Landslide susceptibility mapping is well recognized as an essential element in supporting decision-making activities for preventing and mitigating landslide hazards as it provides information ...regarding locations where landslides are most likely to occur. The main purpose of this study is to produce a landslide susceptibility map of Mt. Umyeon in Korea using an artificial neural network (ANN) involving the factor selection method and various non-linear activation functions. A total of 151 historical landslide events and 20 predisposing factors consisting of Geographic Information System (GIS)-based morphological, hydrological, geological, and land cover datasets were constructed with a resolution of 5 x 5 m. The collected datasets were applied to information gain ratio analysis to confirm the predictive power and multicollinearity diagnosis to ensure the correlation of independence among the landslide predisposing factors. The best 11 predisposing factors that were selected in this study were randomly divided into a 70:30 ratio for training and validation datasets, which were used to produce ANN-based landslide susceptibility models. The ANN model used in this study had a multi-layer perceptron (MLP) structure consisting of an input layer, one hidden layer, and an output layer. In the output layer, the logistic sigmoid function was used to represent the result value within the range of 0 to 1, and six non-linear activation functions were used for the hidden layer. The performance of the landslide susceptibility models was evaluated using the receiver operating characteristic curve, Kappa index, and five statistical indices (sensitivity, specificity, accuracy, positive predictive value (PPV), negative predictive value (NPV)) with the training dataset. In addition, the landslide susceptibility models were validated using the aforementioned measures with the validation dataset and were compared using the Friedman test to check the significant differences among the six developed models. The optimal number of neurons was determined based on the aforementioned performance evaluation and validation results. Overall, the model with the best performance was the MLP model with the logistic sigmoid activation function in the output layer and the hyperbolic tangent sigmoid activation function with five neurons in the hidden layer. The validation results of the best model showed a sensitivity of 82.61%, specificity of 78.26%, accuracy of 80.43%, PPV of 79.17%, NPV of 81.82%, a Kappa index of 0.609, and AUC of 0.879. The results of this study highlight the effectiveness of selecting an optimal MLP model structure for shallow landslide susceptibility mapping using an appropriate predisposing factor section method.
In response to the sharp increase in geological disasters resulting from localized extreme rainfall events in Korea, a landslide early warning methodology is proposed. The method is based on ...sequential applications of the statistical and physically based hazard evaluation approaches and devised by combining the strengths of these two mutually complementary approaches. Following the decision algorithm for five phases of the warning level, the statistical evaluation was set to be applied first by using two different rainfall thresholds and one fixed geo-property (landslide susceptibility) threshold to determine a preliminary conservative warning level. To assess whether higher warning levels with higher certainties should be assigned, the physically based evaluation was set for application in the precondition of the preliminary warning stage. This was accomplished by using a rainfall threshold related to slope instability based on an advanced analysis using the physical modeling of landslide-triggering mechanisms. Consequently, a landslide early warning model based on the sequential evaluation approach was developed. The model ran through from transforming raw rainfall data to generating a series of geographic information system-based landslide early warning level maps. The spatially-discriminating capability and temporal applicability of the model on the distributed landslide events of 2009 were discussed by comparing the simulated results with landslide historical data while contemplating the adequacy of the durations and lead times of early warning levels. As a result, several advantages were identified for both spatial and temporal landslide early warning performances in the proposed methodology.
•A landslide early warning methodology is proposed and verified.•The decision algorithm sequentially applies statistical and physical thresholds.•The proposed methodology enhances the reliability and efficiency of early warnings.•The proposed methodology has several advantages for temporal and spatial performances.
In South Korea, the risk of debris-flow is relatively high due to the country's vast mountainous topographical features and intense continuous rainfall during the summer. Debris-flows can result in ...the loss of human life and severe property damage, which can be made worse due to the poor spatiotemporal predictability of such hazards. Therefore, it is essential to research the preemptive prediction and mitigation of debris-flow hazards. For this purpose, this study developed an ANN model to predict the debris-flow volume based on 63 historical events. By considering the morphology, rainfall, and geology characteristics of the studied area in central South Korea, the data of 15 debris-flow predisposing factors were obtained. Among these data, four predisposing factors (watershed area, channel length, watershed relief, and rainfall data) were selected based on Pearson's correlation analysis to check for significant correlations with the debris-flow volume. To determine the best performing ANN model, a validation testing was carried out involving ten-fold cross-validation with MSE and R2 using both training and validation datasets, which were randomly split into a 7:3 ratio. The model performance validation results showed that an ANN model with two hidden neurons (4×2×1 architecture) had the highest R2 value (0.828) and the lowest MSE (0.022). In addition, in a comparative study with other existing regression models, the ANN model showed better results in terms of adjusted R2 value (0.911) using all datasets. Furthermore, 94% of the observed debris-flow volumes from the ANN model were within 1:2 and 2:1 lines of the predicted volumes. The results of this study have shown the potentiality of the developed ANN model to be a useful resource for decision-making and designing barriers in areas prone to debris-flows in South Korea.
•A prediction model for debris-flow volume based on an artificial neural network (ANN) is proposed and verified.•The systematic procedure of modeling an ANN model for predicting the debris-flow volume is described.•The developed ANN model was compared to other existing regression models.•The ANN model showed better results in terms of predicting the debris-flow volume than the existing regression models.
Hanwoo, an important indigenous and popular breed of beef cattle in Korea, shows rapid growth and has high meat quality. Its yearling weight (YW) and carcass traits (backfat thickness, carcass ...weight- CW, eye muscle area, and marbling score) are economically important for selection of young and proven bulls. However, measuring carcass traits is difficult and expensive, and can only be performed postmortem. Genomic selection has become an appealing procedure for genetic evaluation of these traits (by inclusion of the genomic data) along with the possibility of multi-trait analysis. The aim of this study was to compare conventional best linear unbiased prediction (BLUP) and single-step genomic BLUP (ssGBLUP) methods, using both single-trait (ST-BLUP, ST-ssGBLUP) and multi-trait (MT-BLUP, MT-ssGBLUP) models to investigate the improvement of breeding-value accuracy for carcass traits and YW. The data comprised of 15,279 phenotypic records for YW and 5,824 records for carcass traits, and 1,541 genotyped animals for 34,479 single-nucleotide polymorphisms. Accuracy for each trait and model was estimated only for genotyped animals by five-fold cross-validation. ssGBLUP models (ST-ssGBLUP and MT-ssGBLUP) showed ~19% and ~36% greater accuracy than conventional BLUP models (ST-BLUP and MT-BLUP) for YW and carcass traits, respectively. Within ssGBLUP models, the accuracy of the genomically estimated breeding value for CW increased (19%) when ST-ssGBLUP was replaced with the MT-ssGBLUP model, as the inclusion of YW in the analysis led to a strong genetic correlation with CW (0.76). For backfat thickness, eye muscle area, and marbling score, ST- and MT-ssGBLUP models yielded similar accuracy. Thus, combining pedigree and genomic data via the ssGBLUP model may be a promising way to ensure acceptable accuracy of predictions, especially among young animals, for ongoing Hanwoo cattle breeding programs. MT-ssGBLUP is highly recommended when phenotypic records are limited for one of the two highly correlated genetic traits.
Recently, there has been a growing interest in the genetic improvement of body measurement traits in farm animals. They are widely used as predictors of performance, longevity, and production traits, ...and it is worthwhile to investigate the prediction accuracies of genomic selection for these traits. In genomic prediction, the single-step genomic best linear unbiased prediction (ssGBLUP) method allows the inclusion of information from genotyped and non-genotyped relatives in the analysis. Hence, we aimed to compare the prediction accuracy obtained from a pedigree-based BLUP only on genotyped animals (PBLUP-G), a traditional pedigree-based BLUP (PBLUP), a genomic BLUP (GBLUP), and a single-step genomic BLUP (ssGBLUP) method for the following 10 body measurement traits at yearling age of Hanwoo cattle: body height (BH), body length (BL), chest depth (CD), chest girth (CG), chest width (CW), hip height (HH), hip width (HW), rump length (RL), rump width (RW), and thurl width (TW). The data set comprised 13,067 phenotypic records for body measurement traits and 1523 genotyped animals with 34,460 single-nucleotide polymorphisms. The accuracy for each trait and model was estimated only for genotyped animals using five-fold cross-validations.
The accuracies ranged from 0.02 to 0.19, 0.22 to 0.42, 0.21 to 0.44, and from 0.36 to 0.55 as assessed using the PBLUP-G, PBLUP, GBLUP, and ssGBLUP methods, respectively. The average predictive accuracies across traits were 0.13 for PBLUP-G, 0.34 for PBLUP, 0.33 for GBLUP, and 0.45 for ssGBLUP methods. Our results demonstrated that averaged across all traits, ssGBLUP outperformed PBLUP and GBLUP by 33 and 43%, respectively, in terms of prediction accuracy. Moreover, the least root of mean square error was obtained by ssGBLUP method.
Our findings suggest that considering the ssGBLUP model may be a promising way to ensure acceptable accuracy of predictions for body measurement traits, especially for improving the prediction accuracy of selection candidates in ongoing Hanwoo breeding programs.
Quantitative polymerase chain reaction (qPCR) has been extensively used in the field of mesenchymal stem cell (MSC) research to elucidate their characteristics and clinical potential by normalization ...of target genes against reference genes (RGs), which are believed to be stably expressed irrespective of various experimental conditions. However, the expression of RGs is also variable depending on the experimental conditions, which may lead to false or contradictory conclusions upon normalization. Due to the current lack of information for a clear list of stable RGs in bovine MSCs, we conducted this study to identify suitable RGs in bovine MSCs.
The cycle threshold values of ten traditionally used RGs (18S ribosomal RNA 18S, beta-2-microglobulin B2M, H2A histone family, member Z H2A, peptidylprolyl isomerase A PPIA, ribosomal protein 4 RPL4, succinate dehydrogenase complex, subunit A SDHA, beta actin ACTB, glyceraldehyde-3-phosphate dehydrogenase GAPDH, TATA box binding protein TBP, and hypoxanthine phosphoribosyltrasnfrase1 HPRT1) in bovine bone marrow-derived MSCs (bBMMSCs) were validated for their stabilities using three types of RG evaluation algorithms (geNorm, Normfinder, and Bestkeeper). The effect of validated RGs was then verified by normalization of lineage-specific genes (fatty acid binding protein 4 FABP4 and osteonectin ON) expressions during differentiations of bBMMSCs or POU class 5 homeobox 1 (OCT4) expression between bBMMSCs and dermal skins.
Based on the results obtained for the three most stable RGs from geNorm (TBP, RPL4, and H2A), Normfinder (TBP, RPL4, and SDHA), and Bestkeeper (TBP, RPL4, and SDHA), it was comprehensively determined that TBP and RPL4 were the most stable RGs in bBMMSCs. However, traditional RGs were suggested to be the least stable (18S) or moderately stable (GAPDH and ACTB) in bBMMSCs. Normalization of FABP4 or ON against TBP, RPL4, and 18S presented significant differences during differentiation of bBMMSCs. However, although significantly low expression of OCT4 was detected in dermal skins compared to that in bBMMSCs when TBP and RPL4 were used in normalization, normalization against 18S exhibited no significance.
This study proposes that TBP and RPL4 were suitable as stable RGs for qPCR study in bovine MSCs.
Soil water content is one of the most important physical indicators of landslide hazards. Therefore, quickly and non-destructively classifying soils and determining or predicting water content are ...essential tasks for the detection of landslide hazards. We investigated hyperspectral information in the visible and near-infrared regions (400-1000 nm) of 162 granite soil samples collected from Seoul (Republic of Korea). First, effective wavelengths were extracted from pre-processed spectral data using the successive projection algorithm to develop a classification model. A gray-level co-occurrence matrix was employed to extract textural variables, and a support vector machine was used to establish calibration models and the prediction model. The results show that an optimal correct classification rate of 89.8% could be achieved by combining data sets of effective wavelengths and texture features for modeling. Using the developed classification model, an artificial neural network (ANN) model for the prediction of soil water content was constructed. The input parameter was composed of Munsell soil color, area of reflectance (near-infrared), and dry unit weight. The accuracy in water content prediction of the developed ANN model was verified by a coefficient of determination and mean absolute percentage error of 0.91 and 10.1%, respectively.
Damages caused by landslides have been increasing because of the greater frequency of localized heavy rain. To prevent landslide disasters more efficiently, more studies in relation to predicting the ...initiation areas of debris flows that cause large-scale damage are necessary. The main purpose of this study is to develop a criterion for detecting debris flow initiation areas by using an empirical method that was chosen from several types of approaches to debris flow initiation detection. In this study, ten GIS-based geomorphological datasets were obtained from slide and debris flow initiation areas located in Seoul, Gyeonggi Province and Gangwon Province. The geomorphological characteristics of slide and debris flow occurrence areas were analyzed through a comparative analysis to identify relationships between debris flow initiation and each topographic index. An initiation criterion for debris flows based on the geomorphological characteristics was suggested using an Artificial Neural Network (ANN) model combined with a modified threshold for the relationship between slope and upslope contributing area. To validate the suitability of the initiation criterion for debris flow, sequential applications of slope failure analysis and debris flow initiation analysis were conducted on a case study site to simulate the actual debris flow events. As a result of the debris flow initiation analysis, the prediction capture rate of the debris flow initiation criterion was determined to be 88.9%. Debris flow initiation areas totaling 143,300m2 were identified among 1,935,400m2 of slope failure areas predicted in the slope failure analysis. It is efficient to apply the debris flow initiation criterion in the debris flow simulation because slope failure areas are only partly transformed into debris flows.
•We developed a new criterion to predict the debris flow initiation area.•The criterion composed of an ANN model and a modified slope-UCA threshold was in good agreement with actual cases.•We conducted sequential applications of slope failure analysis and debris flow initiation analysis to validate the criterion.
An ever-increasing trend of extreme rainfall events in South Korea due to climate change is causing shallow landslides and shallow landslide induced debris flows in the mountains that cover 70% of ...the total land area of the nation. These catastrophic, gravity-driven processes cost the government several billion won in losses, and attendant fatalities, every year. The most common type of landslide observed is the shallow landslide occurring at 1–3m depth, which may mobilize into a catastrophic debris flow. A landslide early warning system encompassing different scale-based stages is used to predict potential areas for both the landslide types. Current study focusing on the first stage landslide hazard assessment at regional or medium scale requires the development of spatially evolving landslide hazard maps for both types of landslides based on the real-time rainfall. However, lack of complete landslide inventory data motivates the development of temporal and spatial models as independent components of the landslide hazard. Most of the existing temporal assessment schemes traditionally rooted in recurrence-based concepts does not consider soil factors and are not suitable to be incorporated in to the landslide early warning system since real-time rainfall cannot be considered. This motivated the development of a new probabilistic temporal model termed the extreme rainfall-induced landslide index. The probabilistic index was developed in Gangwon Province through a logistic regression using four factors; namely, continuous rainfall, 20-days antecedent rainfall, saturated hydraulic conductivity, and storage capacity. The developed model exhibited high area under the curve (AUC) values of 82% and 91% obtained for the training and validation curves, exhibiting good performance of the statistical index. Also, a high performance susceptibility model (training and validation AUC values of 96% and 94%, respectively) was developed using a logistic regression analysis, for Deokjeok-ri Creek, located in Gangwon province. Assuming the independence of the hazard components, a dynamic hazard index (DHI) was established through a joint probability of both the well validated models. The DHI was used to study the evolution of landslide hazard for the July 2006 extreme rainfall-induced landslide events in Deokjeok-ri Creek.
•New approach to include spatial and temporal response of slope instability.•Temporal assessment using an index considering rainfall and soil factors.•Database framework for soil parameter estimation over regional or medium scale.•Development of an extreme rainfall-induced landslide index and a dynamic hazard index.
PDF
