that the dominant uncertainty sources vary in space and timeTwo alternative approaches to model selection result in similar projections and uncertainty at 10 kyVariation across alternative models is subtle and initial conditions influence the location of gullies that drain low‐relief areas [ABSTRACT FROM AUTHOR]

Despite considerable community effort, there is no general set of equations to model long‐term landscape evolution. In order to determine a suitable set of landscape evolution process laws for a site where postglacial erosion has incised valleys up to 50 m deep, we generate a set of alternative models and perform a multimodel analysis. The most basic model we consider includes stream power channel incision, uniform lithology, hillslope transport by linear diffusion, and surface‐water discharge proportional to drainage area. We systematically add one, two, or three elements of complexity to this model from one of four categories: hillslope processes, channel processes, surface hydrology, and representation of geologic materials. We apply methods of formal model analysis to the 37 alternative models. The global Method of Morris sensitivity analysis method is used to identify model input parameters that most and least strongly influence model outputs. Only a few parameters are identified as important, and this finding is consistent across two alternative model outputs: one based on a collection of topographic metrics and one that uses an objective function based on a topographic difference. Parameters that control channel erosion are consistently important, while hillslope diffusivity is important for only select model outputs. Uncertainty in initial and boundary conditions is associated with low sensitivity. Sensitivity analysis provides insight to model dynamics and is a critical step in using model analysis for mechanistic hypothesis testing in landscape evolution theory. Key Points: Multiple landscape evolution models are developed for application in a well‐constrained site undergoing locally rapid (4 m/ka) erosionGlobal sensitivity analysis reveals that uncertainty in initial and boundary conditions has low impact on model output metricsGlobal sensitivity analysis finds that channel erosion parameters dominate and that select metrics are sensitive to hillslope diffusivity [ABSTRACT FROM AUTHOR]

We review select mature geomorphic transport laws for use in temperate ridge and valley landscapes and compile parameter estimates for use in applications. This work is motivated by a case study of sensitivity analysis, calibration, validation, multimodel comparison, and prediction under uncertainty, which required bounding values for parameter ranges. Considered geomorphic transport formulae span hillslope sediment transport, soil production, and erosion by surface water. We compile or derive estimates for the parameters in these transport formulae. Additionally, we address a common challenge—connecting changes in precipitation distribution to changes in effective erodibility—by using a simple hydrologic model and a method to estimate precipitation distribution parameters using commonly available data. While some parameters are reasonably well constrained, others span orders of magnitude. Some, such as soil infiltration capacity, have a direct physical meaning but are challenging to measure on geologically relevant timescales. Through the process of compiling these ranges we identify common challenges in parameter determination. The issue of comparable units derives from considering an exponent as an empirically inferred coefficient rather than as an expression of a fundamental relationship. The issue of appropriate timescales derives from the mismatch between human measurement and geologic timescales. This contribution thus serves both as a practical compilation for applications and as a synthesis of outstanding challenges in parameter selection for geomorphic transport laws. Key Points: We provide a compilation of parameter ranges for mature geomorphic transport laws in terrestrial geomorphologyA simple hydrological model connects changes in the precipitation distribution with erodibility coefficientsCommon challenges include incomparable units and mismatch between observation and simulation timescales [ABSTRACT FROM AUTHOR]

We present a multimodel analysis for mechanistic hypothesis testing in landscape evolution theory. The study site is a watershed with well‐constrained initial and boundary conditions in which a river network locally incised 50 m over the last 13 ka. We calibrate and validate a set of 37 landscape evolution models designed to hierarchically test elements of complexity from four categories: hillslope processes, channel processes, surface hydrology, and representation of geologic materials. Comparison of each model to a base model, which uses stream power channel incision, uniform lithology, hillslope transport by linear diffusion, and surface water discharge proportional to drainage area, serves as a formal test of which elements of complexity improve model performance. Model fit is assessed using an objective function based on a direct difference between observed and simulated modern topography. A hybrid optimization scheme identifies optimal parameters and uncertainty. Multimodel analysis determines which elements of complexity improve simulation performance. Validation tests which model improvements persist when models are applied to an independent watershed. The three most important model elements are (1) spatial variation in lithology (differentiation between shale and glacial till), (2) a fluvial erosion threshold, and (3) a nonlinear relationship between slope and hillslope sediment flux. Due to nonlinear interactions between model elements, some process representations (e.g., nonlinear hillslopes) only become important when paired with the inclusion of other processes (e.g., erosion thresholds). This emphasizes the need for caution in identifying the minimally sufficient process set. Our approach provides a general framework for hypothesis testing in landscape evolution. Key Points: Comparison of calibrated models at a well‐constrained field site serves as a formal hypothesis test of landscape evolution theoryRepresenting subtle lithologic contrasts, fluvial incision thresholds, and nonlinear hillslope transport improves simulation results mostModel structure improvements are not always linearly additive [ABSTRACT FROM AUTHOR]

Geomorphologists often rely on simple models of river channel incision for predicting rates of landscape evolution and channel response to perturbations, as well as extracting climatic and tectonic signals from river longitudinal profiles. Recent work has shown that large, hillslope‐derived blocks delivered to rivers may noticeably alter the form and evolution of river profiles from the behavior predicted by the most common models. Here we use a 1‐D model of river reach erosion and hillslope block delivery to explore the conditions under which block delivery strongly influences channel evolution. We use global sensitivity analysis to understand which model parameters most strongly affect the channel longitudinal profile. We explore the effects of blocks on the relationship between erosion rate and channel gradient, and on the erosion rate‐channel steepness exponent ϕ, and find that block effects result in highly variable slope and ϕ over the range of erosion rates and climatic conditions (discharge mean and variability) tested. The influence of blocks on erosion rate‐slope scaling may be approximated by a piecewise model: The erosion threshold imposed by blocks scales linearly with erosion rate when blocks are infrequently mobile and remains constant when blocks are frequently mobile. We explore the implications of this variable‐threshold model for the erosion rate‐channel steepness relationship and find that erosion rate‐dependent thresholds imposed by hillslope‐derived blocks cause significant departures from previous models but may be consistent with existing field data sets. Our work has implications for landscape evolution modeling and the inversion of channel profiles for forcing information. [ABSTRACT FROM AUTHOR]

Erosion processes in bedrock-floored rivers shape channel cross-sectional geometry and the broader landscape. However, the influence of weathering on channel slope and geometry is not well understood. Weathering can produce variation in rock erodibility within channel cross-sections. Recent numerical modeling results suggest that weathering may preferentially weaken rock on channel banks relative to the thalweg, strongly influencing channel form. Here, we present the first quantitative field study of differential weathering across channel cross-sections . We hypothesize that average cross-section erosion rate controls the magnitude of this contrast in weathering between the banks and the thalweg. Erosion rate, in turn, is moderated by the extent to which weathering processes increase bedrock erodibility. We test these hypotheses on tributaries to the Potomac River, Virginia, with inferred erosion rates from ~0.1 m/kyr to >0.8 m/kyr, with higher rates in knickpoints spawned by the migratory Great Falls knickzone. We selected nine channel cross-sections on three tributaries spanning the full range of erosion rates, and at multiple flow heights we measured (1) rock compressive strength using a Schmidt hammer, (2) rock surface roughness using a contour gage combined with automated photograph analysis, and (3) crack density (crack length/area) at three cross-sections on one channel. All cross-sections showed significant ( p < 0.01 for strength, p < 0.05 for roughness) increases in weathering by at least one metric with height above the thalweg. These results , assuming that the weathered state of rock is a proxy for erodibility, indicate that rock erodibility varies inversely with bedrock inundation frequency. Differences in weathering between the thalweg and the channel margins tend to decrease as inferred erosion rates increase, leading to variations in channel form related to the interplay of weathering and erosion rate. This observation is consistent with numerical modeling that predicts a strong influence of weathering-related erodibility on channel morphology. Copyright © 2017 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]