GIS-IDEAS 2026

““Application of Deep Learning for Automated Tree Extraction in Urban Environments Using Remote Sensing Data””

Nguyen Quang Minh, Yashwanth gowda, Harini j, Madhusudhan KM Reddy, Priya Narayanswamy; GIS-IDEAS Oral Presentation

Urban trees are a vital component of sustainable cities, providing ecological, social, and economic benefits. They help mitigate the urban heat island effect, improve air quality, enhance stormwater management, and contribute to the overall livability of urban environments. Accurate mapping and monitoring of trees are therefore essential for effective urban planning and environmental management. However, traditional methods of tree extraction from satellite or aerial imagery are often labor-intensive, time-consuming, and prone to errors due to the complexity of urban landscapes, where vegetation is interspersed with buildings, roads, and other infrastructure. This challenge has motivated researchers to explore advanced computational approaches, particularly deep learning, to automate and improve the accuracy of tree detection and extraction in urban areas.

Deep learning, a subset of artificial intelligence, has shown remarkable success in image recognition and semantic segmentation tasks. Convolutional Neural Networks (CNNs) and advanced architectures such as U-Net and Mask R-CNN are particularly well-suited for distinguishing trees from other urban features in high-resolution remote sensing imagery. These models can learn complex spatial patterns and spectral signatures of vegetation, enabling them to separate tree canopies from impervious surfaces with high precision. By training on annotated datasets, deep learning models can generalize across diverse urban contexts, making them powerful tools for large-scale urban forestry applications.

The proposed study outlines a methodology that integrates deep learning with Geographic Information Systems (GIS) to achieve automated tree extraction. The process begins with data collection from UAV-based aerial photographs and point clouds. Preprocessing steps, including image enhancement and manual annotation, prepare the data for model training. Deep learning models are then applied to classify and segment tree canopies, with performance evaluated using metrics such as accuracy, precision, recall, and F1-score. The extracted tree data is subsequently integrated into GIS platforms, allowing for spatial analysis and visualization of urban greenery. This integration enables planners to assess tree distribution, identify gaps in green coverage, and design strategies for enhancing urban sustainability.

A case study in a rapidly urbanizing city can demonstrate the practical application of this methodology. By comparing deep learning-based extraction results with traditional classification approaches, the study can highlight improvements in efficiency and accuracy. The findings will underscore the potential of deep learning to support evidence-based urban forestry management, providing policymakers and planners with reliable data to guide interventions. Ultimately, this research contributes to the broader agenda of sustainable urban development by offering a replicable framework for monitoring and managing urban trees, ensuring that cities remain resilient, livable, and environmentally balanced.

““Leveraging Geographic Information Systems (GIS) for Quantitative Assessment of Transit-Oriented Development (TOD) Parameters: A Case Study Approach””

Nguyen Quang Minh, Nguyen Viet Hoa, Nguyen Thi Bich Phuong; GIS-IDEAS Poster Presentation

This paper proposes a methodology for applying Geographic Information Systems (GIS) to calculate and evaluate Transit-Oriented Development (TOD) parameters. TOD represents a contemporary planning paradigm that emphasizes compact, mixed-use, and walkable communities organized around high-quality public transit systems. The central idea is to create urban environments where residents can easily access transit services, daily amenities, and employment opportunities without relying heavily on private vehicles. By integrating spatial datasets, GIS provides a powerful tool for measuring TOD indicators with precision, enabling planners and researchers to assess the performance of transit-oriented areas and identify opportunities for improvement. The methodology outlined in this study includes data collection, spatial analysis, and visualization of TOD metrics, and is demonstrated through a case study in Hanoi. The results highlight how GIS-based TOD assessment can support evidence-based urban planning, reveal gaps in transit accessibility, and guide sustainable development policies.

Transit-Oriented Development can be evaluated through a set of parameters that GIS helps quantify and visualize. Accessibility is a core parameter, measured by walking distance to transit stations and the proportion of population and jobs located within service catchments. This indicator reflects how well transit systems are integrated into the urban fabric and whether they are positioned to serve the majority of residents effectively. Density indicators, including population density, employment density, and floor area ratio (FAR), capture the intensity of land use around transit nodes. Higher densities are critical for generating sufficient ridership and supporting vibrant, transit-oriented communities. Land-use mix, often assessed through entropy indices, reflects the diversity of residential, commercial, and public functions within a given area. A balanced mix reduces travel needs, encourages walking and cycling, and fosters lively neighborhoods where daily activities can be accomplished locally.

Urban design and walkability parameters further enrich TOD evaluation. These include measures of street connectivity, block size, intersection density, and the availability of sidewalks and public spaces. GIS-based network analysis can reveal how well the street grid supports pedestrian movement and whether barriers exist that limit access to transit. Environmental sustainability is another essential dimension, gauged through potential reductions in greenhouse gas emissions, provision of green spaces, and energy efficiency in buildings. By integrating transport data with emission models and land cover analysis, GIS can quantify the ecological benefits of TOD. Finally, socio-economic outcomes provide a broader perspective on TOD impacts, including property value changes, job creation, housing affordability, and equity. These indicators ensure that TOD projects contribute not only to mobility but also to inclusive and resilient urban development.

“A Deep Learning Framework for Spatio-Temporal Wind Forecasting and Offshore Wind Potential Assessment in the Gulf of Tonkin”

Nguyễn Văn Thái, Van Truong VU, Hien Nguyen THI; GIS-IDEAS Oral Presentation

Accurate short-term wind power density (WPD) forecasting over complex coastal regions is a critical enabler for offshore wind farm planning and grid integration. However, wind dynamics exhibit highly complex nonlinear spatiotemporal dependencies that conventional models fail to capture effectively. This paper presents a deep learning framework based on Convolutional Long Short-Term Memory (ConvLSTM) networks for simultaneous spatio-temporal prediction of WPD across the Gulf of Tonkin, one of Southeast Asia's most promising offshore wind corridors. Unlike conventional approaches that treat spatial and temporal dependencies separately, the proposed architecture inherently encodes both dimensions within a unified recurrent convolutional structure. Using five years (2020–2024) of ERA5 reanalysis data at 6-hourly intervals as the training set and 2025 data as an independent test set, we conduct an ablation study comparing three architectures: CNN-only, LSTM-only, and the proposed ConvLSTM. Hyperparameters are optimised using two complementary methods - Genetic Algorithm (GA) and Grid Search - enabling both discrete and continuous exploration of the search space. Experimental results demonstrate that ConvLSTM consistently outperforms both unimodal baselines, with the Gridsearch-optimised variant achieving the lowest MAE and RMSE on the unseen 2025 test period. Comprehensive EDA reveals strong spatial correlation structures and persistent temporal autocorrelation, both of which are exploited by the ConvLSTM architecture. The framework produces full spatial WPD maps at each forecast step, making it directly applicable to offshore site assessment.

“A GIS-Based approach for prioritizing urban retention pond locations to mitigate flood risk: A case study of Hanoi, Vietnam”

Chu Manh Hung; GIS-IDEAS Oral Presentation

Urban flooding has emerged as one of the most critical natural hazards in rapidly urbanizing cities, particularly under the combined pressures of climate change, land-use transformation, and increasing population density. Hanoi, the capital city of Vietnam, has experienced frequent and severe urban flooding events in recent years, highlighting the urgent need for effective spatial planning solutions to enhance flood resilience. Urban retention ponds are increasingly recognized as a key component of green infrastructure systems, providing flood regulation, environmental improvement, and social co-benefits. However, identifying optimal locations for such infrastructure remains a complex spatial decision-making challenge involving multiple and often conflicting criteria.
This study proposes a GIS-based spatial multi-criteria approach to prioritize suitable locations for urban retention ponds aimed at mitigating flood risk in Hanoi, Vietnam. A comprehensive set of geospatial indicators was developed and quantified using Geographic Information Systems, including terrain depression derived from ALOS PALSAR DEM data, existing flood-prone areas extracted from Sentinel-1 imagery, vegetation conditions measured by NDVI from Sentinel-2 data, land-use and land-cover characteristics from ESA WorldCover datasets, and population distribution obtained from WorldPop. These criteria were standardized and integrated through a weighted overlay analysis to generate composite suitability maps.
To reflect diverse planning objectives and trade-offs, three alternative spatial scenarios were constructed: (i) a technical–cost-efficient scenario emphasizing natural and hydrological conditions while minimizing land-use conflicts; (ii) an ecological restoration scenario prioritizing environmental quality and green connectivity; and (iii) a community-oriented scenario focusing on accessibility and social benefits. The results reveal distinct spatial patterns of priority areas across the three scenarios, with high-suitability zones distributed in peri-urban low-lying areas as well as selected inner-city locations depending on the planning objective. By overlaying the highest-ranked zones from all scenarios, optimal locations for multifunctional retention pond development were identified.
The findings demonstrate the effectiveness of GIS-based spatial intelligence in supporting urban flood hazard management and green infrastructure planning. This study contributes a transferable geospatial decision-support framework that integrates multi-source Earth observation data with spatial analysis techniques, offering practical implications for urban planners and policymakers in flood-prone cities undergoing rapid urban transformation.

“A GIS-Based Framework for Evaluating Spatiotemporal Traffic Density in Hanoi”

Nguyen Thi Minh Hien; GIS-IDEAS Oral Presentation

Traffic congestion in Hanoi has become a significant obstacle to the city's strategic goals of sustainable urban development. As the urban area undergoes rapid expansion, pressure on existing infrastructure necessitates advanced analytical methods to monitor and mitigate mobility constraints. This study is designed to systematically analyze the current state of traffic, specifically to clarify the complex spatio-temporal distribution patterns of urban traffic density. Unlike traditional studies relying on local sensors and hardware dependence, this research leverages the power of Big Data analytics to provide a more comprehensive and scalable view of urban mobility. The methodological framework begins with the collection of high-precision empirical data. The input data included four levels of traffic density collected from Google Maps, classifying congestion into distinct intensity states, along with real-time movement metrics obtained via the Distance Matrix API. By utilizing these diverse data streams, the study overcame the limitations of traditional measurement devices, such as inductive loops or static cameras, which are often costly to maintain and geographically limited. These datasets were then subjected to spatial analysis techniques, specifically density estimation to visualize vehicle concentrations, hotspot identification to detect statistically significant congestion clusters, and surface interpolation to model continuous traffic conditions across the entire urban area. The output was density maps depicting the development of traffic bottlenecks with high accuracy. These visualizations allow for a better understanding of how congestion changes spatially and temporally, revealing the daily rhythm of the city's transport network. Beyond mere observation, this research establishes a solid scientific foundation for the next generation of Intelligent Transportation Systems (ITS). By providing a quantitative basis for assessing external environmental impacts—such as carbon emissions associated with vehicle stop time—this research serves as a crucial catalyst for infrastructure optimization. Furthermore, the integration of real-time API data lays the groundwork for AI-enhanced urban forecasting, enabling planners to predict congestion before it occurs. These advances are essential to fostering smart, sustainable mobility ecosystems, improving quality of life, and ensuring the long-term sustainability of the modern urban environment.

“A hybrid geostatistical and machine learning approach for lithology prediction of cement limestone deposits”

Vu Dinh Trong, Xuan-Nam Bui; GIS-IDEAS Oral Presentation

Constructing a reliable and accurate geological model has been one of the critical requirements in mine planning and mineral resource management. They help geologists and mining engineers to understand the geological complexity of subsurface deposits and assist them in evaluation and planning to effectively manage and use mineral resources. In most cases, exploration boreholes are considered primary data sources to construct these models. These data are commonly imbalanced and sparse, making the applications of modeling techniques challenging despite the significant advances of geostatistics or machine learning in recent decades. In this study, we proposed a novel hybrid modeling strategy to develop a lithological 3D model for limestone deposits. The strategy combines geostatistics and machine learning to leverage the powers and overcome the limitations of both geostatistics and machine learning approaches in the case of a sparse and imbalanced borehole dataset. A comparison was conducted between a traditional IK model, a standard RF model trained solely on spatial coordinates, and a proposed hybrid RF model trained on coordinates combined with IK probability. The results demonstrate the clear improvement of the hybrid approach in terms of learning geological context and producing practical geological structures. It achieved a higher overall accuracy (93.1%) than the Standard Random Forest (RF) machine learning (80.2%) and Indicator Kriging (IK) (61.0%) models in separation and dealt well with minority rock type units in the dataset, making predictions up to 145% better than the Standard RF model. The feature importance analysis also indicated that the proposed model learned to prioritize the geostatistical probability features and use them as more powerful predictors than raw spatial coordinates alone. Visual validation through 3D models, 2D cross-sections, and validation borehole comparisons further confirmed the Hybrid model 's ability to create better geological detail and predictive accuracy. This result offers significant potential for improving geological model accuracy for resource estimation and mine planning in similar geological settings.

“A uniform sampling and band combination method for land cover mapping”

Trinh Thanh; GIS-IDEAS Oral Presentation

Land cover classification is an important field of object recognition in remote sensing imagery. Machine learning models are widely used for high-resolution land-cover mapping, but their performance strongly depends on input data. In other words, the quantity and quality of samples determine the accuracy of the models and the reliability of land cover maps. In this paper, we first present the band combination method for identifying object types in imagery. Then, we conduct a study on the uniform distribution of object types in a study area. A Sentinel-2 imagery is employed to produce a land-cover map for Van Yen area, Lao Cai province, covering 1,400 km².

Different bands of the imagery were combined to identify five types of land cover: rivers and water surfaces; barren lands; populated areas and roads; sparse vegetation; and dense vegetation. To study the uniform distribution of sample collection within the study area, we divided the area into 3x3 km grids. For each grid, we collected samples of all types to ensure even distribution and coverage within the study area. We obtained 1726 samples for 5 types of objects (equivalent to 1.1 km²).

Two methods, Rule-based and Random Forest (RF), were applied to build land-cover classification models. 70% of the objects were used for training, and the results showed that RF outperformed Rule-based in terms of accuracy. To further investigate data distribution, we conducted further experiments. We first selected a single 3x3km grid with the highest number of objects of different types to build an RF model, and used the remaining objects for testing. RF yielded 92% accuracy. Next, we selected all objects from any three grids for training; the RF result was 96%. Finally, we used 10% of the objects in each grid to train the model, achieving 98% accuracy.

RF results show that the object types labeled using the band-combination method are highly accurate. Moreover, the results show that dividing the study area into grids for object collection is crucial for RF performance. We established a land cover map with only 10% of the objects for training. Furthermore, we compared the established land cover map with official statistical data and found that vegetation types differ by less than 1% from the actual statistical area. These findings confirm that the proposed uniform sampling strategy is highly effective for producing an accurate and reliable land cover map.

“An Automated Approach to Hydrologic Network Generalization: A Case Study in a Selected Area of Tay Ninh Province, Vietnam”

Hanh Nguyen-Thi-Hong; GIS-IDEAS Oral Presentation

Hydrologic network generalization is a fundamental task in cartographic production and spatial data management for multi-scale mapping. It plays a critical role in ensuring that river and stream networks remain legible, structurally coherent, and topologically correct when represented at different levels of detail. However, automated hydrologic network generalization remains a challenging problem due to the need to preserve network connectivity, drainage hierarchy, and essential spatial characteristics while reducing geometric complexity and feature density. In many practical GIS workflows, generalization operations are applied without an explicit prioritization mechanism, which can lead to inconsistent results, loss of important network components, or violations of topological relationships. This study proposes an automated approach to hydrologic network generalization that integrates feature classification and priority weighting prior to the application of built-in tools and algorithms in ArcGIS Pro 3.0.2. Hydrologic features are first classified based on a combination of geometric, topological, and hierarchical criteria, including stream order, length, connectivity, and positional importance within the drainage network. Priority weights are then assigned to individual network elements to explicitly control the sequence, selection, and intensity of generalization operations such as line simplification, feature elimination, and network pruning. This prioritization framework establishes a structured and rule-based workflow that enhances the transparency and reproducibility of the generalization process. The proposed approach is implemented and evaluated using multi-scale topographic datasets at scales of 1:5,000, 1:25,000, and 1:50,000 in a selected area of Tay Ninh Province, Vietnam. The study area represents a typical lowland hydrologic environment characterized by dense river and stream networks with varying hierarchical levels. Experimental results demonstrate that the proposed method effectively generalizes hydrologic networks across multiple scales while preserving essential topological relationships, drainage patterns, and the relative importance of network components. Compared with non-prioritized generalization workflows, the results show improved consistency, better hierarchical preservation, and enhanced cartographic readability. Overall, the proposed approach provides a practical, reproducible, and GIS-oriented solution for automated hydrologic network generalization using widely available software tools. The findings support efficient multi-scale base map production and hydrological thematic mapping and contribute to improved spatial data preparation for land-use planning, environmental management, and other geospatial applications that require consistent and reliable hydrologic representations across different map scales.

“An integrated FAHP, machine learning, and GIS approach for hospital site selection: A case study of five wards in Tuyen Quang province, Vietnam”

Bui Ngoc Tu; GIS-IDEAS Oral Presentation

Hospital site selection is a strategic planning task that directly affects the effectiveness, quality, and equity of health-service delivery, and it is therefore critical to hospital infrastructure development. This study presents an integrated approach combining fuzzy AHP (FAHP), machine learning, and GIS to identify suitable hospital sites in a five-ward study area in Tuyen Quang Province, Vietnam. The workflow captures expert judgment through FAHP and complements it with data-driven learning from machine-learning models to enhance the robustness of multi-criteria suitability assessment. Twelve criteria across economic, social, and environmental dimensions were compiled as thematic GIS layers, including distance to existing hospital, distance to industrial zone, distance to residential area, distance to main road, distance to local road, slope, distance to landfill, distance to cemeterie, distance to historical and cultural site, distance to surface water, distance to power substation, and land use type. The study area was discretized into a 300 × 300 m grid, yielding 2,101 centroid samples, from which predictor values were extracted to construct the training and validation dataset. Three algorithms were evaluated, including Random Forest (RF), Gradient Boosting (GB), and eXtreme Gradient Boosting (XGBoost). The predictive results (R² = 0.86 for RF, 0.95 for GB, and 0.96 for XGBoost) indicate that XGBoost achieved the best predictive performance. Feature importance analysis further suggests that distance to industrial zone has the strongest influence on hospital site suitability, followed by distance to main road, slope and distance to existing hospital. The resulting suitability map delineates five suitability classes and supports the identification and prioritization of candidate locations for hospital planning. Specifically, 33.44 km² (13.83%) of the study area was classified as very highly suitable, 39.77 km² (16.44%) as highly suitable, 73.78 km² (30.51%) as moderately suitable, and the remaining 94.88 km² (39.22%) was classified as low to very low suitability. Overall, the proposed approach provides a robust and practical decision-support tool for hospital siting. It offers a useful reference for local authorities in Tuyen Quang to guide sustainable hospital infrastructure planning and can be adapted to other areas with similar planning contexts.
Keywords: Hospital site selection, Machine learning, GIS, FAHP.

“An Integrated Geoinformatics and Machine Learning Framework for UAV-Based Forest Biomass and Carbon Assessment”

thinnakon angkahad; GIS-IDEAS Oral Presentation

This study highlights the value of integrating field-based forest inventory data with high-resolution UAV imagery and machine learning techniques to enhance the accuracy of forest biomass and carbon sequestration estimation. Conventional approaches for estimating aboveground biomass often depend solely on ground surveys, which are labor-intensive, time-consuming, and spatially limited. The integration of geoinformatics and predictive modeling in this research offers a more efficient and scalable alternative, allowing for comprehensive assessment across larger forest areas with reduced field effort. The application of multiple machine learning algorithms enabled a robust comparison of predictive performance for biomass and carbon estimation. Models such as Random Forest and Support Vector Machine are particularly effective in handling complex, non-linear relationships between vegetation indices and biophysical forest parameters. Meanwhile, Artificial Neural Networks and K-Nearest Neighbors provide strong capabilities for pattern recognition, especially in heterogeneous forest environments. By evaluating several algorithms simultaneously, the study identifies suitable modeling approaches for estimating carbon storage in mixed tropical forest conditions. Model accuracy was assessed using standard statistical indicators, including the coefficient of determination (R²), root mean square error (RMSE), and mean absolute error (MAE), ensuring the reliability and consistency of the predicted results. Vegetation indices derived from UAV imagery, especially NDVI-FC and MSAVI2-FC, were found to be important predictors of aboveground biomass. These indices effectively capture variations in canopy density, vegetation vigor, and photosynthetic activity, which are closely related to biomass accumulation. The fine spatial resolution of UAV data allowed for detailed mapping of biomass and carbon distribution, revealing spatial heterogeneity within the forest that is often overlooked by coarse-resolution satellite imagery. This spatially explicit information is crucial for understanding forest structure and identifying areas with high carbon sequestration potential. Furthermore, the results of this study provide valuable information for sustainable forest management and climate change mitigation. Accurate estimation of forest carbon stocks supports long-term monitoring, conservation planning, and evaluation of forest restoration efforts. The methodology presented can also contribute to national and regional carbon accounting frameworks, including greenhouse gas inventories and carbon offset programs. Overall, the integration of UAV technology, geoinformatics, and machine learning offers a practical, reliable, and replicable framework for forest carbon assessment, particularly in tropical forest ecosystems where efficient and accurate monitoring tools are critically needed.

“An Integration of Spatial-scale Decomposition Method and Gated Recurrent Unit approach for Three-Hour Mean Areal Precipitation Forecasting using Weather Radar Data”

PHUNG DAI KHANH; GIS-IDEAS Oral Presentation

Accurate estimation of mean areal precipitation (MAP) from radar-based quantitative precipitation forecasts (QPFs) is crucial for short-term hydrological forecasting and flood early warning. Although weather radar provides precipitation information with high spatiotemporal resolution, radar-based QPFs remain strongly affected by spatial-scale dependency and increasing uncertainty at longer forecast lead times, which substantially limits the reliability of MAP estimation. To overcome these limitations, this study proposes an integrated framework for generating three-hour MAP from radar-derived QPFs by coupling the Spatial-scale Decomposition Method (SCDM) with a gated recurrent unit (GRU) deep learning model. The proposed framework consists of two main components. First, QPFs are generated from radar precipitation data using SCDM to explicitly represent precipitation characteristics across multiple spatial scales. Second, a GRU model is trained using three-hour QPF sequences as input to capture the temporal evolution of precipitation, thereby enhancing temporal consistency and improving the accuracy of MAP estimates. The performance of the proposed SCDM–GRU framework is evaluated through comparison with a conventional SCDM-only approach for calibrating and generating three-hour MAP forecasts. Their performance are assessed using multiple evaluation metrics, including correlation coefficient (CC), root mean square error (RMSE), bias, critical success index (CSI), and probability of detection (POD), with rainfall thresholds of 3 mm/10 min and 5 mm/10 min. The results demonstrate that the integrated framework significantly enhances MAP estimation skill, achieving CC values greater than 0.77 within the first 60 minutes and maintaining CC values above 0.60 at a forecast lead time of 180 minutes, along with reduced RMSE and bias and increased CSI and POD. Overall, the proposed method consistently outperforms the SCDM-only approach throughout the entire three-hour forecast horizon. These results confirm the effectiveness of integrating radar-based spatial-scale decomposition with deep learning–based temporal modelling and highlight the strong potential of the proposed framework for improving the utilization of radar-derived QPFs in short-term hydrological forecasting and flood early warning applications.

“An Intelligent IoT-Based Poultry Housing System for Real-Time Environmental Control and Precision Management”

Sittichai Choosumrong, Rhutairat Hataitara, Kampanart Piyathamrongchai, Gitsada Panumonwatee, Natima Udon; GIS-IDEAS Oral Presentation

The poultry industry faces increasing challenges related to environmental control, disease outbreaks, rising production costs, and the need for sustainable farm management. Traditional poultry housing systems rely heavily on manual monitoring, which limits real-time responsiveness and reduces overall farm efficiency. This research presents the development of a Smart Poultry House system for real-time management of three-breed chickens, integrating Internet of Things (IoT) technology, artificial intelligence (AI), and a web-based decision support platform to enhance productivity and sustainability. The proposed system deploys multiple environmental sensors inside poultry houses to continuously monitor temperature, humidity, ammonia concentration, light intensity, and feed-water consumption. Sensor data are transmitted through a wireless communication network to a cloud-based server for storage and processing. AI-driven analytics are applied to evaluate environmental patterns, detect abnormal conditions, and generate automated control commands. The system enables intelligent control of ventilation fans, cooling pads, lighting systems, and water supply to maintain optimal microclimate conditions for poultry growth.A web dashboard and mobile interface were developed to provide real-time visualization, alert notifications, and remote management capabilities for farmers. The platform allows young farmers and farm operators to monitor flock health status, review historical data trends, and receive early warning alerts when environmental thresholds are exceeded. This reduces mortality rates, improves feed conversion efficiency, and supports precision livestock farming practices.Field implementation demonstrates that the Smart Poultry House system significantly improves environmental stability within the housing unit and enhances farm operational efficiency. The integration of IoT and AI technologies reduces labor dependency, minimizes energy consumption through optimized control, and supports data-driven decision-making. Furthermore, the system contributes to sustainable agricultural development by promoting resource efficiency and climate-resilient livestock management.This research provides a scalable and practical model for intelligent poultry farm management in Thailand and other developing agricultural regions, aligning with smart farming policies and sustainable development goals.

“An Open-Source Tool for Mapping Landslide Susceptibility: Case Study in Nam Pam, Vietnam”

Doan Huong-Giang, Trinh Thanh; GIS-IDEAS Oral Presentation

Nowadays, Landslide Susceptibility Mapping (LSM) has become a key component in disaster risk reduction strategies for mountainous regions in many countries worldwide. However, many existing LSM studies remain constrained by their dependence on commercial GIS software and by a lack of transparency in the selection and implementation of analytical methods. Vietnam is a country characterized by complex natural conditions, where approximately three-quarters of its territory consists of hilly and mountainous terrain, combined with a long and densely populated coastline. These regions are increasingly affected by climate change impacts, including more frequent extreme rainfall events, prolonged wet seasons, and intensified surface runoff, all of which significantly exacerbate landslide occurrence. In mountainous areas, steep slopes, deeply weathered geological formations, fragmented land use, and rapid infrastructure expansion further increase landslide susceptibility, while limitations in monitoring data and technical resources pose major challenges for effective hazard assessment and risk management. Under such conditions, developing a transparent, reproducible, and adaptable landslide susceptibility mapping approach is essential to support land-use planning, infrastructure development, and disaster risk reduction in Vietnam’s data-scarce mountainous regions.This study introduces OpenLSM - an open-source and programmable methodological framework designed to standardize and enhance the reproducibility of landslide susceptibility assessment workflows within a GIS environment. The integration of GIS and the Analytic Hierarchy Process (AHP) was applied to Nam Pam Commune (Northwestern of Vietnam) using 13 criteria factors related to topography, geology - weathering crust, hydrology, transportation networks, and land use etc. Field surveys and statistical data indicate that the total area affected by historical landslides in the study area is approximately 67 ha. Areas with moderate to high landslide susceptibility occupy a substantial proportion of the region, highlighting the significant influence of both natural and anthropogenic factors on landslide occurrence. Notably, very high susceptibility zones (93,201.04 ha) should be prioritized for land-use planning, early warning, and disaster risk reduction, while the concentration of historical landslides within high and very high susceptibility classes confirms the reliability of the proposed approach in data-scarce mountainous areas.
The proposed framework demonstrates strong potential for application in other mountainous regions of Vietnam and Southeast Asia, where landslide hazards are closely linked to climate change, complex terrain, and limited data availability. The results provide scientific support for sustainable land-use planning and disaster risk mitigation in vulnerable mountainous communities.

“Application of GeoInformatics for Monitoring Agricultural Drought in Gia Lai Province, Vietnam Using MODIS Data and Google Earth Engine”

Ngoc_Anh_Nguyen; GIS-IDEAS Poster Presentation

Agricultural drought is one of the most severe environmental challenges affecting agricultural productivity and livelihood sustainability in Vietnam’s Central Highlands, a region characterized by complex topography, extensive rainfed agriculture, and a pronounced monsoonal climate with a rainy season from May to October and a dry season from November to April. In recent decades, increasing climate variability and prolonged dry periods have intensified drought impacts, highlighting the need for effective spatio-temporal monitoring approaches in data-scarce regions. This study applies a GeoInformatics-based framework to analyze the dynamics of agricultural drought in Gia Lai Province (prior to administrative reorganization) during the period 2005–2024.
The analysis is based on MODIS MOD09A1 surface reflectance data, which provide 8-day composite observations at 500 m spatial resolution and ensure long-term consistency for environmental monitoring. The dataset was processed using the Google Earth Engine (GEE) cloud-computing platform, enabling efficient handling of large multi-temporal satellite archives. After quality screening and cloud masking, a suite of vegetation- and moisture-related indices was derived, including the Normalized Difference Vegetation Index (NDVI), Normalized Difference Water Index (NDWI), Moisture Stress Index (MSI), Normalized Difference Drought Index (NDDI), and Vegetation Condition Index (VCI). Monthly mean values were calculated to characterize seasonal drought variability, while long-term mean composites were generated to examine spatial patterns of drought conditions across the province.
The results reveal a strong and consistent seasonal drought signal. Vegetation vigor and surface moisture conditions decline during the dry season (November–April), with the most severe drought conditions occurring in the late dry months from February to April, as reflected by low NDVI and NDWI values and elevated MSI and NDDI values. In contrast, substantial recovery is observed during the rainy season from May to October, with improved vegetation conditions and moisture availability indicated by higher NDVI, NDWI, and VCI values. Interannual variability is evident throughout the study period, with several years exhibiting persistent drought conditions consistently captured across multiple indices. Spatial analysis highlights recurrent drought-prone areas, particularly in zones dominated by rainfed agriculture and limited irrigation infrastructure.
This study demonstrates that the integration of satellite remote sensing, GIS-based spatial analysis, and cloud-based computing constitutes an effective GeoInformatics approach for agricultural drought monitoring. The proposed framework provides valuable information to support drought assessment, agricultural planning, and climate change adaptation strategies in monsoon-influenced regions of Vietnam.

“Application of GIS and Multi-Criteria Analysis for Forest Fire Susceptibility Mapping under Climate Change Conditions: A Case Study of Muong Chanh Commune, Son La Province, Vietnam”

Cao Dinh Son, Dao Thi Tuyet Mai; GIS-IDEAS Oral Presentation

In mountainous regions of Vietnam, traditional shifting cultivation (slash-and-burn) by ethnic minority communities often leads to vegetation degradation, uncontrolled wildfire spread, and the loss of forest ecological functions (e.g., biodiversity and watershed services). These human agricultural practices-compounded by a warming climate and prolonged dry seasons-exacerbate forest loss and wildfire frequency and intensity, placing increasing pressure on remaining natural forest areas. Muong Chanh commune in Son La Province (northwest Vietnam) still retains substantial forest cover, but it faces rising forest fire threats under these conditions, posing serious challenges for local forest management. In this context, the present study integrates Geographic Information System (GIS) techniques with the Analytic Hierarchy Process (AHP) to map forest fire susceptibility in Muong Chanh under climate change conditions. Nine criteria representing key topographic, vegetative, and anthropogenic factors influencing fire ignition and spread (e.g., slope, elevation, land use/cover, normalized difference vegetation index (NDVI), and distances to roads and settlements) were derived from Sentinel-2 satellite imagery (10 m resolution) and other geospatial data. Each criterion layer was reclassified on a common risk scale and weighted through an expert-based AHP pairwise comparison. A weighted overlay analysis in GIS was then performed to generate a composite fire susceptibility index. The resulting susceptibility map delineates five risk levels across the commune: very low, low, moderate, high, and very high. High-risk and very high-risk zones are predominantly associated with steep terrain and areas adjacent to slash-and-burn agricultural fields, together accounting for a considerable portion of the commune. These findings provide a scientific basis for proactive fire prevention planning, strengthened forest protection measures, and sustainable land use management in climate-sensitive upland regions like Muong Chanh. This integrated approach demonstrates the utility of combining remote sensing data and multi-criteria decision analysis to anticipate and mitigate forest fire hazards amid changing climate conditions. It offers local authorities a valuable tool to prioritize interventions and protect vulnerable forest landscapes from future wildfires

“Application of GIS and Multi-Criteria Analysis in Land Suitability Assessment for Sustainable Agricultural Land Management in Thoi Binh District, Ca Mau Province”

Xuan Dinh, Vu; GIS-IDEAS Poster Presentation

Sustainable agricultural development is a global imperative to ensure food security, conserve natural resources, and mitigate climate change impacts. In Vietnam's Mekong River Delta (MRD), the Thoi Binh district of Ca Mau province exemplifies these challenges, contributing over 50% of national rice production while facing severe land degradation from saltwater intrusion, flooding, subsidence, and urbanization. These factors have reduced crop yields by up to 30%, exacerbated soil salinization, and diminished arable land, underscoring the need for advanced land suitability assessment tools to inform resilient land-use planning. This study integrates Geographic Information Systems (GIS) with Multi-Criteria Analysis (MCA) and the Analytic Hierarchy Process (AHP) to evaluate land suitability for sustainable agriculture in Thoi Binh, focusing on key crops: rice, banana, sugarcane, and coconut. The methodology comprises two phases: (1) natural land suitability assessment using GIS and the Automated Land Evaluation System (ALES), involving overlay analysis of thematic layers (slope, soil depth, soil type, texture, and irrigation potential) derived from field surveys, remote sensing, and secondary data; and (2) sustainable suitability evaluation via group Analytic Hierarchy Process (AHP-GDM), incorporating economic, social, and environmental criteria weighted through expert pairwise comparisons (consistency ratio, CR < 0.1). Data collection combined primary methods (expert interviews, n=27 via Google Forms and in-person) with secondary sources (national soil maps, 2023 land statistics). Suitability indices (S) were computed using the formula S=∑_(i=1)^n▒〖(W_i×X_i)×C〗, classifying areas into high (S1), moderate (S2), and low (S3) suitability classes per FAO guidelines. Results reveal Thoi Binh's 63,630 ha landscape dominated by agricultural land (93%), with acid sulfate soils (82.93%) and saline soils (15.04%). For rice, 15.4% of land is highly suitable (S1), prioritizing flat terrains (<1° slope) and irrigation access; banana and coconut favor deep soils (>100 cm) in 24.1% and 20.5% S1 areas, respectively, in Tan Phu and Tan Loc communes. Sugarcane shows 18.7% S1 suitability, emphasizing texture and low salinity. Weighted criteria highlight irrigation (28-39%) and soil type (32-41%) as dominant factors across crops. Integrated GIS-MCA maps delineate optimal zones, identifying 56% of the district as sustainably viable for diversified cropping, reducing monoculture risks. This framework advances land management in saline-prone MRD regions, offering a replicable model for Vietnam and analogous global deltas. By balancing biophysical and socio-economic factors, it supports policy for adaptive agriculture, enhancing resilience against climate variability and promoting biodiversity conservation.

“Application of GIS and the Gini Coefficient in Assessing Sustainable Green Urban Development: A Case Study of a Commune in Tuyen Quang Province, Vietnam”

Nguyen Thi Tuyet; GIS-IDEAS Oral Presentation

In the context of rapid urbanization and the strong rural-urban transition in Vietnam, especially in midland and mountainous provinces such as Tuyen Quang, urban green spaces are undergoing significant changes in both quantity and ecological quality. Previous studies have mainly focused on measuring green coverage area, while assessments of vegetation quality and the spatial distribution of green spaces for urban planning purposes remain limited. This study proposes an integrated GIS-NDVI framework to assess the current status, quality, and planning orientation of green spaces at the commune level. The research uses multispectral Sentinel-2 satellite imagery with 10 m resolution combined with GIS spatial analysis tools to map green space distribution, calculate the NDVI index to evaluate vegetation health, analyze accessibility to green spaces within a distance threshold of 300-500 m, and identify priority areas for green development. The results show that although Yen Son Commune has a relatively high overall green coverage, the spatial distribution of green spaces is uneven. Residential areas and development zones along transportation corridors exhibit significantly lower NDVI values compared to the hilly areas in the southern and southeastern parts. Based on these findings, the study proposes three main planning orientations: ecological conservation zones, green enhancement and restoration zones, and priority areas for new green space development. The results provide a scientific basis for urban green planning, efficient land use management, and sustainable development in rural-urban transition areas in Vietnam.

“Application of Machine Learning and Geospatial Simulation Models in Predicting Land Use Dynamics: A Case Study of Eastern Ho Chi Minh City”

Nguyen Thi Thuan; GIS-IDEAS Oral Presentation

“Applying a multi-criteria analysis method combined with a Geographic Information System (GIS) to assess land suitability for coffee cultivation in Chieng Ve, Son La province, Vietnam”

Pham Thi Nguyet; GIS-IDEAS Oral Presentation

Son La is the province with the largest coffee cultivation area in the Northwestern region of Vietnam, where coffee plays a crucial role as a perennial cash crop contributing to regional economic development. However, unplanned expansion of coffee cultivation in mountainous areas may lead to land degradation, reduced productivity, and unsustainable livelihood outcomes. Therefore, identifying and spatially delineating suitable areas for sustainable coffee cultivation is essential for effective agricultural land-use planning. This study applies a multi-criteria analysis approach integrated with a Geographic Information System (GIS) to evaluate land suitability for coffee cultivation in Chieng Ve Commune, Mai Son District. The assessment is based on a comprehensive set of natural factors, including soil type, slope, elevation, and climatic conditions, together with socio-economic factors such as current land use, infrastructure accessibility, and population distribution. The relative importance of these criteria was determined using the Analytic Hierarchy Process (AHP), incorporating expert judgment and consistency testing to ensure the reliability of the weighting scheme. All thematic layers were standardized, weighted, and integrated within the GIS environment to generate a spatially explicit land suitability map for coffee cultivation. The results indicate that 23.36% of the study area is classified as highly suitable, 27.62% as suitable, 24.35% as moderately suitable, 13.42% as marginally suitable, and 11.23% as unsuitable for coffee cultivation. Areas with high and very high suitability are widely distributed, indicating favorable conditions for expanding coffee cultivation in a planned and sustainable manner. Importantly, the study emphasizes the role of land suitability assessment in supporting sustainable livelihoods for ethnic minority communities, who constitute a large proportion of the local population and depend heavily on coffee farming as a primary income source. By directing coffee cultivation toward suitable areas and discouraging production on marginal lands, the proposed GIS-AHP framework contributes to improving livelihood stability, reducing environmental risks, and supporting poverty reduction. Overall, the study provides a practical and transferable decision-support tool for sustainable agricultural land-use planning in mountainous regions of Northern Vietnam.

“Assessing changes of mangroves and tidal channel systems from multi-time satellite imagery in the coastal area of Kim Son, Ninh Binh”

Dao Hoang Tung; GIS-IDEAS Oral Presentation

Over the past decades, a massive reduction in mangroves has been recorded worldwide, particularly in Vietnam, leading to significant loss of coastal defence, economic benefits, and social recognition, underscoring the need for mangrove observation and restoration. This study area is the Kim Son mangrove coast, Ninh Binh province, is a speacial ecosystem with high bio-diversity and under a threat of shrinking and degrading due to the impacts of climate change, sea level rise, coastal erosion, expansion of aquaculture, and coastal infrastructure. Therefore, the importance for assessing of changes in mangroves and tidal channel systems along the Kim Son coasts is needed to provide a scientific foundation for integrated coastal solutions for irrigation, forestry, and infrastructure to protect and sustainably develop the Kim Son mangrove ecosystem. The methodology in this study focus on collect and exploit the Landsat 4/5/8 and Sentinel-2 optical satellite images accquired from 1987 to 2025 to distingust the vegetation index (NDVI – Nomarlized Difference Vegetation Index) and water index (MNDWI – Modified Normalised Difference Water Index), repsenting for mangroves and tidal channels. The analytical trending of mangrove and tidal channel changes over the years will be finally presented as the spatial distribution of mangrove forests and tidal channels by combining normalized vegetation index (NDVI ≥ 0.3) and corrected normalized water index (MNDWI < 0). The results show historical trending in mangrove forest area changes from 1987 to 2025, as well as the morphological changes and distribution of tidal channels from 2005 to 2025. Firstly, the coastal area of Kim Son has undergone numerous dike constructions, land reclamation, mudflat conquering, mangrove reforestation, and the development of transportation and irrigation systems, such as: the construction of the Binh Minh I dike from 1959-1960; the construction of the Binh Minh II dike in 1981; the construction of the Binh Minh III dike in 2008; and the commencement of construction of the Binh Minh IV dike in 2020. During the period 1987–1998, the mangrove forest area in Kim Son gradually increased from 579 ha (1987) to 599 ha (1995). It reached its highest level of 909 ha (1996) to 912 ha (1998) during the entire monitoring period, reflecting strong development in both area and ecological structure. However, during 1999–2005, the mangrove forest area decreased significantly from 734 ha (1999) to 435 ha (2005), mainly due to the conversion of forest land to aquaculture (shrimp, clams), the exploitation of mangrove wood and firewood, and the construction of seawalls. From 2005 to 2015, the mangrove forest showed strong recovery, increasing from 257 ha (2005) to 704 ha (2010). As of 2025, the forest area remains at 785 hectares and shows a trend towards more sustainable recovery. Finally, the results clearly reflect the combined impact of natural factors and human activities in the Kim Son coastal area.

“Assessing Environmental Vulnerability to Landslides Using GIS and Unsupervised Clustering: A Case Study of the Non Nuoc Cao Bang Global Geopark, Vietnam”

Phan Hoa; GIS-IDEAS Oral Presentation

Landslides are a major environmental hazard in mountainous regions, frequently causing extensive degradation of land cover, soil stability, hydrological systems, and ecological integrity, thereby posing significant challenges to environmental conservation and sustainable land management. These challenges are particularly pronounced in geopark areas, where geological heritage, biodiversity, and landscape values are closely interconnected and highly sensitive to disturbance. This study develops a GIS-based framework for assessing landslide-induced environmental damage using unsupervised clustering algorithms, with the Non Nuoc Cao Bang Global Geopark in northern Vietnam selected as a representative case study. A comprehensive indicator system was constructed to characterize environmental damage from multiple dimensions, including anthropogenic disturbance, vegetation and land cover degradation, soil surface conditions, hydrological impacts, and ecological sensitivity. The indicators were derived from multispectral remote sensing imagery, digital elevation model–based terrain analysis, and GIS-based spatial proximity assessments, allowing both natural and human-related factors influencing environmental vulnerability to be integrated within a unified analytical framework. Prior to analysis, all indicators were normalized to ensure comparability across different units and value ranges. Unsupervised clustering techniques, specifically K-means and hierarchical clustering, were then applied to classify the study area into zones exhibiting similar levels of environmental damage, enabling spatial patterns to emerge directly from the data without subjective weighting. The clustering results reveal distinct spatial patterns of environmental degradation across the geopark, clearly differentiating areas experiencing severe landslide-induced damage from zones characterized by relatively lower levels of environmental disturbance. These spatial patterns show strong correspondence with documented landslide occurrences and the underlying geomorphological characteristics of the region, indicating that the proposed approach effectively captures the spatial heterogeneity of landslide impacts on environmental systems. In addition, the results provide insights into the interactions between terrain conditions, land cover changes, hydrological responses, and human activities in shaping environmental damage patterns within geopark environments. Overall, the proposed framework offers an objective, reproducible, and scalable method for assessing landslide-induced environmental damage using widely available geospatial datasets, and the findings provide valuable scientific support for landslide risk zoning, environmental monitoring, and sustainable management in geopark and mountainous regions.

“Assessing Factor Effects on Land-Use Plan Implementation Using Explainable AI (XAI) in Son La Province, Vietnam”

Bui Ngoc Tu; GIS-IDEAS Oral Presentation

Bui Ngoc Tu1, Tran Quoc Binh1
1University of Science, Vietnam National University, Thanh Xuan, Hanoi, Vietnam

Email: buingoctu@hus.edu.vn

ABSTRACT
Assessing the factors that influence land-use plan implementation outcomes is essential for improving the effectiveness of land administration and implementation management. However, it remains challenging when the impacts of multiple factor groups must be quantified for individual planned parcels on maps. With the growing availability of multisource datasets and increasing demand for spatially detailed assessments, machine-learning models combined with explainable artificial intelligence (XAI) provide a suitable approach to support objective, transparent, and verifiable evaluation.
This study proposes and tests a machine-learning workflow to assess both the relative importance and the directional influence of factors on local land-use plan implementation outcomes. The dataset was compiled from multiple sources, including thematic maps, technical reports/documents, and expert knowledge, yielding 1,406 parcel-level samples and 28 explanatory features. The features were grouped into: (i) land-category and planned-parcel characteristics, (ii) spatial proximity variables, (iii) topographic conditions, (iv) socio-economic factors, and (v) legal factors. This multisource integration strengthens quantitative representation and enhances objectivity, while still incorporating domain expertise for context-specific factors.
Five models were evaluated: Random Forest (RF), Extremely Randomized Trees (ET), Gradient Boosting (GB), XGBoost, and LightGBM. The results indicate that RF achieved superior overall performance and produced a more plausible distribution of feature importance, better reflecting practical land-use plan implementation conditions. To interpret the RF model, SHAP was employed to provide explainable insights that go beyond ranking feature importance by also characterizing the direction of influence of each factor on implementation outcomes. SHAP results identify the most influential determinants as land category, distance to residential areas, distance to schools, planned-parcel area, the Y-coordinate (capturing a north–south spatial gradient), and the intensity of inspection and compliance monitoring. The SHAP-based interpretation further reveals whether these factors tend to increase or decrease the likelihood of successful plan implementation, thereby supporting spatially explicit explanations of underlying mechanisms rather than relying solely on importance rankings.
Overall, this study advances a novel parcel-level approach for evaluating determinants of land-use plan implementation outcomes, shifting from qualitative judgment to data-driven, transparent, and verifiable assessment. The findings can support local governments in identifying key drivers, prioritizing resources, refining implementation measures, and strengthening management instruments—particularly in spatially sensitive areas—thereby improving objectivity, reliability, and the overall effectiveness of land-use plan implementation in practice.
Keywords: Land-use plan implementation, Explainable artificial intelligence , Son La Province

“Assessing Heat Extremes in Central Vietnam in response to Solar Radiation Management”

Nguyen Dang Xuan Hien; GIS-IDEAS Oral Presentation

Heat extremes are projected to intensify across Central Vietnam under future climate change, posing increasing risks to water resources, agriculture, ecosystems, and human health in this tropical monsoon region characterized by complex topography and strong climate variability. This study assesses the impact of Solar Radiation Modification (SRM) on future temperature extremes in Central Vietnam, using stratospheric aerosol injection (G6Sulfur) simulations available from the Geoengineering Model Intercomparison Project (GeoMIP) under the Coupled Model Intercomparison Project Phase 6 (CMIP6). To enable a detailed regional assessment, CMIP6 climate model outputs are first statistically downscaled to the observational grid through spatial regridding and subsequently bias-corrected using the Quantile Delta Mapping (QDM) technique, which adjusts systematic model biases while preserving the climate change signal in future projections. Five ETCCDI temperature extreme indices—TXx, TNx, TX90p, TN90p, and the Warm Spell Duration Index (WSDI)—are employed to characterize the intensity, frequency, and duration of heat extremes across the region. Model performance after bias correction and downscaling is evaluated against VnGC, a gauge-based gridded climate data product to ensure robustness and consistency across models.
The projected outcomes suggest that under high-emission scenarios, extreme daytime temperatures, warm nights, and heatwave duration are expected to increase substantially across Central Vietnam, indicating heightened risks to both natural and human systems. SRM scenarios are projected to partially mitigate these increases by reducing the intensity of extreme heat and the frequency of hot days and prolonged warm spells relative to high-emission scenarios. Spatial variability in SRM-induced cooling is evident, reflecting the influence of local climatic conditions, elevation gradients, and land–atmosphere interactions. A basin-scale analysis of the Ba River Basin is used as a representative case study to illustrate localized responses within the broader regional context. Overall, the findings indicate that while SRM may reduce future heat extremes in Central Vietnam, it is unlikely to fully offset greenhouse gas–driven warming. The study underscores the critical role of bias correction and statistical downscaling in regional climate impact assessments and highlights the importance of multi-scale evaluations when assessing the potential benefits and limitations of climate intervention strategies in tropical monsoon regions.

“Assessing the robustness of AlphaEarth Foundation models for forest canopy height estimation in Karst and Earth-Mountain Ecosystems: A Google Earth Engine-based approach”

DUC ANH NGO; GIS-IDEAS Oral Presentation

Accurate estimation of forest canopy height (FCH) is crucial for monitoring aboveground biomass and assessing carbon stocks. However, in tropical regions, terrain heterogeneity - particularly in karst ecosystems - poses significant challenges to the accuracy of remote sensing models. This study evaluates the robustness and adaptive mechanisms of embeddings from the AlphaEarth Foundations (AEF) model when fused with Synthetic Aperture Radar (SAR) data (Sentinel-1 VV/VH) and Digital Elevation Models (ALOS DEM). Based on a synchronized field dataset (N=84 for Cuc Phuong and N=66 for Ben En), a modeling pipeline integrating Principal Component Analysis (PCA) and Random Forest (RF) was implemented on Google Earth Engine (GEE) and validated using Leave One Out Cross Validation (LOOCV). The results reveal a distinct divergence in machine learning mechanisms: in the earth-mountain area of Ben En, the model optimized at 6 PCs, prioritizing penetrative radar features to achieve high accuracy (RMSE=1.78 m,R^2=0.33). Conversely, in the complex karst terrain of Cuc Phuong, the model automatically implemented a "deep compression" mechanism at 3 PCs, utilizing elevation data as a primary spatial constraint to mitigate topographic noise and maintaining a reliable error margin (RMSE=3.08 m,R^2=0.25). This research demonstrates that the fusion of foundation models with multi-source geospatial data provides a scalable and resilient solution for forest structure monitoring across diverse tropical landscapes.

“ASSESSMENT OF BIAS CORRECTION BASE ON QUANTILE METHOD TO EXTREME RAINFALL AND TEMPERATURE INDICES BASED ON CMIP6 PROJECTIONS”

Vo Thanh Nhan; GIS-IDEAS Oral Presentation

Global climate models (GCMs) are widely used to simulate long-term climate trends in support of preparedness and mitigation strategies addressing the severe impacts of climate change on socio-economic systems and human life, particularly in vulnerable countries such as Vietnam. However, GCMs suffer from several limitations, including systematic biases and coarse spatial resolution. In this context, statistical downscaling plays an important role in improving climate information at the local scale, especially for the assessment of extreme precipitation and temperature indices. This study employs simulated data from ten global climate models participating in the Coupled Model Intercomparison Project Phase 6 (CMIP6) under the high-emission SSP5–8.5 scenario to evaluate the effects of different quantile-based bias correction methods on extreme precipitation and temperature indices.
Four quantile-based bias correction methods are applied and compared, including Quantile Mapping (QM), Empirical Quantile Mapping (EQM), Quantile Delta Mapping (QDM), and Detrended Quantile Mapping (DQM). Extreme frequency and intensity indices are analyzed to examine the spatial distribution and value ranges of precipitation (Rx1day, Rx5day, R95pTOT, R99pTOT, R10mm, R20mm, CDD, and CWD) and temperature (TXx, TNn, TX90p, TN10p, CSDI, and WSDI). The performance of these methods is evaluated using Percent Bias (PBIAS), Root Mean Square Error (RMSE), the Kolmogorov–Smirnov (KS) test, and the Kling–Gupta Efficiency (KGE) coefficient.
The results indicate that bias correction methods substantially reduce systematic biases in global climate models, with QDM and DQM exhibiting more stable performance in preserving trends and reproducing the distributions of extreme indices. However, all methods introduce distortions in the distribution tails, particularly for precipitation-related variables, and their performance varies across different types of extreme indices.
While the applied bias-correction methods effectively reconstruct the spatial distribution of the CDD index, the CWD index exhibits contrasting behavior. The ranges of the Rx1day and Rx5day indices derived from the DQM, EQM, and QM methods are identical, spanning 122–878 mm and 122–1128 mm, respectively, and remain lower than the observed ranges. In addition, the median values of R99pTOT and R95pTOT are substantially underestimated compared to observations, at approximately 500 mm and 2000 mm, respectively. In contrast, temperature-related indices demonstrate notably better performance for both mean and extreme conditions. The KGE values for maximum temperature (tasmax) reach the acceptable threshold, while those for minimum temperature (tasmin) exceed the threshold at the median level. However, bias-correction methods are unable to compensate for inherent deficiencies in global climate model outputs, as evidenced by the spatial distribution of the TNn index.
This study provides important scientific evidence to support the selection of appropriate statistical downscaling methods for regional-scale studies using CMIP6 climate model data.

“Assessment of Forest Cover Dynamics and Fragmentation in the O Lau Basin Using Remote Sensing and Machine Learning over Two Decades”

Hieu Huu Viet Nguyen; GIS-IDEAS Poster Presentation

The impact of land use activities in areas adjacent to natural forests has been threatening the habitat of animals and plants. Forest fragmentation is one of the main causes of biodiversity loss, degradation of ecosystem services, and loss of natural forests. The study applied Landsat remote sensing images and the Random Forest algorithm to assess forest fragmentation due to land cover changes in the O Lau river basin, Thua Thien Hue province, in the period 2010-2020. The data used included Landsat-7 images in 2010 and Landsat-8 images in 2015 and 2020. The Random Forest classification method was applied to classify 7 main land cover types, including: natural forests, planted forests, shrublands, grasslands and bare land, agricultural land, rural residential land, and water surfaces. The classification results achieved an overall accuracy of 87,53% (κ = 0,856). The morpho-spatial analysis model was also used to analyze the status of forest fragmentation. The results showed that forest cover increased, but the natural forest area decreased from January 2010 to December 2020. The spatial data outputs are very useful for controlling habitat fragmentation, developing sustainable natural forest management plans, and reducing pressure caused by local people in the study area.

“Assessment of Rice Growth at Different Phenological Stages Based on UAV Multispectral and High-Resolution UAV RGB Imagery”

Le Van Canh; GIS-IDEAS Oral Presentation

The Rice growth at different stages play a critical role in determining yield production. Timely and spatially detailed monitoring of rice growth conditions is therefore essential for supporting precision agricultural management and site-specific decision-making. This study investigates the effectiveness of integrating unmanned aerial vehicle (UAV) multispectral imagery and high-resolution UAV RGB imagery to assess rice growth conditions across key developmental stages, including tillering, heading, and milk ripening, within an experimental rice cultivation area. UAV multispectral and RGB data were acquired at the three growth stages. Multispectral imagery was used to derive the Normalized Difference Vegetation Index (NDVI), which serves as an indicator of vegetation vigor, biomass accumulation, and photosynthetic activity. The spatial distribution of NDVI was analyzed at both field and each plot levels to evaluate variations in rice growth conditions across different plots. Beside that, high-resolution UAV RGB orthomosaic images were employed to visually interpret and identify the underlying causes of spatial heterogeneity observed in NDVI values. The results indicate clear variations in NDVI among the examined growth stages. The average NDVI reached 0.84 during the tillering stage, decreased slightly to 0.81 at the heading stage, and further declined to 0.76 during the milk ripening stage, reflecting a gradual reduction in chlorophyll content and photosynthetic activity as rice plants progressed toward maturity. Significant intra-plot variability in NDVI was also observed, indicating non uniform growth conditions within individual fields. The integration of high-resolution UAV RGB imagery proved effective in explaining these spatial NDVI variations by providing detailed visual information on crop structure and field conditions, such as uneven plant density, localized stress, or areas affected by pests or diseases. While multispectral NDVI analysis quantitatively captured overall growth status, RGB imagery complemented this information by enabling qualitative interpretation of specific causes of variability. Compared with the use of multispectral data alone, the integrated approach improves the reliability and interpretability of rice growth assessment. Overall, the combined use of UAV multispectral and RGB imagery provides a detailed, intuitive, and effective method for monitoring rice development at key growth stages, supporting precision crop management and contributing to improved rice yield and production sustainability.

“Assessment of Seasonal Land Surface Temperature Variations in Hanoi Using Remote Sensing Imagery”

Nguyen Trung Nghia; GIS-IDEAS Oral Presentation

Parallel to its rapid urbanization, Hanoi is experiencing an intensified Urban Heat Island (UHI) effect, driven by anthropogenic heat from transportation and industrial activities that elevate Land Surface Temperature (LST). This research aims to identify regions adversely affected by thermal radiation and to provide a quantitative analysis of temperature differentials between areas with high versus low vegetation cover. This research leverages high-resolution (10m) Sentinel-2 satellite imagery to extract the Normalized Difference Vegetation Index (NDVI), integrated with MODIS thermal data selected for its superior temporal resolution on the Google Earth Engine cloud computing platform. Utilizing descriptive statistics and Hotspot Analysis, the study identified high-temperature nuclei (hotspots) with a confidence level exceeding 95%. These areas predominantly coincide with regions of high building density and extensive impervious surface cover. Conversely, coldspots were recorded in suburban districts characterized by significant water bodies, such as the Red River and West Lake, and abundant vegetation. Notably, the statistical results demonstrate a significant inverse correlation between the Normalized Difference Vegetation Index (NDVI) and Land Surface Temperature (LST). In areas with high NDVI values, average surface temperatures were observed to be 2°C to 3°C lower than in sparsely vegetated zones. Seasonal temperature variations further reveal that the Urban Heat Island (UHI) effect peaks during the summer, with the thermal gradient between urban and suburban areas reaching up to 5°C. In contrast, this differential diminishes during winter but remains stable within industrial manufacturing zones. These findings not only corroborate the pivotal role of vegetation cover in thermal regulation but also provide empirical scientific evidence regarding the impact of urbanization on the climate. This study serves as a critical baseline dataset, empowering policymakers to make informed decisions within the framework of sustainable planning. Optimizing green space density based on these data represents a strategic intervention to regulate urban temperatures, thereby bolstering the city’s adaptive capacity and enhancing its resilience against the increasingly complex and extreme manifestations of future climate change.

“Automated Evacuation Route Generation for Citywide Tephra-Fall Evacuation: An Imminent Eruption Scenario in Kagoshima City, Japan”

Andreas Keler; GIS-IDEAS Oral Presentation

Large-scale volcanic eruptions can produce widespread tephra (volcanic ash and pumice) fall that rapidly degrades urban lifelines, including transport, water, and telecommunications. In such conditions, Kagoshima City, with the active Sakurajima volcano, may require a coordinated, citywide evacuation to minimize casualties, particularly under east-wind conditions that threaten densely populated western districts. This study proposes a scalable workflow to operationalize Kagoshima City’s Massive Tephra Fall Response Plan by automatically generating evacuation movements for the full household population (>300,000 households), while integrating background traffic and public transport operations.
As a preprocessing step, we translate the guideline information into explicit spatial scenario inputs by defining evacuation zones and destinations and constructing wind-direction-dependent tephra hazard wedge polygons. These polygons are used to select applicable destinations and corridor constraints, and to generate alternative routes where guideline road-name sequences are incomplete or where disruptions require detours.
Using recent census-based household counts, we synthesize household-level origins and assign each household a passenger-car trip, consistent with local guidance. Evacuation routes and destinations outside city boundaries are generated on a routable OpenStreetMap-derived network, and public transport services are represented using GTFS schedules and stop locations. To translate aggregated census data into network demand, we compare two interpolation strategies for distributing households to road edges: (a) proportional allocation by edge length, and (b) distance-weighted allocation from district centroids. We further incorporate wind-direction-dependent ash-fall considerations to contrast decision-making under an imminent-eruption (pre-emptive) evacuation versus an already ongoing eruption, where disruption and compliance constraints may differ.
Finally, we conduct microscopic traffic-flow simulation with more than 400,000 simulated road users to quantify evacuation performance and identify congestion and conflict-prone locations, enabling a joint discussion of evacuation efficiency and road-traffic safety. The results provide an evidence-based evaluation of the current municipal plan and highlight the traffic control measures required for an organized citywide evacuation.

“Climate Change Impacts On Robusta Coffee Suitability In Central Highland, Vietnam: Insights From Cmip6 Climate Projections”

Nguyễn Nữ Kim Linh; GIS-IDEAS Oral Presentation

Globally, coffee plays an important role in economic development and provides significant public health benefits for society. Viet Nam is one of the world’s major coffee-exporting countries ranking second globally and is the largest exporter Robusta coffee. The Central Highlands is the main growing region, accounting for approximately 95% of Vietnam’s Robusta coffee production. However, coffee is one of the crops most vulnerable to climate change, since even small variations in precipitation and temperature, both in magnitude and spatial distribution, play an important role in determining suitable conditions for coffee growth. Under the impacts of climate change, the natural suitability area of Robusta Coffee is expected to change significantly. This research evaluates the impacts of climate change on the spatial distribution of Robusta Coffee under different future climate scenarios from climate projections based on Shared Socioeconomic Pathways (SSPs). Specifically, we used 4 variables with one of precipition and three of temperature indices obtained from 2 climate models outputs under SSP5-8.5 emission scenarios. These scenarios were simulated using General Circulation Models (GCMs) from the Coupled Model Intercomparison Project Phase 6 (CMIP6). Bias correction and spatial downscaling (BCSD) methods were applied to precipitation and temperature data to generate high-resolution climate data for the Central Highlands. Observed datasets of key environmental factors affecting coffee suitability, including precipitation, temperature, slope, elevation, and soil type, were resampled to a spatial resolution of 1 × 1 km and integrated using an overlay approach. Suitability was then evaluated based on FAO guidelines and the classification framework established by the Ministry of Agriculture and Rural Development of Vietnam. Random Forest classification model was subsequently employed to project coffee suitability under the SSP5-8.5 scenario using outputs from three CMIP6 climate models. The results reveal significant shifts in the spatial distribution of coffee suitability under future climate conditions compared to the historical baseline. Under the SSP5-8.5 scenario, highly suitable areas (S1 and S2) are projected to decline, while marginally suitable (S3) and unsuitable (N) areas expand, particularly in low-elevation regions. Conversely, certain high-altitude areas may become more suitable due to increasing temperatures, indicating a potential upward shift in optimal coffee-growing zones. Although the model demonstrates high overall accuracy, some uncertainties remain due to class imbalance, particularly in representing highly suitable areas (S1). Overall, these findings highlight the vulnerability of coffee production to climate change and provide valuable scientific evidence to support adaptation strategies and sustainable coffee development in the Central Highlands of Vietnam.
Keywords: AHP, Climate change, CMIP6, GIS, Robusta Coffee, SSPs.

“Comparative Assessment of Landsat 8 Level 1 and Level 2 Data for Land Surface Temperature and Urban Heat Island Analysis in Ho Chi Minh City, Vietnam”

Doan Thi Bich Ngoc; GIS-IDEAS Oral Presentation

Rapid urbanization in tropical megacities such as Ho Chi Minh City has increased the surface urban heat island (SUHI) effect, posing challenges to environmental sustainability and urban livability. Remote sensing-based land surface temperature (LST) assessments are widely used to analyze these thermal patterns; however, differences in data processing levels, particularly between Landsat Level 1 (L1) and Level 2 (L2), may cause uncertainties in SUHI estimates, while direct comparative studies between these two data levels in LST, SUHI, land use and land cover (LULC) dynamics remain limited. To clarify this difference, the LULC map was constructed from Sentinel-2 images using the Random Forest algorithm (OA > 93%), showing a marked increase in impermeable surfaces between 2016 and 2026. The findings indicated that both L1 and L2 consistently mirrored the spatial distribution of land surface temperature (LST), exhibiting elevated values in urbanized zones and diminished values in areas characterized by vegetation and water bodies. Nevertheless, L2 produced higher LST values and a broader spectrum of variability. The mean SUHI intensity from L2 is about 1.3-1.60C higher than that of L1, while the maximum values exceed those of L1 by more than 250C, indicating greater sensitivity to thermal extremes. The SUHI variation is closely related to the LULC conversion, in which increasing the impermeable surface increases heat, while vegetation contributes to cooling. The results show that L2 outperforms L1 in SUHI analysis and provides more effective support for environmental monitoring and urban planning.

“Comparative Evaluation of Urban Expansion Mapping Methods in Diriyah Using GHSL, NDBI, and Unsupervised Classification”

Muhannad Mohammed Alfehaid; GIS-IDEAS Oral Presentation

Accurate monitoring of urban expansion in dryland environments remains a persistent challenge in geospatial science due to the strong spectral similarity between built-up surfaces and highly reflective bare soils. This challenge is particularly pronounced in rapidly developing heritage-sensitive regions, where reliable urban mapping is essential for sustainable planning and decision support. This study contributes to the transition from traditional geoinformatics toward predictive geodata science by empirically evaluating and integrating multiple geospatial approaches for urban expansion mapping in Diriyah, Saudi Arabia—an area experiencing accelerated transformation under the national Vision 2030 development agenda.

The research assesses three widely applied built-up mapping methods across three temporal snapshots (2015, 2020, and 2025): the Global Human Settlement Layer (GHSL), the Normalized Difference Built-up Index (NDBI), and unsupervised k-means classification derived from multispectral satellite imagery. These methods represent different paradigms of geospatial analysis, ranging from globally calibrated products to local spectral indices and data-driven clustering. To enhance interpretability and methodological insight, the study introduces a Hybrid Built-up Detection Model (HBDM), which synthesizes the outputs of GHSL, NDBI, and unsupervised classification into a continuous urban intensity surface. Rather than acting as an independent classifier, the hybrid model functions as a diagnostic and integrative layer that highlights spatial convergence and divergence among methods.

Results reveal substantial discrepancies in estimated built-up areas. GHSL provides conservative yet temporally consistent estimates that align with known patterns of urban growth and infrastructure development. In contrast, NDBI and unsupervised classification significantly overestimate built-up extents, largely due to confusion between impervious surfaces and bright bare soils typical of dryland settings. Accuracy assessment based on reference samples confirms GHSL as the most reliable standalone method, while the HBDM layer improves spatial interpretation by balancing global calibration with local spectral and spatial cues.

The study demonstrates how integrating multiple geospatial data sources and analytical techniques can bridge descriptive urban mapping and predictive geodata analysis. From a practical perspective, the findings support urban planning, heritage conservation, and evidence-based decision-making in dryland regions undergoing rapid transformation. Methodologically, the research advances applied geospatial science by presenting a scalable hybrid framework that aligns with emerging GeoAI and predictive modelling paradigms, offering a foundation for future high-resolution, data-driven urban monitoring in similar environments.

“Comparison of Sentinel Imagery Performance Using Continuous Wavelet Transform-based CNN model for Rice-Crop System Mapping in Suphan Buri, Thailand”

Kritchayan Intarat, Nutcha Tuphimai, Phatsachon Kuankhamnuan, and Woraman Jangsawang; GIS-IDEAS Oral Presentation

Rice cropping system mapping is crucial for agricultural monitoring and precision agriculture. Such systems are mainly found in irrigated areas with two or more crops. The reliable of information on cropping intensity and temporal dynamics is required to support cultivation activities. Suphan Buri is a major rice-growing area in Thailand's central plain. It consists of an efficient irrigation system that benefits continuous rice cultivation. Information in this area is essential for rice management. However, cloud-prone areas are a significant issue in the tropical region and affect rice data acquisition. It leads to data gaps and uncertainty when using satellite data alone. Herein, this study integrates multi-temporal remotely sensed data with the continuous wavelet transform (CWT) and a convolutional neural network (CNN) to classify rice crop systems. Such a network improves the extraction of temporal features from satellite time-series data. Sentinel-1 and -2 (S1 and S2) are then incorporated into the experiment to compare the performance of radar- and optical-based inputs under the same classification framework. The VH polarization from S1 and the enhanced vegetation index (EVI) time series data from 2023 to 2025 are acquired. The Savitzky-Golay (SG) filter is then used to transform the 1D signal into 2D scalograms using the Morlet wavelet. The images are then fed into the CNN model and classified into four classes e.g. single-crop (SC), double-crop (DC), two-and-a-half-crop (HC), and triple-crop (TC). The results from S1 and S2 are then tested and evaluated, including accuracy, F1-score, and Kappa coefficient (K). In comparison, the CWT-based CNN model integrated with S2 imagery achieves superior performance compared to S1, reaching an accuracy of 0.892, an F1-score of 0.890, and a K of 0.855. These results reveal that the developed framework significantly improves the classification of rice cropping systems, highlighting its value as an efficient tool for rice cropping system mapping and agricultural monitoring in the cloud-prone region like Thailand.

“DETERMINATION OF GRAVITY POTENTIAL VALUES AT NATIONAL HEIGHT REFERENCE POINTS IN THE INDOCHINA REGION FOR HEIGHT SYSTEM UNIFICATION”

Vu Hong Cuong; GIS-IDEAS Poster Presentation

The unification and transformation of national height systems into regional and international height reference frameworks require the accurate determination of gravity potential values at national height reference points. Gravity potential is a fundamental physical quantity in modern geodesy and forms the theoretical basis for the realization of geopotential-based height systems, including the International Height Reference Frame (IHRF). Therefore, the reliable computation of gravity potential at height datum points is essential for establishing consistent connections among national, regional, and global height systems.
This study presents a methodology for determining gravity potential values at national height reference points using spherical harmonic coefficients of global gravity models. The proposed approach enables the computation of gravity potential in a unified global reference frame and supports the unification and conversion of height systems among different countries. The method is particularly suitable for regions where high-quality and homogeneous terrestrial gravity data are limited.
The methodology is applied to the national height systems of Vietnam, Laos, and Cambodia, representing the Indochina region. Gravity potential values at the national height reference points are computed using two global gravity models, EGM2008 and GOCE-DIR-R6. The calculations are performed based on GPS–leveling (GPS-TC) data, including 235 points in Vietnam, 104 points in northern and central Laos, and 84 points in Cambodia. A computational framework and processing tools were developed using the MATLAB programming environment to implement the theoretical formulation and calculation procedures.
The results indicate that the computed gravity potential value at the Vietnamese national height datum (Hon Dau tide gauge station) using EGM2008 is consistent with previously published studies, confirming the reliability of the proposed method and computational tools. Comparisons between the two global gravity models reveal noticeable differences in the derived gravity potential values, particularly in mountainous regions such as northern and central Laos. The GOCE-DIR-R6 model provides more consistent and physically reliable results, reflecting its higher accuracy in representing long- and medium-wavelength gravity field components.
The obtained gravity potential values allow the estimation of geopotential and height differences among the national height systems of Vietnam, Laos, and Cambodia. These results provide a scientific basis for height system unification and transformation within the Indochina region and for linking national height systems to the International Height Reference Frame. The study contributes practical methodology and reference data for future regional height system integration and supports the transition toward geopotential-based height systems.

“Determining coseismic displacements from GNSS timeseries using physics-based regression model”

Sambuddha Dhar; GIS-IDEAS Oral Presentation

Global Navigation Satellite System (GNSS) measurements continuously monitor the crustal deformation over the inter-, co-, and post-seismic period of an earthquake (e.g., Wang et al., 2012 Nature). During the coseismic event, the GNSS coordinate records an instantaneous discontinuity in the displacement timeseries (hereafter, coseismic offset). Determining such coseismic offset is crucial to understand the coseismic surface deformation as well as earthquake source parameters. The conventional method calculates the coseismic offset by taking the difference between pre- and post-earthquake GNSS coordinate positions. However, the conventional method often does not perform well in the case of missing data, higher positioning error, and lower sampling frequency. To address this issue, here, we developed a new method where the coseismic offset is calculated by fitting a physics-based regression model to the GNSS coordinate timeseries. Our proposed regression model consists of four components: linear term, seasonal variation, coseismic offsets, and transient deformation (Tobita, 2016 EPS; Dhar & Muto, 2023 GRL). The linear term represents the long-term secular velocity due to the continuous elastic deformation of the crust. The seasonal term includes annual and semiannual period variations due to seasonal changes such as surface mass loading and atmospheric changes. The earthquake offset represents the instantaneous response of the coseismic event. For the transient deformation term, we used the physics-based regression model proposed by Dhar & Muto (2023 GRL), which mimics inelastic deformation (viscoelastic stress relaxation of inelastic mantle and afterslip on fault surface) triggered by earthquake. Such deformation processes were inferred by previous laboratory experiments of mantle rock as well as postseismic deformation models (e.g., Muto et al., 2019 Sci. Adv.; Agata et al., 2019 Nat. Commun.). We employed our regression model to determine the coseismic displacements of the M9 2011 Tohoku-oki earthquake from the GNSS observations (1998-2021) provided by the GEONET network. We fit the regression model to the observed timeseries by non-linear least square methods (Freed et al., 2006 JGR). For a good-quality dataset, our results are consistent with those of the previous studies (e.g., Ozawa et al., 2011, Nature). For several coastal GNSS stations with initial missing data, our results show 10-30 cm differences in estimation of coseismic displacement compared to previous estimates (Freed et al., 2017 EPSL).

“Development of a Web-GIS Framework for Real-time PM2.5 Visualization and Monitoring via IoT Technology”

Natima Udon; GIS-IDEAS Oral Presentation

In recent decades, Thailand has experienced a critical air pollution problem, particularly fine particulate matter with an aerodynamic diameter of less than 2.5 µm (PM₂.₅), which has been identified as a major threat to public health, especially affecting the respiratory and cardiovascular systems. Addressing PM₂.₅ pollution under the constraints of climate change requires sustainable and technology-driven monitoring approaches that support continuous, real-time, and spatially explicit air quality assessment. This study presents the development of a Web-based Geographic Information System (Web-GIS) framework for real-time PM₂.₅ visualization and monitoring by integrating Internet of Things (IoT) technology with renewable energy systems.
The proposed framework is structured into three primary layers. First, a prototype sensing device was designed using an ESP8266 microcontroller integrated with particulate matter sensors and RS485-based temperature and relative humidity sensors. To ensure continuous and autonomous operation, a solar energy harvesting system was implemented to support energy-efficient and sustainable monitoring in remote or resource-limited areas. Second, the network and data management layer was developed using the HTTPS protocol to ensure secure data transmission. Sensor data were processed and stored in a cloud-based PostgreSQL database, enabling efficient data management and scalability. Third, the application layer was implemented as a responsive web-based platform to provide seamless access across multiple devices, including desktop computers, tablets, and mobile phones. An interactive online mapping module based on Leaflet.js was incorporated to support real-time spatial visualization and geospatial analysis of PM₂.₅ concentrations.
System performance evaluation demonstrates that the proposed Web-GIS framework provides stable and reliable continuous data acquisition, transmission, and visualization. The system effectively processes georeferenced PM₂.₅ data with high spatial accuracy and temporal consistency. The results indicate that the framework is suitable for deployment as a community- and local-scale air quality monitoring system, supporting evidence-based public health planning and spatial environmental management. Furthermore, the framework can be extended to integrate PM₂.₅ mitigation or control systems, offering a scalable foundation for future smart environmental management applications.

“Estimation of Rice Leaf NPK Content Using UAV-Based Hyperspectral Imagery and Machine Learning Techniques”

Tong Si Son; GIS-IDEAS Oral Presentation

Applications of modern sensing technologies and machine learning for monitoring crop nutrient status are essential to precision and smart agriculture. Accurate and timely information on vegetation nutrients supports optimized fertilizer management, improved yields, and reduced environmental impacts. This study investigates the integration of unmanned aerial vehicle (UAV)–based hyperspectral remote sensing and machine learning models to estimate nitrogen (N), phosphorus (P), and potassium (K) concentrations in rice leaves. Hyperspectral imagery was collected using a UAV-mounted hyperspectral camera with 122 spectral bands covering wavelengths from 350 to 950 nm and a ground sampling distance of 2 cm. Data acquisition was conducted during the panicle initiation stage of rice growth. Field sampling was performed at 78 locations evenly distributed across 24 rice parcels, each with a size of 10 m × 10 m, to obtain reference measurements of leaf NPK content. Spectral information extracted from the hyperspectral images was used as input variables for model development. Three machine learning regression algorithms, including Decision Tree Regression (DT), Random Forest Regression (RF), and Neural Networks (NN), were implemented and compared to model the relationship between spectral reflectance and rice leaf nutrient concentrations. Model performance was evaluated using accuracy metrics to identify the most suitable approach for each nutrient. The results indicate that the Neural Network model achieved the highest performance for nitrogen and potassium estimation, with accuracy rates of 67.2% and 88.67%, respectively. In contrast, the Decision Tree Regression model outperformed the other algorithms for phosphorus estimation, achieving an accuracy of 92.87%. These findings demonstrate that model effectiveness varies depending on the target nutrient and data characteristics. Overall, this study highlights the strong potential of combining UAV-based hyperspectral imagery, machine learning techniques, and GIS-based analysis for efficient monitoring of rice nutrient status. The proposed approach provides a valuable decision-support tool for precision agriculture and nutrient management in Vietnam rice systems.

“Evaluation of Horizontal Positional Accuracy in UAV-Based Mapping: A Case Study in Flat Urban and Rural Areas of Southern Vietnam”

Le Van Tinh; GIS-IDEAS Oral Presentation

Unmanned aerial vehicle (UAV) technology is widely applied in land surveying, topographic and thematic mapping, and geospatial data acquisition. Examples of mapping applications are cadastral mapping, topographic mapping, and current land-use mapping. In any of these applications, the accuracy of the imaging process is crucial and requires careful assessment. Therefore, the accuracy evaluation of the UAV imaging results has attracted considerable attention from many researchers. Although numerous studies have evaluated the accuracy of orthomosaics and digital elevation model (DEM) under different conditions, including variations in UAV platforms, technical parameters, and terrain characteristics, the lack of specific studies investigating the relationship between aerial survey technical factors such as flight altitude, the number of ground control points (GCPs), and terrain conditions of flat urban and rural areas of southern Vietnam remains. Therefore, the applying of UAV technology for producing cadastral mapping in this region remains challenge and limited. The accuracy of UAV-based mapping methods depends strongly on flight platforms, onboard cameras, photogrammetric control schemes, terrain characteristics, and image processing workflows. Thus, this study aims to evaluate the horizontal positional accuracy of orthomosaics generated from UAV imagery acquired using a DJI Phantom platform. The experiments were conducted in four representative deltaic areas, including two areas in Ho Chi Minh City (Tan Son Hoa Ward and An Khanh Ward) and two areas in Tra Vinh Province (Tra Cu Town and Kim Son Commune). Horizontal accuracy was assessed by comparing planimetric coordinates extracted from the orthomosaics with GCPs measured using a GNSS-CORS system. Accuracy metrics were quantified using the root mean square error (RMSE). The results indicate that the achieved horizontal positional accuracy ranges from 2.1 cm to 4.8 cm, and it meets the technical requirements for the accuracy of cadastral mapping scale 1:1000 as specified in Circular No. 26/2024/TT-BTNMT (Vietnamese standard: technical regulations on cadastral surveying and mapping). The proposed workflow and experimental results provide practical guidance for UAV-based mapping and contribute valuable empirical evidence on the positional accuracy evaluation practice of UAV photogrammetry in flat delta regions of southern Vietnam.

“Evaluation of New Global Gravity Models Using GNSS-Leveling Data in Viet Nam”

trung thanh; GIS-IDEAS Oral Presentation

This study evaluates the performance of five recent global gravity models, namely EGM2008, EIGEN-6C4, GECO, SGG-UGM-2, and XGM2019e_2159, using 818 GNSS-leveling points distributed across the territory of Viet Nam. The assessment was carried out using two approaches: an absolute evaluation based on the differences between observed height anomalies and those derived from the global models, and a relative evaluation based on height anomaly differences between point pairs at different distance intervals. The absolute assessment shows that the models provide different levels of agreement with the GNSS-leveling data, with root-mean-square error values of 0.835 m for EGM2008, 0.502 m for EIGEN-6C4, 0.489 m for GECO, 0.466 m for SGG-UGM-2, and 0.461 m for XGM2019e_2159. The relative assessment also confirms the superior performance of XGM2019e_2159 over the evaluated distance ranges. As the evaluated distance increases, the root-mean-square deviation per route tends to increase for all models, whereas the root-mean-square deviation per kilometer remains more stable and better reflects the comparative performance of the models. Among the tested models, XGM2019e_2159 provides the best overall performance, while EGM2008 gives the weakest performance. The results indicate that XGM2019e_2159 is the most suitable model among the tested global gravity models for applications related to height anomaly determination and geoid studies in Viet Nam.
Keywords: global gravity model, GNSS–leveling, height anomaly, relative assessment, Viet Nam

“Evaluation of Relationship Between Land Surface Temperature and Urban Spatial Density in Hanoi, Vietnam”

Takumi Dowaki; GIS-IDEAS Poster Presentation

In recent years, intense heat and frequent heatwaves driven by climate change have been observed worldwide, posing serious threats to human health and urban sustainability. In addition, the rapid expansion of urban areas and the consequent increase in energy consumption have further intensified the urban heat island effect, leading to higher ambient temperatures in cities. These temperature increases exacerbate heat-related risks and place additional pressure on urban infrastructure and energy systems. Therefore, mitigating urban temperature rise is essential not only for improving the quality of urban living environments but also as an effective measure for climate change adaptation and mitigation. Such efforts are indispensable for achieving sustainable urban development and long-term planning strategies.
　This study focuses on Hanoi, Vietnam, a city undergoing rapid urbanization. It uses satellite data to measure surface temperatures and evaluates their correlation with urban spatial density, calculated from factors such as topography, land use, buildings, infrastructure, and population. Vietnam's urban structure has been significantly influenced by land alterations and urban development during the French colonial era. This study quantitatively assesses such historically shaped urban structures using the new indicator of urban spatial density. Specifically, urban spatial density is calculated from three perspectives: 1) topography (including past land modifications), 2) land use, and 3) society (buildings, infrastructure, population, etc.), and its relationship with surface temperature is evaluated. This indicator, which quantitatively represents a city's structure and functions while considering historical impacts, enables the extraction of regional characteristics and impact assessment.
　This approach is expected not only to clarify the causes of recent surface temperature increases and understand thermal environmental systems but also to provide guidelines for future urban development and planning. Furthermore, this research is expected to contribute to evaluating the adaptability of cities toward future smart cities and compact cities in Vietnam. Moreover, as it is a methodology applicable to other major cities in Southeast Asia with similar social and historical backgrounds, it possesses not only creativity but also sufficient applicability across various fields.

“Evaluation of Urban Activity Distribution and Excursion Based on 3D City Models and Pseudo People Flow”

Konno Momoka; GIS-IDEAS Poster Presentation

Background and Objectives
In contemporary urban planning, the shift toward Evidence-Based Policy Making (EBPM) is essential to address structural issues such as population decline and the diversification of mobility. This study focuses on the downtown area of Sakai City, Osaka, which is defined by a "dual-center" structure between the Sakai and Sakai-Higashi station districts. A long-standing challenge in this region is the "straw effect," where the high accessibility to central Osaka leads to a drainage of local activity, resulting in fragmented urban excursion. The closure of the Takashimaya Sakai department store in January 2026 has further intensified the need for a scientific understanding of how urban structures influence stay patterns. The primary objective of this research is to establish a quantitative foundational study that evaluates the causal relationship between physical urban environments and human activity distribution.
Methodology and Data Integration
The methodology integrates two high-resolution data sources: Project PLATEAU 3D city models and synthetic population data (pseudo people flow). A significant departure from traditional 2D GIS analysis is the adoption of "Building Volume" as a key explanatory variable. By calculating 3D volumes from PLATEAU's LOD1 semantics, the research represents the city's actual capacity and the visual density experienced at eye level. Furthermore, accessibility is modeled using the inverse of the distance to the nearest station to accurately capture the non-linear decay of urban attraction. The use of synthetic population data allows for the analysis of detailed activity logs—including shopping, dining, and leisure—while maintaining privacy and providing specific agent attributes like age and purpose.
Predictive Modeling via Random Forest
To analyze the distribution of these activities, a machine learning framework using a Multi-Output Random Forest Regressor was implemented. The model incorporates 34 input variables, ranging from block area and perimeter to building counts and volumes categorized across 13 usage codes (e.g., 401 for office, 402 for commercial). The output consists of 18 variables representing both the headcount and the total stay duration (in seconds) for eight different activity purposes. Random Forest was selected due to its robustness in handling non-linear interactions and complex dependencies between diverse spatial features, allowing for a more stable estimation than simple linear models.
Verification of Results and Discussion
The model’s performance was rigorously evaluated using R-squared ($R^2$) and Mean Squared Error (MSE). The results demonstrated high predictive reliability for activities closely tied to specific building functions, such as "at-home" and "office work." Conversely, activities such as "shopping" and "leisure" yielded lower $R^2$ scores, suggesting that these behaviors are determined not only by the characteristics of a single block but also by broader spatial interactions and facility agglomerations. These findings serve as foundational research that clarifies the physical determinants of stay duration, providing a quantitative basis for future urban design and walkable policy assessments in Sakai’s urban center.

“Extraction of Slope Failure Features by Topographic Surface Changes”

Mitsunori Ueda; GIS-IDEAS Oral Presentation

Slope failures are geological hazards that occur in almost all regions of the world. Accordingly, slope failure inventory maps play an important role in various fields such as environmental conservation, urban development and disaster mitigation. The inventory maps provide crucial information on the locations, scales, and geometric shape of the slope failure areas, which is essential for analyzing their mechanisms and dynamics. The slope failure areas can be characterized by the changes in the surface morphology before and after the event. However, the slope failure features are not included in many inventory maps. The lack of information about slope failure features decreases the efficacy of the inventories, as the scarp in the area contains information about the slide surface and movement direction, while the main body in slope failure contains information about the volume of material and movement type. Therefore, it is necessary to develop a method to extract the slope failure features from relative changes in the topographical surface.
In this study, we introduce a concept of the topographic surface frame to identify geometric changes in the surface. This frame is a mathematical concept that expresses the relative relationship between surfaces before and after the event. The topographic surface frame describes geometric changes in terms of slope, relative dip and relative direction, regardless of their orientation. Therefore, the morphological changes in slope failure areas are quantitatively characterized by four variables: the elevation changes, the slope angle before the event, the relative dip, and the relative direction. In this study, an algorithm is presented to extract main scarps and the top surface of main bodies in the slope failure areas using the characterized morphological changes. The extraction algorithm was applied to slope failures caused by heavy rainfall. It extracts the changes in the topographic surface, scarp and top surface of the main body in the slope failure area. This study revealed relative changes in the morphology of slope failure areas and enabled robust and straightforward extraction.
The method contributes to generating novel slope failure inventory maps that include information on location, scale, morphology, shape and flow direction. Furthermore, the algorithm can improve the slope failure modeling for understanding and predicting future events.

“From 2D to 3D Cadastre: A Hybrid 3D Cadastral Database Model for Managing Complex Urban Spaces in Vietnam”

Le Van Tinh; GIS-IDEAS Oral Presentation

The rapid growth of high-rise buildings and underground infrastructures in Vietnam has revealed fundamental limitations of the current two-dimensional (2D) cadastral system, which primarily represents land parcels on the surface. Complex vertical overlaps of land use rights, building units, and subsurface utilities cannot be consistently represented, queried, or visualized in a 2D environment, resulting in legal ambiguity, inefficient land management, and limited support for land valuation and urban planning. Numerous studies have demonstrated that three-dimensional (3D) cadastre models, particularly those based on the Land Administration Domain Model (LADM), provide an effective framework for integrating legal rights, restrictions, and responsibilities (RRRs) with physical spatial objects. However, studies addressing the specific institutional and technical context of Vietnam remain limited. This paper aims to propose a hybrid 2D/3D cadastral database model that can be gradually implemented without disrupting the existing 2D cadastral system. The study uses a three-phase methodological framework: (i) analysis of international 3D cadastre concepts and implementation experiences; (ii) design of a conceptual, logical, and physical 3D cadastral database model using UML and GML 3.2.2; and (iii) implementation and testing of the proposed model through a prototype system in a dense urban area of Ho Chi Minh City. The hybrid 3D cadastral database model incorporates data classes including 2D land parcels, 3D cadastral objects, terrain surfaces, and legal spaces. The proposed model organizes cadastral information into nine interrelated sub-models representing legal and physical components and supports explicit 3D parcel representations using GM_Solid geometry and associated semantic classes. The results were conducted on a 3D model named VN3DLIS that effectively integrates 2D and 3D cadastral information, supports advanced 3D queries and visualization of overlapping rights, and provides a scalable pathway toward 3D cadastre implementation in Vietnam. The proposed hybrid model enhances land administration agencies' capacity to manage complex urban spaces and supports multi-purpose land information services in rapidly urbanizing environments.
Keywords: 3D cadastre; cadastral database; hybrid 2D/3D model; land administration

“Geostatistical Analysis and Variogram Simulation Techniques for Assessing Earthquake Risk between the Myanmar and Thailand Areas.”

Chaiwiwat Vansarochana; GIS-IDEAS Oral Presentation

Earthquakes are one of the natural disasters that cause damage to lives, property, buildings, and the economy, especially in Southeast Asia, which is located on the boundary of the Eurasian and Indo-Australian tectonic plates. The upper region of Myanmar, particularly Mandalay, is considered one of the areas with a high rate of earthquakes due to its proximity to the Sagaing Fault, a large, active fault with continuous movement. The tremors from earthquakes can affect areas in northern Thailand. Therefore, studying the distribution of earthquake intensity is crucial for risk assessment and disaster management planning. However, due to limited and unevenly distributed data, estimating intensity in areas without data is difficult. Currently, geostatistical analysis techniques such as Kriging's method, which considers the relationship between distance and direction, are being applied. This provides more accurate results than conventional estimation methods such as Inverse Distance Weighting (IDW).Although Kriging applications are popular in geostatistics and environmental studies, comparisons of the performance of different Kriging models (Ordinary and Universal) and various variogram models (Spherical, Exponential, and Gaussian) in simulating earthquake intensity are limited, particularly in the context of Southeast Asia, which has complex geology and diverse fault lines. This study, therefore, focuses on applying Kriging techniques to model the Community Determined Intensity (CDI) distribution of earthquake intensity from a large earthquake near Mandalay, Myanmar, using USGS data from 176 sites in Myanmar and Thailand. To create a map showing seismic attenuation patterns affecting Thailand, this research aims to (1) compare the performance of Kriging models (Ordinary and Universal) in simulating earthquake intensity from the Mandalay epicenter, (2) evaluate the suitability of Variogram models (Spherical, Exponential, Gaussian) for spatial simulation, and (3) create an intensity distribution map and an earthquake hazard map to be used as supporting information for disaster risk assessment and appropriate land use planning.
The analysis was performed using geostatistical techniques to create intensity maps and hazard maps. The results showed that the Ordinary Kriging – Spherical Variogram model
provided the most accurate results, with an R² value of 0.8679, a low RMSE, and a Nugget/Sill of 14.73%, indicating a high level of spatial correlation. It can realistically simulate seismic attenuation. The resulting map shows continuous attenuation of intensity with distance. The highest-intensity earthquakes were observed around Mandalay, decreasing to moderate to low levels in northern Thailand. The study results can be used as baseline data for assessing earthquake risk, identifying risk areas, and planning infrastructure in the region, which is very interesting.

“GIS Applications in Managing Cultural Heritage Tourism: A Case Study of Can Tho City, Vietnam”

HienLe; GIS-IDEAS Poster Presentation

Cultural heritage tourism in Southeast Asia has increasingly embraced spatial technologies to enable data-driven planning, sustainable conservation, and enhanced visitor experiences amid rapid urbanization and climate pressures. Can Tho City, a centrally administered municipality and the largest urban center in Vietnam's Mekong Delta, boasts a rich tapestry of cultural heritage assets spanning historical pagodas, French colonial architecture, traditional markets, archaeological sites, and unique ecological landscapes shaped by the region's intricate riverine system. Despite this wealth, spatial data on these assets remains fragmented across municipal departments, poorly digitized, and underutilized in tourism management frameworks, leading to inefficient resource allocation, missed economic opportunities, and risks to site preservation.

This study addresses these challenges by developing a comprehensive, integrated spatial database using Geographic Information Systems (GIS) to revolutionize cultural heritage tourism management in Can Tho. Methodologically rigorous, we conducted extensive field surveys across 12 urban and rural districts, documenting 187 heritage sites through GPS ground-truthing, photographic inventories, and stakeholder interviews with local tourism authorities and community leaders. Leveraging open-source QGIS software, we performed advanced geospatial analyses including site classification by official designation status (national, provincial, district-level), chronological age (pre-colonial, colonial-era, post-1945), architectural typology, ecological significance, and multi-modal transport accessibility (road networks, waterways, public transit).

Results unveil stark spatial disparities: urban Ninh Kieu and Binh Thuy districts concentrate 62% of accessible, high-value sites, while rural Vinh Thanh and Thot Not lag with 78% of sites facing poor connectivity and degradation risks. Accessibility analysis reveals that 41% of provincial-level heritage assets lie beyond 5 km of paved roads or major waterways, severely limiting tourism potential. Hotspot mapping identifies cluster concentrations along the Can Tho River corridor, suggesting prime zones for themed heritage trails. Network analysis demonstrates that integrating underutilized inland waterways could boost site connectivity by 35%, transforming them into sustainable tourism corridors.

GIS proves transformative, enabling interactive 3D visualizations, multi-criteria decision analysis for conservation prioritization, and scenario modeling for tourism carrying capacity. The resulting web-based decision-support platform facilitates real-time query capabilities for planners, integrating layers for visitor flows, environmental vulnerability, and economic impact projections.

Strategic recommendations include: (1) developing integrated waterway heritage tourism routes linking Cai Rang floating market with upstream pagoda clusters; (2) launching community-based homestay networks in rural districts to distribute tourism benefits; (3) establishing a municipal heritage GIS portal with public API access; and (4) implementing adaptive management protocols using annual spatial monitoring. These evidence-based interventions promise to elevate Can Tho's cultural tourism from fragmented site visits to immersive, sustainable experiences.

This research contributes a replicable methodological framework for other Mekong Delta cities and rapidly urbanizing riverine municipalities across Southeast Asia. By mainstreaming spatial technologies in heritage governance, Can Tho can position itself as a model for balancing economic development, cultural preservation, and environmental sustainability in climate-vulnerable urban ecosystems, ultimately enhancing community livelihoods through responsible tourism growth.

“GIS-BASED FLOODWATER DEPTH MAPPING WITH FWDET: VALIDATION AND DEM INTERCOMPARISON FOR THE 2025 VU GIA–THU BON FLOOD”

Hang Ha; GIS-IDEAS Poster Presentation

Determining flood depth from remote sensing data and a Digital Elevation Model (DEM) using the Floodwater Depth Estimation Tool (FwDET) is the latest cutting-edge approach in flood hazard management in recent years. The accuracy of these flood inundation depth maps depends mainly on the reliability of the DEM data. Despite the increasing availability of global DEM products, limited research has systematically evaluated and recommended suitable DEM sources for flood-inundation depth mapping, particularly for freely available datasets with identical spatial resolution. In this study, flood-inundation depth maps were generated and validated for the Vu Gia–Thu Bon river basin during the historic flood event in October 2025 using multiple freely available DEM datasets, including FABDEM, ALOS, and SRTM. All selected DEMs have a spatial resolution of 30 meters, allowing a consistent comparison of their performance in flood depth estimation. The flood boundary was determined from SAR Sentinel-1 imagery data at the exact time of the flood event in October 2025 in the study area. Then, this flood extent data was combined with each DEM using the Floodwater Depth Estimation Tool (FwDET) on the Google Earth Engine platform to derive spatially distributed flood depth maps. Historical flood traces were used to verify and assess the accuracy of the flood inundation depth maps in this study. The obtained results indicate that FABDEM reflects the depth quite closely with the actual measured depth value, so it was chosen as the most suitable DEM data in this study. The results of this study provide a scientific basis and contribute to improving the accuracy and practical applicability of flood inundation depth maps in implementing response and relief strategies during flood seasons. Overall, this approach demonstrates strong potential for operational flood inundation depth mapping in areas frequently facing flooding and data-scarce regions.

“Integrating GIS and Machine Learning for Agricultural Crop Zoning and Suitability Assessment: A Case Study in Nong Truong Moc Chau Town, Son La Province, Vietnam”

Bui Ngoc Tu, Doan Huong-Giang; GIS-IDEAS Oral Presentation

Land suitability assessment for specialty crops in highland Vietnam remains largely dependent on traditional knowledge and empirical practice, with systematic, evidence-based spatial planning frameworks still limited at the local scale. This study presents an integrated Geographic Information System Multi Criteria Analysis and Machine Learning (GIS-MCA-ML) workflow for identifying suitable zones for concentrated Japanese plum (Prunus salicina) cultivation in Nong Truong Moc Chau, Son La Province Vietnam's largest plum-growing area with approximately 3,500 hectares of orchards and over 100,000 tons of provincial output in 2025. Eleven spatial criteria spanning terrain, climate, hydrology, soil properties, land cover, and infrastructure were standardized, reclassified and weighted using the Analytic Hierarchy Process (AHP; Consistency Ratio CR<0.10). Weighted overlay analysis in GIS produced a five-class land suitability index map across approximately 107,000 ha of the study area. Results show that highly suitable (S1) zones account for less than 1% of the study area (about 964.8 ha). The AHP-derived suitability map was subsequently used as labeled training data for a Random Forest (RF) classifier trained on an 80/20 stratified split (N = 1,070 pixels; optimal hyperparameters: n_estimators = 200, max_depth = 7, max_features = 5), achieving an overall accuracy of ~88-92% and Cohen's Kappa of ~0.85-0.90 on the independent test set, confirming the internal consistency of the expert-based weighting scheme. SHAP (SHapley Additive exPlanations) analysis using the TreeExplainer algorithm and permutation-based importance confirmed that slope (SLP) and distance to surface water (DTW) are the dominant drivers of suitability outcomes, exerting the largest mean absolute SHAP values and accounting for the greatest accuracy drop under permutation consistent with their top AHP priority weights (0.180 and 0.060, respectively) and with the biophysical requirements of Prunus salicina. The alignment between RF feature importance scores and AHP-derived weights (Pearson r ≈ 0.87) validates the expert judgment underpinning the multi-criteria framework. These findings provide the first systematic GIS-based land suitability map for plum cultivation in Moc Chau and offer an interpretable, replicable ML-assisted framework for spatial agricultural planning in highland Vietnam.

“Integrating GIS in Climate Change Education: A Case Study at Hanoi University of Natural Resources and Environment”

Ngoc_Anh_Nguyen, Bùi Thị Phương Thùy; GIS-IDEAS Poster Presentation

The rapid development of geospatial technologies has significantly transformed approaches to climate change research, shifting from traditional descriptive methods toward data-driven and predictive geospatial science approaches. In this context, the integration of Geographic Information Systems (GIS) into university teaching is increasingly recognized as a critical solution for fostering spatial thinking and enhancing students’ analytical capacity. Developing these competencies is particularly essential for students in climate change–related disciplines, where the ability to analyze spatial patterns, trends, and interactions plays a central role in scientific research and evidence-based decision-making.

This study was conducted to evaluate the effectiveness of integrating GIS into university teaching in improving spatial analysis skills for climate change research among undergraduate students majoring in Climate Change and Sustainable Development at Hanoi University of Natural Resources and Environment (HUNRE). The research adopted a quasi-experimental design involving 30 students from the Faculty of Climate Change and Sustainable Development. Students participated in two different learning modalities: (i) a course integrating GIS applications into climate change analysis, and (ii) a course delivered using traditional teaching methods without GIS integration.

The GIS-integrated instructional content focused on key competencies, including spatial data visualization, thematic map interpretation, and the assessment of spatial differentiation of climatic conditions. These activities were designed to enhance students’ ability to connect theoretical knowledge with spatial evidence and to support analytical reasoning in climate change research contexts.

Research data were collected through a structured questionnaire survey aimed at assessing students’ perceptions of learning effectiveness, spatial thinking development, data analysis skills, and readiness to apply GIS in climate change research. Quantitative data were analyzed using SPSS software, employing descriptive statistics, reliability testing of measurement scales, and comparative statistical analyses.

The findings indicate that the integration of GIS into university teaching has a positive and significant impact on students’ ability to read, interpret, and analyze spatial data, as well as on their confidence in applying GIS tools in academic research. The study confirms the important role of GIS-based teaching approaches and provides empirical evidence to support curriculum innovation in higher education toward strengthening the application of geospatial technologies in climate change education.

“Integrating harmonized Land use dynamics and Biodiversity models for Global agricultural impacts”

Can Trong Nguyen; GIS-IDEAS Oral Presentation

Title: Integrating harmonized Land use dynamics and Biodiversity models for Global agricultural impacts

Speaker: Can Trong Nguyen

Global biodiversity is declining rapidly, with agricultural expansion and intensification representing dominant anthropogenic drivers of ecosystem degradation. This study develops a high-resolution geospatial framework integrating harmonized land use data and biodiversity models to estimate global biodiversity intactness index (BII), thereby quantifying biodiversity impacts from agricultural production. Multiple datasets of geospatial big data sources, including global land-use reconstructions, satellite-derived land-cover products, global pasture, and forest information, were combined to produce land use map in 2020. Biodiversity integrity was quantified by BII, estimated from spatial models of species abundance and compositional similarity trained by the PREDICTS biodiversity database. Agricultural biodiversity impacts were quantified as the area of ecosystems experiencing irreversible biodiversity loss and allocated to specific crops. Results show that over 60% of global terrestrial ecosystems are characterized by moderately intact and degraded biodiversity states. Global agricultural production causes 2.1 million km2 of permanent biodiversity loss. One-third of these impacts occur in the Asia-Pacific region, followed by Sub-Saharan Africa and the Americas. Staple crops such as wheat, maize, rice, and soybean account for the largest total impacts, while other crops including fruits, coffee, and other cereals exhibit higher biodiversity loss per cultivated area. These findings provide a spatially explicit basis for assessing biodiversity footprints in global food systems.

Keywords: biodiversity integrity; biodiversity intactness index (BII); agricultural impacts; spatiotemporal dataset; land use harmonization.

“Integrating Multi-Sensor Satellite Imagery and Machine Learning to Assess the Impact of Urban Expansion on Air Quality and Surface Urban Heat Islands”

Nguyen Manh Hung; GIS-IDEAS Oral Presentation

Nguyen Manh Hung1,2,, Vu Anh Tuan1, Nguyen Hong Quang3, Dao Duy Toan4, Le Mai Son1, Le Thi Thu Hang1
1 Vietnam National Space Center, Vietnam Academy of Science and Technology, Hanoi, Vietnam
2 Graduate University of Science and Technology, Vietnam Academy of Science and Technology, Hanoi, Vietnam
3 Technology Multimedia Technology Department, Multimedia Faculty, Posts and Telecommunications Institute of Technology
4 Faculty of Bridge and Road, Hanoi University of Civil Engineering, Hanoi, Vietnam
E-mail:nmhung@vnsc.org.vn

Rapid urbanization in emerging megacities is intensifying dual environmental crises: localized microclimate shifts and air quality degradation. This study establishes an integrated analytical framework using multi-sensor satellite data and machine learning to decode the synergistic nexus between urban sprawl, surface urban heat islands (SUHI), and atmospheric pollution in Hanoi, Vietnam. Leveraging the computational power of Google Earth Engine (GEE), we synthesized a multi-temporal dataset including Landsat (for LST, SUHI, and LULC), MODIS (for long-term Aerosol Optical Depth - AOD trends), and Sentinel-5P TROPOMI (for NO2 and SO2 concentrations). A Random Forest (RF) algorithm was implemented for land cover classification, incorporating spectral bands alongside NDVI, NDWI, NDBI, and Digital Elevation Model (DEM) data to mitigate topographic effects and enhance classification robustness. The model achieved superior performance with an overall accuracy (OA) exceeding 98% and a Kappa coefficient above 0.95. Spatiotemporal results reveal that the uncontrolled expansion of impervious surfaces, primarily in western and southern Hanoi, acts as the primary driver for thermal and pollution "hotspots." The conversion of agricultural land and water bodies into high-density built-up areas has led to a significant increase in Land Surface Temperature (LST), with peak SUHI intensities observed within industrial clusters and along radial transport axes. Statistical analysis confirms a strong positive correlation between built-up density, SUHI intensity, and pollutant accumulation. These findings highlight the severe consequences of green space loss and "heat-trapping" effects within dense urban structures, where stagnant atmospheric conditions exacerbate pollutant concentrations. Furthermore, the integration of Sentinel-5P data identified clear spatial overlaps between thermal anomalies and NO2 emission plumes, particularly in the vicinity of coal-fired power plants and key transport corridors. These empirical findings provide a robust scientific basis for the "Urban-Heat-Air" nexus. The study advocates for the strategic implementation of nature-based solutions, such as green infrastructure and ventilation corridors, to enhance ecological resilience and mitigate adverse urban microclimate impacts in Hanoi amidst global climate change.
Keywords: Urban expansion, Air Quality, SHUI, Machine Learning, Multi-Sensor Satellite.

“International Cooperation for Climate Change Response under the SDGs Framework: Evidence from Vietnam”

Doan Thi Xuan Huong; GIS-IDEAS Oral Presentation

Climate change poses significant challenges to sustainable development, particularly for countries highly vulnerable to environmental and socio-economic impacts such as Vietnam. Achieving Sustainable Development Goal 13 (Climate Action) requires not only strong national commitments but also effective international cooperation in policy coordination, technology transfer, capacity building, and information sharing. In recent years, the growing availability of geospatial and climate-related data has highlighted the need to move beyond static information systems toward more predictive, forward-looking approaches to climate governance and decision-making. Within this context, international cooperation plays a critical role in enhancing national capacities to utilize geospatial information and climate data for sustainable development planning.
This paper examines the role of international cooperation in supporting Vietnam’s response to climate change within the framework of the United Nations Sustainable Development Goals (SDGs) during the period 2015–2030. The study adopts a qualitative research approach based on policy and document analysis, reviewing key international agreements, national strategies, and cooperation programs related to climate change mitigation and adaptation in Vietnam. An analytical framework is applied to explore the linkages between climate governance, international cooperation mechanisms, and SDG implementation outcomes, with particular attention to information sharing, knowledge exchange, and decision-support processes.
The results indicate that international cooperation has made important contributions to strengthening institutional capacity, enhancing access to climate-related and geospatial data, and supporting pilot initiatives for climate change adaptation. These efforts have contributed to improving evidence-based planning and supporting anticipatory approaches to climate risk management. However, several challenges persist, including fragmented stakeholder coordination, limited integration of social dimensions into cooperation frameworks, and insufficient utilization of shared information in local-level decision-making processes. These gaps constrain the effective translation of available geospatial information into predictive and actionable insights for climate governance.
The paper discusses the gaps between global climate commitments and national implementation practices and proposes policy-relevant recommendations to improve the effectiveness of international cooperation for climate action. Emphasis is placed on strengthening information exchange, promoting multi-stakeholder engagement, and aligning international support with local development needs. The study contributes to ongoing discussions on climate governance and provides practical insights for advancing the implementation of SDG 13 while supporting the transition toward predictive geodata-informed decision-making in developing countries.

“Landfill Site Suitability Assessment Using Machine Learning and ANP Supermatrix: Evidence from Hai Phong City, Vietnam”

Sang Nguyen; GIS-IDEAS Oral Presentation

Land suitability assessment for specialty crops in highland Vietnam remains largely dependent on traditional knowledge and empirical practice, with systematic, evidence-based spatial planning frameworks still limited at the local scale. This study presents an integrated Geographic Information System Multi Criteria Analysis and Machine Learning (GIS-MCA-ML) workflow for identifying suitable zones for concentrated Japanese plum (Prunus salicina) cultivation in Nong Truong Moc Chau, Son La Province Vietnam's largest plum-growing area with approximately 3,500 hectares of orchards and over 100,000 tons of provincial output in 2025. Eleven spatial criteria spanning terrain, climate, hydrology, soil properties, land cover, and infrastructure were standardized, reclassified and weighted using the Analytic Hierarchy Process (AHP; Consistency Ratio CR<0.10). Weighted overlay analysis in GIS produced a five-class land suitability index map across approximately 107,000 ha of the study area. Results show that highly suitable (S1) zones account for less than 1% of the study area (about 964.8 ha). The AHP-derived suitability map was subsequently used as labeled training data for a Random Forest (RF) classifier trained on an 80/20 stratified split (N = 1,070 pixels; optimal hyperparameters: n_estimators = 200, max_depth = 7, max_features = 5), achieving an overall accuracy of ~88-92% and Cohen's Kappa of ~0.85-0.90 on the independent test set, confirming the internal consistency of the expert-based weighting scheme. SHAP (SHapley Additive exPlanations) analysis using the TreeExplainer algorithm and permutation-based importance confirmed that slope (SLP) and distance to surface water (DTW) are the dominant drivers of suitability outcomes, exerting the largest mean absolute SHAP values and accounting for the greatest accuracy drop under permutation consistent with their top AHP priority weights (0.180 and 0.060, respectively) and with the biophysical requirements of Prunus salicina. The alignment between RF feature importance scores and AHP-derived weights (Pearson r ≈ 0.87) validates the expert judgment underpinning the multi-criteria framework. These findings provide the first systematic GIS-based land suitability map for plum cultivation in Moc Chau and offer an interpretable, replicable ML-assisted framework for spatial agricultural planning in highland Vietnam.

“MACHINE LEARNING-BASED MISSING DATA IMPUTATION FOR HOURLY AIR QUALITY MONITORING DATA FROM AUTOMATED STATIONS IN VIETNAM”

Lê Quang Tú; GIS-IDEAS Oral Presentation

Air quality monitoring data, acquired through automatic monitoring stations, serves as a cornerstone for effective environmental management, public health impact assessment, and air pollution forecasting. In the specific context of Vietnam, these automated stations are responsible for providing hourly measurements of critical pollution parameters, including PM₁, PM₁₀, NO₂, and CO. Nevertheless, the datasets gathered from these sources frequently suffer from severe data missingness. These interruptions are primarily caused by sensor malfunctions, data transmission errors, and routine system maintenance processes. Such data deficiency significantly compromises the reliability of the dataset and restricts the efficacy of subsequent analytical tasks, including comprehensive air quality assessment and predictive modeling. Traditional handling methods, such as mean substitution or linear interpolation, often prove inadequate. They are particularly limited when addressing the complex, non-linear fluctuations and the strong time-dependent nature inherent in atmospheric pollutants. Addressing these limitations, this study focuses on the imputation of missing hourly air quality data by leveraging advanced machine learning-based methods. To ensure practical applicability, real-world datasets collected from automatic air quality monitoring stations across Vietnam are utilized to simulate a variety of missing data scenarios. These scenarios encompass both random missing data patterns and long, consecutive data gaps, reflecting actual operational challenges. A diverse range of machine learning models are trained to rigorously exploit the temporal dependencies and the inter-correlations among different air quality parameters to accurately reconstruct the missing values. The performance of the proposed methods is systematically evaluated and compared against traditional statistical imputation techniques using standard error metrics. Experimental results demonstrate that machine learning-based imputation methods significantly outperform traditional approaches in reconstructing missing air quality data. Notably, the proposed models successfully preserve key temporal features, including diurnal (day-night) cycles and seasonal variations, while maintaining the statistical distribution properties of the original datasets. Although challenges remain when dealing with exceptionally long data gaps, the findings confirm that machine learning imputation is a robust and reliable solution for enhancing the completeness of air quality data. This improvement in data quality establishes a solid foundation for critical downstream applications, including air pollution forecasting, the accurate estimation of the Air Quality Index (AQI), and data-driven decision-making for environmental management in Vietnam.

“Mapping Above-Ground Biomass and Carbon Stock of Melaleuca Forests (Melaleuca cajuputi Powell) Using PlanetScope Imagery in the Gao Giong Ecotourism Site/Nature Reserve, Vietnamese Mekong Delta”

Nguyen Ho; GIS-IDEAS Oral Presentation

Quantifying above-ground biomass (AGB) of Melaleuca forests is essential for assessing carbon storage potential and supporting evidence-based wetland management in the Vietnamese Mekong Delta (VMD). Recent advances in high-resolution satellite data and machine-learning methods enable spatially explicit biomass estimation by integrating remotely sensed predictors with field inventory measurements. In this study, we mapped AGB and corresponding above-ground carbon stocks of Melaleuca cajuputi at the Gao Giong Ecotourism Site/Nature Reserve (Dong Thap Province, VMD) by combining PlanetScope imagery with plot-based forest inventory data.
We collected field measurements from 30 sample plots distributed across representative Melaleuca stands. PlanetScope multispectral imagery – characterized by fine spatial detail suitable for heterogeneous wetland forests – was processed to derive spectral predictors (e.g., band reflectance and vegetation indices) and used as input features for statistical learning. We compared candidate models and found that a Random Forest approach provided the most reliable performance for AGB estimation. The resulting model produced a high-accuracy AGB prediction and enabled wall-to-wall mapping across the study area. The mean estimated AGB of Melaleuca forests was 161.25 ± 36.17 t ha⁻¹, indicating substantial biomass and carbon storage capacity at Gao Giong. Based on the AGB outputs, we derived an above-ground carbon stock surface using a standard biomass-to-carbon conversion approach. Map validation indicated an overall mapping accuracy of 81.25%, supporting the robustness of PlanetScope-based biomass and carbon mapping for local-scale monitoring.
Our results demonstrate that PlanetScope imagery, when coupled with field plots and machine-learning regression, has strong potential for operational AGB and carbon-stock mapping in Melaleuca-dominated wetlands – not only in Dong Thap but also across the broader VMD. The study contributes practical geospatial evidence for carbon accounting, conservation planning, and climate-mitigation strategies aligned with predictive geodata science.
Keywords: Remote sensing; Carbon stock; PlanetScope; Random Forest; Melaleuca forest; Mekong Delta

“Mission-Centric Task Planning for Heterogeneous UAV Swarms in Large-Scale Topographic Mapping”

Luu Hai Au; GIS-IDEAS Oral Presentation

The rapid development of unmanned aerial vehicles (UAVs) has enabled their widespread use in large-scale topographic mapping and geospatial data acquisition. Compared to single-platform operations, UAV swarms offer significant advantages in coverage efficiency, redundancy, and operational flexibility. However, task planning for heterogeneous UAV swarms remains a challenging problem due to differences in sensor payloads, flight endurance, mobility constraints, and mission priorities. Existing task allocation approaches often overlook mission-level objectives or assume homogeneous agent capabilities, limiting their effectiveness in complex real-world scenarios.
The aim of this study is to develop a mission-centric task planning framework for heterogeneous UAV swarms that maximizes overall mapping efficiency while respecting individual platform constraints and mission requirements. The proposed concept emphasizes global mission objectives rather than local task optimization, enabling coordinated decision-making across the swarm.
The approach integrates mission decomposition, capability-aware task allocation, and adaptive scheduling. First, the large-scale mapping mission is decomposed into spatially and functionally independent subtasks based on terrain characteristics and required data resolution. Each UAV is modeled according to its sensing capability, energy capacity, flight dynamics, and communication range. A multi-objective optimization strategy is then employed to assign tasks to UAVs by balancing coverage completeness, energy consumption, and mission time. To enhance robustness, the framework incorporates dynamic reallocation mechanisms that respond to UAV failures or environmental changes.
Simulation-based experiments are conducted on realistic topographic scenarios to evaluate the effectiveness of the proposed method. The results demonstrate that the mission-centric planning strategy significantly improves coverage efficiency and reduces overall mission time compared to conventional homogeneous or static task allocation methods. Moreover, the heterogeneous swarm shows enhanced adaptability and resilience under uncertain operational conditions.
In conclusion, this research presents an effective and scalable task planning framework tailored for heterogeneous UAV swarms in large-scale topographic mapping missions. The main contribution lies in introducing a mission-centric perspective that systematically integrates platform heterogeneity into task allocation and execution. The proposed approach provides practical insights for the design and deployment of intelligent UAV swarm systems in geospatial and environmental applications.
Key words: Swarm UAV, mission planning, survey area partitioning, UAV, topographic mapping.

“Optimizing Green Supply Chain Management for Vietnamese Coffee Exports: A Digital MRV Framework Integrating AI and Blockchain for Carbon Emission Transparency”

thamhua; GIS-IDEAS Poster Presentation

In 2025, Vietnam’s coffee exports are expected to reach a historic record, with export revenue estimated to exceed US $8.9 billion and export volume around 1.59 million tons by year-end, driven by high global coffee prices, strong demand for Robusta, and a strategic shift toward deep processing and value-added products. This growth reflects not only increased global demand but also the industry’s investment in quality upgrade strategies, specialty coffee segments, and sustainability certification, which have strengthened Vietnam’s position as a leading Robusta exporter in global markets, especially in Europe and North America.
The substantial increase in export value is underpinned by rising international prices for coffee beans and processed products, as well as more robust supply chain transparency and traceability mechanisms that align with international buyers’ sustainability expectations. However, emerging regulatory challenges, such as the European Union’s Deforestation-free Regulation (EUDR) and the Carbon Border Adjustment Mechanism (CBAM), as well as US tariff measures, require Vietnamese exporters to rapidly adapt their compliance systems to meet stringent traceability, due diligence, and environmental standards. Scientific research shows that such regulatory pressures are reshaping global coffee flows, making comprehensive supply chain transparency and verified origin data mandatory prerequisites for continued access to EU markets. Academic literature highlights that supply chain coordination toward sustainability is a critical driver for exporters, as institutional regulations, market forces, and competitive pressures encourage firms to integrate green practices and certification into their business models. Furthermore, systematic reviews of the coffee supply chain emphasize the importance of digital traceability technologies such as blockchain, IoT, and data analytics; to ensure environmental compliance, enhance transparency, and support circular economy objectives. Against this backdrop, this research systematically charts the green label certification process and outlines the key steps exporters must undertake to meet European regulatory requirements, thereby minimizing risks such as container demurrage at destination ports. The author also proposes a digital framework for green supply chain management, integrating traceability and compliance documentation to help enterprises more efficiently demonstrate adherence to sustainability standards when interacting with European certification bodies and customs authorities

“Pavement crack detection and segmentation using lightweight deep learning models”

Nguyen Hong Quang; GIS-IDEAS Oral Presentation

Pavement is a significant part of the road system, providing comfort and necessary space for rudimentary vehicles and pedestrians. However, it is often overloaded and distracted by human-induced and natural impacts like water erosion and tree roots. The cracks on the pavements are a manifestation of damage and degradation. Hence, early detection and segmentation of those cracks will help to regularly inspect, repair, and most importantly, prevent accidents for traffic participants. Deep learning in computer vision has proved its robustness in many aspects of application, including pavement study. In this study, we conducted on-site image and video capturing for pavement crack observations to prepare a training dataset of around 25 thousand images and masks, which are split into 70%, 22%, and 10% for training, validation, and testing datasets. Although deep learning models have been developed to advance with accuracy approaching perfection. The applications of small (lightweight) architectures are still limited as they often generate lower accuracy compared to large networks. However, light networks offer comfortable and affordable computational resources. Therefore, the YOLO-NAS, a super lightweight model, is tuned for rapid crack detection, and the U-Net(VGG16) architecture is used for the segmentation task. While the transfer learning technique is used for fine-tuning the YOLO-NAS model, the U-Net(VGG16) is inferred directly with adjustment of probability thresholds to generate accurate crack masks. Both models have demonstrated their super-fast speed and precision of above 87%. The results of the models are nicely delineated and mapped for purposes visulization. Even though there are some small false positives and negatives of crack segmentations, the model performances are robust and at a high level of accuracy. The combination of these models is recommended for real-time operations and is integrated into portable devices for on-site implementations. Finally, the future work is suggested to focus more on training data improvement, both in quality and quantity.

“Quantifying the contribution of topographic and spectral indices to LULC classification accuracy: a Random Forest-based feature selection study”

Nguyen Manh Hung; GIS-IDEAS Oral Presentation

Quantifying the contribution of topographic and spectral indices to LULC classification accuracy: a Random Forest-based feature selection study
Vu Anh Tuan1, Nguyen Manh Hung1,2,, Ngo Duc Anh1,2, Phạm Công Giang1, Nguyen Thi Phuong Hao1
1 Vietnam National Space Center, Vietnam Academy of Science and Technology, Hanoi, Vietnam
2 Graduate University of Science and Technology, Vietnam Academy of Science and Technology, Hanoi, Vietnam
E-mail: nmhung@vnsc.org.vn
Abstract:
Accurate Land Use and Land Cover (LULC) mapping is essential for sustainable urban planning and environmental monitoring. However, spectral similarity between diverse land cover types, such as barren land and impervious surfaces, or urban shadows and water bodies, often leads to significant classification inaccuracies when relying solely on raw optical bands. This study proposes a robust analytical framework to quantify the individual and collective contributions of spectral indices and topographic variables to LULC classification performance using the Random Forest (RF) algorithm. Leveraging the Google Earth Engine (GEE) cloud-computing platform, we synthesized a multi-temporal dataset for Hanoi, Vietnam, integrating Landsat-8 optical imagery with a 30-meter Digital Elevation Model (DEM). The methodology employed a systematic feature selection approach by evaluating three distinct scenarios: (1) raw spectral bands; (2) integration of spectral indices (NDVI, NDWI, NDBI); and (3) a comprehensive fusion of spectral, index, and topographic data (elevation and slope). To quantify variable contribution, the Random Forest model’s internal importance mechanism, based on Mean Decrease Gini and Mean Decrease Impurity, was utilized to rank the significance of each input feature. Our results demonstrate that the synergy of spectral indices and topographic data significantly enhances classification robustness. The optimized RF model achieved a superior Overall Accuracy (OA) of 98.3% and a Kappa coefficient of 0.96. Feature importance analysis revealed that while raw spectral bands provide the foundational data, the NDBI and Elevation variables emerged as the most influential contributors to model precision. Specifically, NDBI was crucial for accurately delineating high-density built-up areas, while the inclusion of DEM and slope effectively mitigated topographic-induced spectral confusion in the city’s peripheral hilly regions. This research underscores that a multi-feature approach, prioritizing variables that minimize spectral overlap, is essential for high-precision LULC mapping in complex, rapidly urbanizing landscapes. These findings provide a scientifically rigorous template for selecting optimal feature combinations, improving the reliability of Earth observation products using accessible optical and terrain data without the need for active sensor integration.
Keywords: LULC classification, Random Forest, Feature importance, Spectral indices, DEM.

“QUANTIFYING THE UNCERTAINTY OF CLIMATE CHANGE IMPACTS ON HYDROLOGICAL REGIME OF THE LO RIVER BASIN, VIETNAM”

Nguyen Ba Thanh Son; GIS-IDEAS Oral Presentation

The Red–Thai Binh River system is the second-largest river basin in Viet Nam and plays a fundamental role in long-term water security, agricultural productivity, and socio-economic development for over 30 million people. This river system comprises three major tributaries, including the Da River, the Thao River and the Lo River. Among them, the Lo and Thao rivers together contribute approximately 50% of the total annual discharge of the Red River. Tran Anh Quan et al. (2023), have indicated that the Red River Delta region is likely to experience a substantial increase in temperature, particularly in extreme heat events, while future precipitation projections remain highly uncertain, despite an overall tendency toward increased rainfall intensity and altered dry–wet cycle characteristics.

This study aims to assess the impacts of future climate on hydrological characteristics of the Lo River, up to the Vu Quang gauging station. With a large contributing catchment area, Vu Quang station serves as a downstream flow control point, providing representative hydrological information for the Lo River basin. To quantify the uncertainty of future hydrological changes, this study employs the Soil and Water Assessment Tool (SWAT) model forced with five GCMs that have relativiely high performance for precipitation across Vietnam (IPSL-CM6A-LR, NorESM2-MM, FIO-ESM-2-0, CESM2-WACCM, and EC-Earth3-Veg) under three climate change scenarios available under CMIP6—SSP2-4.5, SSP3-7.0, and SSP5-8.5—up to the year 2100. Future climate projections from the selected GCMs were bias-corrected and downscaled prior to their integration into the hydrological model to assess projected changes in annual mean streamflow, dry- and wet-season streamflow, and hydrological extremes, represented by flood and low-flow indices, including Flood Frequency Analysis (FFA), Flood Frequency Index (FFI), and the low-flow indicator Q95 derived from cumulative distribution functions (CDF). The results provide a scientific basis for water resources management, reservoir operation, and disaster risk reduction in the downstream areas of the Hong –Thai Binh River system.

Keywords: CMIP6, uncertainty quantification, climate change, SWAT-Model, Hong –Thai Binh River Basin

“RESEARCH ON THE DEVELOPMENT OF SOLUTIONS FOR PROCESSING GPS NETWORK DATA FOR HORIZONTAL DISPLACEMENT MONITORING OF ENGINEERING STRUCTURES”

Tran Thuy Linh; GIS-IDEAS Poster Presentation

This paper presents a study on the development and improvement of methods for processing GPS control network data used for monitoring horizontal displacements of engineering structures in construction areas. Based on an analysis of the characteristics of GPS network establishment and the requirements for processing GPS data in horizontal displacement monitoring, the study focuses on the rational determination of the reference frame of the GPS monitoring network at the observation cycle under investigation, in comparison with the data obtained from the initial monitoring cycle.

The research is initiated from the application of the free-network adjustment method for the adjustment of deformation monitoring networks. Particular attention is paid to network processing under different assumptions regarding the stability of reference points within the network. Numerous studies have indicated that the key to solving this problem lies in the identification of unstable reference points. To address this issue, several existing studies have proposed iterative procedures to gradually detect and eliminate unstable points. However, such approaches lack generality and may lead to biased results. Therefore, this study investigates and proposes a more general processing method by examining all possible combinations of hypotheses concerning unstable reference points that may occur in the network.

Following this approach, the study synthesizes existing scientific publications, selects and applies information technology techniques to propose an algorithm, and simultaneously develops a consistent system of formulas for processing GPS networks used in deformation monitoring of engineering structures. An optimal computational procedure compatible with modern computational conditions is also established, ensuring the most reliable results. The new proposals presented in this paper are validated through experimental data processing for real engineering projects, thereby demonstrating that the proposed processing method is general, rigorous, and convenient for implementation through computational programming. On this basis, it can be affirmed that the proposed processing method and computational procedure are rational and highly effective for processing GPS network data in horizontal displacement monitoring of engineering construction projects.

“SAR-Based Mapping of flood inundation and associated impacts from extreme Typhoon-induced flooding”

Pham Thi Thanh Hoa, Nguyen Tri Anh Khoa; GIS-IDEAS Poster Presentation

Vietnam is highly vulnerable to extreme flooding due to its location in an active cyclonic region and the increasing impacts of climate change. In September 2024, Super Typhoon Yagi caused widespread flooding in the Red River Delta. This event is considered an extreme natural disaster in history, which resulted in severe socio-economic losses and exposed limitations in conventional flood monitoring approaches. A critical challenge during the immediate response phase was the lack of synoptic, near-real-time visual data. Remote sensing offers effective benefits in disaster studies with its image acquisition repeat cycle and wide coverage capabilities. However, passive optical remote sensing satellites, such as Landsat 8/9 and Sentinel-2, rely on solar illumination and cloud-free atmospheric conditions, which severely limit their effectiveness during extreme weather events. This limitation underscores the indispensable role of Synthetic Aperture Radar (SAR) - systems transmit their own signals and record the backscatter, allowing them to penetrate clouds, and to operate day or night. This study conducts a flood assessment in the Red River Delta, using archived Sentinel-1 C-band Synthetic Aperture Radar (SAR) data, which enables reliable flood detection under persistent cloud cover and extreme weather conditions. Google Earth Engine was employed for data processing, including SAR image noise correction and multi-temporal change detection between pre- and post-flood conditions, enabling accurate delineation of flooded areas. The main objectives of this study are to: (1) delineate the spatial extent of inundation after the Yagi flood event; (2) estimate flooded areas and (3) estimate damage to cropland and built-up areas using the European Space Agency (ESA) WorldCover 10 m data. The results demonstrate the effectiveness of SAR-based flood mapping for rapid inundation assessments. This study provides critical spatial evidence to support flood risk zoning and informs flood mitigation and land-use planning strategies in the Red River Delta during the development period.

“Satellite-Based Assessment of Wildfire Risk Using Fuel Condition Indicators : A Case Study of Mae Yom National Park”

punnaporn lueangworakit; GIS-IDEAS Oral Presentation

Wildfire management remains a major challenge in many regions due to limited personnel, constrained budgets, and the need for rapid decision-making. Conventional geoinformatics approaches primarily emphasize mapping and monitoring after fire events, which reduces their effectiveness for proactive control. This study proposes a predictive geodata science framework that integrates neural networks (NN), satellite-based remote sensing, and spatial analysis to support early wildfire risk assessment and faster fire suppression.

The primary objective of this research is to identify fire-prone areas by analyzing surface fuel conditions derived from satellite imagery, with particular attention to vegetation color, dry leaf accumulation, and environmental characteristics influencing ignition potential. These factors are interpreted within the context of the wildfire triangle—fuel availability, heat, and oxygen—to evaluate spatial fire susceptibility. A neural network model is employed to first classify land surface areas based on environmental and vegetation characteristics. Subsequently, color-based spectral features are extracted to estimate relative fuel quantity and dryness levels associated with burnable materials.

Historical wildfire occurrence data were collected and used as reference points for model training and validation. Multispectral satellite images covering the high-risk dry season (January to May) over a five-year period were analyzed. Image processing, spatial visualization, and temporal comparisons were conducted using QGIS, enabling the identification of consistent spectral patterns linked to past fire incidents. The trained neural network model was then applied to separate areas with similar environmental conditions and predict zones with higher ignition probability.

The results indicate that combining neural network-based area classification with color and spectral analysis significantly enhances the ability to detect potential wildfire risk zones. The generated outputs were visualized through a web-based satellite imagery platform, allowing operational officers to rapidly interpret risk levels and prioritize surveillance and fire control actions. This approach is designed to be computationally efficient and practical for real-world implementation, particularly in regions with limited resources.

This study contributes to the advancement of predictive geodata science by demonstrating how geoinformatics, remote sensing, and machine learning can be integrated into an operational wildfire risk prediction system. The proposed framework supports faster response times, improved resource allocation, and a transition from static geospatial analysis toward intelligent, predictive wildfire management.

“Seasonal Variation Assessment of Land Surface Temperature Based on Land Cover Types in Gia Lai Province, Vietnam”

Eaint Chit Phone; GIS-IDEAS Oral Presentation

Seasonal variation in land surface temperature (LST) is a key indicator of land-atmosphere interactions, reflecting the combined influence of climate conditions and land cover characteristics. This study examines seasonal land surface temperature variations across land cover types in a tropical highland region of Vietnam using Landsat 8 OLI/TIRS data acquired during the dry and wet seasons of 2024. The findings reveal pronounced land cover–driven thermal contrasts, demonstrating significant spatial and seasonal land surface temperature variation across Gia Lai Province that is primarily determined by land cover characteristics rather than regional climate alone. The greater spatial temperature variability is observed during the dry season, indicating differing land cover responses to climatic conditions. In both seasons, artificial surfaces consistently exhibit higher LST, while natural surfaces provide more effective thermal regulation. The results indicate that LST exhibits substantial seasonal variability that differs markedly among land cover types. The greatest seasonal LST variations are observed in agricultural and bare soil areas. Agricultural land, which occupies the largest proportion of the study area, shows markedly higher LST during the dry season than during the wet season, primarily due to reduced soil moisture, crop seasons, decreased vegetation cover, and increased surface exposure. In contrast, natural forest areas dominated by dense vegetation consistently exhibit lower and more stable LST during the wet season than in the dry season, reflecting the enhanced cooling effects of a healthy vegetation canopy through shading and evapotranspiration. Meanwhile, land cover types with relatively stable surface properties, such as built-up areas and water bodies, show minimal seasonal thermal variation, highlighting their strong thermal regulation capacity. These findings underscore the dominant role of land cover in controlling surface thermal dynamics and emphasize the importance of preserving natural surfaces, particularly vegetation and water bodies, to mitigate heat stress and enhance climate resilience in tropical highland environments.

“Spatial Heterogeneity and Determinants of the MODIS AOD-PM2.5 Relationship in a Tropical Megacity: A Case Study of Hanoi, Vietnam”

Lê Thị Thu Hằng; GIS-IDEAS Oral Presentation

Satellite‑derived Aerosol Optical Depth (AOD) is essential for estimating ground‑level PM2.5, yet the AOD-PM2.5 relationship exhibits substantial spatial heterogeneity in complex urban environments, especially in tropical monsoon megacities where research remains limited. This study quantifies such heterogeneity and identifies its key driving factors in Hanoi, Vietnam, using six years (2019-2025) of MODIS MAIAC AOD at 1 km resolution, ground‑based PM2.5 from ten monitoring stations, and AERONET data at Nghia Do station for validation (55 collocated pairs). Site‑specific linear regression with 1000‑iteration bootstrap revealed marked variability in correlation strength (R²: 0.338-0.650) and sensitivity (slope: 2.50-11.39×10⁻³ per μg/m³), with intercept ranging from -0.076 to 0.271, reflecting variable background aerosol contributions. Ordinary kriging spatial interpolation indicated stronger linkages in the dense urban core (e.g., DHBKParabol, R²=0.650) and weaker correlations in vegetated areas (e.g., CongVienNhanChinh, R²=0.338). Correlation analysis combined with ERA5 meteorological variables (PBLH, RH) and Landsat‑derived surface indices (NDBI, NDVI) showed that NDBI and NDVI were associated with slope and R² variability, while PBLH and RH exhibited secondary seasonal influences. A distinct urban-rural gradient emerged, with slope and R² decreasing progressively from the high‑density city center to suburban areas, reflecting the diminishing influence of local emission sources. Seasonal analysis confirmed consistently higher slopes in the dry season (mean 6.27 vs. 4.62×10⁻³), consistent with lower PBLH and RH. These findings demonstrate that the AOD-PM2.5 relationship is intrinsically location‑specific within the city, precluding a single city‑wide conversion model. We propose a zoning framework integrating urban density (NDBI), vegetation cover (NDVI), and meteorological factors (PBLH, RH) to improve satellite‑based PM2.5 estimation in tropical urban environments, a contribution with high practical relevance for air quality management and public health protection

“Spatio-Temporal Traffic Accident Risk Analysis in Regional Transit Hubs: A Graph Neural Network Approach with H3 Indexing in Phitsanulok, Thailand”

Sittichai Choosumrong, Rhutairat Hataitara, Kampanart Piyathamrongchai, Natima Udon; GIS-IDEAS Oral Presentation

Traffic accidents remain a critical public safety challenge in Thailand, particularly in regional transit hubs like Phitsanulok province, where high-speed highways intersect with dense urban grids. Traditional geospatial hotspot analysis often fails to capture the underlying network connectivity and the dynamic shift of risk between different time periods. To address these limitations, this study proposes a novel Spatio-Temporal prediction framework integrating the H3 Hierarchical Spatial Index with Graph Neural Networks (GNN) to model the non-linear relationships between road infrastructure and accident severity. The methodology consists of three key phases: (1) Network-Based Spatial Discretization: The province's road network is tessellated into hexagonal cells (H3 Resolution 9). Topological metrics, including road sinuosity, segment density, and intersection complexity, were extracted to serve as dynamic proxies for urban morphology. (2) Temporal Disaggregation: To capture human mobility dynamics, the historical accident dataset was partitioned into Day (06:00-17:59) and Night (18:00-05:59) scenarios. Separate GNN models were trained to identify time-specific risk patterns. (3) Graph Learning: A Graph Convolutional Network (GCN) coupled with a weighted loss function was employed to overcome the zero-inflated nature of accident data, allowing the model to effectively propagate risk information across adjacent network cells. Experimental results reveal a highly significant "Spatio-Temporal Risk Shift" driven by urban functions. During the daytime, accident risks are heavily concentrated in the urban core, specifically around government complexes, educational institutions, and their primary connecting arteries. Conversely, nighttime risk exhibits a distinct spatial shift towards commercial corridors dominated by nightlife and entertainment venues. Notably, major highway intersections in peri-urban areas consistently emerge as persistent high-risk zones regardless of the time of day. The proposed framework not only accurately identifies historical hotspots but also discovers latent risk zones, offering a scalable, data-driven tool for local authorities to optimize proactive traffic safety management.
Keywords: Graph Neural Networks (GNN); H3 Spatial Index; Spatio-Temporal Analysis; Road Topology; Phitsanulok.

“Spatiotemporal Assessment of Seasonal Vegetation Cover Dynamics in Hanoi”

Lê Thị Hồng Nhung; GIS-IDEAS Oral Presentation

In the context of escalating climate change and rapid urbanization, monitoring vegetation cover dynamics has emerged as a critical imperative for maintaining ecological equilibrium, mitigating the urban heat island effect, preserving biodiversity, and enhancing the overall environmental quality for the resident community. This reseach focuses on analyzing changes in vegetation cover using the NDVI index with Sentinel-2 seasonal imagery in the Hanoi area. The cloud computing platform (Google Earth Engine) is used to filter cloud noise through the QA60 band, normalize surface reflectivity and calculate the NDVI vegetation index which varies with the seasons. The research provided a system of seasonal vegetation cover maps with high spatial resolution (10m) and showed a clear differentiation between areas of protected forests with stable indices, while agricultural production areas and inner-city green spaces showed significant variation due to the direct impact of specific weather conditions. The results demonstrate a pronounced spatial heterogeneity governed by complex land-use configurations. Protected forest regions in the West, most notably the Ba Vi mountain range, maintain peak and consistent NDVI values, signifying the robust photosynthetic health of natural silvicultural ecosystems. Conversely, agricultural zones within the Dong Anh and Gia Lam districts exhibit high-amplitude spectral variability, primarily attributed to winter crop phenology and distinct harvesting cycles. Notably, the urban core records the minimum NDVI thresholds due to intensive sclerotization and a high prevalence of impervious surfaces, where vegetative presence is restricted to fragmented 'green pockets' within public parks and lacustrine systems. General assessment indicates that the integration of Sentinel-2 data enables not only the detection of spatial micro-variations but also the identification of zones undergoing severe ecological degradation under anthropogenic urban pressure. This continuous monitoring framework yields critical quantitative datasets, empowering policymakers to optimize green infrastructure allocation and formulate resilient climate adaptation strategies, ultimately advancing the capital’s trajectory toward sustainable and green urban development.

“STUDY OF GRAVITY VARIATIONS OVER VIETNAM AND SURROUNDING REGIONS USING TIME-VARIABLE GRAVITY MODELS”

Ngo Thi Men Thuong; GIS-IDEAS Poster Presentation

Gravity variations directly reflect internal dynamic processes of the Earth as well as the redistribution and transport of mass on and near the surface, and therefore play an important role in studies of geodynamics, geodesy, climate change, and resource management. In the field of geodesy, the determination of temporal gravity variations is essential for maintaining national reference systems, updating Geoid models, and correcting errors in high-precision measurements. In Vietnam, research on gravity variations remains limited due to resource constraints for data acquisition and the assessment of time-dependent gravity signals. The advent of satellite missions such as GRACE and GRACE-FO has opened a new era for observing gravity variations from space with high temporal resolution.
This paper presents the results of a study and analysis of gravity field variations over Vietnam and the East Sea using time-variable gravity field models. The study employs time-variable gravity models derived from GRACE/GRACE-FO satellite gravity data provided by the International Centre for Global Earth Models (ICGEM). These Earth’s time-variable gravity field models are represented as a sequence of monthly spherical harmonic coefficients. The monthly gravity solutions enable the identification of short-term fluctuations as well as the analysis of long-term trends in gravity field variations.
The methodology focuses on the processing of multi-temporal spherical harmonic coefficients, the computation of gravity field differences, and trend analysis using geospatial data analysis techniques. Model-derived gravity variations are compared with ground-based gravity observations at selected sites, based on repeated measurements conducted during the period 2016–2020. The correlation analysis between international satellite-based data and locally acquired ground gravity measurements plays an important role in proposing strategies for integrating and exploiting multiple data sources, thereby improving the effectiveness of gravity variation studies under the current resource conditions in Vietnam.
The results indicate that the gravity field over Vietnam and surrounding regions exhibits clear temporal variations, characterized by pronounced long-term trends and distinct spatial contrasts between continental and marine areas. The findings provide valuable data and a methodological foundation for future studies on the prediction of mass redistribution and lithospheric dynamics, contributing to the modernization of the spatial data infrastructure and national geodetic reference systems.

“Transformation of Charted Depth from Bathymetric Depth Based on Mean and Lowest Sea Surface Models: A Case Study of the Quang Ngai Coastal Area, Vietnam”

Đặng Thu Hằng; GIS-IDEAS Poster Presentation

Abstract: Charted depth is a fundamental element in nautical chart production and maritime safety, serving as the primary reference for safe navigation and marine spatial management. In practical marine surveying and hydrographic data management, bathymetric depth data are often acquired and stored with respect to different vertical reference surfaces, such as mean sea level, instantaneous sea surface, or geodetic ellipsoids. In contrast, charted depth must be referred to a prescribed tidal datum (chart datum), typically associated with the lowest sea level expected under astronomical conditions. This inconsistency between vertical reference systems poses significant challenges for data integration, accuracy assurance, and the modernization of digital nautical charts.
This paper presents a method for transforming bathymetric depth into charted depth based on mean sea surface and lowest sea surface models. The proposed approach is founded on the determination of the vertical offset between these two sea surface representations, which reflects the regional relationship between mean sea level and chart datum. A regional depth correction factor is derived from this offset and applied to convert bathymetric depth into charted depth in a consistent and systematic manner. The method is designed to be practical and adaptable, particularly in coastal regions where long-term tidal observations and comprehensive hydrodynamic models are limited.
The proposed methodology is validated using an experimental dataset from the coastal waters of Quang Ngai Province, central Vietnam. The dataset includes in situ bathymetric measurements and interpolated mean and lowest sea surface models representative of the study area. The results indicate that the derived depth correction factor exhibits relatively stable spatial behavior within the investigated region. After applying the correction, the resulting charted depths show good agreement with reference values and satisfy the accuracy requirements for medium-scale nautical charting.
The findings demonstrate that the proposed method provides a feasible and reliable solution for charted depth transformation in data-limited coastal environments. From a marine geodetic perspective, the method can be interpreted as a regional vertical datum transformation based on sea surface modeling. The approach has potential applications in marine surveying, hydrographic chart production, and the integration of multi-source bathymetric data, contributing to the ongoing transition toward standardized and modernized hydrographic information systems in Vietnam and similar coastal regions worldwide.

Keywords: Charted depth; Bathymetric depth; Mean sea surface model; Lowest sea surface model.

“Variations of Urban Ecosystem Service Value in Ho Chi Minh city due to Land Use and Land Cover Changes”

HienLe; GIS-IDEAS Poster Presentation

Urban ecosystem services (ES) deliver essential benefits that underpin human well-being, including provisioning services like food and water, regulating services such as climate moderation and air purification, cultural services fostering recreation and social cohesion, and supporting services like biodiversity maintenance and nutrient cycling. These services are particularly vital in rapidly urbanizing megacities facing compounded pressures from climate change, environmental degradation, and population growth. Despite their recognized importance for enhancing quality of life and resilience in urban environments, the monetary valuation of ecosystem service values (ESVs) remains largely overlooked in urban planning frameworks and policy development, leading to suboptimal land-use decisions that prioritize short-term economic gains over long-term sustainability.

This study addresses this critical gap by quantifying the spatiotemporal dynamics of land use and land cover (LULC) changes in Ho Chi Minh City (HCMC)—Vietnam's largest metropolis and economic hub—from 2017 to 2024, and rigorously assessing their cascading impacts on total ESV. Leveraging high-resolution Sentinel-2 satellite imagery at 10m spatial resolution, we classified LULC into seven primary categories: built-up areas, vegetation (forests and green spaces), cropland, water bodies, bare land, wetlands, and grasslands. The classification achieved an overall accuracy exceeding 89%, enabling precise mapping of urban expansion patterns.

Key findings reveal accelerated urbanization, with built-up areas surging by 42.75% over the seven-year period—an average annual growth rate of 4.43%. This expansion primarily occurred through the conversion of ecologically valuable lands, including a 23.4% loss in cropland, 15.2% reduction in water bodies, and 18.7% decline in vegetation cover. Consequently, total ESV diminished from 4,309.569 million 2020 USD in 2017 to 4,080.13 million 2020 USD in 2024, representing a net loss of 229.44 million USD (5.33%). Vegetation emerged as the dominant contributor to ESV (48.2% of total value), followed by water bodies (26.5%) and cropland (17.8%), underscoring their disproportionate regulatory and provisioning roles.

To elucidate sensitivity, we conducted elasticity analysis, demonstrating that a 1% shift in LULC composition induces approximately a 0.33% change in total ESV, highlighting the system's responsiveness to land conversions. Directional sprawl analysis further identified preferential eastward and southwestern expansion from the central business district, with Shannon's entropy metrics confirming compact yet dispersive growth patterns.

These empirical insights provide a robust scientific foundation for evidence-based policymaking. By integrating ESV assessments into HCMC's master planning (2021-2030, vision to 2050), authorities can prioritize green infrastructure, urban forests, and wetland preservation to mitigate ESV losses. The methodology offers a scalable template for other rapidly urbanizing Asian cities grappling with similar trade-offs between development imperatives and ecological sustainability. Ultimately, this study advocates for ESV mainstreaming in urban governance to foster resilient, livable cities amid escalating anthropogenic pressures.

“Vegetation Classification Using Time-Series Sentinel-1 Data Integrated with Sentinel-2 Imagery in a Mountainous Area of Northern Vietnam: A Case Study of Chieng Ken Commune, Lao Cai Province”

Huyen Oanh; GIS-IDEAS Oral Presentation

Vegetation classification plays a critical role in identifying and analyzing the spatial distribution, structure, and condition of land cover, thereby providing a scientific basis for ecological studies, assessing the status and potential of vegetation resources, and monitoring their environmental impacts. This study was conducted on the Google Earth Engine (GEE) platform to classify five major vegetation classes in Chieng Ken Commune, Lao Cai Province, including natural forest, planted forest, cassava, maize, and rice paddies, using the Random Forest algorithm based primarily on time-series Sentinel-1 Synthetic Aperture Radar (SAR) data, integrated with Sentinel-2 optical imagery and topographic variables. Specifically, rice paddies were identified based on temporal variations in VH (Vertical transmit-horizontal receive) backscatter derived from Sentinel-1 data. The remaining vegetation classes, including major agricultural crops (cassava and maize) and forest types (natural forest and planted forest), were classified using a combination of the Radar Vegetation Index (RVI), SAR-derived texture features, canopy density indicators extracted from Sentinel-2 imagery, and topographic factors, namely elevation and slope. The results indicate that natural forest is the dominant vegetation class, covering approximately 12,285 ha (64.54% of the total vegetated area), followed by planted forest with 2,976 ha (15.63%). Agricultural crops occupy smaller proportions, with cassava, maize, and rice accounting for 12.49%, 6.31%, and 1.03% of the total vegetated area, respectively. The classification achieved high accuracy, with an overall accuracy of 94.30% and a Kappa coefficient of 0.913. Overall, the findings provide a reliable scientific basis for planning and decision-making in forest and agricultural resource management, and demonstrate the strong potential of integrating Sentinel-1 time-series SAR and Sentinel-2 optical data for sustainable vegetation classification and monitoring in cloud-prone mountainous regions.

“Wildfire Susceptibility Mapping in the Context of Climate Change Using an MCDA-GIS Integrated Approach: A Case Study of Pu Nhung Commune”

Cao Dinh Son; GIS-IDEAS Oral Presentation

Mountainous regions in Northwest Vietnam are highly vulnerable to forest fires due to complex terrain, steep slopes, prolonged dry seasons, and increasing climate variability. In recent years, rising temperatures and irregular rainfall patterns have further intensified wildfire risks, particularly in Pú Nhung commune, Điện Biên province. This study aims to assess spatial wildfire susceptibility using an integrated GIS and Multi-Criteria Decision Analysis (MCDA) approach based on the Analytic Hierarchy Process (AHP).A set of key influencing factors, including topography (slope, elevation), land cover, climate conditions (temperature, rainfall), and human-related variables (distance to roads and settlements), were selected and standardized within a GIS environment. The relative importance of each factor was determined through AHP pairwise comparisons, ensuring consistency (CR < 0.1). These weighted factors were then combined using a weighted overlay method to generate a wildfire susceptibility map.The results classify the study area into five susceptibility levels, highlighting zones with high and very high fire risk concentrated in areas with dense vegetation, steep terrain, and strong human interaction. The findings provide a scientific basis for forest management, fire prevention planning, and resource allocation. This study also demonstrates the applicability of GIS-AHP as a practical and transferable tool for wildfire risk assessment in similar mountainous regions.