Retreat Rates Research Articles

Hydrological effects of multiphase water transformation in Three-River Headwaters Region, China

The present study focuses on the influence of multiphase water transformation on the hydrology of the Three-River Headwater Region. Climatic factors, accelerated transformation of water bodies, and changes in runoff patterns were evaluated based on long-term field observations. Subsequently, we estimated that changes in climatic factors have prolonged the ablation period and accelerated warming since 1990. Additionally, the glacier retreat rate during 2001–2013 was accelerated by approximately 16.23% compared with that during 1976–2001, whereas the ratio of snowfall in annual precipitation has decreased by 2.37% since 1990. The beginning of the freezing cycles of lakes along with the complete cycles were delayed. Moreover, snow sublimation significantly increased after 1990. Amplitudes of frost days (number of days with daily minimum temperature below 0°C) during 1991–2019 were eight times higher than those during 1960–1990. Furthermore, the contribution of moisture recycling to precipitation increased by 9.74% from1979 to 1990 to 1991–2019. The variations in actual evaporation increased from −1.257 mm/a during 1980–1990 to 0.865 mm/a during 1991–2017 and the runoff increased by 66.83×108 m3/10a during 1991–2012. July experienced the highest rainfall and runoff. During the multiphase water transformation, the runoff recharge sources changed significantly, thus, changing the historical trend of recharge in the headwaters of rivers only by the glacier snow meltwater by providing a new recharge mode. However, the accelerated multiphase water transformation caused by climate warming increased the challenges in water resource management and flood control in alpine regions.

Impacts of Engineered Diversions and Natural Avulsions on Delta‐Lobe Stability

AbstractReduced sediment supply and rising sea levels are driving land submergence on deltas worldwide, motivating engineering practices that divert water and sediment to sustain coastal landforms. However, lobe response following channel abandonment by diversions has not been constrained by field‐scale studies. Herein, avulsion and engineered diversion scenarios are explored for the Huanghe delta (China), where three lobes were abandoned in the last 40 yr. Two lobes were completely cut off by diversions, and one naturally by an avulsion. Shoreline retreat rates are strikingly different: ∼400 m/yr for diverted lobes and ∼90 m/yr for avulsed lobe. We hypothesize that this variability is linked to vegetal cover across lobes, and therefore the capacity to buffer hydrodynamic reworking of shoreface sediment. Furthermore, the vegetal cover is related to lobe salinity and elevation, which vary by abandonment style. We offer this as a case study to inform about the efficacy of future delta diversions.

Effects of Beach Nourishment Project on Coastal Geomorphology and Mangrove Dynamics in Southern Louisiana, USA

Relative sea-level (RSL) rise associated with decreased fluvial sediment discharge and increased hurricane activity have contributed to the high rate of shoreline retreat and threatened coastal ecosystems in Port Fourchon, Louisiana, USA. This study, based on QuickBird/drone images (2004–2019) and LIDAR data (1998–2013), analyzed the impacts of shoreline dynamics on mangroves (Avicennia germinans) and marshes before and after the initiation of a beach nourishment project in 2013. The coastal barrier and dune crest migrated landward between 1998 and 2013. Meanwhile, the dune crest height increased between 1998 and 2001, then decreased in 2013, probably due to hurricane impacts. The total sediment volume along this sandy coastal barrier presented an overall trend of decline in the 1998–2013 period, resulting in a wetlands loss of ~15.6 ha along 4 km of coastline. This has led to a landward sand migration onto muddy tidal flats occupied by Avicennia germinans (1.08 ha) and Spartina (14.52 ha). However, the beach nourishment project resulted in the advancement of the beach barrier from Nov/2012 to Jan/2015, followed by a relatively stable period between Jan/2015 and Mar/2019. Additionally, both the dune crest height and sediment volume increased between 2013 and 2019. This set of factors favored the establishment and expansion of mangroves (3.2 ha) and saltmarshes (25.4 ha) along the backbarrier environments after 2013, allowing the tidal flats to keep pace with the RSL rise. However, waves and currents caused shoreline erosion following the beach nourishment project between Oct/2017 and Nov/2019, threatening wetlands by resuming the long-term process of shoreline retreat.

The Role of Frost Processes in the Retreat of River Banks

The rate of bank retreat was measured using erosion pins on the alluvial banks of the rivers in the Podhale region (the boundary zone between Central and Outer Carpathians) during the hydrological year 2013/2014. During the winter half-year (November–April), the bank retreat was mainly caused by processes related to the freezing and thawing of the ground (swelling, creep, downfall). During the summer half-year (May–October), fluvial processes and mass movements such as lateral erosion, washing out, and sliding predominated. The share of fluvial processes in the total annual amount of bank retreat (71 cm on average) was 4 times greater than that of the frost phenomena. Erosion on bank surfaces by frost phenomena during the cold half-year was greatest (up to 38 cm) on the upper parts of banks composed of fine-grained alluvium, while fluvial erosion during the summer half-year (exceeding 80 cm) mostly affected the lower parts of the banks, composed of gravel. The precise calculation of the relative role of frost phenomena in the annual balance of bank erosion was precluded at some stations by the loss of erosion pins in the summer flood.

An application of three different field methods to monitor changes in Urumqi Glacier No. 1, Chinese Tien Shan, during 2012–18

AbstractThis study deploys RTK-GNSS in 2012, TLS in 2015 and UAV in 2018 to monitor the changes of Urumqi Glacier No. 1 (UG1), eastern Tien Shan, and analyzes the feasibility of three technologies in monitoring the mountain glaciers. DEM differencing shows that UG1 has experienced a pronounced thinning and mass loss for the period of 2012–18. The glacier surface elevation change of −0.83 ± 0.57 m w.e. a−1 has been recorded for 2012–15, whereas the changes of glacier tongue surface elevation in 2015–18 and 2012–18 were −2.03 ± 0.95 and −1.34 ± 0.88 m w.e. a−1, respectively. The glacier area shrunk by 0.07 ± 0.07 × 10−3 km2 and the terminus retreat rate was 6.28 ± 0.83 m a−1 during 2012–18. The good agreement between the glaciological and geodetic specific mass-balances is promising, showing the combination of the three technologies is suitable to monitor glacier mass change. We recommend application of the three technologies to assess each other in different locations of the glacier, e.g. RTK-GNSS base stations, ground control points, glacier tongue and terminus, in order to avoid the inherent limitations of each technology and to provide reliable data for the future studies of mountain glacier changes in western China.

Tree ring evidence of the retreat history of the Zepu glacier on the southeastern Tibetan Plateau since the Little Ice Age

Glaciers on the southeastern Tibetan Plateau, known as the “Asian Water Tower”, directly affect the utilization of water resources in downstream regions. The currently available instrumental climatic dataset covers only a short period, and the resolution of proxy-based climate reconstructions is low, and these limitations restrict the accurate exploration of glacial changes on the southeastern Tibetan Plateau within the context of long-term climate change. Here, annual-resolution tree ring samples were collected from the oldest pioneer trees (Hippophae rhamnoides subsp. yunnanensis), from the bottom of a glacial valley on the southeastern Tibetan Plateau, and used to reconstruct the glacial changes based on the relationship between the ages of the trees and their distances from the glacier terminal. Results indicate that Zepu glacier has been retreating since 1778 CE, with a total retreat distance of 2757 m and an average retreat rate of 11 m yr−1. The rate of retreat has varied over time, with a maximum of 40 m yr−1 between the 1950s CE and 1970s CE, and glacial retreat has been modulated mainly by temperature variability over interdecadal to centennial timescales. This study provides a new way to reconstruct the history of glacial retreat.

Monitoring the Spatiotemporal Difference in Glacier Elevation on Bogda Mountain from 2000 to 2017.

The difference in glacier surface elevation is a sensitive indicator of climate change and is also important for disaster warning and water supply. In this paper, 25 glaciers on Bogda Mountain, in the eastern Tianshan Mountains, are selected as the study object as they are typical of glaciers in arid or semi-arid areas with importance for water supply. The Repeat Orbit Interferometry (ROI) method is used to survey the surface elevation of these glaciers using Sentinel-1A Radar data from 2017. Using data from the Shuttle Radar Topography Mission (SRTM) and a Digital Elevation Model (DEM), the difference in the glacier surface elevation between 2000 and 2017 is obtained. A scheme to evaluate the accuracy of estimated variations in glacier surface elevation is proposed in this article. By considering the surfaces of lakes in the study region as ideal horizontal planes, the average standard deviation (SD) value of the lake elevation is taken as the error caused by the radar sensor and observing conditions. The SD of the lake elevation is used as an index to evaluate the error in the estimated variation of the glacier surface elevation, and the obtained SD values indicate that the result obtained using the ROI method is reliable. Additionally, the glacier surface elevation variation pattern and a Logarithmic Fitting Model (LFM) are used to reduce the error in high-altitude glacial accumulation areas to improve the estimation of the difference in the glacier surface elevation obtained using ROI. The average SD of the elevation of the 12 lakes is ±2.87 m, which shows that the obtained glacier surface elevations are reliable. This article concludes that, between 2000 and 2017, the surface elevation of glaciers on Bogda Mountain decreased by an average of 11.6 ± 1.3 m, corresponding to an average decrease rate of 0.68 m/a, and glaciers volume decreased by an average of 0.504 km3. Meanwhile, the surface elevations of the lakes increased by an average of 8.16 m. The decrease of glacier surface elevation leads to the expansion of glacial lakes. From the north slope clockwise to the south slope, the glacier elevation variation showed a decreasing trend, and the elevation variation gradually increased from the south slope to the north slope. With the increase of glacier altitude, the variation of glacier surface elevation gradually changed from negative to positive. The findings of this article suggest that the rate of glacier retreat on Bogda Mountain increased from 2000 to 2017.

Mass balance and morphological changes of Dokriani Glacier (1992–2013), Garhwal Himalaya, India

Glacier mass balance measurements is an important element in understanding the processes of accumulation and ablation over a period of time that depend on the climatic conditions and geometry of the glacier. Glacier change can be through field, ground and aerial photographic measurements and high-resolution remote sensing at local, regional and global scale. In this study we presents a long time series ground observations of the mass balance and length change of Dokriani Glacier for the period 1992–2013. A medium sized glacier (6.6 km2) characterised by large accumulation area with narrowing ablation is located in the monsoon dominated zone of Central Himalaya, India. The spatial variability in the mass balance gradient profile of the two different study periods (i.e. 1992-2000 and 2007–2013) show consistency, but has shift upwards in the recent year. The glacier has higher value of AARn (67%) which attributed to a flat and broader accumulation area. Although, having a larger accumulation area compared to the ablation area, the glacier has strong mass wasting with an average annual ablation of −1.85 m w. e. a−1 as compared to the residual average annual accumulation of 0.44 m w. e. a−1. The recession rate of 21 m a−1 during 2007–2013 slightly higher than the retreat rate of 18 m a−1 during 1992–2000 and 16 m a−1 during 2000–2007. In addition, the state of Dokriani Glacier, based on accumulation-area ratio (i.e. AARn and AAR0) suggests that glacier in disequilibrium having decrease its accumulation area. The progressive retreat rate and continuous negative mass balance of the glacier affects its extension and volume. The overall morphological changes that includes reduction ∼14% of total glacier area and enhancement of debris over the ablation zone (∼4%) during the study period is significant. The present study is a part of long ground based observations on retreat of glacier in the Himalayan region.

Determining the Shoreline Retreat Rate of Australia Using Discrete and Hybrid Bayesian Networks

AbstractEvaluating shoreline retreat rate (SRR) on different spatial‐temporal scales is critical for effective coastal management. Large‐scale evaluations typically rely on data‐driven methods such as Discrete Bayesian networks (BNs). However, these BNs require discretization of continuous variables which can lead to information loss. Here, we propose a new method, the Hybrid BN to incorporate continuous variables without discretization. Both Discrete and Hybrid BNs were developed and compared to evaluate large‐scale (continental scale) SRR in Australia, using Digital Earth Australia data set. These BNs used forcing parameters (e.g., waves, tide, sediment sink/source, and sea level rise [SLR]) and geomorphic settings (e.g., geomorphology, backshore profile, and surfzone slope) to predict SRR. Validation of the BNs showed that Hybrid BNs, which provide a more realistic assessment of the range of SRR, outperform in predicting continuous variables, when compared with Discrete BNs. However, Discrete and Hybrid BNs provide consistent qualitative findings for the SRR of Australia. Among forcing parameters, the sediment sink/source was found to be the most informative variable to indicate the shoreline retreat, followed by tide, SLR rate, and wave processes. In the scenario of an increased SLR rate, tropical tidal flats were predicted as the most at risk coasts in Australia. We found that BNs can reflect the impact of different factors on coastal evolution, and predict future shoreline change by exploring historical data. The performance of these models can be further improved when more data sets become available.

Automating coastal cliff erosion measurements from large-area LiDAR datasets in California, USA

Quantifying coastal cliff erosion is critical for improved predictions of coastal change and coastal management. However, few studies have been conducted at a scale (>100 km) and resolution (~1 m) sufficient to constrain regional change. Here, we quantified cliff erosion for 866 km of the California coastline using airborne LiDAR data collected in 2009–2011 and 2016. A semi-automated method was used to map cliff faces. Negative (volume loss) and positive (volume gain) change objects were created by grouping adjacent cells using vertical and areal change thresholds and surface optical signatures. We assessed the performance of five machine learning algorithms to separate erosion and deposition from other changes within the cliff face, notably vegetation loss and growth, and found that discriminant analysis performed best. After applying the classification method to the entire cliff change dataset, the results were visually inspected for quality control, producing a final dataset comprised of 45,699 erosion and 1728 deposition objects. The net volume loss from 2009–2011 to 2016 was 1.24 × 107 m3, equivalent to an erosion rate of 2.47 m3 yr−1 per meter of coastline, and an average cliff retreat rate of 0.06 m yr−1. Eroded volumes ranged from 6.43 m3 to 7.52 × 105 m3 and fit a power-law frequency distribution (β = 0.80; r2 = 0.99). Over this study period, 7% of eroded material remained on the cliff face. Cliff retreat rates varied spatially with the highest rates in Humboldt County (0.18 m yr−1) and the lowest in Orange County (0.003 m yr−1).

Spatio-temporal changes in the Machoi glacier Zanskar Himalaya India using geospatial technology

We present the temporal changes of Machoi glacier, Zanskar region of the north-western Himalaya, India using multi-temporal Landsat satellite data from 1973 to 2018. The results suggest that the Machoi glacier has retreated consistently from last 45 years with an uneven retreat rate. The temporal fluctuations since 1973 AD reveal that the glacier snout has retreated ∼563 m at an average of ∼12.51 m yr−1. However, the snout has retreated at variable rates during different time intervals e.g., 80 m (average of ∼11.43 m yr−1) from 1973 to 1980 AD, 120 m (average of ∼12.0 m yr−1) from 1980 to 1990 AD, 123 m (average of ∼12.3 m yr−1) from 1990 to 2000 AD, 128 m (average of ∼12.8 m yr−1) from 2000 to 2010 AD and 112 m (average of ∼14 m yr−1) from 2010 to 2018AD. The highest rate of snout (∼14 m yr−1) retreat during 2010–2018 AD is linked to the changing climate in the region as observed from the meteorological data. The analysis of meteorological data suggests that during this period temperature as well as the liquid precipitation have increased due to global warming whereas the extent of solid precipitation has decreased which might be the possible causes of higher retreat of Machoi glacier.

Caves and bats: Morphological impacts and archaeological implications. The Azé Prehistoric Cave (Saône-et-Loire, France)

The Prehistoric Cave at Azé (France) was divided in two parts by calcite flowstone. This isolated the innermost parts of the cave from the entrance, which remained connected to the surface. Since the closure, bats have not been able to access the cave beyond the flowstone blockage. They have been present only in the cave entrance. The byproducts released by bats had a considerable effect on this entrance part. Gaseous exhalations have changed the composition of the surrounding atmosphere, leading to the development of dome-shaped cupolas and other wall weathering features. Guano deposits built up causing the formation of aerosols and acidic leachates. The resulting corrosion of the limestone has led to the formation of phosphate deposits and biogenic karst. The rate of retreat of the cave walls has been estimated at 5 to 7 mm/ka. Corrosion has also affected archaeological artefacts, as well as any traces left on the walls by humans or animals. Bioglyphs linked to bears as well as anthropic graffiti have completely disappeared from the part of the cave used by bats.

Marine ice-cliff instability modeling shows mixed-mode ice-cliff failure and yields calving rate parameterization

Marine ice-cliff instability could accelerate ice loss from Antarctica, and according to some model predictions could potentially contribute >1 m of global mean sea level rise by 2100 at current emission rates. Regions with over-deepening basins >1 km in depth (e.g., the West Antarctic Ice Sheet) are particularly susceptible to this instability, as retreat could expose increasingly tall cliffs that could exceed ice stability thresholds. Here, we use a suite of high-fidelity glacier models to improve understanding of the modes through which ice cliffs can structurally fail and derive a conservative ice-cliff failure retreat rate parameterization for ice-sheet models. Our results highlight the respective roles of viscous deformation, shear-band formation, and brittle-tensile failure within marine ice-cliff instability. Calving rates increase non-linearly with cliff height, but runaway ice-cliff retreat can be inhibited by viscous flow and back force from iceberg mélange.

Self‐Organization of Coastal Barrier Systems During the Holocene

AbstractUnderstanding the response of coastal barrier systems to sea level rise is a crucial societal need. Despite the problem having been studied extensively, major knowledge gaps remain. For example, neither the sedimentary record nor existing numerical models have been conclusive in explaining the formation of barrier islands. Here I present a comprehensive 2D model that seamlessly couples cross‐shore and along‐shore transport, tidal transport, storm surges, and wind waves, and use it to simulate an idealized passive margin during the last 7,000 years. In the early Holocene, when sea level was rising ∼20 mm/yr, shoals and ephemeral barrier islands formed, periodically drowned, and then formed again at a landward location. Shoal emergence was triggered by the disequilibrium of the recently submerged shelf, especially for large waves and mild shelf slopes. About 5,000 years ago, as sea level rise slowed down to ∼1 mm/yr, barriers stabilized and even prograded seaward. The combination of excess sediment in the nearshore and storm surges allowed barriers to accrete above mean high water. When barriers eventually equilibrated to the new sea level rise rate and started to retreat, their retreat rate was highly variable in space and time due to autogenic processes such as inlet formation and backbarrier channel interception. This variability also included multi‐decadal periods of localized progradation. Both lag dynamics and autogenic processes confound the relationship between barrier retreat and sea level rise rate.

Glacier Retreat in Iceland Mapped from Space: Time Series Analysis of Geodata from 1941 to 2018

In recent decades, glaciers outside Greenland and Antarctica have shown increasingly rapid rates of mass loss and retreat of the ice front, which is associated with climatic and oceanic warming. Due to their maritime location, Icelandic glaciers are sensitive to short-term climate fluctuations and have shown rapid rates of retreat and mass loss over the last decade. In this study, historical maps (1941–1949) of the US Army Map Service (AMS series C762) and optical satellite imagery (Landsat 1, Landsat 5, Landsat 7, Landsat 8, and Sentinel-2) are used to study the Langjökull, Hofsjökull and Vatnajökull ice caps. By the help of the Normalized Difference Snow Index (NDSI), the glacier terminus fluctuations of the ice caps from 1973 to 2018 and the Equilibrium Line Altitude (ELA) from 1973 to 2018 are analyzed. The results are compared with climate data, especially with mean summer temperatures and winter precipitation. Due to the negative temperature gradient with increasing altitude, bivariate histograms are generated, showing the glaciated area per altitude zone and time, and providing a prediction of the future development until 2050 and beyond. The results indicate that Langjökull, Hofsjökull and Vatnajökull are retreating and advancing over the study period in correlation with the mean summer temperature, with a steady decrease over time being the clearest and most significant trend. The lower parts of the glaciers, thus, will probably disappear during the next decades. This behaviour is also evident by an exceptional increase of the ELA observed on all three glaciers, which leads to a reduction of the accumulation zone.

Seasonal to decadal scale shoreline changes along the Cameroonian coastline, Bay of Bonny (1986 to 2020)

The Bay of Bonny, located along the Cameroon coastline, is home to diverse ecosystems. It is under significant pressure from large human activity, but remains very little studied, like much of the Gulf of Guinea. In order to understand its long-term shoreline variations and the role the wave regime plays in the evolution of the coastline, a study was conducted on the basis of optical image archives from Landsat 5∕7∕8 and Sentinel-2A/2B satellite missions acquired between October 1986 and May 2020, coupled with daily ERA-Interim wave re-analysis data covering the period from January 1986 to August 2019. Overall, the results show that the evolution of the coast is highly variable in space and time, as indicated by different levels of erosion (30.55 %), and accretion (27.7 %) on the decadal-scale, with the most significant variations occurring in estuarine areas. Nevertheless, 41.75% of the Cameroon shoreline remains stable during the study period. Three main periods (1986–1994; 1995–2005; 2006–2020) during which the coast underwent significant changes a different location were identified, reaching a retreat rate of up to −10 m/year in the northern section during the first period 1986–1994. The annual trend of significant wave heights anomaly along the coast (−5.6 to −4.1 mm/year) with wave height maxima estimated at 1.46 ± 0.65 m, where observed during the summer months (July–August). Monthly shoreline changes are inversely correlated with wave climate in some segments. Eigenvalue orthogonal decomposition analyses (Mode 1, Empirical Orthogonal Function) show that 76.3% of the observed variability would be due to the relatively strong local influence of erosion and accretion. This local influence can be related to wave regime at shorter timescale (monthly to seasonal) and the sediment variability from the source inland to the coast at longer timescale (seasonal to decadal). These observations explain a complex pattern of shoreline changes with an almost continuous retreat shoreline during the period 1986–2013 and a reversal in trend towards accretion during the period 2013–2020. EOF mode 2 explains 23.7% of variability, which can be potentially associated with the synergetic relationship between tidal currents and wave-induced longshore current, tidal currents and possibly pockets of human activity. This is attributed to an overall decrease in the supply of sediment via the coastal transport system that prevails in the Gulf of Guinea.

Utilizing airborne LiDAR data to quantify marsh edge morphology and the role of oyster reefs in mitigating marsh erosion

Marsh habitats, experiencing accelerated change, require accurate monitoring techniques. We developed methods to quantify marsh edge morphology using airborne LiDAR data. We then applied these methods within the context of oyster reef restoration within the shallow coastal bays of Virginia, USA, by comparing retreat and morphology quantified at paired reef-lined and control marsh edges at 10 different marsh sites. Retreat metrics were analyzed between 2002 and 2015, utilizing a LiDAR derived edge for the year 2015 from points of maximum slope and aerial imagery pre-2015. Retreat was also compared before and after oyster reef restoration to determine if reefs slow erosion. We found that slope statistics from airborne LiDAR elevation data can accurately capture marsh edge morphology. Retreat rate, measured at edges typically found near the vegetation line, was not significantly different between reef-lined and control marshes and ranged from 0.14 to 0.79 m yr-1. Both retreat rate (ρ = -0.90) and net movement (ρ = -0.88) were strongly correlated to marsh edge elevation. Exposed control marshes had significantly greater mean and maximum slope values compared to reef-lined marshes. The mean edge slope was 11.4° for exposed marshes and 6.0° for reef-lined marshes. We hypothesize that oyster reefs are causing an elongation of the marsh edge by reducing retreat at lower elevations of the marsh edge. Therefore, changes in marsh edge morphology may be a precursor to changes in marsh retreat rates over longer timescales and emphasizes the need for repeated LiDAR measurements to capture processes driving marsh edge dynamics.

How does land use/cover influence gully head retreat rates? An in‐situ simulation experiment of rainfall and upstream inflow in the gullied loess region, China

AbstractLand use/cover plays a crucial role in gully head retreat (GHR). However, little is known about how land use/cover influences GHR rates. An in situ simulation experiment of rainfall and upstream inflow was conducted in a gullied loess region to investigate hydraulic erosion, mass movements, and linear GHR processes under three types of land use/cover (bare land, grassland, and shrub‐grass land). The results showed that the average linear GHR rates on grassland and shrub‐grass land were 87–89% and 72–81% lower than that on bare land (0.99–2.06 cm min−1), respectively. Gully heads retreat by hydraulic and gravitational erosion. In the case of hydraulic erosion, upstream runoff incision is dominant on bare land, while undercutting by on‐wall and jet flow dominates on grassland and shrub‐grass land. In the case of mass movement, collapse dominates with a frequency of 62–100%, of which gully sidewall collapse is most common, especially on bare land and it acts to widen the gully. Gully headwall collapse dominates on grassland and shrub‐grass land to cause retreat of the gully head. Overall, on bare land, upstream runoff incision dominates GHR. However, on grassland/shrub‐grass land GHR is mainly driven by the undercutting of on‐wall and jet flow and subsequent gully headwall collapses. As a consequence, the GHR length on bare land exponentially increases over time, while on grassland or shrub‐grass land, it discretely increases as an analogous step function. Moreover, the average linear GHR rate on grassland was considerably lower than that on shrub‐grass land, implying that an optimized vegetation restoration pattern enhances GHR control.

The validation of a gully headcut retreat model in short-term scale based on an in-situ experiment in dry-hot valley

How to predict gully erosion rates was one of the most important issues at present. A model implemented in the Channel Hillslope Integrated Landscape Development (CHILD) model based on the shape factor (Sf) of the plunge pools, which successfully predicted the headcut retreat rates over a long timescale. But whether this model was also valid in the short-term scale? What factors affected the prediction effects of the model? These issues still need evaluation. Four gully head plots with consistent initial topography and soil types were constructed for conducting in-situ scouring experiments with different flow discharges. Eighteen tests were conducted in each plot to monitor the plunge pool growth and headcut retreat rates using an unmanned aerial vehicle (UAV) to create high-precision topographic data. The topographic parameters of the plunge pools generally increased with scouring duration and were strongly correlated with the total runoff energy consumption (0.55 < R2 < 0.92, P < 0.01). In contrast, headcut retreat was an intermittent process that suddenly migrated in a short time after relatively long stable periods lasting from 20 to 130 min. The relationships between the headcut retreat rates and total energy consumption were weaker than those with the plunge pools. As the timescale shortened, the contribution of mass failure due to gravity to headcut migration significantly increased, leading to the pool effects of the model to influence the prediction of the gully headcut retreat rates, which assumed that flow hydraulics were the key dynamics.

VEdge_Detector: automated coastal vegetation edge detection using a convolutional neural network

ABSTRACT Coastal communities, land covers, and intertidal habitats are vulnerable receptors of erosion, flooding or both in combination. This vulnerability is likely to increase with sea level rise and greater storminess over future decadal-scale time periods. The accurate, rapid, and wide-scale determination of shoreline position, and its migration, is therefore imperative for future coastal risk adaptation and management. This paper develops and applies an automated tool, VEdge_Detector, to extract the coastal vegetation line from high spatial resolution (Planet’s 3 to 5 m) remote-sensing imagery, training a very deep convolutional neural network (holistically nested edge detection), to predict sequential vegetation line locations on annual to decadal timescales. Red, green, and near-infrared (RG-NIR) was found to be the optimum image spectral band combination during neural network training and validation. The VEdge_Detector outputs were compared with vegetation lines derived from ground-referenced positional measurements and manually digitized aerial photographs, which were used to ascertain a mean distance error of <6 m (two image pixels) and >84% producer accuracy (P A) at six out of the seven sites. Extracting vegetation lines from Planet imagery of the rapidly retreating cliffed coastline at Covehithe, Suffolk, United Kingdom, has identified a landward retreat rate >3 m year−1 (2010–2020). Plausible vegetation lines were successfully retrieved from images in The Netherlands and Australia, which were not used to train the neural network, although significant areas of exposed rocky coastline proved to be less well recovered by VEdge_Detector. The method therefore promises the possibility of generalizing to estimate retreat of sandy coastlines from Planet imagery in otherwise data-poor areas, which lack ground-referenced measurements. Vegetation line outputs derived from VEdge_Detector are produced rapidly and efficiently compared to more traditional non-automated methods. These outputs also have the potential to inform upon a range of future coastal risk management decisions, incorporating future shoreline change.