Optimizing Machine Learning Algorithms for Landslide Susceptibility Mapping along the Karakoram Highway, Gilgit Baltistan, Pakistan: A Comparative Study of Baseline, Bayesian, and Metaheuristic Hyperparameter Optimization Techniques.

Algorithms for machine learning have found extensive use in numerous fields and applications. One important aspect of effectively utilizing these algorithms is tuning the hyperparameters to match the specific task at hand. The selection and configuration of hyperparameters directly impact the performance of machine learning models. Achieving optimal hyperparameter settings often requires a deep understanding of the underlying models and the appropriate optimization techniques. While there are many automatic optimization techniques available, each with its own advantages and disadvantages, this article focuses on hyperparameter optimization for well-known machine learning models. It explores cutting-edge optimization methods such as metaheuristic algorithms, deep learning-based optimization, Bayesian optimization, and quantum optimization, and our paper focused mainly on metaheuristic and Bayesian optimization techniques and provides guidance on applying them to different machine learning algorithms. The article also presents real-world applications of hyperparameter optimization by conducting tests on spatial data collections for landslide susceptibility mapping. Based on the experiment's results, both Bayesian optimization and metaheuristic algorithms showed promising performance compared to baseline algorithms. For instance, the metaheuristic algorithm boosted the random forest model's overall accuracy by 5% and 3%, respectively, from baseline optimization methods GS and RS, and by 4% and 2% from baseline optimization methods GA and PSO. Additionally, for models like KNN and SVM, Bayesian methods with Gaussian processes had good results. When compared to the baseline algorithms RS and GS, the accuracy of the KNN model was enhanced by BO-TPE by 1% and 11%, respectively, and by BO-GP by 2% and 12%, respectively. For SVM, BO-TPE outperformed GS and RS by 6% in terms of performance, while BO-GP improved results by 5%. The paper thoroughly discusses the reasons behind the efficiency of these algorithms. By successfully identifying appropriate hyperparameter configurations, this research paper aims to assist researchers, spatial data analysts, and industrial users in developing machine learning models more effectively. The findings and insights provided in this paper can contribute to enhancing the performance and applicability of machine learning algorithms in various domains.

Climatic and topographic controls on glacial changes (1973-2020) in Shigar Basin, Central Karakoram, Northern Pakistan.

The Hindukush, Karakorum, and Himalaya (HKH) mountains are often referred to as the "Third Pole" because of high snow, being a major freshwater resource and early indicator of climate change. Therefore, research on the dynamics of glacier changes and their relationship with climate and topographic variability is essential for sustainable water resource management and adaptation strategies in Pakistan. In this contribution, we delineated 187 glaciers and examined these glacier changes in the Shigar Basin from 1973 to 2020 using Corona, Landsat Operational Land Imager/Enhanced Thematic Mapper Plus/Thematic Mapper/Multispectral Scanner System (OLI/ETM/TM/MSS), Alaska Satellite Facility (ASF), and Shuttle Radar Topography Mission Digital Elevation Model (SRTM DEM) imageries. The total glacier area decreased from 2796.31 ± 132 km2 in 1973 to 2756.27 ± 63 km2 in 2020 at an average rate of - 0.83 ± 0.03 km2yr-1. Specifically, during the period of 1990-2000, these glaciers shrank most heavily at an average rate of - 2.372 ± 0.08 km2yr-1. In contrast, an increased rate of 0.57 ± 0.02 km2yr-1 in total glacier area was observed during the recent decade (2010-2020). Moreover, the glaciers with gentle slopes retreated less heavily than the steep ones. There was reduction in glacier coverage and length for all slope classes, and a small reduction was observed with gentle slopes, while higher losses were observed on steep slope gradients. The transition of glaciers in the Shigar Basin may be attributed by the direct influence of glacier size and topographical characteristics. By comparing with climate records, our findings suggest that the overall reduction in glacier area from 1973 to 2020 was associated with declining precipitation (- 0.78 mmm/year) and rising temperature (0.045 °C/year) trends in the region, and glacier advances in recent decade (2010-2020) were likely to be driven by increased winter and autumn precipitation.

SBAS-InSAR based validated landslide susceptibility mapping along the Karakoram Highway: a case study of Gilgit-Baltistan, Pakistan.

Geological settings of the Karakoram Highway (KKH) increase the risk of natural disasters, threatening its regular operations. Predicting landslides along the KKH is challenging due to limitations in techniques, a challenging environment, and data availability issues. This study uses machine learning (ML) models and a landslide inventory to evaluate the relationship between landslide events and their causative factors. For this, Extreme Gradient Boosting (XGBoost), Random Forest (RF), Artificial Neural Network (ANN), Naive Bayes (NB), and K Nearest Neighbor (KNN) models were used. A total of 303 landslide points were used to create an inventory, with 70% for training and 30% for testing. Susceptibility mapping used Fourteen landslide causative factors. The area under the curve (AUC) of a receiver operating characteristic (ROC) is employed to compare the accuracy of the models. The deformation of generated models in susceptible regions was evaluated using SBAS-InSAR (Small-Baseline subset-Interferometric Synthetic Aperture Radar) technique. The sensitive regions of the models showed elevated line-of-sight (LOS) deformation velocity. The XGBoost technique produces a superior Landslide Susceptibility map (LSM) for the region with the integration of SBAS-InSAR findings. This improved LSM offers predictive modeling for disaster mitigation and gives a theoretical direction for the regular management of KKH.

Assessment of glacier status and its controlling parameters from 1990 to 2018 of Hunza Basin, Western Karakorum.

Ice masses and snow of Hunza River Basin (HRB) are an important primary source of fresh water and lifeline for downstream inhabitants. Changing climatic conditions seriously put an impact on these available ice and snow masses. These glaciers may affect downstream population by glacial lake outburst floods (GLOF) and surge events due to climatic variation. So, monitoring of these glaciers and available ice masses is important. This research delivers an approach for dynamics of major glaciers of the Hunza River Basin. We delineated 27 major glaciers of HRB and examined their status by using Landsat (OLI, ETM+, ETM, TM), digital elevation model (DEM) over the period of 1990-2018. In 1990, the total area covered by these glaciers is about 2589.75 ± 86 km2 and about 2565.12 ± 68km2 in 2018. Our results revealed that from 2009 to 2015, glacier coverage of HRB advanced with a mean annual advance rate of 2.22 ± 0.1 km2 a-1. Conversely, from 1994 to 1999, the strongest reduction in glacier area with a mean rate of - 3.126 ± 0.3 km2 a-1 is recorded. The glaciers of HRB are relatively stable compared to Hindukush, Himalayan, and Tibetan Plateau region of the world. The steep slope glacier's retreat rate is more than that of gentle slope glaciers, and the glaciers below an elevation of 5000 m above sea level change significantly. Based on climate data from 1995 to 2018, HRB shows a decreasing trend in temperature and increasing precipitation. The glacier area's overall retreat is due to an increase in summer temperature while the glacier advancement is induced possibly by winter and autumn precipitation.

Expansion of Shishper Glacier lake and recent glacier lake outburst flood (GLOF), Gilgit-Baltistan, Pakistan.

Shishper lake is an ice-dammed lake in northern Pakistan that has drained twice within one (1) year. The parameters evaluated in this paper are the lake's area, volume, peak discharge, and its outburst events using various satellite images from November 2018 to June 2020. Based on satellite imagery and empirical approaches, the lake formed in November 2018 and reached a maximum of 0.34 km2 till its first breach that occurred on 22 June 2019. Since June 2019, the lake drained till September 2019. After that, the flow was blocked again, and the lake expanded to an area of 0.27 km2 till its second outburst event that happened on May 29, 2020. Eight cross-sectional profiles of Hassanabad ravine are generated based on peak discharge in the lake's rapid outburst. The results indicate that, the peak discharge for both 2019 and 2020 was more than 4500 m3 s-1. Delineation of downstream Hassanabad ravine shows that more than 1000 buildings and 2000+ population is prone to flood. However, the lake drain twice steadily, but it has a high potential to cause severe damages if it bursts abruptly.

The role of sense of place, risk perception, and level of disaster preparedness in disaster vulnerable mountainous areas of Gilgit-Baltistan, Pakistan.

The study endeavored to analyze the risk perception, sense of place, and disaster preparedness in response to landslide disaster-prone mountain areas of Gilgit-Baltistan, Pakistan. To this end, we surveyed 315 rural residents of two vulnerable landslide districts (Hunza and Nagar) of Gilgit-Baltistan. To explore the relationships between the dimensions of risk perception, sense of place, and disaster preparedness, we used partial least squares (PLS) structural equation modeling (SEM) to test the hypotheses. The results derived from PLS-SEM have implied that there is a significant negative relationship between risk perception (apprehension and unidentified) with a sense of place (bond with society and place dependence). It was observed that the residents usually overestimate the risks of disasters due to their limited scientific knowledge regarding disaster occurrence, which reduces their dependencies on the place. We revealed that disaster preparedness enhances the place attachment and reduces the apprehension of landslides in the study area. This study devotes to government and relevant agencies to devise policies that can help relocate the vulnerable rural settlements, develop, and educate the masses on disaster mitigation and prevention strategies, and help prepare a suitable landslide management plan.

The dilemma of natural disasters: Impact on economy, fiscal position, and foreign direct investment alongside Belt and Road Initiative countries.

This paper investigates the damages and population affected by natural disasters based on percentile rankings, and analyzes the impact on the economy, per capita, fiscal balance, and foreign direct investment using novel panel algorithms including; Generalized Method of Moment (GMM), Crossectionally augmented Autoregressive Distributed Lags (CS-ARDL), and Driscoll & Kraay (DK) in Belt and Road initiative countries (B&RIC) over 1990-2018. The results indicate that severe natural disasters have negatively influenced economic growth with an average size of -0.016, which is transmitted to fiscal balance (-0.011) and foreign direct investment (-0.0271) in the long-run. The results also imply that the intensity of severe disasters on the fiscal position of the B&RIC countries is negative with an average effect of -0.011; however, the trade-openness, FDI, and economic activities support to improve the fiscal balance in the long-run. The outcomes of the study further revealed that foreign direct investment is more elastic in response to natural disasters in these countries. Therefore, it is recommended that the policymakers in B&RI countries should integrate the economic impacts of natural disasters in long-term economic planning. This would help the policymakers for better fiscal decisions, attracting FDI inflows and preparedness aftermath of natural disasters.