Search

This paper aims to apply and compare the performance of the three machine learning algorithms&ndash, support vector machine (SVM), bayesian logistic regression (BLR), and alternating decision tree (ADTree)&ndash, to map landslide susceptibility along the mountainous road of the Salavat Abad saddle, Kurdistan province, Iran. We identified 66 shallow landslide locations, based on field surveys, by recording the locations of the landslides by a global position System (GPS), Google Earth imagery and black-and-white aerial photographs (scale 1: 20,000) and 19 landslide conditioning factors, then tested these factors using the information gain ratio (IGR) technique. We checked the validity of the models using statistical metrics, including sensitivity, specificity, accuracy, kappa, root mean square error (RMSE), and area under the receiver operating characteristic curve (AUC). We found that, although all three machine learning algorithms yielded excellent performance, the SVM algorithm (AUC = 0.984) slightly outperformed the BLR (AUC = 0.980), and ADTree (AUC = 0.977) algorithms. We observed that not only all three algorithms are useful and effective tools for identifying shallow landslide-prone areas but also the BLR algorithm can be used such as the SVM algorithm as a soft computing benchmark algorithm to check the performance of the models in future.

In recent years, the intensification of drought and unsustainable management and use of water resources have caused a significant decline in the water level of the Urmia Lake in the northwest of Iran. This condition has affected the lake, approaching an irreversible point such that many projects have been implemented and are being implemented to save the natural condition of the Urmia Lake, among which the inter-basin water transfer (IBWT) project from the Zab River to the lake could be considered an important project. The main aim of this research is the evaluation of the IBWT project effects on the Gadar destination basin. Simulations of the geometrical properties of the river, including the bed and flow, have been performed, and the land cover and flood map were overlapped in order to specify the areas prone to flood after implementing the IBWT project. The results showed that with the implementation of this project, the discharge of the Gadar River was approximately tripled and the water level of the river rose 1 m above the average. In April, May, and June, about 952.92, 1458.36, and 731.43 ha of land adjacent to the river (floodplain) will be inundated by flood, respectively. Results also indicated that UNESCO’s criteria No. 3 (“a comprehensive environmental impact assessment must indicate that the project will not substantially degrade the environmental quality within the area of origin or the area of delivery”) and No. 5 (“the net benefits from the transfer must be shared equitably between the area of origin and the area of water delivery”) have been violated by implementing this project in the study area. The findings could help the local government and other decision-makers to better understand the effects of the IBWT projects on the physical and hydrodynamic processes of the Gadar River as a destination basin. Validerad;2020;Nivå 2;2020-01-09 (johcin)

It is presently well-known that more than 50% of total water quality impairment originates from non-point sources (NPS) of pollution. As an important NPS pollutant, runoff from agricultural lands contributes to water quality problems more than other non-point sources. Therefore, the identification and quantification of runoff generation areas is crucial for source water protection and nutrient management. Runoff generation is a complicated landscape process, affected by various factors in different seasons of the year. In this research, monitoring and modeling were selected as two important approaches to identify the mechanisms of runoff generation, runoff generating areas (RGAs) and its variability in time and space in a small agricultural watershed in southern Ontario. A wireless sensor network (WSN) was designed to monitor runoff generating areas in the study watershed with the ability to measure the depth of surface runoff and soil moisture over ten minutes time intervals. Eight pressure and soil moisture sensors were installed at the outlet of eight fields in the watershed. Data from eighteen natural rainfall events for the period from July 2008 to April 2009 were analyzed to study the spatial and temporal variability of runoff generation areas in the study watershed. The results showed that runoff generating areas in the watershed are highly dynamic in summer, fall and spring with differences in 100%-contribution-status persistency. The results also indicated that 15% of the watershed generates 75% of surface runoff during summer, 100% in fall and 45% during spring. In spite of the dynamic nature of RGAs, the sensitivity of different fields in the watershed in response to rainfall events remained constant, such that some specific fields responded first in all three seasons. This finding led to the introduction of a Slope/Area index for the identification of sensitive fields in the study watershed. Statistical analyses of the factors affecting RGAs indicated that the factors affecting the spatial and temporal variability of RGAs in three seasons vary; however, the soil moisture and rainfall intensity played important roles in the runoff generation mechanism and variability of contributing areas in all seasons. Based on monitored results and field observations, a hydrological model was developed to simulate runoff generating area and to classify the sensitivity of the fields to runoff generation on the basis of the modified Soil Conservation Service Curve Number approach. The model was able to identify the fields that generate runoff and classify the sensitivity of the fields in the watershed. The developed model could simulate RGAs for the summer season with higher degree of accuracy than fall. The developed model needs further improvements for simulation of runoff generating area in the spring season.