Artificial afforestation and natural recovery from abandoned cropland are two typical recovery types on the Loess Plateau, China. However, few studies have investigated the difference of natural secondary vegetation restoration and man-made plantation in soil aggregate physicochemical properties and soil aggregate stability. Therefore, we have selected natural restoration grassland and Chinese red pine plantation to study the differences of soil aggregate size distributions, aggregate carbon (C) and nitrogen (N) distributions, soil aggregate stability index (fractal dimension, D; mean weight diameter, MWD; geometric mean diameter, GMD; percentage of aggregation destruction, PAD) as well as their relationships. The results showed that after ~ 15 years restoration from abandoned cropland, natural restoration grassland had higher soil organic carbon (SOC), total nitrogen (TN), ammonium nitrogen (AN), microbial biomass nitrogen (MBN) and MWD compared to Chinese red pine forest, but Chinese red pine forest had higher aggregate C and N, D, GMD and PAD. In addition, SOC positively correlated with MWD in natural restoration grassland but opposite in Chinese red pine forest. In detail, the differences of soil general properties and aggregate size fraction percentages between two land use types were found mainly in 2–5 mm, 1–2 mm, 0.25 mm and clay water-stable aggregate size fractions. The results suggested that higher C content would further contribute the soil aggregate stability in natural restoration grassland, and higher N content would be more important in Chinese red pine plantation.

Soil aggregate stability and aggregate-associated carbon and nitrogen in natural restoration grassland and Chinese red pine plantation on the Loess Plateau

This study sought to produce an accurate multi-hazard risk map for a mountainous region of Iran. The study area is in southwestern Iran. The region has experienced numerous extreme natural events in recent decades. This study models the probabilities of snow avalanches, landslides, wildfires, land subsidence, and floods using machine learning models that include support vector machine (SVM), boosted regression tree (BRT), and generalized linear model (GLM). Climatic, topographic, geological, social, and morphological factors were the main input variables used. The data were obtained from several sources. The accuracies of GLM, SVM, and functional discriminant analysis (FDA) models indicate that SVM is the most accurate for predicting landslides, land subsidence, and flood hazards in the study area. GLM is the best algorithm for wildfire mapping, and FDA is the most accurate model for predicting snow avalanche risk. The values of AUC (area under curve) for all five hazards using the best models are greater than 0.8, demonstrating that the model's predictive abilities are acceptable. A machine learning approach can prove to be very useful tool for hazard management and disaster mitigation, particularly for multi-hazard modeling. The predictive maps produce valuable baselines for risk management in the study area, providing evidence to manage future human interaction with hazards.

/pdf/a-machine-learning-framework-for-multi-hazards-modeling-and-36mqqqvsw1.pdf

A machine learning framework for multi-hazards modeling and mapping in a mountainous area.

Summary The presented study explores the prediction of soil water retention and its variability from soil texture and bulk density measurements, using a physically-based scaling technique. Specifically, the Arya–Paris (AP) physico-empirical model is applied to two soil datasets that are collected from two catchments located in different areas of Southern Italy. Laboratory-measured soil water retention functions are scaled to characterize soil variability. The laboratory-measured and AP-predicted reference water retention functions are compared by evaluating the lognormal distribution of derived scaling factors, relative to the mean reference retention function. Since the scaling theory assumes geometric similitude for the investigated soils, successful application of using particle-size distribution to estimate soil water retention requires separation of soils with different textures, using variance analysis. We conclude that variability in soil water retention can be determined from limited soil water retention data using the scaling approach when combined with particle-size distribution measurements. This method can potentially be used as an effective tool for identifying soil hydrologic response at catchment scales.

Scaling soil water retention functions using particle-size distribution

Information about the soil–water retention curve is necessary for modeling water flow and solute transport processes in soils. Soil spectroscopy in the visible, near-infrared, and shortwave infrared (Vis-NIR-SWIR) range has been widely used as a rapid, cost-effective and nondestructive technique to predict soil properties. However, less attention has been paid to predict soil hydraulic properties using soil spectral data. In this paper, spectral reflectances of soil samples from the Zanjanrood watershed, Iran, were measured in the Vis-NIR-SWIR ranges (350–2500 nm). Stepwise multiple linear regression coupled with the bootstrap method was used to construct predictive models and to estimate the soil–water retention curve. We developed point and parametric transfer functions based on the van Genuchten (VG) and Brooks-Corey (BC) soil hydraulic models. Three different types of transfer functions were developed: (i) spectral transfer functions (STFs) that relate VG/BC hydraulic parameters to spectral reflectance values, (ii) pedotransfer function (PTFs) that use basic soil data as input, and (iii) PTFs that consider spectral data and basic soil properties, further referred to as spectral pedotransfer functions (SPTFs). We also derived and evaluated point transfer functions which estimate soil–water contents at specific matric potentials. The point STFs and SPTFs were found to be accurate at low and intermediate water contents (R² > 0.50 and root mean squared error [RMSE] < 0.018 cm³ cm⁻³), while the point PTFs performed better close to saturation. The parametric STFs and SPTFs of both the VG and BC models performed similarly to parametric PTFs in estimating the retention curve. The best predictions of soil–water contents were obtained for all the three transfer functions when the VG and BC retention models were fitted to the retention points estimated by the point transfer functions. Overall, our findings indicate that spectral data can provide useful information to predict soil—water contents and the soil–water retention curve. However, there is a need to extend and validate the derived transfer functions to other soils and regions.

A Comparative Study of Multiple Approaches for Predicting the Soil–Water Retention Curve: Hyperspectral Information vs. Basic Soil Properties

Mountainous areas are highly prone to a variety of nature-triggered disasters, which often cause disabling harm, death, destruction, and damage. In this work, an attempt was made to develop an accurate multi-hazard exposure map for a mountainous area (Asara watershed, Iran), based on state-of-the art machine learning techniques. Hazard modeling for avalanches, rockfalls, and floods was performed using three state-of-the-art models—support vector machine (SVM), boosted regression tree (BRT), and generalized additive model (GAM). Topo-hydrological and geo-environmental factors were used as predictors in the models. A flood dataset (n = 133 flood events) was applied, which had been prepared using Sentinel-1-based processing and ground-based information. In addition, snow avalanche (n = 58) and rockfall (n = 101) data sets were used. The data set of each hazard type was randomly divided to two groups: Training (70%) and validation (30%). Model performance was evaluated by the true skill score (TSS) and the area under receiver operating characteristic curve (AUC) criteria. Using an exposure map, the multi-hazard map was converted into a multi-hazard exposure map. According to both validation methods, the SVM model showed the highest accuracy for avalanches (AUC = 92.4%, TSS = 0.72) and rockfalls (AUC = 93.7%, TSS = 0.81), while BRT demonstrated the best performance for flood hazards (AUC = 94.2%, TSS = 0.80). Overall, multi-hazard exposure modeling revealed that valleys and areas close to the Chalous Road, one of the most important roads in Iran, were associated with high and very high levels of risk. The proposed multi-hazard exposure framework can be helpful in supporting decision making on mountain social-ecological systems facing multiple hazards.

/pdf/multi-hazard-exposure-mapping-using-machine-learning-26014224bm.pdf

Multi-Hazard Exposure Mapping Using Machine Learning Techniques: A Case Study from Iran

Assessment of soil structure and soil aggregation remains a challenging task. Routine methods such as dry- and wet-sieving approaches are generally time consuming and tedious, which calls for a robust, fast, and nondestructive method of soil aggregate characterization. Over the last two decades, diffuse reflectance spectroscopy (DRS) has emerged as a rapid and noninvasive technique for soil characterization. Combined with chemometric and data-mining algorithms, it provides an effective way of measuring several soil attributes and has the added advantage of being amenable to a remote sensing mode of operation. The objective of this study was to determine if the DRS approach could be used as a rapid, noninvasive technique to estimate soil aggregate characteristics. The DRS approach was examined for the estimation of soil aggregate characteristics such as the geometric mean diameter and two statistical parameters of the lognormal aggregate size distribution (ASD) functions using 910 soil samples from India representing three important soil groups. Results showed that the geometric mean diameter and the median aggregate size parameter provided excellent predictions, with ratio of performance deviation (RPD) values ranging from 1.99 to 2.28. The RPD value for the standard deviation of the ASD ranged from 1.36 to 1.72, suggesting moderate prediction. It was further observed that soil aggregates influence the incident electromagnetic radiation on soils primarily in the visible region and to some extent the shortwave- and near-infrared regions. Electronic transitions of Fe-bearing minerals, clay minerals, and C–H functional groups of organic matter may be responsible for modifying the spectral reflectance from soils in addition to the self-shadowing effects of surface roughness. The results of this study suggest that the chemometric approach may be combined with DRS to estimate soil aggregate size characteristics.

/pdf/diffuse-reflectance-spectroscopic-approach-for-the-4cvi1p2iht.pdf

Diffuse Reflectance Spectroscopic Approach for the Characterization of Soil Aggregate Size Distribution

Estimation of weathering indices using spectral reflectance over visible to mid-infrared region

Rapid and Noninvasive Assessment of Atterberg Limits Using Diffuse Reflectance Spectroscopy

Among the various natural disasters, landslides are considered to be one of the serious geological hazards that are triggered due to intensive rainfall, earthquakes, deforestation, mining, floods, etc. Landslides result in devastating impacts causes thousands of deaths and injuries, damage to properties. The changing pattern of landslide hazard zones every year forces the need to safeguard people and properties in the respective areas. Weighted linear combination (WLC) is used to prepare the landslide hazard zonation map in the Nilgiris district of Tamil Nadu, India. Nine layers, namely slope, aspect, lineament density, rainfall, distance of roads, elevation, distance from rivers, landuse/landcover, geology, are used in overlay analysis. Supply chain facilities are widely used in the field of transportation of goods to the consumers with the reduction of transportation costs. The implementation of supply chain mechanism along with GIS in disaster management could help to save numerous lives during disaster events.

Risk Mapping Analysis With Geographic Information Systems for Landslides Using Supply Chain

Identification of similar soils is an important step in different branches of soil hydrology. Generally, a characteristic length in porous media is used for estimating scaling factors in a similarity analysis. Such an approach amounts to the use of the first moment of the pore size distribution (PSD). Previous research has suggested that the σ of PSD of similar soils should also be similar. Authors of this research proposed a new criterion in terms of the CV for σ (CVσ) with the suggestion that the “geometrically similar” soils should satisfy the condition of CVσ < 10%. Here, we validate this result using the aggregate size distribution (ASD) data of 1375 soils collected from 10 different zones with varying scale and properties. The lognormal distribution function and the physically based scaling (PBS) approach were used to scale the ASD datasets. The effectiveness of scaling was determined by the RMSE between the scaled ASD curve and the reference ASD curve. Four out of 10 soil datasets had CVσ < 12%, and their corresponding RMSE values were an order of magnitude lower than those for the remaining soil groups. A plot between the RMSE and CVσ values showed that CVσ < 10% may be used as a practical limit for identifying similar soils.

Biswajita Mohanty

Papers

Diffuse Reflectance Spectroscopic Approach for the Characterization of Soil Aggregate Size Distribution

Estimation of weathering indices using spectral reflectance over visible to mid-infrared region

Rapid and Noninvasive Assessment of Atterberg Limits Using Diffuse Reflectance Spectroscopy

Risk Mapping Analysis With Geographic Information Systems for Landslides Using Supply Chain

Variance of Aggregate Size Distribution as a Criterion for Soil Similarity