Digital Elevation Model Derivatives Research Articles

Digital Mapping of Key Static Soil Attributes of Tamil Nadu, India using Legacy Soil Information

Abstract Acquiring spatial soil information is pivotal for land resource management, environmental and soil modelling. Digital soil mapping approach helps in prediction of spatial soil properties by establishing the relationship between soil and environmental covariates. In the current study, we attempted to predict and map spatial distribution of soil depth, coarse fragments (CF), and soil textural classes over 0.13 million sq km area of Tamil Nadu, India. About 2105 samples were used for the prediction of soil attributes viz., soil depth and coarse fragments using random forest (RF) regression model, multiple linear regression (MLR), and support vector machine (SVM), while the same set of soil data was used to predict the soil textural classes as categorical variables using Random Forest classifier. Different environmental covariates such as derivatives of digital elevation models, IRS LISS-III data and bioclimatic variables were related for predicting the soil properties. The predicted soil depth and CF ranged from 46-200 cm and 1-42 per cent, respectively. The RF model outperformed by explaining the variability (R2) of 43% for soil depth and 21% for coarse fragments with root mean square error (RMSE) of 38 cm and 13%, respectively, whereas, MLR and SVM could achieve the maximum variability of R2 of 0.20 and 0.25 for soil depth and R2 of 0.07 and 0.09 for CF. With respect to soil textural classes, RF classifier performed well with overall accuracy of 63.8% and kappa index of 0.43. Variable importance ranking of Random Forest model showed that elevation, multiresolution valley bottom flatness (MrVBF), multiresolution ridge top flatness (MrRTF) and remote sensing variables (NDVI & EVI) displayed decisive role in prediction of the soil depth, coarse fragments and soil textural classes. In this study, 250 m resolution detailed soil depth, CF and textural class maps were prepared which will be useful for different environmental modelling and proper agricultural management purposes.

Impacts of Resampling and Downscaling Digital Elevation Model and Its Morphometric Factors: A Comparison of Hopfield Neural Network, Bilinear, Bicubic, and Kriging Interpolations

The digital elevation model (DEM) and its derived morphometric factors, i.e., slope, aspect, profile and plan curvatures, and topographic wetness index (TWI), are essential for natural hazard modeling and prediction as they provide critical information about the terrain’s characteristics that can influence the likelihood and severity of natural hazards. Therefore, increasing the accuracy of the DEM and its derived factors plays a critical role. The primary aim of this study is to evaluate and compare the effects of resampling and downscaling the DEM from low to medium resolution and from medium to high resolutions using four methods: namely the Hopfield Neural Network (HNN), Bilinear, Bicubic, and Kriging, on five morphometric factors derived from it. A geospatial database was established, comprising five DEMs with different resolutions: specifically, a SRTM DEM with 30 m resolution, a 20 m resolution DEM derived from topographic maps at a scale of 50,000, a 10 m resolution DEM generated from topographic maps at a scale of 10,000, a 5 m resolution DEM created using surveying points with total stations, and a 5 m resolution DEM constructed through drone photogrammetry. The accuracy of the resampling and downscaling was assessed using Root Mean Square Error (RMSE) and mean absolute error (MAE) as statistical metrics. The results indicate that, in the case of downscaling from low to medium resolution, all four methods—HNN, Bilinear, Bicubic, and Kriging—significantly improve the accuracy of slope, aspect, profile and plan curvatures, and TWI. However, for the case of medium to high resolutions, further investigations are needed as the improvement in accuracy observed in the DEMs does not necessarily translate to the improvement of the second derivative morphometric factors such as plan and profile curvatures and TWI. While RMSEs of the first derivatives of DEMs, such as slope and aspect, reduced in a range of 8% to 55% in all five datasets, the RMSEs of curvatures and TWI slightly increased in cases of downscaling and resampling of Dataset 4. Among the four methods, the HNN method provides the highest accuracy, followed by the bicubic method. The statistics showed that in all five cases of the experiment, the HNN downscaling reduced the RMSE and MAE by 55% for the best case and 10% for the worst case for slope, and it reduced the RMSE by 50% for the best case of aspect. Both the HNN and the bicubic methods outperform the Kriging and bilinear methods. Therefore, we highly recommend using the HNN method for downscaling DEMs to produce more accurate morphometric factors, slope, aspect, profile and plan curvatures, and TWI.

Predictive map of soil texture classes using decision tree model and neural network with features of geomorphology level

This study aims to compare decision tree (DT) and artificial neural network (ANN) models, in addition, the efficiency of geomorphic surface attributes in predicting soil texture classes. The study area is located in the north of Chaharmahal and Bakhtiari province, central west Iran, and covers 6875 ha. Ninety-six pedons were excavated on separated geoforms. Soil samples of top soil (A horizon) were analyzed for clay, sand, and silt contents. Totally 57 auxiliary variables, including the derivatives of digital elevation model (DEM), Landsat 8 images, geomorphic surface map, geology map, and land-use map, were used to predict both soil texture classes and soil particle size fractions. Root-mean-square error (RMSE), R² or the coefficient of determination ( R_square), overall accuracy, and Kappa coefficient were selected as criteria for evaluating model performance. The R-square coefficients of clay, silt, and sand fractions for both models, respectively, were 0.41, 0.25, and 0.63 for ANN and 0.52, 0.62, and 0.75 for DT. According to RMSE, R-square, overall accuracy, and Kapa coefficient of validation data, the DT model produced better prediction fits to the both soil particle-size fraction and soil texture classes and was the most accurate classifier model. The parameters were 0.59, 0.09, 0.66, and 0.24 for ANN and 0.41, 0.75, 0.76, and 0.60 for DT models, respectively. The accuracy of each individual soil texture class was generally dependent upon the number of soil texture observations in each texture class. According to this fact, both models had better prediction for silty clay loam and clay loam texture classes.

Estimation of Soil Characteristics from Multispectral Sentinel-3 Imagery and DEM Derivatives Using Machine Learning.

In this paper, different machine learning methodologies have been evaluated for the estimation of the multiple soil characteristics of a continental-wide area corresponding to the European region, using multispectral Sentinel-3 satellite imagery and digital elevation model (DEM) derivatives. The results confirm the importance of multispectral imagery in the estimation of soil properties and specifically show that the use of DEM derivatives improves the quality of the estimates, in terms of R2, by about 19% on average. In particular, the estimation of soil texture increases by about 43%, and that of cation exchange capacity (CEC) by about 65%. The importance of each input source (multispectral and DEM) in predicting the soil properties using machine learning has been traced back. It has been found that, overall, the use of multispectral features is more important than the use of DEM derivatives with a ration, on average, of 60% versus 40%.

Transferability of Covariates to Predict Soil Organic Carbon in Cropland Soils

Precise knowledge about the soil organic carbon (SOC) content in cropland soils is one requirement to design and execute effective climate and food policies. In digital soil mapping (DSM), machine learning algorithms are used to predict soil properties from covariates derived from traditional soil mapping, digital elevation models, land use, and Earth observation (EO). However, such DSM models are trained for a specific dataset and region and have so far only allowed limited general statements to be made that would enable the models to be transferred to different regions. In this study, we test the transferability of SOC models for cropland soils using five different covariate groups: multispectral soil reflectance composites (satellite), soil legacy data (soil), digital elevation model derivatives (terrain), climate parameters (climate), and combined models (combined). The transferability was analyzed using data from two federal states in southern Germany: Bavaria and Baden-Wuerttemberg. First, baseline models were trained for each state with combined models performing best in both cases (R2 = 0.68/0.48). Next, the models were transferred and tested with soil samples from the other state whose data were not used during model calibration. Only satellite and combined models were transferable, but accuracy declined in both cases. In the final step, models were trained with samples from both states (mixed-data models) and applied to each state separately. This process significantly improved the accuracies of satellite, terrain, and combined models, while it showed no effect on climate models and decreased the models based on soil covariates. The experiment underlines the importance of EO for the transfer and extrapolation of DSM models.

Assessment of the capability of SWAT model to predict surface runoff in open cast coal mining areas.

The hydrological response of watersheds affected by large-scale coal mining activities is complex and difficult to simulate. The present study aims to bridge this gap by simulating the effects of land-use and topographical changes due to coal mining on surface runoff in the Jamunia basin of Jharkhand, India. The derivatives of digital elevation model (DEM) have been used to simulate the changes in topography of the study area and the runoff has been calculated using Soil and Water Assessment Tool (SWAT) hydrological model. The study results revealed significant increase in surface runoff (mm) during the simulation period. The findings of this study established that unplanned mining activities can reduce the water holding capacity of downstream reservoirs and increase the runoff. The outcome of the study will be helpful for mine planners to design sustainable mining operations which will have less adverse impact on the hydrological regime of the watershed.

A novel method using explainable artificial intelligence (XAI)-based Shapley Additive Explanations for spatial landslide prediction using Time-Series SAR dataset

As artificial intelligence (AI) techniques are becoming more popular in landslide modeling, it is important to understand how decisions are made. Fairness, and transparency becomes ever more vital due to ethical concerns and trust. Despite the popularity of machine learning (ML) algorithms in landslide modeling, the explainability of these methods are often considered as black box. This paper aims to propose an explainable artificial intelligence (XAI) for landslide prediction using synthetic-aperture radar (SAR) time-series data, NDVI (normalized difference vegetation index) time-series data and other geo-environmental factors such as DEM (digital elevation model) derivatives. We employed a Shapley Additive Explanations (SHAP) approach to understand how and what decisions ML-based models are making. 37 features were extracted from various sources such as ALOS-PALSAR (ALOS Phased Array type l-band Synthetic Aperture Radar), ALOS-2 (SAR), Landsat-8, topographic maps, and DEM for landslide susceptibility mapping in a landslide prone area in Chukha, Bhutan as a test site. The result was then compared using two standard ML methods: random forest (RF) and support vector machine (SVM). As per results, the RF model outperformed (0.914) the SVM. Moreover, the higher reliability of the RF model was proved by the area under the curve (AUC) of 0.941. XAI results revealed, features like altitude, aspect, NDVI-2014, NDVI-2017, and NDVI-2018 were the most effective features for landslide prediction by both models. Interestingly, among those features, NDVI-2014, aspect, and NDVI-2017 negatively correlated with the landslide prediction; whereas positively correlated when SVM was utilized. This interpretation ability indicates the advantages of XAI over the conventional methods as it measures the impact, interaction and correlation of conditioning factors within a model. The current research finding can provide more transparency and explainability when working with MLs in landslide studies. This could help to build trust among the geoscientists and decision-makers while making geohazard prediction.

Predictive soil mapping in the Boreal Plains of Northern Alberta by using multi-temporal remote sensing data and terrain derivatives

As Canada's vast Boreal Plains are extensively managed, predictive soil mapping could be used as an effective tool to generate high-resolution soil information for the region to inform sustainable resource management. This study aimed to investigate the use of multi-temporal remote sensing data and terrain derivatives to map soil types in the region. A method of constraining subgroup and great-group soil-type predictions based on the predictions at higher-order levels (great-group and order, respectively) was tested. Sentinel time series median values obtained by using Google Earth Engine were tested in combination with first- and second-order digital elevation model derivatives for use as predictor variables in the predictive models. A recursive feature selection process was implemented to reduce the number of predictor variables used in model training. Soil classes were predicted at the order, great-group, and subgroup levels and two approaches were tested. In the first approach, models were unconstrained based on previous predictions. In the second approach, models were constrained to predict only soil great-group classes that occur within the predicted soil order for a given location and similarly predict only soil subgroup classes that occur within the predicted soil great group for a given location. Determined through independent validation testing, the most probable predicted soil maps had overall accuracies ranging from 42% to 68% and kappa scores ranging from 0.33 to 0.48. Overall, the constrained models had the best performance of the approaches tested.

Predicting soil property in hilly regions by using landscape and multiscale micro-landform features

To assess the high-resolution digital soil mapping method for small watersheds in hilly areas, we explored the role of landscape classification and multiscale micro-landform features in predicting soil pH, soil clay content (SCC), and cation exchange capacity (CEC). Geomorphons (GM) terrain classification method was used to create landform units. The traditional digital elevation model (DEM) derivatives and remote sensing variables were employed for different combinations with landscape and micro-landform classification variables, with further compa-rison and analysis being conducted. In addition, three machine learning techniques, including support vector machine (SVM), partial least squares regression (PLSR), and random forest (RF), were used to build prediction models. The best method was then selected, and then combined with regression kriging by modeling spatial structure of the model residuals. The results showed that the application of landscape and multiscale micro-landform classification variables effectively improved the prediction accuracy of pH, SCC, and CEC by 18.8%, 8.2% and 8.7%, respectively. The map of landscape classification that contained vegetation coverage information had greater model contribution than land use data. The GM classification map with 5 m resolution was more suitable for high-precision DSM than those with lower resolution. The composite model of RF performed the best in predicting SCC, while the pH and CEC were not suitable for adding the residual regression kriging on the basis of RF model. Finally, the combination of landscape and multiscale micro-landform classification variables, DEM derivatives and remote sensing variables had the highest prediction accuracy for all the three soil properties. This result indicated that multivariable contained more effective soil information than single data source for rolling areas. The landscape variables composed of GM and surface classified data explained about 40% of the spatial variation of tested soil attributes in hilly area. Therefore, multi-resolution GM and landscape classified variables could be included into the construction of prediction model in research of soil mapping.

Digital mapping of soil classes using spatial extrapolation with imbalanced data

Digital mapping of soil classes using the extrapolation approach is timesaving, economically cheap, and helps collecting soil data from areas with difficult sampling. However, it has not been explored widely enough for digital mapping of soil classes. This study seeks to evaluate and compare several machine learning and regression algorithms for the extrapolation of soil sub-groups. Also the issue of imbalanced number of observations was addressed and oversampling technique was applied on the minority soil class to improve the models performance. The study area is located in central north Iran with 84 and 72 soil profiles sampled in the donor and recipient areas, respectively. A set of various environmental covariates including remotely sensed data, digital elevation model derivatives and geomorphology map were used as explanatory variables for predicting soil classes. Results showed that among eleven investigated models, C5.0 decision tree (DT), random forest (RF) and multinomial logistic regression (MNL) had the highest overall accuracy of 46%, 42% and 38%, respectively, for the extrapolation of soil classes. Also the Kappa statistic values for these models were 0.30, 0.24 and 0.22, respectively. Oversampling of the minority soil class led to an increase in overall accuracy for some of the models with the highest ones being DT = 53% and RF = 50%. Also, the Kappa value for DT and RF models increased to 0.39 and 0.35, respectively. In addition, oversampling of the minority soil class led to the prevention of losing this class in the final map.

Geological and terrain attributes for predicting soil classes using pixel- and geographic object-based image analysis in the Brazilian Cerrado

One of the great challenges in the geosciences is mapping targets and phenomena that are not fully in direct range of imaging instruments such as remote sensors. Thus, soil maps are still scarce, especially at the detail level for large regions, with the few official maps being restricted to small spatial cutouts, resulting in an increasing demand for pedological information. The aim of this work was to analyze the use of covariates related to geological and terrain attributes on the performance of soil class predictive models for digital soil mapping with pixel- and geographic object-based image analysis (GEOBIA) approaches in the context of sub-humid tropical environments of the Cerrado biome, in central-western Brazil, to reproduce a conventional map. A reference soil map at the scale 1:50,000 with twelve soil classes was used. Randomly stratified samples and the Random Forest (RF) classifier was performed with seventeen digital elevation model derivatives and the geology categorical data (lithostratigraphic units) as covariates for the prediction of soil classes in both approaches. The main results demonstrated that GEOBIA is a promising method for mapping soil classes, with overall accuracy of 0.70, which is 5% higher than pixel-based. The most important covariates for soil class mapping were related to Geology and terrain attributes: Altitude, Vertical Distance to Channel Network, Vector Terrain Ruggedness, Relative Slope Position, Topographic Height, Profile Curvature, Roughness, Slope, and Valley Depth. The integration of the geological covariate improved the models in both approaches. Integrated evaluation of the geological units and terrain attributes along with the use of machine learning techniques, such as the RF algorithm, is a promising and relevant strategy for digital predictive mapping of soil classes that can contribute toward the greater reproducibility of conventional soil maps.

DIGITAL ELEVATION MODELS FROM STEREO, VIDEO AND MULTI-VIEW IMAGERY CAPTURED BY SMALL SATELLITES

Abstract. Small satellites play an increasing role in earth observation. This article evaluates different possibilities of utilizing data from Planet’s SkySat and PlanetScope satellites constellations for derivation of digital elevation models. While SkySat provides high resolution image data with a ground sampling distance of up to 50 cm, the PlanetScope constellation consisting of Dove 3U cubesats provide images with a resolution of around 4 m. The PlanetScope acquisition strategy was not designed for stereo acquisitions, but for daily acquisition of nadir viewing imagery. Multiple different products can be acquired by the SkySat satellites: Collects covering an area of usually 12 by 6 km, tri-stereo collects and video products with a framerate of 30 Hz. This study evaluates DSM generation using a Semi-Global Matching from multi date stereo pairs for SkySat and PlanetScope, and the dedicated Video and tri-stereo SkySat acquisitions. DSMs obtained by merging many PlanetScope across track stereo pairs show an normalized median deviation against LiDaR first pulse data of 5.2 meter over diverse landcover at the test sites around the city of Terrassa in Catalonia, Spain. SkySat tri-stereo products with 80 cm resolution reach an NMAD of 1.3 m over Terrassa.

Retrieval of digital elevation models from Sentinel-1 radar data – open applications, techniques, and limitations

Abstract With the launch of Sentinel-1 in 2014, a new era of openly accessible spaceborne radar imagery was begun, and its potential has been demonstrated throughout all fields of applications. However, while interferometric approaches to detect surface deformations are continuously being published, only a few studies address the derivation of digital elevation models (DEMs) from Sentinel-1 data. This is mainly because of the narrow orbital tube, which was primarily designed for subsidence measurements using differential interferometry. Nonetheless, the technical conditions are provided for successful applications involving DEM generation. These are outlined in the first part of this article with a focus on potential error sources and the impact of the most important constraints, namely, temporal and perpendicular baselines. The second part evaluates 21 studies on this topic, their aims, and how they dealt with error sources and the necessity of validation. These studies are then discussed based on the main challenges and potentials including how these can be tackled in the future to lay a solid foundation for scientific discourse.

Identification of soil piping-related depressions using an airborne LiDAR DEM: Role of land use changes

The detection of soil piping features is still a challenging issue that requires further studies, especially at larger scale. It seems that field mapping as a time-consuming method can be supported by the new technologies. This study has exploited the freely available airborne LiDAR (Light Detection and Ranging) digital elevation model (DEM) to explore piping-related depressions. The aim was to test the possibility of digital detection of piping forms using sink-fill method and different derivatives of airborne LiDAR DEM (slope, Topographic Position Index, Topographic Ruggedness Index, Positive and Negative Openness Index) in order of its utility in regional scale studies. As geological structure, soils and climate are similar in the whole study area, the analysis was extended to include the changes in land use/land cover (LULC) in the last two centuries (1862–2017). The results were verified by the detailed geomorphological mapping in the field. The study was conducted in a small catchment in the Bieszczady Mts. (SE Poland) covered by forests and grasslands/pastures, where in the past transition from pastures to forests was observed. The findings have revealed the great impact of LULC changes on soil piping, thus also on the possibility of digital mapping which is often based only on topographic information. The success rate in recognition was satisfactory on grasslands/pastures (76% for individual forms and 80% for piping systems). Areas covered by forests since the mid-19th century are unlikely places to develop piping forms (just 2 forms found) in the study area, whereas places where transition from pastures to forest has been observed may be prone to soil piping. However, the digital identification of piping forms in areas with rough topography is difficult (a success rate of 45% for individual forms and 50% for piping systems). The digital mapping of piping forms cannot rely only on the topographic information but should be always accompanied by an analysis of factors controlling piping erosion.

Using floristic gradient mapping to assess seasonal thaw depth in interior Alaska

AbstractQuestionsIs it possible to map floristic gradients in heterogeneous boreal vegetation by using remote‐sensing data? Does a continuous vegetation map enable the creation of a spatially continuous map of seasonal permafrost soil thaw depth?LocationBonanza Creek LTER, Fairbanks, Alaska, USA.MethodsVegetation records are subjected to an ordination to extract the predominant floristic gradient. The ordination scores are then extrapolated using Sentinel 2 imagery and a digital elevation model (DEM). As the relation between vegetation pattern and seasonal thaw depth was confirmed in this study, the spatial distribution of ordination scores is then used to predict seasonal thaw depth over the same area.ResultsThe first dimension of the ordination space separates species corresponding to moist and cold soil conditions from species associated with well‐drained soils. This floristic gradient was successfully mapped within the sampled plant communities. The extrapolated thaw depths follow the typical distribution along a topographical and geomorphological gradient for this region. Besides vegetation information also DEM derivatives show high contributions to the thaw depth modeling.ConclusionWe demonstrate that floristic gradient mapping in boreal vegetation is possible. The accuracy of the thaw depth prediction model is comparable to that in previous analyses but uses a more parsimonious set of predictors, underlining the efficacy of this approach.

Surface expressions of morphostructural features at Hasandağ stratovolcano on DEM datasets

Although the young pyroclastic cover tends to hide the structural features, derivatives of Digital Elevation Models (DEM) help reveal the morphological elements exposing the structure of the volcanoes. While the shaded relief maps improve the visuality of the DEM data, they are not the best derivatives to express the detailed morphological characteristics of the terrain. Advanced morphometric DEM derivatives, such as Red Relief Image Maps and Openness images, provide improved terrain legibility especially on volcanic landforms. Satellite-derived mid-resolution DEM and very-high-resolution aerial stereo-photogrammetric summit model of Hasandag was used to investigate the structural features of the volcano. The mid-resolution topographic derivatives evidently mapped structural features which have had important role in the evolution of the volcano. Effect of fault systems on the development of the volcano, evident and probable topographic rims of collapse structures, avalanche and lava flow surface morphologies and characteristics revealed by the advanced derivatives of DEMs were evaluated. Summit of Greater Hasandag volcano was modeled using Structure-from-Motion/Multiview stereo-photogrammetry using acquired drone imagery; the model presents the summit geology in high-resolution and precision.

Quantitative Analysis of Different Environmental Factor Impacts on Land Cover in Nisos Elafonisos, Crete, Greece.

Land Cover monitoring is an essential task for a better understanding of the ecosystem’s dynamicity and complexity. The availability of Remote Sensing data improved the Land Use Land Cover mapping as it is routine work in ecosystem management. The complexity of the Mediterranean ecosystems involves a complexity of the surrounding environmental factors. An attempt to quantitatively investigate the interdependencies between land covers and affected environmental factors was conducted in Nisos Elafonisos to represent diverse and fragile coastal Mediterranean ecosystems. Sentinel-2 (MSI) sensor and ASTER Digital Elevation Model (DEM) data were used to classify the LULC as well as to draw different vegetation conditions over the designated study area. DEM derivatives were conducted and incorporated. The developed methodology is intended to assess the land use land cover for different practices under the present environmental condition of Nisos Elafonisos. Supervised classification resulted in six different land cover clusters and was tested against three different environmental clusters. The findings of the current research pointed out that the environmental variables are independent and there is a vertical distribution of the vegetation according to altitude.

A Method for Extracting Some Key Terrain Features from Shaded Relief of Digital Terrain Models

Detection of terrain features (ridges, spurs, cliffs, and peaks) is a basic research topic in digital elevation model (DEM) analysis and is essential for learning about factors that influence terrain surfaces, such as geologic structures and geomorphologic processes. Detection of terrain features based on general geomorphometry is challenging and has a high degree of uncertainty, mostly due to a variety of controlling factors on surface evolution in different regions. Currently, there are different computational techniques for obtaining detailed information about terrain features using DEM analysis. One of the most common techniques is numerically identifying or classifying terrain elements where regional topologies of the land surface are constructed by using DEMs or by combining derivatives of DEM. The main drawbacks of these techniques are that they cannot differentiate between ridges, spurs, and cliffs, or result in a high degree of false positives when detecting spur lines. In this paper, we propose a new method for automatically detecting terrain features such as ridges, spurs, cliffs, and peaks, using shaded relief by controlling altitude and azimuth of illumination sources on both smooth and rough surfaces. In our proposed method, we use edge detection filters based on azimuth angle on shaded relief to identify specific terrain features. Results show that the proposed method performs similar to or in some cases better (when detecting spurs than current terrain features detection methods, such as geomorphon, curvature, and probabilistic methods.

High-resolution digital mapping of soil organic carbon and soil total nitrogen using DEM derivatives, Sentinel-1 and Sentinel-2 data based on machine learning algorithms

Soil organic carbon (SOC) and soil total nitrogen (STN) are important indicators of soil health and play a key role in the global carbon and nitrogen cycles. High-resolution radar Sentinel-1 and multispectral Sentinel-2 images have the potential to investigate soil spatial distribution information over a large area, although Sentinel-1 and Sentinel-2 data have rarely been combined to map either SOC or STN content. In this study, we applied machine learning techniques to map both SOC and STN content in the southern part of Central Europe using digital elevation model (DEM) derivatives, multi-temporal Sentinel-1 and Sentinel-2 data, and evaluated the potential of different remote sensing sensors (Sentinel-1 and Sentinel-2) to predict SOC and STN content. Four machine-learners including random forest (RF), boosted regression trees (BRT), support vector machine (SVM) and Bagged CART were used to construct predictive models of SOC and STN contents based on 179 soil samples and different combinations of environmental covariates. The performance of these models was evaluated based on a 10-fold cross-validation method by three statistical indicators. Overall, the BRT model performed better than RF, SVM and Bagged CART, and these models yielded similar spatial distribution patterns of SOC and STN. Our results showed that multi-source sensor methods provided more accurate predictions of SOC and STN contents than individual sensors. The application of radar Sentinel-1 and multispectral Sentinel-2 images proved useful for predicting SOC and STN. A combination of Sentinel-1/2-derived predictors and DEM derivatives yielded the highest prediction accuracy. The prediction accuracy changed with and without the Sentinel-1/2-derived predictors, with the R2 for estimating both SOC and STN content using the BRT model increasing by 12.8% and 18.8%, respectively. Topographic variables were the main explanatory variables for SOC and STN predictions, where elevation was assigned as the variable with the most importance by the models. The results of this study illustrate the potential of free high-resolution radar Sentinel-1 and multispectral Sentinel-2 data as input when developing SOC and STN prediction models.

Optimizing collapsed pipes mapping: Effects of DEM spatial resolution

Finding a digital elevation model (DEM) of suitable spatial resolution is vital to investigate piping erosion using aerial remote-sensing platforms like unmanned aerial vehicles (UAV). Previous studies have implied that the best spatial resolution is a DEM with the most detail. This study evaluates piping-affected areas with five DEMs (1, 5, 10, 20, and 30 m resolutions) with three trained machine-learning methods: support vector machine (SVM), maximum entropy (ME), and boosted regression tree (BRT). This method enables the identification of the specific impacts caused by changing pixel resolution to guide the selection of the most effective DEM. This study employs piping morphometry data to predict the locations of completely collapsed pipes. The performance of the methods for mapping of pipes was assessed against a piping inventory map. The results demonstrate that the finest resolution DEM is not always the most useful. Though 1 m-resolution DEMs show the most detail, the best performance was the 5 m-resolution DEM when tested for all three mapping models. The 5 m-resolution DEM-SVM combination was the best predictor of known piping sites (AUC = 81.0%). The 5-m DEM-ME was second most effective model (AUC = 75.8%). And 5-m DEM-BRT was third (AUC = 72.9%). Applying more DEM derivatives may increase confidence in the selection of the most appropriate resolution.