High-resolution Digital Elevation Models Research Articles

Gully Head-Cuts Inventory and Semi-Automatic Gully Extraction Using LiDAR and Topographic Openness—Case Study: Covurlui Plateau, Eastern Romania

The Covurlui Plateau, a subunit of the Moldavian Plateau located in eastern Romania, possesses a high natural agricultural potential, significantly impacted by soil erosion, particularly gully erosion. The only inventory in the Moldavian Plateau that comprises approximately 9000 gullies extracted from topographical maps was conducted during the 90s. Nowadays, with the advent of advanced techniques and geodata, such as GIS software, aerial photograms, high-resolution satellite images, and high-resolution digital elevation models, we aim to achieve an (1) up-to-date comprehensive inventory of gully head-cuts and (2) a very detailed mapping of the spatial distribution of gullied lands. Firstly, the gully head-cuts were inventoried for the entire region using platforms such as Google, Esri, and Bing, through the QuickMapService plugin within QGIS 3.16 software, with the assistance of Landsat and Sentinel satellite images. Secondly, the automatic mapping of gullies was carried out using a 5 m high-resolution Digital Elevation Model and the Topographic Openness module offered by SAGA GIS software through QGIS software. As a result, we identified 5868 gully head-cuts for the Covurlui Plateau, with an average density of 2.57 gully head-cuts per square kilometer. Additionally, the identified gullies occupy over 3570 hectares, representing 1.57% of the total area. Overall, the topographic openness index proves to be an efficient tool for the semi-automatic extraction of gullies from high-resolution digital elevation models.

Improving filtering methods based on the Fast Fourier Transform to delineate objective relief domains: An application to Mare Ingenii lunar area

A recent study has proven that high-pass filtering (HPF) based on the Fast Fourier Transform (FFT) is a rapid and efficient computational method for the semi-automated detection of geomorphic features from high-resolution digital elevation models (DEM). Although this new approach shows great potential for cartographic purposes using remote sensing data, some methodological improvements are still required in the following areas: (i) to develop a robust criteria for filter radius selection; (ii) to test the relationship between filter vectors and landscape form, and explore how DEM artefacts (vegetation, anthropic structures, etc.) can interfere with landform detection; and (iii) to explore filter response regarding generalisation and blurring effects when working with landscapes composed of landforms of different scales that are superimposed on one another. These topics are addressed here through two experiments (Experiment_1 and Experiment_2) with synthetic digital relief models inspired in the lunar landscape. Finally, the improved methodology was applied on the Mare Ingenii lunar relief (Experiment_3) using the Lunar Orbiter Laser Altimeter DEM and the results were tested against ground truths (GTs) developed using the extensive database available at Astropedia website and an ad hoc crater map. The analysis of existing frequencies in a 2D DEM signal through the true magnitude-true frequency plot provides an objective method for filter radius selection, and the use of a Butterworth transference function enables a more versatile filtering. Experiment_1 demonstrates a close correspondence between vectors obtained by filtering called Filtered Geomorphic References (FGRs) and the synthetic landform selected. The accuracy indicators from Experiment_1 and 2 show the good results obtained in the correspondence between FGRs and crater depressions, either from flat-bottomed to bowl shapes. Experiments 2 and 3 confirm that in landscapes generated by superimposed geomorphic features of different sizes, the smaller the crater, the better the filters detect its boundaries. Moreover, the spatial repeatability of FGRs can be used as a cartographic criterion in the identification of crater shape depressions or hills. Besides, the criterion is useful to assess true reality mapped in the GT employed. Finally, the objective geomorphic units obtained by combining the FGRs demonstrate their usefulness for the objective characterisation of the moonscape. Using the synthetic landscapes, the FGRs identify those relief domains composed of depressions and hills.

A multi-terrain feature-based deep convolutional neural network for constructing super-resolution DEMs

Scale conversion between DEMs is an important issue in geomorphometry. There are many mature studies on the generation of low-resolution(LR) DEMs from high-resolution(HR) DEMs. However, as an important and convenient means of obtaining HR DEMs, traditional super resolution (SR) methods have shown insufficient consideration of the terrain features embedded in DEMs. Therefore, this article investigates the combination of terrain features and the use of convolutional neural networks (CNN) to reconstruct HR DEMs, and proposes a multi-terrain feature-based deep CNN for super-resolution(SR) DEMs (MTF-SR). In the experiments, from the perspective of vector and raster terrain features, we fuse raster terrain features in the input and loss functions, and fuse vector terrain features in the optimization of the output of the model. The results show that the MTF-SR model has a 30–50 % reduction in mean absolute error (MAE) compared with interpolation methods, has the lowest slope and aspect error and has a 10 to 30 % improvement in streamline matching rate (SMR). These results point to the advantages of the method in overall elevation accuracy and the preservation of terrain features.

MORPHOMETRIC ANALYSIS OF LLAP RIVER WATERSHED (KOSOVO)

The Llap River is main tributary of Sitnica River, having a watershed of 780.23 km2 which extends in northeast part of Republic of Kosovo. In this study, GIS and a high-resolution Digital Elevation Model have been utilized for estimation and analysis of t

Fault structure and segmentation of the Punta Montalva Fault in southwestern Puerto Rico using high-resolution digital elevation models

The Punta Montalva Fault is an active fault that slipped during the 2019–2023 Puerto Rico seismic sequence. It lies in southwestern Puerto Rico within the ∼250-km-wide deformation zone of the North American-Caribbean plate boundary. The fault cuts Miocene limestone and contains surface features that suggest left-lateral displacement. Previous work suggested that the fault connects with the North Boquerón Bay Fault for a length of 33 km and poses a large seismic hazard. However, not much is known about the structure, segmentation, and spatial extent of the fault as it is obscured under thick vegetation. This study used high-resolution topographic data to map the structure, segmentation, and spatial extent of the fault. The results would allow us to better assess the seismic potential of the Punta Montalva Fault by constraining its spatial extent and how strain has been accommodated. A 0.5 m resolution bare-earth digital elevation model (DEM) derived from a 2018 airborne LiDAR survey and another 1 m resolution 2018 DEM showing shallow bathymetry were used to map the Punta Montalva Fault. High-resolution mapping revealed that the Punta Montalva Fault is comprised of three main fault segments on land, a possible additional scarp, and three fault scarps on the seafloor to the SE. No evidence suggests the Punta Montalva Fault extends further NW of Punta Montalva into La Parguera in Lajas. The DEM revealed that the fault cuts a shallowly dipping sedimentary sequence of the Ponce Limestone that has karst topography in places. On land, the fault is conspicuous where it forms scarps, drags and cuts sedimentary layers, and forms a duplex structure with a predominantly left-lateral offset. A left-lateral offset of an intermittent stream has been refined to 133.6 ± 7.3 m. The fault scarps along a marine shelf strike E-W and show a maximum normal offset of 4.8 m. Another sub-parallel fault scarp in the seafloor strikes NW-SE and displays a left-lateral offset of 98.3 m. The total length of the fault segments is ∼2.67 km long. The surface segmentation suggests that the fault would have a lower seismic hazard than if it were continuous and connected to the North Boquerón Bay Fault. However, if the Punta Montalva Fault is continuous beneath the surface, its total fault length would be approximately 7.97 km long and would pose a larger seismic hazard than that suggested by the surface segmentation. Strike-slip, folding, and normal offset suggest that the area is being deformed by transtentional deformation, which supports seismic analyses and tectonic models.

Study of a Steady-State Landscape Using Remote Sensing and Topographic Analysis

The current limited approaches to calculating hillslope erosion rate hamper the study of the relationships among the rates of hillslope erosion, river incision, and tectonic uplift and hence the discussion of steady-state landscape evolution. In this paper, we use remote sensing and geochronological methods to calculate the upper and lower bounding hillslope erosion rates in the Qilian Shan range, Tibet. Our analysis focuses on five convex landslide sediment units derived from the weathered hillslopes at Qingyang Mountain on the tectonically active northeastern Tibetan Plateau. These sediment units range in thickness from 5.5 to 12.8 m and in volume from 119 × 103 to 260 × 103 m3. Based on field observations, measurements extracted from high-resolution DEMs, and optical stimulated luminescence (OSL) ages on fluvial terraces, we obtain lower and upper bounding rates of 0.13 ± 0.03 and 0.21 ± 0.04 mm/yr, respectively. Finally, we calculate incision rates, ranging from 0.21 ± 0.02 to 0.39 ± 0.01 mm/yr, from heights of a dated fluvial terrace above the present river and the time of abandonment of the associated bedrock strath estimated from OSL ages. The rates of hillslope erosion and river incision at Qingyang Mountain and the tectonic uplift of the Qilian Mountains are estimated to be within a factor of two over the past 117 ka, suggesting that a state of dynamic equilibrium has likely existed on this timescale.

Groundwater flooding hazard assessment in a semi-urban aquifer through probability modelling of surrogate data

Many areas of the world are experiencing groundwater flooding (GF) events with serious impacts on the environment and human health. In Europe, GF has received specific attention with the Directive 2007/60/EC, which requires Member States to map GF hazard and risk and propose mitigation measures. In Italy, despite the commitment assigned by the Legislative Decree 49/2010 to River Basin District Authorities, no models have been developed for groundwater flooding hazard (GFH) evaluation also for the lack of standard methods based on commonly available data. In this paper, a new method has been developed through a specific procedure that accounts for i) probability of GF events estimation based on available time series of groundwater levels and ii) probability mapping through a newly developed MATLAB™ code, high-resolution topographic data and groundwater table model. The methodology has been tested and validated in the Metropolitan City of Naples (Italy), affected by GF since 2010, for which i) sixty-two-year long time series of groundwater levels, ii) high-resolution Digital Elevation and Digital Terrain Models and iii) basement/building geometry, are available. The results show that about 59% of the buildings present in the study area have potentially floodable underground structures. In the 2010–2021 period, only 2% of these buildings were flooded and most of them were built in the period 1980–2000, when the piezometric levels were deeper due to over-pumping of the groundwater. The obtained GFH map shows that only 2% of the built area is involved by 1- to 2- year events, while 6% is affected by 1- to 10- year events. A good spatial correspondence between the return period and locations of GF events occurred during 2010–2021 period was observed, thus indicating the significant performance of the model. The obtained GFH map output can support decisions of authorities and policy-makers to reduce the GF risk.

Geoinformatics Perspective of Landslide and Catastrophic Flash Floods in Dhauliganga, Uttarakhand, India

On 07 February 2021, around 10:30 hrs local time catastrophic flash flood occurred in the Dhauliganga River (a tributary of the Ganga River) near Rini village at 2000 m above MSL (mean sea level) (Chamoli District), which killed 79 people and about 125 people were missing. Part of the area belongs to Nanda Devi Biosphere Reserve, which is completely protected from human interventions. Further, on the Dhauliganga River, two run-of-river hydroelectric power Rishiganga Small Hydro (13.2 MW) at 1975 m above MSL and Tapovan Vishnugad (520 MW) at 1795 m above MSL projects were also severely damaged due to the devastating flash flood. More than 150 workers were also trapped in the under-construction power tunnel of the Tapovan Vishnugad project. Initial assessment on the day of the event suggested that there was a glacial burst. Later, it was evaluated through time series of high spatial resolution remote sensing images of various satellites that a large part of a north-facing triangular-shaped slope at 5540 m above MSL had failed, which was also supporting a small hanging glacier. This landslide and, consequently, massive debris flow into the Raunthi Gadhera initially blocked the flow of Dhauliganga near Rini village [at 2000 m above MSL (mean sea level)], which later failed around 10:30 hrs on 07 February 2021 and brought a catastrophic flash flood in the Dhauliganga river. Further, remote sensing images acquired around 10:33 hrs of 07 February 2021 revealed a large dust cloud which clearly unravels the sequence of events from a high-altitude landslide, collapse of a small hanging glacier, and snow avalanche to catastrophic flooding. Even after the catastrophic flash flood of 07 February 2021, an elongated lake was created due to the blocking of the flow of the Rishi Ganga River. For detailed analysis, the calculation of dimension, area and volume of the failed slope was done using the high-resolution satellite images and digital elevation model using the RAMMS modelling technique. The north-facing triangular shape had a base of about 660 m and 1100 m height and the estimated total volume calculated was 20 million cubic meters, including rocks, snow, and ice. The debris flow runout simulation of the event was performed using the RAMMS debris flow model to calculate flow depth, flow velocity and maximum pressure. Also, from high-resolution satellite images, the dimensions of the artificial Rini Lake were estimated to have a length of about 800 m, a width at the front of about 100 m and a depth of about 46m, including freshly deposited debris and silt of about 10 m. To calculate the volume of the lake, simulation of lake was done in ArcView software using digital elevation model, and it came out to be ~5 million cubic meters. The paper also emphasizes monitoring of such vulnerable areas based on high-resolution time series satellite images, which are available on a regular basis to avoid the loss of human lives in the future.

Holocene Activity of the Wudaoliang–Changshagongma Fault of the Eastern Tibetan Plateau

The Wudaoliang–Changshagongma fault is one of the NW-trending faults located within the southern Bayan Har Block of the Tibetan Plateau in China. In this paper, we used high-resolution imagery and digital elevation model data to study the geomorphological and geological characteristics of the fault. Furthermore, the result also determined the fault trace and estimated the average horizontal slip rate of the fault since the late Quaternary to have been 2.6 ± 0.6 mm/a. This slip rate is approximately equivalent to that of the Awancang, Madoi–Garde, and Dari faults, which are also located within the block. Furthermore, the slip rates of these faults obtained by remote sensing and geological methods are consistent with GPS observations. It indicates that tectonic deformation within the block is continuous and diffuse. Using trenching study results and sedimentary radiocarbon dating, we identified four paleoearthquake events that occurred at 42,378–32,975, 33,935–20,663, 5052–4862, and after 673–628 cal BP, respectively. The recurrence intervals of large earthquakes on the faults within the block are much longer than those of the boundary faults, and the slip rates are also smaller, indicating that faults within the block play a regulatory role in the tectonic deformation of the Bayan Har Block.

Quantification of Microtopography in Natural Ecosystems Using Close-Range Remote Sensing

Microtopography plays an important role in various ecological, hydrologic, and biogeochemical processes. However, quantifying the characteristics of microtopography represents a data-intensive challenge. Over the last decade, high-resolution or close-range remote sensing data and techniques have emerged as powerful tools to quantify microtopography. Traditional field surveys were mostly limited to transects or small plots, using limited sets of observations but with the decrease in the cost of close-range remote sensing technologies and the increase in computing performance, the microtopography even in forested environments can be assessed. The main objective of this article is to provide a systematic framework for microtopographic studies using close-range remote sensing technologies. This is achieved by reviewing the application of close-range remote sensing to capture microtopography and develop microtopographic models in natural ecosystems. Specifically, to achieve the main objectives, we focus on addressing the following questions: (1) What terrain attributes represent microtopography in natural ecosystems? (2) What spatial resolution of terrain attributes is needed to represent the microtopography? (3) What methodologies have been adopted to collect data at selected resolutions? (4) How to assess microtopography? Current research, challenges, and applicability of close-range remote sensing techniques in different terrains are analyzed with an eye to enhancing the use of these new technologies. We highlight the importance of using a high-resolution DEM (less than 1 m2 spatial resolution) to delineate microtopography. Such a high-resolution DEM can be generated using close-range remote sensing techniques. We also illustrate the need to move beyond elevation and include terrain attributes, such as slope, aspect, terrain wetness index, ruggedness, flow accumulation, and flow path, and assess their role in influencing biogeochemical processes such as greenhouse gas emissions, species distribution, and biodiversity. To assess microtopography in terms of physical characteristics, several methods can be adopted, such as threshold-based classification, mechanistically-based delineation, and machine learning-based delineation of microtopography. The microtopographic features can be analyzed based on physical characteristics such as area, volume, depth, and perimeter, or by using landscape metrics to compare the classified microtopographic features. Remote sensing techniques, when used in conjunction with field experiments/data, provide new avenues for researchers in understanding ecological functions such as biodiversity and species distribution, hydrological processes, greenhouse gas emissions, and the environmental factors that influence those parameters. To our knowledge, this article provides a comprehensive and detailed review of microtopography data acquisition and quantification for natural ecosystem studies.

Nieuwerkerk - Emperor of China (NEC) Seamounts (Banda Sea): A multibeam seafloor imagery analysis

A new multibeam dataset of Nieuwerkerk and Emperor of China (NEC) seamounts was acquired during the Jalacitra 2-2022 ”Banda” Expedition in the Banda Sea region. This study discusses a morphometric analysis of the seamounts based on a high-resolution bathymetry digital elevation model (DEM) and describes the morphometric features, namely: slope, profile curvature, and aspect. Morphological results indicate that the Nieuwerkerk seamount has an area of 2,416 km2 with a maximum peak elevation of approximately 3,460 m above the seabed and 357 m below the water surface. In the southwest of the Nieuwerkerk seamount, two new seamounts were discovered, which are still unnamed and seem morphologically connected to the Nieuwerkerk seamount. Each of the seamounts covers an area of 148 km2 and 220 km2, respectively, with a maximum peak elevation of approximately 2,338 m and 2,053 m above the seabed. Additionally, both seamounts have an elevation of 1,878 m and 2,017 m below the sea surface. Meanwhile, the Emperor of China seamount is located southwest of the three formerly mentioned seamounts, with the maximum peak seats at 2,165 m below the sea surface. Seamount C is located in the west of the surveyed area and it covers an area of 95.6 km2 with an elevation of 1,834 m above the seabed. The results of regression analysis indicate that all of the features could be used for the morphological classification.

Soil Erosion Estimation Using Revised Universal Soil Loss Equation Integrated with Geographic Information System by Different Resolution Digital Elevation Model Data in Weyto Sub-Basin, Southern Ethiopia

Soil erosion is a global environmental challenge for developing countries including Ethiopia that require regular monitoring to take corrective measures. In this context, this study was focused on estimating soil erosion using the Revised Universal Soil Loss Equation (rusle) integrated with Geographical Information System (gis) technique for which it applied 30 m and 200 m resolution Digital Elevation Model (dem) data to generate slope gradient and length. Rainfall erosivity, soil erodibility, land cover/use and management factors data were obtained from existing studies and field-based assessments where the data were used to estimate the soil erosion using rusle model in ArcMap under two different dem resolution scenario. The model estimated an average of 1.38 and 1.86 million tons of annual soil loss by water using 200 and 30 meters resolution dem data, respectively, while keeping other factors constant. The erosion estimated using higher (30 m) resolution dem data was more realistic than low (200 m) resolution data , as the higher resolution dem data allowed less generalization. In high resolution dem data, the slopes generated were also more in line with ground reality. Based on the case study of Weyto sub-basin in Southern Ethiopia, we thus conclude that the gis technique and remote sensing data can be used in rusle based erosion risk prediction for large areas even at basin, sub-basin and macro watershed level. We suggest that the accuracy of the prediction can be improved by using high resolution (large scale) input data disaggregated by micro- and sub-watersheds.

Combining remote sensing and social media data for flood mapping: a case study in Linhai, Zhejiang Province, China

The rapid mapping of flood inundation is essential for timely assessment of damage and post-disaster recovery. Remote sensing technology provides clear spatial information for mapping flood inundation, but its real-time (RT) images are often not available due to the severe weather conditions during the disaster events; the combination of RT data can better compensate for the deficiency. In this study, a near RT (NRT) flood inundation probability mapping method based on post-event NRT remote sensing data combined with RT volunteer geographic information (VGI) was proposed for mapping the 2019 flood in Linhai City, Zhejiang Province. First, a probabilistic index distribution (PID) layer was constructed from high-resolution digital elevation model data using an inverse distance-weighted height filter based on each VGI point. Then, a quality evaluation method for non-reference data was introduced to evaluate the validity of the VGI data. The final flood probability map was generated by PID weighting. The results show that, by fusing NRT images and RT data, the proposed model for mapping flood inundation probability is more robust and enhances the spatial characteristics of the flood inundation probability index, allowing it to enable emergency responders to quickly identify areas requiring urgent attention.

Photogrammetry Technology by Using DJI Phantom 4 RTK in Batang Mahat, Lima Puluh Kota Regency West Sumatera

Technological developments in the field of engineering have increased rapidly. The use of geospatial information becomes one of the bases for decision-making related to planning, managing, and evaluating a field, especially in engineering, which includes the field of measurement and making topographic maps or Digital Elevation Model (DEM). High-resolution geospatial information can create a high-resolution Digital Elevation Model (DEM) map that can be very useful in making models or maps of flood-affected areas. This research will use the photogrammetry technology with DJI Phantom 4 RTK to create the DEM data of Batang Mahat located in Lima Puluh Kota Regency, West Sumatra Province. 28 flights at predetermined points had carried out to measure the area of about 9 km2. All the results of photos captured by the drone will be imported to Agisoft Metashape 1.8.2 software to create the DEM data. The DEM was then compared with DEM data from DEMNAS and the measurement data of Batang Mahat by the SDABK office. The result shows that photogrammetry technology using DJI Phantom 4 RTK compared to measurement data and DEMNAS have a similar ground elevation but different river elevation. This is because the photogrammetry by using DJI Phantom 4 RTK is not equipped with LiDAR. An alternative to solve the problem is by combining the DEM data by DJI Phantom 4 RTK with data from terrestrial surveys. However, this method is quite expensive and ineffective in analyzing flood inundation compared to DEM data (8 m resolution) from DEMNAS.

Distribution and morphometry of pingos, western Canadian Arctic, Northwest Territories, Canada

Pingos of the western Canadian Arctic (WCA) coastal plain are among the most abundant and well-studied in the northern hemisphere. However, the abundance and morphometric variation of these pingos has not been quantitatively examined with respect to surficial and physiographic settings, likely because they have never been systematically catalogued with precision. We assess the distribution and morphology of pingos in the WCA using the High Resolution Digital Elevation Model (HRDEM) and demonstrate how this population compares to others in the Arctic, and how it varies in relation to surficial and physiographic settings. Of the 2363 pingos identified in the WCA (900 more than previous estimates), relief ranges from a few 10s of centimetres to 46.7 m for Ibyuk Pingo, which is the archetype of closed-system (hydrostatic) pingos in the region. Nearly 18 % (272) are taller than 10 m, although about 35 % of the pingos (830) are <2 m tall, which is the minimum height limit of most other studies. Using this minimum cut-off to compare populations, the mean height of WCA pingos >2 m (n = 1533) is 6.5 m, which is greater than other lowland pingos in the Arctic. WCA pingos are about 10 % larger in radius than pingos in North Alaska but have comparable mean slopes. Nevertheless, about 72 % of WCA pingos >2 m tall are “high-slope” pingos with a mean slope of 5° or greater. Within the WCA, 95 % of all pingos occur within the Tuktoyaktuk Lowlands in relation to the distribution of underlying Pleistocene sands, specifically the extent of braidplain deposits from the paleo-Peel and Anderson rivers known as the Kidluit Formation (Marine Isotope Stage 3). However, the spatial density of pingos within the Tuktoyaktuk Lowlands varies with physiography, and the highest densities relate to ice-rich, topographically-complex terrains that have a history of glaciation and thermokarst. In contrast, pingos within the modern Mackenzie Delta, where the low-lying, Holocene-age terrain is not amenable to lake drainage that promotes closed-system pingo formation, have a low density (5 % of total population) and are smaller. Nearly 30 % of WCA pingos reside at <5 m asl. Consequently, ongoing relative sea level rise and the effects on coastal change are likely to erode some pingos within the current population but may also trigger lake drainage and formation of new pingos.

Reconstructing terrestrial ice sheet retreat dynamics from hummocky topography using multiscale evidence: An example from central Ireland

The research reported here combines high-resolution digital elevation models (DEMs) derived from airborne LiDAR with ground-penetrating radar (GPR) surveys and macro- and micro-scale sedimentological analyses to study the geomorphology and internal structure and composition of ridges and mounds within an area of hummocky topography in the Brosna basin, central Ireland. Our evidence indicates that much of the hummocky topography consists of fragmented mega-scale glacial lineations (MSGLs) indicating a phase of accelerated ice flow, overlain by groups of small ridges composed of subglacially derived till and sediment gravity flow deposits. Subtle differences in ridge morphology indicate they may be multi-genetic and formed both ice-marginally as moraines and subglacially as small ribbed moraine. We interpret MSGL-ridge associations as a subglacial bedform continuum reflecting the evolution from a deforming bed to brittle deformation, due to changing thermal and hydrological conditions at the ice-bed interface during ice sheet retreat. Adjacent glaciofluvial landforms indicate initial en- and supra-glacial meltwater drainage, possibly directly related to changing basal thermal/hydrological characteristics and subglacial ridge formation. Subsequently a subglacial conduit system evolved, and later retreat involved formation of ice-marginal ridges. The partially preserved landsystem reflects formation at a polythermal to active temperate ice marginal zone, rather than the stagnating temperate margin previously assumed for this area. Our work demonstrates the usefulness of hummocky topography in identifying changes in ice sheet bed thermal/hydrological characteristics during deglaciation and the importance of combining multiple evidence strands in reconstructing the processes involved in glacial landform construction.

Terrain Self-Similarity-Based Transformer for Generating Super Resolution DEMs

High-resolution digital elevation models (DEMs) are important for relevant geoscience research and practical applications. Compared with traditional hardware-based methods, super-resolution (SR) reconstruction techniques are currently low-cost and feasible methods used for obtaining high-resolution DEMs. Single-image super-resolution (SISR) techniques have become popular in DEM SR in recent years. However, DEM super-resolution has not yet utilized reference-based image super-resolution (RefSR) techniques. In this paper, we propose a terrain self-similarity-based transformer (SSTrans) to generate super-resolution DEMs. It is a reference-based image super-resolution method that automatically acquires reference images using terrain self-similarity. To verify the proposed model, we conducted experiments on four distinct types of terrain and compared them to the results from the bicubic, SRGAN, and SRCNN approaches. The experimental results show that the SSTrans method performs well in all four terrains and has outstanding advantages in complex and uneven surface terrains.

Effects of complex terrain on the shortwave radiative balance: a sub-grid-scale parameterization for the GFDL Earth System Model version 4.1

Abstract. Parameterizing incident solar radiation over complex topography regions in Earth system models (ESMs) remains a challenging task. In ESMs, downward solar radiative fluxes at the surface are typically computed using plane-parallel radiative transfer schemes, which do not explicitly account for the effects of a three-dimensional topography, such as shading and reflections. To improve the representation of these processes, we introduce and test a parameterization of radiation–topography interactions tailored to the Geophysical Fluid Dynamics Laboratory (GFDL) ESM land model. The approach presented here builds on an existing correction scheme for direct, diffuse, and reflected solar irradiance terms over three-dimensional terrain. Here we combine this correction with a novel hierarchical multivariate clustering algorithm that explicitly describes the spatially varying downward irradiance over mountainous terrain. Based on a high-resolution digital elevation model, this combined method first defines a set of sub-grid land units (“tiles”) by clustering together sites characterized by similar terrain–radiation interactions (e.g., areas with similar slope orientation, terrain, and sky view factors). Then, based on terrain parameters characteristic for each tile, correction terms are computed to account for the effects of local 3D topography on shortwave radiation over each land unit. We develop and test this procedure based on a set of Monte Carlo ray-tracing simulations approximating the true radiative transfer process over three-dimensional topography. Domains located in three distinct geographic regions (Alps, Andes, and Himalaya) are included in this study to allow for independent testing of the methodology over surfaces with differing topographic features. We find that accounting for the sub-grid spatial variability of solar irradiance originating from interactions with complex topography is important as these effects led to significant local differences with respect to the plane-parallel case, as well as with respect to grid-cell-scale average topographic corrections. We further quantify the importance of the topographic correction for a varying number of terrain clusters and for different radiation terms (direct, diffuse, and reflected radiative fluxes) in order to inform the application of this methodology in different ESMs with varying sub-grid tile structure. We find that even a limited number of sub-grid units such as 10 can lead to recovering more than 60 % of the spatial variability of solar irradiance over a mountainous area.

Spatial variability in melting on Himalayan debris-covered glaciers from 2000 to 2013

Glaciers covered by debris are widespread in the Himalayas. Debris layer makes the melting process of glaciers complicated, and the correlation between the spatial melt differences of glacier and its surface features is still not clear. The aim of this study was to identify possible causes of thickness changes for Himalayan debris-covered glaciers. Firstly, three high resolution Digital Elevation Model (DEM) datasets including SRTM DEM, AW3D and TanDEM-X DEM were used to estimate glacier thickness changes from 2000 to 2013. By checking and removing geometric displacements, systematic deviations and elevation-related bias, as well as correction of radar wave penetration depth in the clean glacier areas, reliable thickness change information in this complex glaciated region were provided. Then debris thicknesses on the glaciers were estimated. The surface lowering caused by ice cliffs or supraglacial ponds significantly reduces debris thickness. Finally, some surface features including debris thicknesses, spatial patterns of supraglacial ponds and ice cliffs extracted from high-resolution satellite images in different times, were comprehensively used to evaluate the main drivers of glacier melting differences. Results demonstrate that spatial variability of glacier melting in Himalayas is caused by the combined effects of debris thicknesses, ice cliffs and supraglacial lakes, with different dominant factors for different regions. The thickened glacier region, which mostly appears above 5300 m a.s.l., with minimal debris cover, displayed large patches of thickening and partial serious thinning in the debris-free areas. Although the thick debris layer (> 0.20 m) would slow down the glacier melting at the low-elevations, it leads to the heavy glacier thinning for the regions located 5000–5300 m a.s.l., where there are dense ice cliffs. In some areas with >1 m of average debris thickness, glaciers have thinned slowly because of erosion from the ice cliffs or large supraglacial ponds. Area and number of cliffs have proven to be key factors in controlling regional ablation for debris-covered glaciers, and supraglacial lakes serve as assistance.

Super-resolution for terrain modeling using deep learning in high mountain Asia

High Mountain Asia (HMA) is characterized by some of the most complex and rugged terrain conditions in the world. However, high resolution terrain data are not easy to quickly acquire from the area due to difficulties in accessing the region. In this study, we trained a modified super-resolution residual network (MSRResNet) to develop super-resolution (SR) digital elevation models (DEMs) in the HMA areas by using freely available DEM data from the HMA region and limited high resolution (HR) DEMs from other areas to train the model. In this network, a new loss function was constructed that considered the terrain parameters of slope and curvature to constrain the network learning and convergence. The proposed method was applied to and validated by data from the Hengduan Mountains in the southeastern part of HMA, which is a world-famous longitudinal belt of mountains and canyons. A comparative analysis between the current and existing methods (i.e., SRGAN and Bicubic interpolation) was conducted to assess the effectiveness of the proposed approach. The experimental results were also investigated and evaluated by visual inspection and analysis of the terrain parameters. The results demonstrate that the proposed MSRResNet super-resolution process can achieve highly accurate terrain data by downscaling DEMs in HMA. This SR process also outperforms the other comparable methods. Compared to the Bicubic interpolation method, the RMSE and MAE accuracy are improved by 32.17% and 33.97%, and compared to the SRGAN method, the RMSE and MAE accuracy are improved by 39.15% and 32.47%. The HR DEM generated by the new method is more conducive to improving the accuracy of extracted terrain features, such as stream networks. It is promising to apply this model on other areas on Earth or even other planets with terrain similar to that of HMA.