Morphological Profiles of Sand Dunes from ICESat-2 Geolocated Photons

Aeolian process leads to the transportation and accumulation of sand particles that result in sand dune landforms. The structure and shape of the sand dunes are driven by the parameters of interacting wind force and the material composition of sand within. Cross-section profiles over the sand dunes will essay the geomorphological parameters through which the steady state and rate of sand transport can be computed. National Aeronautics and Space Administration’s novel satellite namely Ice, Cloud and land Elevation Satellite-2 (ICESat-2) hosts a solo sensor namely Advanced Topographic Laser Altimeter System (ATLAS) which is a photon counting instrument that measures the round-trip time of the light pulse being emitted and reflected back from the surface determines the true height of the topographic feature on the Earth. In this article, cross-section profiles generated from the beams of ICESat-2 ground-tracks acquired over sand dunes of the Thar Desert region were analysed for detecting the geomorphological parameters. Observations from the cross-section profiles have resulted in giving unprecedented details about the shapes and morphological settings of various types of sand dunes like barchanoids, parabolic, longitudinal, and transverse dunes. Morphological parameters of sand dunes like the length of the stoss slope, crest height, slip face details, inter-arms spacing, height of the trailing arms, length of the depositional lobes, and sinuosity of the recurring crest lines were retrieved with ease from the Level-2A data product namely ATL03 of ICESat-2/ATLAS.


Introduction
Sand dunes occur in two distinct habitats; one on the barren waterless floors of How to cite this paper: Dandabathula, G., Sitiraju, S. R., & Jha, C. S. (2021). Morphological Profiles of Sand Dunes from ICE-state of sand dunes is controlled by the parameters of the agitating medium like wind vector (like velocity and direction) and the material composition of sand itself (Tsoar, 1985). Aeolian process leads to the transportation and accumulation of sand that result in sand dune formations that ranges from 1 meter to tens of kilometres in length and few tens of centimetres to over 300 meters the height (Bristow, 2009). Sand particles of various sizes will participate in the aeolian transport mechanism in several modes like saltation, reptation, and suspension (Bagnold, 1971;Kok et al., 2012). The shape of a sand dune depends entirely on sand deposit structure that gets influenced by the strength and direction of the interacting wind; as either transverse to the wind, longitudinal or lengthwise in the direction of the wind, or solitary dunes related to the varying wind direction (Edgell, 2006;Lorenz and Zimbelman, 2014). Wasson and Hyde (1983) recognized four types of dunes based on wind regime and sand availability as barchans, transverse, longitudinal and star dunes; barchans occur where there are little sand and almost unidirectional winds, transverse dunes occur where sand is abundant and the wind is moderately variable, longitudinal dunes occur where the winds are more variable but there is little sand, and star dunes occur where sand is abundant and wind variability is at a maximum. Pye and Tsoar (1990) have considered the influence of vegetation (phytogenic) and further introduced parabolic dunes which have U or V shape with trailing arms. Crosssection profile analysis over sand dunes provides information regarding their shapes, heights, lengths, slopes and angle of repose of the structure. Tsoar (1985) while studying the steady state of sand dunes has emphasized the importance of cross-section profiles and appreciated its role in computing the rate of sand accumulation or sand loss. While utilizing the cross-section profiles over sand dunes, Zimbelman et al. (2012) have concluded that the profile information is a useful tool for evaluating possible mechanisms of the origin for aeolian-based features. In their studies, Courrech du Pont (2015) and Gao et al. (2015) have used the parameters from the cross-sections to study the dune morphodynamics and growth rates.
As most of the desert sand dunes are located in inaccessible regions, Earth observation systems help in providing a synoptic perspective to characterize the dune field patterns and boosted the understanding of arid zone geomorphology (Gutiérrez and Gutiérrez, 2013). Advances in remote sensing and spatial analysis have enabled better understanding of evolution, pattern development, morphometry and activity of sand dunes (Hugenholtz et al., 2012). Researchers have utilized remote sensing data to perform geomorphology studies over extra-terrestrial sand dunes Hayward et al., 2007;Al-Masrahy and Mountney, 2013;Bourke et al., 2010). Microwave remote sensing data has been used to retrieve geometrical characteristics of sand dunes (Stephan and Long, 2005) and to identify active/stabilized landforms in the driest deserts (Abdelkareem et al., 2020). Substantial advancements in the generation of Digital Elevation Models (DEM) and its availability to the scientific fraternity enabled research and reali-G. Dandabathula et al. Journal of Geoscience and Environment Protection zation of 3D aspects of sand dune geomorphology (Bubenzer and Bolten, 2008;Hugenholtz and Barchyn, 2010;Hugenholtz et al., 2012).
Advances in Light Detection And Ranging (LiDAR) data have expanded the possibilities to provide qualitative information about the morphological signatures and also enabled to monitor short term dune activities like volumetric changes (Hugenholtz et al., 2012). Satellite laser altimeter transmits laser pulses (energy in the order of several tens of mJ per pulse) towards target surfaces with a typical pulse rate of 10 -30 Hz. The instrument measures the round-trip time of the light pulse between the emitted and the returned surface reflection to determine the distance of the spacecraft with respect to the surface and yields the true height of the topographic features on the ground (Falkner and Schulz, 2015). Space-based laser altimeters are effective in providing topographic measurements that are useful to understand the formation and early evolution of planetary bodies. Using laser altimetry data, geophysical parameters can be generated which provides significant insights related to the shape, structure, and evolution of the subject body (Cole, 1997). For subject areas like sand dunes, the surface measurements retrieved from consecutive orbits over a time period can be used to understand the morphological parameters and quantify the amount of sand transport. Integrating various remote sensing, photogrammetry and LiDAR are highly useful to monitor the changes in sand dune landforms.

Background to the ICESat Mission
National Aeronautics and Space Administration (NASA) has initiated space based laser altimeter mission namely Ice Cloud and Land Elevation Satellite (ICESat) that was designed to measure the changes on polar ice using on-board sensor titled Geoscience Laser Altimeter (GLAS) (Schutz et al., 2005). GLAS carried 532 nm (visible) and 1064 nm (near infrared) laser light to satisfy both atmospheric and topographic investigations. GLAS has operated by sending laser pulses at 40 Hz, and the transmitted laser pulse illuminates a spot on the Earth's surface with a diameter of 65 m and successive spots are separated on the Earth's surface by 172 m. The waveform data from GLAS contains transmitted time and echo received time that is used to compute the range measurements between the sensor platform and the Earth's surface (Martino et al., 2019). Measurements from the first ICESat satellite, which operated from 2003 to 2009 have provided vital information on the response of the Earth's frozen surface (Cryosphere) with respect to the changes of atmosphere and ocean parameters (Markus et al., 2017). ICESat mission has provided huge altimetry based measurements not only to the Cryosphere scientific community, but also to the ocean, terrestrial and atmospheric scientific communities. Potts et al. (2008) used ICESat data and a DEM from shuttle Radar Topography Mission (SRTM) for recognizing the seasonal dune changes in the Namib Desert. Dabboor et al. (2013) has successfully used ICESat laser altimetry data to track the sand dune migration in the Rub Al-Khali region. But, certain limitations with respect to its spatial resolutions  (Abdalati et al., 2010;Markus et al., 2017).

Photon Counting Sensor
ICESat-2 is a follow-up to the ICESat mission was launched in September 2018 and started releasing the data products with version 1 in May, 2019. ICESat-2 spacecraft carries a single instrument namely Advanced Topographic Laser Altimeter System (ATLAS) which sends pulses of laser light to the ground, collects the returning packets of energy called photons, and records the photon travel time (Blumenfeld, 2019). Photons of the trillion in number, starts the journey from ATLAS sensor, travelling through the atmosphere to reflect off the Earth's surface, return through the atmosphere and back into the ATLAS telescope. Of these, only few of the photons complete the journey and are recorded by on-board instruments (Markus et al., 2017). In comparison with the scientific concepts that was

Objectives of the Study
A major objective of this study is to utilize the height information from the geolocated photon data obtained from the beams of ICESat-2 ground tracks acquired over various types of sand dunes; along-track sampled elevation is to be used to reconstruct the cross-section profiles to retrieve the geomorphological parameters pertaining to various sand dunes types. Towards this, samples of study areas falling in the Thar Desert region have been selected. Studies done on the sand dunes by earlier scholars in the Thar Desert region have been selected to infer and compare with the field level assessments.
Challenges exist to study sand dune morphology using optical remote sensing data and also photogrammetric techniques due to the homogeneous nature of sand dunes and shadows. The research in this article aims to bridge the gap due to the existing remote sensing techniques with the data from ICESat-2's along-track laser altimetry.

Study Area
The Thar Desert region is spread between the Aravalli hill ranges of India and  (Kar, 2014).
Details about genesis, orientation, and distribution of sand dunes in the Thar Desert region were studied and reported by earlier scholars like Singh et al. (1972), Kar (1993), Kar (1996), Dhir and Singhvi (2012) and Wadhawan (2020). Kar (1996) has elaborated the morphological details of various active sand dunes in the Thar Desert region. Breed et al. (1979) and Moharana et al. (2013) used satellite data to delineate and map various sand dune types in the Thar Desert region. In this study, six sample sites from the Thar Desert region were selected to analyse the cross-section profile from the ICESat-2 LiDAR data. The extents of these six sample sites were represented in Figure 1. The selection of these study areas was chosen so that

Data Sets
The ICESat-2 Global Geolocated Photon Level-2A data product identified as ATL03 provides the latitude, longitude and ellipsoidal height of photons detected by the ATLAS instrument (Neumann et al., 2019b). The ATL03 product is used as the foundation for other surface-specific geophysical data products such as sea ice, land ice, and vegetation canopy height. All ICESat-2 data products are provided in the Hierarchical Data Format (HDF) and are available to the scientific fraternity from the web portal maintained by the National Snow and Ice Data Center (NSIDC-https://nsidc.org/data/icesat-2). Alternatively ICESat-2 data sets are available to visualize and download from the Geoportal titled OpenAltimetry (https://openaltimetry.org/). The footprint pattern formed by ICESat-2 laser beams contains six consecutive Ground Tracks (GT) that are defined from left to right in the direction of travel and termed as GT1L, GT1R, GT2L, GT2R, GT3L, and GT3R.
When ATLAS sensor makes forward orientation then these GTs contains the weak beams at the left side of the beam pair (GT1L, GT2L, and GT3L) and when the orientation is backwards then the relative positions of weak and strong beams changes. The ATL03 algorithm discriminates erroneous photon events and actual signal events; high confidence photons can be considered by setting the parameter "signal_conf_ph" to the value of 4 (Neumann et al., 2019a).
ATL03 data product undergoes numerous corrections with respect to solid earth tides, ocean loading, solid earth pole tide, ocean pole tide, and the wet and dry atmospheric delays; finally the data product contains height above the ellipsoid, time and geodetic latitude and longitude for individual photons (Neumann et al., 2020). In this study, selected ground tracks obtained over different sand dunes types have been used to generate the cross-section profiles. Figure 2 represents the six ground tracks of ICESat-2 geolocated photons data overlaid on a sand dune system.

Methodology
Selected study areas contain barchanoids, megabarchanoids, parabolic, longitudinal, and transverse types of sand dunes; Level-2A data product titled ICESat-2 ATL03 photon elevation data acquired over these study areas were used to gen-  analysis for all the six study sites belonging to this investigation.

Results and Discussion
ICESat-2 photon data has the ability to quantify canopy height and provides an indication of the vegetation structure (Neuenschwander and Magruder, 2019).
Sand sheets are planar and featureless surfaces with strata of continuous deposits of windblown fine-grained particles may be sparsely anchored with vegetation (Grotzinger et al., 2005;Lorenz and Zimbelman, 2014). Study area earmarked as S1 contains sand sheet with sparse vegetation and barchan chain of sand dunes. Figure 3 (Tirsch, 2015). Solitaire barchans and barchanoids usually have gently dipping windward slopes and steeply dipping leeward slopes (slip face). From Figure 3(b), it can be noted that the morphological parameters for the barchanoids (earmarked as B1 and B2) like length of the windward base and height of the slip face can be retrieved from the cross-section profiles. These morphological parameters play pivotal role while computing the rate of sand transport when using temporal data. The rate of barchan advance is directly related to the rate of sand transport over the brink and inversely related to the brink height (Pye and Tsoar, 1990;Haff and Presti, 1995;Ould Ahmedou et al., 2007). The cross-section profile for the sand sheet with sparse vegetation is also reconstructed in Figure 3  upwind-facing linguoid element; the concept discussed earlier by researchers (Inman et al., 1966;Pye and Tsoar, 1990). Earlier researchers have done extensive field work and mapped the barchanoids in this study area and recorded their height range to be 10 to 15 m (Kar, 1993;Singhvi and Kar, 2004). The cross-section profiles retrieved from the ICESat-2 LiDAR data to confirm the height range of these barchanoids in these study area to be ranging between 10 to 15 m.
Compound dunes consist of two or more dunes of the same type combined by overlapping or being superimposed (McKee, 1979). Sample of the study area (earmarked as S3 in Figure 1) falling in the contiguous part of India and Pakistan contains compound barchanoids which were termed as megabarchanoids by Kar (1990)     ICESat-2 LiDAR data enabled to detect these type of sand dunes with the geolocated photon data that will able to detect height changes in the order of centimetres.
Parabolic dunes are hairpin shaped (or tend to be U-shaped or V-shaped) landforms that have three basic features: a depositional lobe on the downwind side, trailing arms or ridges, and a deflation basin between the trailing arms (Goudie, 2011). Figure 6(a) shows the extent of typical parabolic dunes that exists in the Thar Desert region and the area under investigation (sample of study area earmarked as S4) to retrieve the cross-section profile from the ICESat-2 LiDAR data. Figure 6(b) and Figure 6(c) represent the cross-section profiles from the pair of weak and strong beams of ICESat-2 ground track. Parabolic dune arms are almost always stabilized by some form of vegetation (May, 2015) and are evident in Figure 6 Figure 6(c)). Earlier, scholars like Kar (1996) and Srivastava et al.
(2019) explored this region and mentioned that the inter-arm spacing of these parabolic dunes ranges between 30 to 650 m of length, dunes in the middle of the parabolic cluster will be of longer arms, and presence of gradually shorter arms on each side of it; the same has been essayed by the cross-section profiles generated by ICESat-2 data which is being represented in Figure 6(b) and Figure 6(c). In this sort of study area, the height of the ridges being in the range of 10 to 50 m and the same has been reported earlier by Kumar et al. (1993), Kar (1996, and Singhvi and Kar (2004). Cross-section profiles retrieved from ICE-Sat-2 LiDAR data over parabolic dunes gives the dimensions of blowout hollow (deflation basin) through which basin morphology and wind flow patterns can be studied (Hesp, 2011).
Longitudinal dunes or linear dunes are elongated sand ridges of aeolian sediment forming approximately parallel to the resultant vector of long term prevailing wind direction (Radebaugh et al., 2015). Longitudinal dunes are also termed as seif dunes in Africa and Saudi Arabia. Details about the evolution of longitudinal dunes in the Thar Desert region were explored and reported by Kar (1993), Kumar et al. (1993), Wadhawan (1994), and Singhvi and Kar (2004). Kumar et al. (1993) attributed the reason for development of gullies in between the linear dunes is because of reactivation and break-down of a transverse sand ridge coalescing the trailing arms of the parabolic dunes through erosion by strong winds. In general, the length of longitudinal dunes is much greater than their width, and their long axes are parallel to the time-averaged vector sum of winds (Radebaugh et al., 2015).    Kar (1993) and Kumar et al. (1993) independently reported the height of the crest lines of these longitudinal dunes ranging from 5 to 50 m; through this study, similar dimensions have been retrieved form the ICESat-2 LiDAR data. The morphological parameters like height of crest lines, angle of linearity, and inter-spacing pattern between the crest lines can be used to deduce the wind regimes history and dune field state of development (Reffet et al., 2010). The pattern of slip faces occurring on either side of the dune long axis will give information about the recent winds regimes (Radebaugh et al., 2015).
Transverse dunes resemble barchan dunes in being a product of unidirectional winds, with steep slip faces oriented normal to wind direction and may occur in a gradational sequence (Fitzsimmons, 2015). Evolution details and structural parameters of transverse dunes have been reported by McKee (1979). Transverse dunes develop where sand is abundant and attain maximum height where there is a balance between strength of the wind and available sand (McKee, 1979;Kar, 1990;Tsoar and Blumberg, 2002). Multiple apices can form in the presence of adjacent dunes or landforms which influence wind orientation on a local scale (Bagnold, 1971).

Conclusion
Sand dunes are the geomorphological records of the aeolian processes and their shapes represent the wind regime and history of sand transport. Profiles from space based laser altimeters like ICESat-2 are effective in providing the topographic measurements and profiling the landform shapes. In this article, samples of sand dunes located in the Thar Desert region have been studied using geolocated photons from ICESat-2 data. Cross-section profiles generated from various beams of the ICESat-2 ground tracks have shown that the ability to distinguish between sand sheets and sand dunes in the semi-arid regions. Morphological parameters of the sand dunes like the length of the windward side, height of slip face, and crestline can be retrieved from the cross-section profiles generated from the photon elevation data of ICESat-2 due to its ability to generate dense along-track sampling (~ 70 cm). Importantly, ICESat-2 will give the elevation estimates over high sloped area and thus, mensuration over features like slip face will enable us a better understanding of the sand dune structures. Also, studying the distances between the crests of successive barchanoids (wavelength) is useful to quantify the saltation. High sampling feature of ICESat-2 photon events with centimeter level accuracies are able to fetch the details of minor elevations carved by the superimposed features on the main structure in compound dunes. Structural details like distance between inter-arm spacing, height of the trailing arms, shape of deflation basin, and pattern of slip faces can be retrieved using cross-section profiles for parabolic and longitudinal sand dunes. Unprecedented details about the shapes and morphological settings of the sand dunes have been reported through the cross-section profiles generated from the ICESat-2 photon data.
Future work can include the integration of ICESat-2 data with optical remote sensing data for automatic recognizing of sand dunes. However, limitation in the study is the availability of the cross-section profile direction that restricts to study landform of the sand dune in its entire form.