Search Results (113)

Droughts are increasingly threatening ecological balance, agricultural productivity, and socio-economic resilience—especially in semi-arid regions like the Inner Mongolia segment of China’s Yellow River Basin. This study presents a hybrid drought forecasting framework integrating machine learning (ML) and deep learning (DL) models with high-resolution historical and downscaled future climate data. TerraClimate observations (1985–2014) and bias-corrected CMIP6 projections (2030–2050) under SSP2-4.5 and SSP5-8.5 scenarios were utilized to develop and evaluate the models. Among the tested ML algorithms, Random Forest (RF) demonstrated the best trade-off between accuracy and interpretability and was selected for feature importance analysis. The top-ranked predictors—precipitation, solar radiation, and maximum temperature—were used to train a Long Short-Term Memory (LSTM) network. The LSTM outperformed all ML models, achieving high predictive skill (R² = 0.766, CC = 0.880, RMSE = 0.885). Scenario-based projections revealed increasing drought severity and variability under SSP5-8.5, with mean PDSI values dropping below −3 after 2040 and deepening toward −4 by 2049. The high-emission scenario also exhibited broader uncertainty bands and amplified interannual anomalies. These findings highlight the value of hybrid AI–climate modeling approaches in capturing complex drought dynamics and supporting anticipatory water resource planning in vulnerable dryland environments. Full article

Urban flood damage is increasing globally, particularly in major cities. Factors contributing to flood risk include urban environmental changes, such as watershed development and precipitation variations caused by climate change. Rapid urbanization and weather anomalies further complicate flood management and damage mitigation. Additionally, detailed analyses at small spatial units (e.g., roads, buildings) remain insufficient. Hence, urban flood analysis considering such spatial variations is required. This study analyzed flood risk in Ulsan, Korea, under a severe flood scenario. Land cover changes from the 1980s to 2010s were examined in 10-year intervals, along with the frequency of heavy rainfall and high river water levels that trigger severe floods. Flood risk was structured as a matrix of likelihood and impact. The results revealed that land cover changes, influenced by development policies or regulations, had a minimal impact on urban flood risk, which is likely because effective drainage systems and stringent urban planning regulations mitigated their effects. However, the frequency and intensity of extreme precipitation events had a substantial effect. These findings were validated using a comparative analysis of an inundation damage trace map and flood range simulated by a physical model. The 10 m grid resolution and time-series likelihood-and-impact framework used in this study can inform budget allocation, resource mobilization, disaster prevention planning, and decision-making during disaster response efforts in major cities. Full article

Understanding precipitation variability is essential for assessing climate dynamics and their impacts on agriculture, water resources, and infrastructure. This study analyzes subseasonal precipitation patterns in Rio Grande do Sul, Brazil, using 45-day accumulated intervals over a 17-year period (2006–2022), a timescale critical for understanding drivers of extreme events like the catastrophic floods of 2024. A total of 138 precipitation fields were generated from 670 spatial points. Spatial analysis revealed median precipitation values ranging from 130 to 329 mm/45 days, with the northeast showing the highest accumulations and the southwest showing the driest conditions. Temporal variability was marked by abrupt anomalies, with median peaks up to 462 mm and minima of 33 mm. Significant temporal autocorrelation (lag-1, 45 days) was identified in the central and northern regions, while lag-2 (90 days) showed inverse patterns in the south (correlation coefficient ≈ −0.45). Principal component analysis (KMO = 0.909; Bartlett’s χ² = 187,990.945; p < 0.05) identified seven dominant modes, with PC1 explaining 26% of total variance and highlighting extremely wet anomalies (e.g., SPI > 2.0). Correlation with the Oceanic Niño Index revealed heterogeneous responses to ENSO phases, with strong El Niño episodes (2009, 2015–2016) associated with precipitation peaks up to 966 mm/45 days. These results underscore the importance of subseasonal scales for understanding climate anomalies and support the development of regional forecast strategies and water management policies under increasing climate variability. Full article

The Xiaobaihegou fluorite deposit is located in the southwest of the Altyn-Tagh Orogen, NW China. However, the provenance, thermodynamic properties, and enrichment mechanisms of the ore-forming fluids in this deposit remain unclear. Fluorite mineralization primarily occurs in the vicinity of the contact zone between the granite and the wall rocks. The zircon U-Pb age of the alkali-feldspar granite in the Xiaobaihegou fluorite deposit is 482.3 ± 4.1 Ma. The ore-hosting lithologies are mainly calcareous rock series of the Altyn Group. The ore bodies are controlled by NE-trending faults and consist primarily of veined, brecciated, massive, and banded ores. The ore mineral assemblage is primarily composed of calcite and fluorite. The rare earth element (REE) patterns of fluorite and calcite in the Xiaobaihegou deposit exhibit right-dipping LREE enrichment with distinct negative Eu anomalies, which closely resemble those of the alkali-feldspar granite. This similarity suggests that the REE distribution patterns of fluorite and calcite were likely inherited from the pluton. The ore-forming process can be divided into an early stage and a late stage. The massive ores formed in the early stage contain mainly gas-rich two-phase fluid inclusions and CO₂-bearing three-phase inclusions, with homogenization temperatures ranging from 235 °C to 426 °C and salinities from 28.59% to 42.40% NaCl equivalent. In the late stage, brecciated and stockwork ores were formed. They host liquid-rich two-phase and gas-rich two-phase fluid inclusions, with homogenization temperatures ranging from 129 °C to 350 °C and salinities from 0.88% to 21.61% NaCl equivalent. The results of hydrogen and oxygen isotope studies indicate that the ore-forming fluids were derived from a mixture of magmatic–hydrothermal and meteoric water. Fluorite precipitation in the early stage was mainly due to the mixing of magmatic–hydrothermal solution and meteoric water, as well as a water–rock reaction. In the late stage, fluid mixing further occurred, resulting in a decrease in temperature and the formation of brecciated and stockwork ores. The ⁸⁷Sr/⁸⁶Sr and ¹⁴³Nd/¹⁴⁴Nd ratios of fluorite from the deposit range from 0.71033 to 0.71272 and 0.511946 to 0.512073, respectively, indicating that the ore-forming material originates from the crust. Based on the ore-forming characteristics, it is proposed that Ca may be primarily leached from the strata formation, while F may predominantly originate from magmatic–hydrothermal solutions. The formation of fluorite deposits is closely related to the transition of the Central Altyn-Tagh Block and Qaidam Block from a compressional orogenic environment to an extensional tectonic environment. Full article

The Amazon Basin plays a crucial role in the global hydrological cycle, where seasonal and interannual variations in terrestrial water storage (TWS) are essential for understanding climate–hydrology coupling mechanisms. This study utilizes data from the Gravity Recovery and Climate Experiment (GRACE) satellite mission and its follow-on mission (GRACE-FO, collectively referred to as GRACE) to investigate the spatiotemporal dynamics of hydrological mass changes in the Amazon Basin from 2002 to 2021. Results reveal pronounced spatial heterogeneity in the annual amplitude of TWS, exceeding 65 cm near the Amazon River and decreasing to less than 25 cm in peripheral mountainous regions. This distribution likely reflects the interplay between precipitation and topography. Vertical displacement measurements from the Global Navigation Satellite System (GNSS) show strong correlations with GRACE-derived hydrological load deformation (mean Pearson correlation coefficient = 0.72) and reduce its root mean square (RMS) by 35%. Furthermore, the study demonstrates that existing hydrological models, which neglect groundwater dynamics, underestimate hydrological load deformation. Principal component analysis (PCA) of the Amazon GNSS network demonstrates that the first principal component (PC) of GNSS vertical displacement aligns with abrupt interannual TWS fluctuations identified by GRACE during 2010–2011, 2011–2012, 2013–2014, 2015–2016, and 2020–2021. These fluctuations coincide with extreme precipitation events associated with the El Niño–Southern Oscillation (ENSO), confirming that ENSO modulates basin-scale interannual hydrological variability primarily through precipitation anomalies. This study provides new insights for predicting extreme hydrological events under climate warming and offers a methodological framework applicable to other critical global hydrological regions. Full article

Evapotranspiration (ET) crucially regulates water storage dynamics and is an essential component of the terrestrial water cycle. Understanding ET dynamics is fundamental for sustainable water resource management, particularly in regions facing increasing drought risks under climate change. In regions like southwestern China, where extreme drought events are prevalent due to complex terrain and climate warming, ET becomes a key factor in understanding water availability and drought dynamics. Using the SWAT model, this study investigates ET dynamics and influencing factors in the Jizi Basin, Yunnan Province, a small basin with over 71% forest coverage. The model calibration and validation results demonstrated a high degree of consistency with observed discharge data and ERA5, confirming its reliability. The results show that the annual average ET in the Jizi Basin is 573.96 mm, with significant seasonal variations. ET in summer typically ranges from 70 to 100 mm/month, while in winter, it drops to around 20 mm/month. Spring ET exhibits the highest variability, coinciding with the occurrence of extreme hydrological events such as droughts. The monthly anomalies of ET effectively reproduce the spring and early summer 2019 drought event. Notably, ET variation exhibits significant uncertainty under scenarios of +1 °C temperature and −20% precipitation. Furthermore, although land use changes had relatively small effects on overall ET, they played crucial roles in promoting groundwater recharge through enhanced percolation, especially forest cover. The study highlights that, in addition to climate and land use, soil moisture and groundwater conditions are vital in modulating ET and drought occurrence. The findings offer insights into the hydrological processes of small forested basins in southwestern China and provide important support for sustainable water resource management and effective climate adaptation strategies, particularly in the context of increasing drought vulnerability. Full article

Groundwater storage refers to the water stored in the pore spaces of underground aquifers, which has been increasingly affected by both climate change and anthropogenic activities in recent decades. Therefore, monitoring their changes and the factors that affect it is of great importance. Although the influence of natural factors on groundwater is well-recognized, the impact of human activities, despite being a major contributor to its change, has been less explored due to the challenges in measuring such effects. To address this gap, our study employed an integrated approach using remote sensing and the Google Earth Engine (GEE) cloud-free platform to analyze the effects of various anthropogenic factors such as built-up areas, cropland, and surface water on groundwater storage in the Lake Urmia Basin (LUB), Iran. Key anthropogenic variables and groundwater data were pre-processed and analyzed in GEE for the period from 2000 to 2022. The processes linking these variables to groundwater storage were considered. Built-up area expansion often increases groundwater extraction and reduces recharge due to impervious surfaces. Cropland growth raises irrigation demand, especially in semi-arid areas like the LUB, leading to higher groundwater use. In contrast, surface water bodies can supplement water supply or enhance recharge. The results were then exported to XLSTAT software2019, and statistical analysis was conducted using the Mann–Kendall (MK) non-parametric trend test on the variables to investigate their potential relationships with groundwater storage. In this study, groundwater storage refers to variations in groundwater storage anomalies, estimated using outputs from the Global Land Data Assimilation System (GLDAS) model. Specifically, these anomalies are derived as the residual component of the terrestrial water budget, after accounting for soil moisture, snow water equivalent, and canopy water storage. The results revealed a strong negative correlation between built-up areas and groundwater storage, with a correlation coefficient of −1.00. Similarly, a notable negative correlation was found between the cropland area and groundwater storage (correlation coefficient: −0.85). Conversely, surface water availability showed a strong positive correlation with groundwater storage, with a correlation coefficient of 0.87, highlighting the direct impact of surface water reduction on groundwater storage. Furthermore, our findings demonstrated a reduction of 168.21 mm (millimeters) in groundwater storage from 2003 to 2022. GLDAS represents storage components, including groundwater storage, in units of water depth (mm) over each grid cell, employing a unit-area, mass balance approach. Although storage is conceptually a volumetric quantity, expressing it as depth allows for spatial comparison and enables conversion to volume by multiplying by the corresponding surface area. Full article

An ancient repertoire of ultraviolet (UV)-absorbing pigments which survive today in the phylogenetically oldest extant photosynthetic organisms, the cyanobacteria, point to a direction in evolutionary adaptation of the pigments and their associated biota; from largely UV-C absorbing pigments in the Archean to pigments covering ever more of the longer wavelength UV and visible regions in the Phanerozoic. Since photoprotection is not dependent on absorption, such a scenario could imply selection of photon dissipation rather than photoprotection over the evolutionary history of life, consistent with the thermodynamic dissipation theory of the origin and evolution of life which suggests that the most important hallmark of biological evolution has been the covering of Earth’s surface with organic pigment molecules and water to absorb and dissipate ever more completely the prevailing surface solar spectrum. In this article we compare a set of photophysical, photochemical, biosynthetic, and other inherent properties of the two dominant classes of cyanobacterial UV-absorbing pigments, the mycosporine-like amino acids (MAAs) and scytonemins. We show that the many anomalies and paradoxes related to these biological pigments, for example, their exudation into the environment, spectral coverage of the entire high-energy part of surface solar spectrum, their little or null photoprotective effect, their origination at UV-C wavelengths and then spreading to cover the prevailing Earth surface solar spectrum, can be better understood once photodissipation, and not photosynthesis or photoprotection, is considered as being the important variable optimized by nature. Full article

How can the shape of biodiversity inform us about cities’ ecoclimatic fitness and guide their development? Can we use species as the harbingers of climatic extremes? Eco-climatically sensitive species carry information about hydroclimatic change in their distribution, fitness, and preferential gradients of habitat suitability. Conversely, environmental features outside of the species’ fitness convey information on potential ecological anomalies in response to extremes to adapt or mitigate, such as through urban parks. Here, to quantify ecosystems’ fitness, we propose a novel computational model to extract multivariate functional ecological networks and their basins, which carry the distributed signature of the compounding hydroclimatic pressures on sentinel species. Specifically, we consider butterflies and their habitat suitability (HS) to infer maximum suitability gradients that are meaningful of potential species networks and flows, with the smallest hydroclimatic resistance across urban landscapes. These flows are compared to the distribution of urban parks to identify parks’ ecological attractiveness, actual and potential connectivity, and park potential to reduce hydroclimatic impacts. The ecosystem fitness index (EFI) is novelly introduced by combining HS and the divergence of the relative species abundance (RSA) from the optimal log-normal Preston plot. In Shenzhen, as a case study, eco-flow networks are found to be spatially very extended, scale-free, and clustering for low HS gradient and EFI areas, where large water bodies act as sources of ecological corridors draining into urban parks. Conversely, parks with higher HS, HS gradients, and EFIs have small-world connectivity non-overlapping with hydrological networks. Diverging patterns of abundance and richness are inferred as increasing and decreasing with HS. HS is largely determined by temperature and precipitation of the coldest quarter and seasonality, which are critical hydrologic variables. Interestingly, a U-shape pattern is found between abundance and diversity, similar to the one in natural ecosystems. Additionally, both abundance and richness are mildly associated with park area according to a power function, unrelated to longitude but linked to the degree of urbanization or park centrality, counterintuitively. The Preston plot’s richness–abundance and abundance-rank patterns were verified to reflect the stationarity or ecological meta-equilibrium with the environment, where both are a reflection of community connectivity. Ecological fitness is grounded on the ecohydrological structure and flows where maximum HS gradients are indicative of the largest eco-changes like climate-driven species flows. These flows, as distributed stress-response functions, inform about the collective eco-fitness of communities, like parks in cities. Flow-based networks can serve as blueprints for designing ecotones that regulate key ecosystem functions, such as temperature and evapotranspiration, while generating cascading ecological benefits across scales. The proposed model, novelly infers HS eco-networks and calculates the EFI, is adaptable to diverse sensitive species and environmental layers, offering a robust tool for precise ecosystem assessment and design. Full article

A variety of variables, such as organic matter input, redox conditions, depositional rates, and terrigenous input, affect the deposition of black shale. Furthermore, because of the significant regional variations in paleodepositional environments, these factors have a complex role in organic matter enrichment. Global geological events influenced sedimentary conditions, organic enrichment, and the development of organic-enriched shales during the Late Ordovician to Early Silurian. The Wufeng–Longmaxi Formation black shales in Southeastern Chongqing were analyzed for X-ray diffraction (XRD), major and trace element geochemistry, and total organic carbon (TOC) data; this led to further analysis of the relationship between the depositional environment and organic matter aggregation and rock type evolution. The primary minerals found in the Wufeng–Longmaxi shale are quartz, feldspar, carbonatite (calcite and dolomite), and clay. The high index of compositional variability (ICV) values (>1) and the comparatively low chemical index of alteration (CIA) values (52.6–72.8) suggest that the sediment source rocks are juvenile and are probably experiencing weak to moderate chemical weathering. The selected samples all show negative Eu anomalies, flat heavy rare earth elements, and mildly enriched light rare earth elements. The ratios of La/Th, La/Sc, Th/Sc, ΣREE-La/Yb, TiO₂-Ni, and La/Th-Hf suggest that acidic igneous rocks were the main source of sediment, with minor inputs from ancient sedimentary rocks. The correlations of paleoclimate proxies (Sr/Cu, CIA), redox proxies (V/Cr, V/Ni, V/(V + Ni), Ni/Co, U/Th), paleoproductivity proxies (Ba_xs, Cu_EF, Ni_EF), and water mass restriction proxies (Mo/TOC, U_EF, Mo_EF) suggest a humid–semiarid, anoxic, moderate–high paleoproductivity, and moderate–strongly restricted environment. On the basis of the aforementioned interpretations, the paleoenvironment of the Wufeng–Longmaxi Formations was established, with paleoredox conditions and restricted water masses likely being the primary factors contributing to organic matter enrichment. Full article

On 23 January 2024, a M_S7.1 earthquake struck Wushi County, Xinjiang Uygur Autonomous Region, marking the largest seismic event in the Southern Tianshan (STS) region in the past century. This study investigates the relationship between hydrothermal fluid circulation and seismic activity by analyzing the chemical composition and origin of fluids in natural hot springs along the Maidan Fracture (MDF). Results reveal two distinct hydrochemical water types (Ca-HCO₃ and Ca-Mg-Cl). The δD and δ¹⁸O values indicating spring water are influenced by atmospheric precipitation input and altitude. Circulation depths (621–3492 m) and thermal reservoir temperatures (18–90 °C) were estimated. Notably, the high ³He/⁴He ratios (3.71 Ra) and mantle-derived ³He content reached 46.48%, confirming that complex gas–water–rock interactions occur at fracture intersections. Continuous monitoring at site S13 (144 km from the epicenter of the Wushi M_S7.1 earthquake) captured pre-and post-seismic hydrogeochemical fingerprints linked to the Wushi MS7.1 earthquake. Stress accumulation along the MDF induced permeability changes, perturbing hydrogeochemical equilibrium. At 42 days pre-Wushi M_S7.1 earthquake, δ¹³C DIC exceeded +2σ thresholds (−2.12‰), signaling deep fracture expansion and CO₂ release. By 38 days pre-Wushi M_S7.1 earthquake, Na⁺, SO₄²⁻, and δ¹⁸O surpassed 2σ levels, reflecting hydraulic connection between deep-seated and shallow fracture networks. Ion concentrations and isotope values showed dynamic shifts during the earthquake, which revealed episodic stress transfer along fault asperities. Post-Wushi M_S7.1 earthquake, fracture closure reduced deep fluid input, causing δ¹³C DIC to drop to −4.89‰, with ion concentrations returning to baseline within 34 days. Trace elements such as Be and Sr exhibited anomalies 12 days before the Wushi M_S7.1 earthquake, while elements like Li, B, and Rb showed anomalies 24 days after the Wushi M_S7.1 earthquake. Hydrochemical monitoring of hot springs captures such critical stress-induced signals, offering vital insights for earthquake forecasting in tectonically active regions. Full article

In this study, survey methods including seismic techniques and controlled-source audio-frequency magnetotelluric, drilling, and pumping tests were employed to investigate the geothermal systems and their formation mechanisms in northern Shanxi Province, China. The following characteristics were observed: (1) Geothermal resources in northern Shanxi Province are primarily located in Archean metamorphic rocks and fracture zone aquifer groups. The direct heat source is likely uncooled magma chambers in the middle-upper crust, whereas the overlying layers consist of Quaternary, Neogene, and Paleogene deposits. (2) The high-temperature geothermal system is of the convective-conductive type: atmospheric precipitation and surface water infiltrate pore spaces and fault fractures to reach thermal storage, where they are heated. Hot water then rises along the fracture channels and emerges as shallow hot springs, and ongoing extensional tectonic activity has caused asthenospheric upwelling. The partial melting of the upper mantle forms basic basaltic magma, which ascends to the middle-upper crust and forms multiple magma chambers. Their heat is transferred to the shallow subsurface, causing geothermal anomalies. (3) Borehole YG-1 findings revealed that these geothermal resources are primarily static reserves. Our findings provide a foundation for further geothermal development in the region, including the strategic deployment of wells to improve geothermal energy extraction. Full article

Agricultural practices in the cold desert region of the Himalayas are frequently affected by climate-induced uncertainty in the past few decades. This research work aimed to examine the following questions: (a) Are there any significant climatic changes in the cold desert region of Himachal Himalayas? (b) How do the local farmers perceive climate change? (c) What and how indigenous and modern climate sensitive resilience measures/practices are being adapted by farmers for risk mitigation? A modified Mann–Kendall (m-MK) test and anomaly index were used to examine the changes in climatic variables over the cold desert region. Data on the observed changes in climatic variables were investigated through gridded products provided by the Indian Meteorological Department (IMD) and farmer perception, and their adaptation measures were collected by an extensive primary survey using a semi-structured questionnaire. The results indicate that farmers’ perceptions of changing rainfall, temperature, and seasons were consistent with historical climatic data. The drying water resources and crop damage were the most pressing concerns for farmers due to climate change activity. The farmers are adapting to climate change by altering their farming practices for agricultural risk management. The binary logistics regression (BLR) model was used to investigate the influence of different variables on the adopting farmer’s decision. The result revealed that various factors like landholding size, accessibility of transport, awareness of climate change, availability of water, and distance from market were responsible for choosing suitable climate resilience adaptation measures. This research contributes to recalibrating appropriate strategies across the cold desert region for designing sustainable agricultural practices. Full article

Droughts, intensified by climate change and human activities, pose a significant threat to winter wheat cultivation in the Huang-Huai-Hai (HHH) region. Soil moisture drought indices are crucial for monitoring agricultural droughts, while challenges such as data accessibility and soil heterogeneous necessitate the use of numerical simulations for their effective regional-scale applications. The existing simulation methods like physical process models and machine learning (ML) algorithms have limitations: physical models struggle with parameter acquisition at regional scales, while ML algorithms face difficulties in agricultural settings due to the presence of crops. As a more advanced and complex branch of ML, deep learning algorithms face even greater limitations related to crop growth and agricultural management. To address these challenges, this study proposed a novel hybrid monitoring system that merged ML algorithms with a physical process model. Initially, we employed the Random Forest (RF) regression model that integrated multi-source environmental factors to estimate soil moisture prior to the sowing of winter wheat, achieving an average coefficient of determination (R²) of 0.8618, root mean square error (RMSE) of 0.0182 m³ m⁻³, and mean absolute error (MAE) of 0.0148 m³ m⁻³ across eight soil depths. The RF regression models provided vital parameters for the operation of the Water Balance model for Winter Wheat (WBWW) at a regional scale, enabling effective drought monitoring and assessments combined with the Soil Moisture Anomaly Percentage Index (SMAPI). Subsequent comparative analyses between the monitoring system-generated results and the actual disaster records during two regional-scale drought events highlighted its efficacy. Finally, we utilized this monitoring system to examine the spatiotemporal variations in drought patterns in the HHH region over the past two decades. The findings revealed an overall intensification of drought conditions in winter wheat, with a decline in average SMAPI at a rate of −0.021% per year. Concurrently, there has been a significant shift in drought patterns, characterized by an increase in both the frequency and extremity of drought events, while the duration and intensity of individual drought events have decreased in a majority of the HHH region. Additionally, we identified the northeastern, western, and southern areas of HHH as areas requiring concentrated attention and targeted intervention strategies. These efforts signify a notable application of multi-source data fusion techniques and the integration of physical process models within a big data context, thereby facilitating effective drought prevention, agricultural management, and mitigation strategies. Full article

Critical National Infrastructure includes large networks such as telecommunications, transportation, health services, police, nuclear power plants, and utilities like clean water, gas, and electricity. The protection of these infrastructures is crucial, as nations depend on their operation and stability. However, cyberattacks on such systems appear to be increasing in both frequency and severity. Various machine learning approaches have been employed for anomaly detection in Critical National Infrastructure, given their success in identifying both known and unknown attacks with high accuracy. Nevertheless, these systems are vulnerable to adversarial attacks. Hackers can manipulate the system and deceive the models, causing them to misclassify malicious events as benign, and vice versa. This paper evaluates the robustness of traditional machine learning techniques, such as Support Vector Machines (SVMs) and Logistic Regression (LR), as well as Artificial Neural Network (ANN) algorithms against adversarial attacks, using a novel dataset captured from a model of a clean water treatment system. Our methodology includes four attack categories: random label flipping, targeted label flipping, the Fast Gradient Sign Method (FGSM), and Jacobian-based Saliency Map Attack (JSMA). Our results show that, while some machine learning algorithms are more robust to adversarial attacks than others, a hacker can manipulate the dataset using these attack categories to disturb the machine learning-based anomaly detection system, allowing the attack to evade detection. Full article