FOSS4G-Asia 2024

Large Language Models (LLMs) have revolutionized numerous aspects of modern life, demonstrating remarkable capabilities in language processing, code generation, and knowledge synthesis. Their potential extends to supporting the achievement of various Sustainable Development Goals (SDGs), offering innovative approaches to tackling complex global challenges. One promising application area is the mapping and monitoring phenomena measurable through Earth Observation (EO) data. It's estimated that around 40 of the 169 SDG targets and 30 of the 232 SDG indicators could benefit from the insights provided by the EO data analysis. The use of artificial intelligence (AI) for EO data analysis can further improve the number of SDG indicators that can be monitored with a higher accuracy and frequency.
In this context, research is underway to develop multimodal LLMs capable of directly processing EO data. However, these models are often computationally expensive to train, develop, and maintain, making them less feasible for low-capacity, high-risk countries that urgently need technological solutions for disaster mitigation. To address this challenge, we introduce SATGPT (accessible at satgpt.net), an innovative solution that leverages the current capabilities of LLMs and integrates them with cloud computing platforms and EO data. SATGPT represents a fully functional, innovative spatial decision support system designed for rapid deployment, particularly in resource-limited contexts.
This talk presents an instance of SATGPT configured for flood mapping, as an example. It simplifies the process with a user-friendly interface requiring only a prompt specifying flood duration and location. SATGPT leverages LLMs to generate GEE code dynamically, access historical databases, or perform unsupervised classification to detect flooded areas. This innovative integration of LLMs with GEE enhances the speed, accessibility, and real-time capabilities of flood mapping, making it more accessible to non-specialists and supporting resilient disaster management practices.
Furthermore, to build the capacity to use these technologies effectively, the talk discusses the development of an online, free, and self-paced course titled "Introduction to Geospatial Data Analysis with ChatGPT and Google Earth Engine." This course introduces participants to the fundamentals of ChatGPT and the Earth Engine Code Editor platform, empowering them to process and interpret geospatial data effectively outside the SATGPT. The innovative aspect of the course is developing the Geo-prompt engineering (GPE) concept, which focuses on using spatial, temporal, and satellite sensor- specific information in the prompt engineering process. The course aims to foster broader adoption of SATGPT and similar tools, equipping users with the knowledge and skills needed to leverage advanced technologies in disaster management. This talk is structured to provide a comprehensive overview of SATGPT and its contribution to enhancing flood mapping and disaster management in the Asia-Pacific region.

The global climate crisis poses significant challenges, particularly for urban areas. Nebula, a specialist in resilient and regenerative city development, advocates for a paradigm shift towards creating "happy future cities." Unlike traditional approaches that focus on rebuilding from scratch, Nebula emphasizes fostering an “environmentally enhancing, restorative relationship” with nature. This explores how cities can play a crucial role in addressing environmental challenges through strategic urban planning.

We pioneer a transformative approach to urban development, creating future cities that are not just smart, but regenerative, resilient, and centered on the well-being of both people and the planet.

The research assessment is traditionally based on the evaluation of criteria such as the number of peer-reviewed publications, impact factor, and number or amount of grant funding. Unfortunately, this approach has been proved to deeply influenced the way of conducting research that focus on quantity rather then quality and not . To maximize the impact of research as a practical mean to address societal challenges, a new approach, named Open Science, has been endorsed worldwide by major funding agencies over the last decade as the new research pathway. Quality and impact, collaboration and sharing, diversity and equity, transparency and efficiency has become the new paradigms to be pursued.
To foster the adoption of the Open Science funding agencies are acting on two fronts: on one hand, by influencing policies and requiring the adoption of open science practices as a condition for funding access (Open Access, Open Data and Citizen Science), and on the other hand, by focusing on incentives and exploring new methods for evaluating scientific results.
It is clear that in the near future, the current “publish or perish” aphorism is shifting towards “open or perish” to describe the required work to succeed in an academic career. But how should a modern researcher act and comply with this new paradigm? It essential for her/him to understand the best practices that guarantee the recognition of her/his achievements by connecting the researcher, the publications, the software and the data. In this talk, an introduction of these best practices is addressed with the aim of sustain the Open Science adoption with particular reference to Open Software. Finally, new possible approaches envisioned for the evaluation of project proposals, career advancements and institutions assessment are presented and discussed.

My presentation will explore the landscape of open data in Thailand, focusing on laws, regulations, policies, and the fundamental right to access information. It will assess how open data initiatives can foster transparency, innovation, and public engagement, while addressing challenges and proposing solutions. Special emphasis will be placed on how open data can play a crucial role in reducing inequality by empowering citizens with greater access to information, enabling more equitable participation in decision-making processes, and driving inclusive social development.

Over recent years, Mysore, a district in Karnataka, India, has seen remarkable urban growth and infrastructural development, transforming its landscape significantly. This study examines how this urban expansion influences property values, using data from 2014 and 2024 to forecast property values for 2034 with a Random Forest regression model. We focus on 110 key locations, looking at factors such as closeness to the central business district, railway station, bus stand, and local amenities like schools and hospitals. By finding the strongest correlations between these elements, we establish a relationship between property values and these factors to predict future values.

Our findings highlight Mysore's vibrant economic growth and its potential for sustained progress. These insights are crucial for the real estate market, providing valuable information to make informed decisions about future property values amid ongoing urban development. By analyzing how urban growth impacts property values through sophisticated statistical models, this study sheds light on how infrastructural improvements and strategic locations drive real estate trends. The expected significant rise in property values by 2034 underscores Mysore's economic dynamism and its appeal as an emerging urban hub.

We conducted a thorough analysis of various factors affecting property values, focusing on proximity to essential services and transportation hubs. These elements significantly influence property desirability and accessibility. Our use of the Random Forest regression model enables accurate predictions of future property values by understanding complex relationships between these variables. The strong correlation between guideline values and market values provides a reliable basis for predicting future real estate trends. This correlation is essential for stakeholders, including developers, investors, and policymakers, as it supports strategic decision-making based on market projections.

The expected significant rise in property values indicates that Mysore is poised for considerable growth, driven by strategic developments and improved infrastructure. By understanding these trends, stakeholders can make informed decisions to capitalize on Mysore’s ongoing urban expansion, ensuring that investments and development strategies align with the city's projected economic vitality and growth potential. Our analysis highlights that proximity to the central business district, bus stops, and railway stations are key determinants of property values, greatly influencing market prices. We project a significant increase in property values, estimating a 118% rise by 2034.

To visualize these future values, we employ Voronoi polygons, which offer a clear spatial representation of the predicted property value distribution. This approach provides stakeholders, including developers, investors, and policymakers, with valuable insights into future market trends. By understanding the impact of these location factors, they can make informed decisions regarding investments and development strategies. The anticipated rise in property values underscores the ongoing urban development and economic growth in Mysore, highlighting its potential as a thriving urban center.

In summary, this study provides an in-depth analysis of the relationship between urban growth and property values in Mysore. By employing advanced regression models and detailed location-based data, we have developed a robust forecast for property values in 2034. Our findings indicate a significant projected increase in property values, highlighting Mysore's continuous development and potential for future growth. These insights are essential for the real estate market, offering valuable guidance for future investments and development strategies in Mysore. The study emphasizes the impact of key factors such as proximity to the central business district, bus stops, and railway stations on property values. By understanding these dynamics, stakeholders, including developers, investors, and policymakers, can make informed decisions to navigate the evolving real estate landscape. The anticipated rise in property values underscores Mysore’s economic vitality and its promise as a thriving urban center, driven by strategic infrastructure development.

Keywords: Mysore, Urban growth, Property values, Regression model, Infrastructure, Stakeholders

In this session, I will introduce mago3DTiler (https://github.com/Gaia3D/mago-3d-tiler), an open-source OGC 3D Tiles creator that has gained global popularity thanks to its robust features, high performance, and user-friendly interface. Initially unveiled at FOSS4G-Asia 2023 in Seoul, mago3DTiler supports over ten different 3D data formats, including 3DS, OBJ, FBX, glTF, Collada DAE, BIM (IFC), LAS, LAZ, and SHP. One of its standout features is on-the-fly Coordinate Reference System (CRS) conversion during the 3D Tiles creation process. Additionally, it allows users to convert 2D data with height attributes into extruded 3D Tiles.

During this session, I will also demonstrate how to create a digital twin using mago3DTiler in just a few minutes. This tool makes complex geospatial tasks more manageable, especially for users looking to integrate diverse data formats seamlessly into 3D projects.

The Meteorological Assimilation from Galileo and Drones for Agriculture (MAGDA) project, funded by EUSPA in the framework of the Horizon Europe program, aims to develop a comprehensive toolchain for atmosphere monitoring, weather forecasting, and advisory services related to severe weather, irrigation, and crop monitoring. By integrating GNSS, Copernicus Sentinel, Meteodrone, ground-based weather radar, and in-situ weather and soil observations into open source weather and hydrological models, MAGDA seeks to provide valuable information to agricultural operators. Measured data, model results, and warnings/advisories are delivered to farmers through a dedicated dashboard or by interfacing with existing Farm Management Systems. The technical and methodological components developed within MAGDA will form the basis for services supporting agricultural operations.
The project is based on the concept that continuous monitoring, combined with advanced prediction models, is essential for effective resource management. As extreme weather events, such as droughts and heatwaves, become more frequent due to climate change, farmers need to leverage technology to mitigate disasters, conserve resources, and enhance productivity. A system that can automatically collect and process measurements of key parameters significantly reduces economic losses. When this data is presented clearly and usefully to end users, it can significantly enhance agricultural efficiency.
MAGDA unites seven partners from seven European countries (Austria, France, Italy, Romania, Spain, The Netherlands, and Switzerland) and is inherently interdisciplinary, drawing on expertise from various sectors to develop a system tailored to agricultural meteorological and hydrological forecasts.
The selected demonstrator areas in Italy (Cuneo), France (Burgundy), and Romania (Braila) target different crops/cultures and allow for the gathering of different user needs and feedback through direct interactions with farmers. Deployment includes nine low-cost, dual-frequency GNSS stations, along with fifteen low-cost in-situ sensor stations, and three Meteobases to fly meteodrones.
Severe weather cases were identified to test the open source WRF meteorological model's performance: in Italy, the focus was on rainfall events, while in France and Romania, hail events were prioritized. Water balance simulations were conducted to support an operational irrigation advisory service, using the open source SPHY hydrological model across the pilot areas in France, Italy, and Romania.
All data used in the MAGDA project are open for research applications, and the GNSS processing software utilized in this project leverages the goGPS open source software. The MAGDA dashboard for result visualization uses Leaflet as a web mapping tool and OpenStreetMap data as a background layer. The results presented here are derived from the currently ongoing MAGDA demonstrators, showcasing the project's impact on weather forecasting and water management for agricultural operations.

This talk presents a research case of an open source implementation of a task management system based on 3D spatial information web service to efficiently conduct and manage tasks in the field of environmental impact assessment (EIA), which combines very diverse specialties.

Vector tiles are changing the way we create maps. Client-side rendering offers endless possibilities to the cartographer and has introduced new map design tools and techniques. Let’s explore an innovative approach to modern cartography based on simplicity and a comprehensive vector tiles schema. Take a visual tour of vector tiles cartography, and learn how to possibly adapt the map design for an asian audience.

Global remotely sensed evapotranspiration (RS-ET) products are increasingly pivotal in enhancing the accuracy and scope of hydrological modeling, particularly in regions where traditional ground-based streamflow data are sparse or non-existent. These products play a pivotal role in understanding the dynamics of the climate-soil-vegetation system, where evapotranspiration constitutes a substantial portion of water loss following precipitation events. Their extensive spatial coverage and accessibility have significantly expanded the capability to predict hydrological dynamics in ungauged basins, offering insights that were previously inaccessible through in-situ observations alone.

Despite their benefits, RS-ET products are tempered by inherent uncertainties, primarily stemming from biases that vary across datasets and geographical regions. These biases manifest as either overestimation or underestimation compared to ground truth measurements, posing challenges for the accurate calibration of hydrological models. Traditional approaches in hydrological modeling commonly utilize absolute ET values directly derived from RS-ET products for model calibration, without accounting for potential biases. However, the reliability of such direct calibrations is contingent upon the quality and accuracy of the RS-ET data, which remains uncertain in many cases.

To address these challenges, this study shifts the focus from absolute ET values to utilizing spatial structural information embedded within RS-ET data, particularly emphasizing spatial autocorrelation, which refers to the tendency of ET values at nearby locations to exhibit similarities. Employing the local Moran's I index, a spatially weighted autocorrelation statistic that is insensitive to biases, we capture the spatial structure of ET data across sub-basins. Additionally, a composite Kling-Gupta Efficiency (KGE) metric, integrating absolute ET values and spatial autocorrelation information in a weighted manner, is employed for calibrating hydrological models.

Three calibration schemes are thus designed to analyze the effectiveness of spatial autocorrelation in hydrological modeling: one focusing solely on absolute ET values, another solely on spatial autocorrelation, and a combined approach. Testing these schemes for hydrological modeling with four RS-ET products in the Meichuan basin—MOD16, GLASS, and SSEBop with large biases, and PMLV2 with minimal bias—the study demonstrates varying effectiveness across these schemes.

For RS-ET products with substantial biases, hydrological modeling using spatial autocorrelation proved to be the optimal solution. It achieved a higher KGE and lower Percent Bias (PBIAS) on simulated streamflow compared to using solely the absolute ET value or the combined approach. Conversely, for RS-ET products with minimal biases, hydrological models calibrated using both the combined approach were considered the preferred solutions. This approach can result in a high KGE, similar to that obtained from spatial autocorrelation information alone, while maintaining a reasonable PBIAS. Therefore, we recommend calibrating hydrological models using both absolute ET values and spatial autocorrelation information in regions where ground ET observations are available. This approach enhances the robustness and reliability of hydrological predictions, mitigating the influence of biases inherent in RS-ET products.

In contrast, in scenarios where the quality of RS-ET products is unknown, we suggest calibrating using only spatial autocorrelation information, thereby circumventing potential biases and improving model accuracy under such circumstances. Moreover, methodologically, the study contributes by demonstrating the efficacy of the local Moran's I index in capturing the spatial structure of ET data within hydrological sub-basins. This geostatistical measure not only quantifies spatial autocorrelation but also identifies clusters and patterns of ET values, thereby enriching our understanding of spatial variability in hydrological processes.

Furthermore, the comprehensive analysis of the composite KGE index underscores the significant contribution of spatial autocorrelation information to hydrological modeling, surpassing the influence of absolute ET values in enhancing model performance. In conclusion, the spatial autocorrelation-based approach presented in this study represents a significant advancement in the application of global RS-ET products for hydrological modeling. By leveraging spatial structural information and mitigating biases inherent in RS-ET data, this approach not only improves the accuracy of hydrological predictions but also enhances the practical utility of RS-ET products in diverse hydrological contexts. Future research directions may explore additional spatial statistical techniques and incorporate a broader array of RS-ET datasets to further refine and validate these findings across different geographical settings and hydrological conditions.

Mapping land suitability is a critical approach for identifying appropriate land use for site selection and land-use planning. However, climate change exacerbates water shortages and droughts, significantly affecting land suitability and resulting in decreased crop yields, especially for sugarcane. While land suitability is typically evaluated based on multiple criteria such as soil properties, topography, climate, and socioeconomic factors, it is essential to incorporate drought conditions into land suitability mapping to mitigate climate change influences on crop yields. Therefore, this study aimed to map sugarcane land suitability using fuzzy AHP and multi-criteria evaluation approaches in the Northeast region of Thailand.

The study selected six significant criteria for sugarcane land suitability mapping: the ETDI as an agricultural drought index, slope, soil texture, distance from the river, distance from the road, and distance from the sugar mill. The ETDI was assessed by calculating the difference between spatial Potential Evapotranspiration (PET) and actual Evapotranspiration (AET). Spatial PET was analyzed using a PET estimation model based on integrated GNSS-derived Precipitable Water Vapor, processed with goGPS open-source software, along with the MODIS land surface temperature product. Concurrently, the spatial AET was derived from the SEBAL model, utilizing GRASS GIS software.

Subsequently, land suitability for sugarcane cultivation was evaluated by integrating fuzzy AHP and multi-criteria evaluation approaches. The results indicated that two primary factors affected sugarcane cultivation: the ETDI and distance from the river. The ETDI was the most significant factor, with an average weight of 0.66, while the distance from the river had an average weight of 0.34. Other factors, including slope, soil texture, distance from the road, and distance from the sugar mill, did not influence land suitability. The spatial distribution of these factors was consistent throughout the study area.

Suitable areas for sugarcane were predominantly found in the moderately suitable class (S2; 49.6%), followed by marginally suitable (S3; 36.0%) and highly suitable (S1; 11.2%). Actual sugarcane cultivation areas were mainly in the S3 class (49.0%), followed by S2 (43.2%) and S1 (6.7%). S3 class areas were concentrated in Wang Sam Mo district, Udon Thani province (129 km²), with a sugarcane yield of approximately 60.6 tons/ha. S2 class areas, primarily in Phu Khiao district, Chaiyaphum province (178 km²), yielded about 62.5 tons/ha, while S1 class areas in Phimai district, Nakhon Ratchasima province (30 km²), achieved a higher yield of 63.6 tons/ha.

S2 class areas could potentially be enhanced through irrigation systems and small ponds to mitigate drought risks. Limiting the distance from the river to within 2 km could increase sugarcane yields and promote areas to the S1 class, expanding S1 areas by 2.7 times and raising yields by approximately 1.1 tons/ha (1.8% of S2 yield). Areas classified as S1 exhibit significant potential for sugarcane cultivation expansion due to their underutilization. A total area of 6,519 km² within the S1 class was analyzed for suitability. Nakhon Ratchasima province has the greatest potential (2,272 km², 35%), followed by Khon Kaen (725 km², 11%), Chaiyaphum (592 km², 9%), Udon Thani (519 km², 8%), and Surin (441 km², 7%).

Encouraging a shift from currently cultivated crops (rice, corn, and cassava) to sugarcane in these potential areas is essential for optimal resource utilization. However, farmers often continue rice cultivation due to its traditional significance and shorter growth period, providing quicker income. Government policies should support participatory knowledge transfer on sugarcane cultivation, ensure price guarantees, and facilitate access to credit. Additionally, the high price of sugarcane could incentivize farmers to expand sugarcane cultivation to meet increasing domestic and export demand. Further research on a larger scale, covering the entire country, is necessary to enhance the accuracy of land suitability maps in addressing challenges posed by global climate change.

The range of AI applications in the geospatial information field is diverse, object detection, area extraction, change detection, and super-resolution. In our industry, we collectively refer to these technologies as GeoAI.

However, these technologies remain primarily confined to specialized groups, making it difficult to consider them as universal technologies ready for everyday use.

In light of this, we sought to explore areas that could be easily accessible to the general public. Consequently, we developed 'magoGPT', a application enabling users to manipulate maps and interact with 3D objects using natural language in a digital twin environment.

Based on a 3D FOSS architecture composed of CesiumJS, it uses 3DTiles buildings, Terrain from high-resolution DEMs, and multiple layers visualisation. It is also related to artificial intelligence techniques such as Large Language Models (LLM), Speech-to-Text (STT), and Natural Language Processing (NLP).

In this presentation, we'd like to introduce magoGPT, the technology behind it, and the development process.

In October 2018, the art world witnessed an unprecedented event when Banksy’s “Girl with Balloon” partially shredded itself immediately after being auctioned at Sotheby’s. This bold act challenged traditional perceptions of art, value, and the role of the artist. Drawing inspiration from this event, I developed a unique QGIS plugin that shreds layers.

This plugin shreds the input data into shredded pieces and is implemented with a simple pyQGIS API.
It can be shedded with either vector or raster layers, and the fineness of the shredding can be set.

This plugin may seem useless at first glance, but it can be used when you want to mess up your data, when you have created data that you cannot show to others, when you just want to relieve stress, or when you want to feel like Banksy.

Most importantly, while shredding physical documents produces waste, shredding data creates no physical waste, making it a truly sustainable practice.

Drought is a recurring issue in South Asia (SA) caused by extreme climate events, posing ongoing challenges for food management, sustainable agricultural practices, and livelihoods, especially in frequently affected areas. Earth Observation (EO) data provides valuable information for long-term drought monitoring across wide regions. However, there remains a need to map and monitor spatial drought events over large regions and extended periods, particularly in the Chi River Basin. In this region, droughts have increasingly occurred with high frequency, leading to low water-holding capacity and adversely affecting agricultural production and productivity.

In this context, this study aimed: i) to identify spatial drought from 2001 to 2020 over the Chi River Basin, Thailand using MODIS image time series via Google Earth Engine (GEE); ii) analyse the correlation between Temperature Vegetation Dryness Index (TVDI), Standardized Precipitation Index (SPI), and Streamflow Drought Index (SDI); and compare severe drought areas to land use maps provided by the Land Development Department (LDD).

In this study, we collected MODIS data using the MYD09Q1 (250m spatial resolution) and MOD11A1 products (1000m pixel size) across the h27v07 and h28v07 tiles. Image pre-processing was implemented, consisting of image resampling, data compositing, and image mosaicking through the GEE platform. Subsequently, the TVDI index was generated for both dry and wet seasons to map and monitor the spatial distribution of drought events over 21 years. The spatial drought based on TVDI and meteorological datasets (SPI and SDI) was determined to identify their relationship. Additionally, this study compared the spatiotemporal distribution of drought to land use groups.

Historical droughts were most frequent during the dry seasons of 2005 (82%), 2013 (80%), and 2004 (78%), and appeared in the wet seasons of 2019 (41%), 2017 (41%), and 2009 (38%). The TVDI drought map had a slightly low coefficient of determination (R²) for SPI and SDI, ranging from 0.12 to 0.22. However, these findings showed similar trends of drought across all study years, with drought events predominantly occurring in the central and northeast parts of the region. In comparison, the spatial drought map in 2021 showed severe droughts, mostly impacting cassava and rice fields during the dry season and urban areas during the wet season. Our proposed workflow is reliable and robust, providing spatial drought areas with confidence in the accuracy and validity of our results.

This study produced spatial drought maps using MODIS image time series datasets. The mapped results were well smoothed and effectively distributed drought areas across large regions. The highly severe spatial droughts in 2005 align with Thailand's extreme drought due to the El Niño event, demonstrating high severity compared to other years. This confirms that the TVDI index provides excellent and efficient map results for mapping the spatial distribution of droughts in cloudy regions and complex landscapes of the Chi River Basin. The proposed workflow can generate drought maps in cloudy regions and complex landscapes over large or national regions, particularly in zones like Thailand. Our findings can be used to manage future droughts and serve as a significant tool for drought mitigation planning and management, as well as for warning systems, providing an integrated model under climate change conditions.

Keywords: Drought; earth observation; Temperature Vegetation Dryness Index (TVDI); Land use; Google Earth Engine

This study aims to analyze the relationship between PM2.5 concentrations, derived from Aerosol Optical Depth (AOD), and solar power generation a solar farm owned by A. Co., Ltd.(alias), in Samut Prakan Province, Thailand, This research utilizes PM2.5 data from pollution monitoring stations of the Pollution Control Department, AOD data from the MCD19A2.061 product, and solar power generation data from the Electricity Generating Authority of Thailand in 2022. The results indicate in summer season a negative correlation between PM2.5 concentrations and solar power generation R² = -0.7, meaning that as PM2.5 levels increase, solar power generation decreases. A regression equation used for power prediction achieved an accuracy of R² = 0.97. In contrast, a positive correlation is observed during the winter season R² = 0.6, indicating that as PM2.5 levels increase, solar power generation increases, with a prediction accuracy of R² = 0.93. No significant correlation is found during the rainy season, which may be due to other influencing factors. Predicting solar power generation in other areas should consider the different physical factors unique to each location.
Keywords: PM2.5, Aerosol Optical Depth, Solar Power, Solar Farm

In recent years, the availability of geospatial data has significantly increased, with aerial photographs and Digital Elevation Models (DEMs) becoming widely accessible as open data. Additionally, the acquisition of high-resolution image data through drones has become more feasible and commonplace. However, extracting meaningful information from these vast datasets remains a labor-intensive process, often requiring significant time and resources.

While various deep learning techniques have been employed to address this challenge, they typically demand extensive effort in collecting and preparing training data. In light of these constraints, the Segment Anything Model (SAM) has emerged as a promising solution. SAM, a recent development in the field of fundamental models, offers the advantage of zero-shot classification without the need for specific training. Moreover, its Apache-2.0 license ensures accessibility for a wide range of applications.

This presentation aims to demonstrate the potential of SAM in the realm of geospatial information processing. We will showcase practical applications of SAM in analyzing geospatial data, with a focus on two critical areas: landslide detection and forest canopy mapping. These case studies will illustrate how SAM can efficiently process and extract valuable insights from complex geospatial datasets, potentially enhancing the efficiency and effectiveness of our approach to environmental monitoring and disaster risk assessment.

According to the United Nations World Water Development Report, groundwater accounts for 26% of the world's renewable freshwater, with around 2.5 billion people relying primarily on it for basic water needs. The most realistic and cost-effective strategy to increase universal access to clean water, meet the 2030 sustainable development goals (SDG), and minimize climate change impacts is broad exploitation and management of groundwater. The study area is Nigeria's capital Abuja, generally characterized by moderate precipitation and few surface water sources. The water treatment plant, designed with a capacity of 10,000 cubic meters per hour of treated water, was aimed at supporting a population of 500,000 people 34 years ago. However, due to population growth and urbanization, the water supply is no longer meeting demand. Groundwater demand and consumption in Abuja have increased significantly over the last decade due to fast population expansion, urbanization, and industrialization. Understanding groundwater potential and aquifer vulnerability is critical for sustainable resource management.

Geologically, Abuja is underlain by Precambrian rocks of the Nigerian Basement Complex, which cover approximately 85% of the land surface, and sedimentary rocks, which cover approximately 15%. In the study area, four significant lithologic units are visible; these include the Older Granites, the Metasediments/Metavolcanics, the Migmatite-Gneiss Complex, and the Nupe sandstones of the Bida Basin, which occupies the southwestern region of the territory.

This study aims to map groundwater potential and aquifer vulnerability zones using Hydrogeophysical method, which incorporates geoelectrical resistivity through vertical electrical sounding (VES) and geographic information system (GIS) approaches. With a maximum current electrode separation (AB/2) of 100m, the Schlumberger electrode configuration was used to acquire the field resistivity data in 823 locations across the study area using a DC resistivity meter (Campus Ohmega Ω).

The resistivity method works by passing an electric current into the ground through two electrodes and measuring the consequent potential difference across two other electrodes. The electrode spacing gradually increases while the electrode array's center point remains fixed. However, as the current electrode spacing grows, the current penetrates deeper into the ground, and the apparent resistivity reflects the resistivity of the deeper layers as well. The resistance is estimated as the ratio of potential difference to current in ohms(Ω). Using a global positioning system (GPS), the absolute coordinates of the survey points (VES) were determined.

Three to five subsurface geoelectrical layers were identified in the research area with the aid of IPI2Win software. Vertical electrical sounding (VES) data are often interpreted using the IPI2Win software, which is a user-friendly geophysical software designed to process resistivity data and generate one-dimensional models of subsurface layers Layer resistivity and thickness were estimated using the software by iterating the model with the observed field data acquired using the Schlumberger array. The H-type sounding curve is the most dominant among the identified curve types.

The interpreted data were used to determine parameters including Depth to Bedrock, Transverse Resistance, Longitudinal Conductance, Reflection Coefficient, and Layer resistivity. Using scaling criteria, the longitudinal conductance was used to determine the aquifer protective Capacity (Vulnerability), and the result revealed the dominance of moderate vulnerability across the study area.

The groundwater potential zones in the research area were characterized based on the following criteria as established by previous authors in this field: Areas with overburden thickness ≥ 30m and reflection coefficient < 0.8 were classified as very high groundwater potential; Areas with overburden thickness ≥13m and reflection coefficient < 0.8 were classified as High groundwater potential; Areas with overburden thickness ≥ 13m and reflection coefficient ≥ 0.8 were classified as moderate potential while areas with overburden thickness <13m and reflection coefficient ≥ 0.8 were classified as low potential, and finally, areas with overburden thickness >13m and reflection coefficient < 0.8. were classified as very low potential. These criteria were written as Python codes that classify the area into five groundwater potential zones. The area covered by each zone was calculated after the geospatial analysis: the very high GPZ occupies about 19.70% of the study area, high 20.30%, moderate 20.0%, low 19.74%, and very low 20.31%.

Ordinary Kriging (OK) interpolation algorithm was used to generate the layer resistivity map, layer thickness map, depth to bedrock map, aquifer vulnerability map, and groundwater potential zone maps using the smart-map QGIS plugin. Smart-Map is a QGIS plugin that allows the generation of interpolated maps in the QGIS environment. Kriging is an unbiased linear interpolation technique that uses a weighted average of nearby samples to estimate unknown values in specific areas. It is deemed the best interpolation method for spatially varying data. For this study, the resistivity (VES) data was randomly distributed over a large area, and the sampling distance between one VES data and the other ranged from 0.5km to 10km.

This study evaluated groundwater parameters in the study area based on the geo-electric properties of the earth material. The results reveal that weathered/fractured basement and sandstone formations in the study area are substantial aquifer systems that host potable water. Data from some drilled boreholes across the study area were used to cross-validate the VES results against borehole log records. This knowledge aided in a better understanding of aquifer disposition, vulnerability, and potential consequences. The study's findings will provide a geo-database for groundwater potential zones in the Federal Capital Territory (Abuja), with significant implications for sustainable groundwater resource design and management.

In this session, we will introduce Re:Earth Visualizer, a 3D WebGIS platform, covering its overview, features, technical components, and use cases.
Re:Earth Visualizer is open-source tool that allows non-engineers to visualize data on maps without coding and publish it on the Web. It also includes a plugin system that enables users to develop and add custom functions like QGIS.

We will explain the system architecture, use of advanced web technologies, and integration with CesiumJS. Finally, we will show several use cases demonstrating Re:Earth applications.

Automating land use surveys in rural areas using advanced AI techniques can significantly enhance the efficiency and accuracy of identifying various land features. This project focuses on utilizing the YOLOv8 framework for land use detection through image segmentation and object detection.
Traditional land use surveys in rural areas are time-consuming and often prone to inaccuracies due to manual methods. Leveraging artificial intelligence, particularly deep learning models, presents a promising solution to streamline this process and improve data reliability. The project addresses the challenge of automating land feature identification, which includes detecting houses, rivers, roads, and vegetation from visual data captured by satellite/drone. Accurate identification of these features is crucial for effective rural planning and development, as it helps in resource allocation and infrastructure development. The limitations of conventional methods, such as the need for extensive human labor and susceptibility to human error, further highlight the necessity for innovative solutions like AI-driven land use surveys.
The primary aim of this study is to develop and train an AI model capable of accurately detecting and segmenting various land features in rural landscapes. By doing so, the project seeks to demonstrate the applicability of AI in enhancing rural development planning and management. The specific objectives include creating a reliable dataset of annotated aerial images, optimizing a deep learning model for high accuracy, and evaluating the model's performance across different types of land features. Ultimately, the project aims to provide a scalable and efficient tool that can assist policymakers, researchers, and rural development planners in making informed decisions.
The methodology involved several key steps to ensure the robustness and accuracy of the AI model. First, a diverse dataset of aerial images was collected, encompassing various rural landscapes with distinct features such as houses, rivers, roads, and farms/vegetation. These images were meticulously annotated using specialized tools to create ground truth data for training and validation. Data augmentation techniques, including rotations, flips, and color adjustments, were employed to expand the dataset and improve the model's generalization capabilities.
The YOLOv8 model was selected for its state-of-the-art performance in object detection and segmentation tasks. YOLOv8's architecture is well-suited for real-time applications due to its balance between accuracy and speed. The model was trained using the annotated dataset, with hyperparameters optimized to enhance its detection and segmentation performance. Training was conducted on a high-performance computing setup, leveraging GPU acceleration to expedite the process.
The results demonstrated the model's high precision in detecting and segmenting land features. The YOLOv8 model achieved notable accuracy metrics across various classes. The segmentation masks generated by the model closely matched the ground truth annotations, indicating its effectiveness in distinguishing different land features.
The findings of this study underscore the potential of AI in transforming rural development practices. The successful application of the YOLOv8 model for land use detection highlights its capability to deliver precise and actionable insights. The practical implications of this project are significant, offering a scalable solution for land survey automation, which can greatly assist policymakers and rural planners. The integration of such AI-driven methodologies can lead to more informed decision-making, efficient resource allocation, and ultimately, the betterment of rural communities.
The study also highlights several challenges and limitations encountered during the project. Data collection in rural areas can be logistically challenging, often requiring collaboration with local authorities and stakeholders. Ensuring the diversity and quality of the dataset is crucial, as biased or insufficient data can affect the model's performance. Additionally, the model's accuracy is dependent on the quality of annotations, which requires meticulous effort and expertise.
Despite these challenges, the project demonstrates that AI can significantly enhance the accuracy and efficiency of land use surveys. The use of deep learning models like YOLOv8 can reduce the reliance on manual methods, providing a more reliable and scalable solution. However, continuous efforts are needed to improve the dataset, address potential biases, and refine the model to handle more complex scenarios.
In conclusion, this project not only advances the field of AI in rural development but also sets a precedent for future studies aiming to leverage AI for similar applications. The integration of AI in land use surveys can revolutionize the way rural areas are planned and developed, leading to more sustainable and efficient outcomes. The success of this project inspires further research and development in AI-driven solutions for rural development, with the potential to make a lasting positive impact on rural communities worldwide.

This research aims to estimate PM2.5 concentrations from Aerosol Optical Depth (AOD) and meteorological data and study the spatial distribution patterns of PM2.5 in Saraburi Province. The estimation of PM2.5 levels is conducted using AOD data combined with meteorological data through a Multiple Linear Regression (MLR) method. The estimated values are then used to analyze the distribution patterns of PM2.5. The study found that in 2018, the average monthly PM2.5 concentration ranged from 0 to 74.1 μg/m³, with high-value clustering (hot spots) covering approximately 421.43 km², or 12.04% of the provincial area. In 2019, the average monthly PM2.5 concentration ranged from 0 to 41.4 μg/m³, with hot spots covering approximately 509.29 km², or 14.55% of the provincial area. In 2020, the average monthly PM2.5 concentration ranged from 0 to 50.0 μg/m³, with hot spots covering approximately 648.37 km², or 18.53% of the provincial area. In 2021, the average monthly PM2.5 concentration ranged from 0 to 55.3 μg/m³, with hot spots covering approximately 562.93 km², or 16.09% of the provincial area. In 2022, the average monthly PM2.5 concentration ranged from 0 to 57.3 μg/m³, with hot spots covering approximately 615.97 km², or 18% of the provincial area. The most of high-value clusters were in the western part of the province, where agricultural activities are prevalent, contributing to higher PM2.5 levels. In contrast, low-value clusters (cold spots) were primarily found in the eastern part of the province, which is largely forested.

Clouds, atmospheric disturbances, and sensor failures influence the quality of Earth observation (EO) data and satellite images, in particular. Many modeling techniques and statistical analysis, to be applied to EO data, require the detection and removal of such aberrations. However, the data gap created after removing the involved pixels needs to be imputed with numerical values that resemble the expected noncorrupted ones. Several imputation, or gap-filling, methods available in literature are based on time-series reconstruction, working only on the temporal dimension of each pixel to input the missing values. Such methods, compared to alternative ones that also consider spatial neighbor pixels or data fusion with other sensors, have the advantage of maintaining the same spatial resolution and spectral consistency in the imputed data.
In contrast with methods that only work with a local temporal window, some of these methods take advantage of the whole time series of each pixel to reconstruct each missing value, allowing the full reconstruction of each gappy time-series. Nevertheless, such methods, like most recent image propagation or linear interpolation, often require an iterative search of available values along the time-series. When the time-series is composed of several samples and/or the involved number of pixels is large, the application of such methods leads to prohibitive computational costs.

We present in this work a computational framework based on discrete convolution that numerically approximates such methods and does not require iterating over the time-series to be applied [1]. In addition, the framework flexibility allows the application of different time-series reconstruction methods by only adapting the convolution kernel. The framework has been used to reconstruct the PetaByte scale Landsat Analysis Ready Data (ARD) collection provided by the Global Land Analysis and Discovery team (GLAD) [2]. New research fronts include the extension of the method to data-fusion approaches that combine time-series of multiple sensors to maintain the highest spatial resolution while also using the temporal information provided by all the sensors. The code, developed in Python with a C++ backend to guarantee usability and high computational efficiency, is openly available at https://github.com/openlandmap/scikit-map.

[1] Consoli, Davide & Parente, Leandro & Simoes, Rolf & Murat, & Tian, Xuemeng & Witjes, Martijn & Sloat, Lindsey & Hengl, Tomislav. (2024). A computational framework for processing time-series of Earth Observation data based on discrete convolution: global-scale historical Landsat cloud-free aggregates at 30 m spatial resolution. 10.21203/rs.3.rs-4465582/v1.

[2] Potapov, Peter & Hansen, Matthew & Kommareddy, Anil & Kommareddy, Anil & Turubanova, Svetlana & Pickens, Amy & Adusei, Bernard & Tyukavina, Alexandra & Ying, Qing. (2020). Landsat Analysis Ready Data for Global Land Cover and Land Cover Change Mapping. Remote Sensing. 12. 426. 10.3390/rs12030426.

“toeng.net”, which we announced at FOSS4G-ASIA 2023 Seoul, will be launched as “MSpace.E”.
MSpace.E is a comprehensive urban environment analysis platform using 3D city models. This platform integrates simulations of shadows, building surface shadows, noise, and wind, providing an approach to urban planning and environmental assessment.

Key features include 3D visualization using the Re:earth platform, environmental analysis integrated with user-provided construction data in IFC, FBX, GLB, and 3D Tiles formats, and newly added functionality for group management and sharing of analysis results. Users can analyze the environmental factors within a selected area, and by utilizing the PLATEAU 3D city model, more accurate analysis is possible by taking into account existing building structures.

In the case of shadow analysis, users can select analysis options, specify the range, set parameters, and receive the results in the CMZL file format. In addition, users can upload their own construction data and visualize it in 3D on Re:earth. Group management and result sharing functions make team collaboration easier. We also discuss a comparison of the implementation and performance of MSpace.E with the toeng.net prototype published last year. Application areas include urban planning for optimal building placement and public space design, environmental impact assessment of architectural projects, and energy-saving strategy planning at the urban scale.

MSpace.E is a powerful platform for multi-faceted analysis and visualization of complex urban environments. It enables comprehensive urban environment simulation, strongly supporting decision-making for sustainable urban development.
Join us at mspace.apptec.co.jp!

Urban planning and mobility analysis have traditionally been studied through observation or questionnaires, which can be time-consuming and costly. However, the rapid advancement of technology has enabled tracking devices to be installed in individual vehicles, allowing the measurement of various values, particularly global positioning system (GPS) signals.

The location data collected is accurate, regularly updated, and can offer valuable insights into people's movements and behavior. Because the amount of trajectory data is substantial and continues to increase over time, specialized platforms and skills are needed for its analysis.

In this study, we developed large-scale analysis toolkits to extract insights, including trip statistics, origin–destination analysis, and hotspot identification from vehicle-based GPS trajectories. The toolkits are specifically designed to handle large-scale datasets using Apache Spark, an analytics engine capable of processing large volumes of data by distributing tasks across a Hadoop cluster for efficient processing.

Algorithms for the analytics model were created to reconstruct trips based on their type of mobility, and trip locations were mapped using open data such as administrative boundaries and points of interest. We then verified our approach using real-world taxi data from Bangkok, Thailand.

The results revealed that taxis had more vacant trips than busy trips, and the travel time and distance taken to search for passengers were longer than those taken to pick them up and drop them off. Taxi activity was concentrated in the city center and nearby areas, particularly those within the vicinity of transport-connecting hubs. Taxi stay hotspots were mainly areas near tourist attractions and parking hubs.

Furthermore, we found that the processing performance of the proposed approach increased with the number of executor cores. This study comprehensively presented information on taxi travel patterns, service availability, hotspots, and processing performance using the developed trip analysis toolkits.

Disasters like typhoons, earthquakes, and flooding are inevitable, especially as climate change intensifies. This heightens the need for effective information sharing, which some governments have addressed by sending out timely digital alerts. While large organizations work to prepare communities for disasters, local knowledge often gets overlooked despite its critical role in understanding a hazard’s impact. Residents possess intimate knowledge of their surroundings, which becomes invaluable during emergencies for identifying evacuation routes and understanding the landscape. In creating a disaster risk assessment, selecting data sources is crucial, as demonstrated by the Metro Manila Climate and Disaster Risk Atlas. This atlas, created using QGIS and OSM, assesses hazards in Metro Manila—a region prone to a potential 7.2 magnitude earthquake due to the West Valley Fault System. Leveraging these tools, provides a comprehensive view of risks, empowering communities with vital information about the vulnerabilities and resilience of their locales. The project exemplifies how integrating local and digital knowledge fosters a safer, more prepared society.

This session will explore how combining the powerful 3D mapping capabilities of Cesium with the intuitive and efficient web framework Svelte can simplify the development of 3D map applications. We will demonstrate practical examples, such as data binding and custom stores, to show how these tools can make working with Cesium more straightforward and intuitive. Participants will gain new insights into using Svelte and Cesium together, making complex 3D geospatial projects more accessible and manageable. This session is ideal for frontend engineers looking to enhance their development experience in the growing field of 3D mapping.

The integration of vast data analysis and AI technologies in urban planning represents a significant advancement in managing the complexities of modern cities. Through this presentation, it show the vast potential of these technologies to enhance decision-making, optimize resource allocation, and improve urban sustainability. By understanding and applying these technologies, future urban planners can develop smarter, more efficient, and environmentally friendly cities. The practical applications discussed, including traffic injury risk assessment, human mobility analysis, and carbon emission estimation, demonstrate the tangible benefits of leveraging Big Earth Data and AI in urban planning.

Yen Bai Province in northern Vietnam, especially Mu Cang Chai (MCC) and Van Yen (VY) districts, are highly susceptible to slope failure due to rugged terrain, high rainfall and anthropogenic activities . In this research MCC was used as an area for training and testing the machine learning models, while VY serves for model validation due to similar topographic and geological conditions.

The methodology treats the slope failure prediction as a binary classification task (landslide/no-landslide). A balanced dataset of 286 landslide and 286 non-landslide points in MCC, along with 16 contributing factors, including topographic, geologic, hydrologic, anthropogenic and vegetation factors calculated from open data sources and made use from existing databases and from previous research on the area. Principal Component Analysis (PCA) and Pearson Correlation Coefficients refine the dataset by evaluating correlated factors and removing the least important ones, the size of the training dataset can be reduced while ensuring the performance of the ML models. Four ML models: Random Forest (RF), Support Vector Machine (SVM), Logistic Regression (LR), and Extreme Gradient Boosting (XGBoost) are trained and evaluated to select the best hyperparameter tuning for each model. Model accuracy is assessed via confusion matrices, accuracy score, ROC (Receiver operating characteristic) curves and AUC (Area under the ROC Curve).

Results show the models perform effectively in MCC with the average accuracy of all models being 0.74. The trained ML models with tuned hyper-parameters after running on MCC data, was validated on datasets for VY. The VY data also consists of 16 factors, with a data set of 308 landslide/non-landslide points. RF and XGBoost have the highest accuracy for both training and testing area (MCC) and Validation area (VY), with XGBoost showing a slightly higher accuracy score of 0.83 while RF scores 0.80.

The XGBoost model produces good results and could be further optimized to achieve even better zonation in future studies. The machine learning workflow can be applied on other areas that are prone to slope failures. Other geologic and weathering factors could be included in the analysis to further improve the model.

As the maturity of digital twin technology increases over time, there is an increasing demand to implement various services (especially those related to decision support) that utilize invisible phenomena/analysis information by visualizing various types of sensor data or the results of professional analysis/prediction together, from three-dimensional visualization using only buildings/terrain.

In particular, Korea aims to provide city-scale housing with a quality living environment, and one of the most complained about items is noise, and there is a strong desire to utilize noise prediction analysis to derive planning/design results in urban planning/design work.

In this presentation, we will introduce a digital twin system that combines 3D spatial information and noise predictive modeling using "OGC standard CityGML" and "open source mago 3DTiler and mago 3DTerrainer developed by Gaia3D" to support urban noise analysis and decision-making.

Effective data management is essential for maximizing the value of spatial data in today’s data-driven landscape. This presentation provides an overview of implementing ETL (Extract, Transform, Load) processes using NDJSON (Newline Delimited JSON) for efficient spatial data integration. We will discuss the importance of robust data management, the benefits of using NDJSON for handling large and complex spatial datasets, and the practical applications of this approach.

Key topics include the steps of the ETL process with NDJSON, from extracting spatial data from various sources, transforming it into usable formats, to loading it into databases such as MongoDB and Elasticsearch. We will highlight the efficiency gains and flexibility provided by NDJSON in streaming and processing spatial data. Additionally, we will cover real-world use cases and best practices for optimizing spatial data integration with NDJSON.

Attendees will gain practical insights into the strategic and technical aspects of utilizing NDJSON in ETL processes, enabling them to implement effective spatial data integration within their organizations.

This research aims to develop a low-cost GNSS receiver device for positioning agricultural tractors, incorporating Differential GPS (DGPS) technology for enhanced accuracy using free and open source software. Integrated with IoT technology, the device was tested to receive GNSS data and other relevant information, including geographic coordinates (latitude and longitude), tractor speed, tractor direction, date, time, and the number of satellites receiving signals. The DGPS setup involves using one receiver as a base station and another on the tractor, where the base station provides correction data to improve positioning accuracy. The data collected by the receiver is transmitted to a signal processing device for mapping the coordinates, creating a route of the tractor's movement that is displayed on a real-time Web Map Application. This process includes error correction to ensure high accuracy. The IoT device was installed on the left rear wheel of the agricultural tractor. Test results show that the data from the developed device has an average accuracy of 22 centimeters, which is acceptable and sufficient for agricultural tractor positioning applications. Furthermore, this system enables real-time monitoring of the tractor's operations.

Geospatial technologies are being widely used to address societal needs like land use, demographics, and natural resource management. Spatial data analysis play a crucial role in how we understand and interact with our environment. They involve using maps, GPS, and satellite images to collect, analyse, and display data about the world. Teaching these skills to school children has become increasingly important.

Geospatial techniques help students develop a better understanding of the world around them. Maps, for instance, are not just tools for finding directions; they tell stories about our environment, culture, and history. By learning to read and create maps, children begin to see the connections between different places and the events that shape them. Understanding these concepts early helps them develop a broader view of the world and how places are interconnected This kind of knowledge fosters a global perspective, encouraging students to think beyond their immediate surroundings and consider how their actions can affect the world.

As part of its mission to build awareness of the importance of Earth Science in daily life, the team at ‘The Centre for Education and Research in Geosciences (CERG), India conducts various activities aimed at laymen and schoolchildren. These events are conducted throughout the year. One such program is a workshop titled “Maps & Me”, which is focused on giving a basic understanding to school & college children on the geospatial world and open-source mapping tools. In the ‘Maps & Me’ workshop we explore the basics of maps, satellite images and digital maps and how to navigate using these tools.

The workshop is hands-on and interactive, covering key map elements like latitude, longitude, and scale, along with a session on using QGIS, a popular open-source mapping software. Participants are introduced to the fundamentals of remote sensing, satellite imagery, photo recognition, and digital mapping. After that, they get to create their own maps using QGIS.

At CERG we believe that such skills are important because they help students make sense of real-world issues, like climate change, urban planning, and natural resource management. By learning how to read and interpret maps, for example, young students can see how their local environment fits into the larger world. It also encourages them to think critically and solve problems creatively, skills that are valuable in all areas of life.

The integration of Internet of Things (IoT) technology with environmental monitoring systems has significantly enhanced the ability to collect real-time data from diverse and remote locations. However, the challenge lies in efficiently analyzing and interpreting this vast amount of data to make informed decisions. This paper explores the application of generative AI in IoT data analysis for environmental monitoring. Generative AI, with their advanced natural language processing capabilities, offer a novel approach to processing and understanding complex data patterns. By leveraging Generative AI, it is possible to automate the identification of critical environmental changes, predict trends, and provide actionable insights with unprecedented accuracy. This study demonstrates how Generative AI can enhance data analytics in environmental monitoring through case studies that highlight improvements in air quality assessment (e.g. PM 2.5). The findings suggest that Generative AI not only streamline the data analysis process but also enhance the reliability and responsiveness of environmental monitoring systems. Consequently, this research underscores the potential of Generative AI to transform IoT-based environmental monitoring, promoting more proactive and effective environmental management practices.

This talk will explore the transition from proprietary nature to PWAGIS QGIS plugins, focusing on the limitations experienced with previous solutions of proprietary software, the challenges encountered during development, and the innovative functionalities introduced in the new plugins. Attendees will gain insights into the practical aspects of moving from a legacy system to an open-source solution, highlighting both the obstacles and the opportunities this transition presents.

Introduction

With the spread of IoT and the increasing resolution of observation sensors, the total amount of geospatial information data is increasing exponentially daily. On the other hand, the increase in resolution of display devices used to analyze and visualize these data is reaching its limit due to various physical constraints. The maximum resolution of commercially available display devices is 8K; 4K or 5K is considered the upper limit for desktop use.
Using the OS's multi-display function or a tiled display driver provided by the GPU manufacturer, it is possible to create a display environment with an even larger area and higher resolution. However, the middleware provided by the GPU manufacturer currently has a maximum resolution limitation of 16K [1], which is the maximum resolution that can be achieved on a single PC.
However, even if these mechanisms are used to create an ultra-high-resolution display environment, it is only possible to render data within the web browser's heap memory limit in the case of WebGIS applications. For example, the 3DWebGIS viewer provided by the Tokyo Digital Twin Project [2] cannot render 3DTiles [3] building data for all 23 wards of Tokyo at once (textured building data is used for areas provided with texture).
In this paper, we introduce ChOWDER, a web-based tiled display driver that enables distributed rendering of 3DWebGIS content across multiple web browsers, as a solution to the above problems and report the results of memory load balancing experiments using ChOWDER for distributed rendering.

Proposal of a distributed rendering method for 3DWebGIS

One possible solution to the above problems is to distribute the display of one WebGIS content across multiple PCs (multiple web browsers). This makes it possible to display a WebGIS at a resolution that exceeds the upper limit of a single PC (web browser) and distributes the memory load required to display the content across each PC (web browser).
The scalable display system ChOWDER[4][5], jointly developed by RIKEN Center for Computational Science and Kyushu University, is an open-source tiled display driver that can create an ultra-high-resolution pixel space by arranging multiple displays that display a web browser in full-screen mode in tiles. It also supports distributed rendering of 3DWebGIS.
This function uses iTowns[6], an open-source 3DWebGIS, as middleware. iTowns uses Three.js as a WebGL rendering library, and Three.js has an API that can offset the view frustum[7].
The view frustum must be split appropriately to split and display 3D content on multiple display devices. ChOWDER uses the view frustum offset API of Three.js to split a single iTowns content into multiple view frustums, enabling multiple web browsers to split and render 3DWebGIS content [8].
However, at the time of the previous report [8], when iTowns executed a 3DTiles load command, it loaded all the data without judging whether it was inside or outside the view frustum range, so distributed rendering did not improve memory utilization efficiency. Since then, the 3DTiles load process was improved in iTowns Release 2.42.0; in this paper, we measured the amount of heap memory consumed by each browser when iTowns content was distributed and rendered using ChOWDER on multiple web browsers and confirmed the memory load distribution achieved by this method.

Experimental procedures and results

The experimental data used was the textured building data for Chiyoda, Minato, and Chuo wards in Tokyo, from the 3DTiles data distributed by the PLATEAU project [9] of the Ministry of Land, Infrastructure, Transport and Tourism of Japan.
The experiment first displayed the 3DTiles building data for the above three wards in full screen on a single 4K resolution display using iTowns on ChOWDER. The heap memory size of the web browser at this time was 268MB.
Next, the same content was displayed on a ChOWDER distributed display consisting of four 4K displays arranged in two horizontal and two vertical rows. Each display had a full-screen web browser. The heap memory sizes of each web browser were 133MB, 188MB, 68.3MB, and 37.7MB.
Finally, we conducted an experiment using nine 4K displays arranged in three rows and three columns. The heap memory sizes of each web browser were 66.8MB, 122MB, 140MB, 84.3MB, 87.2MB, 56.9MB, 41.2MB, 38.4MB, and 33.6MB.
From these experimental results, it can be said that distributed rendering of 3DWebGIS using ChOWDER achieves memory load balancing.
During distributed rendering, the heap memory size of each web browser is different because the amount of 3DTiles data contained in each responsible drawing area is different. Also, the total heap memory size of all browsers is larger than when rendering in a single browser because iTowns loads 3DTiles data that is wider than its view frustum, and data loading in overlapping areas occurs during distributed rendering.

Future work and conclusion

In this experiment, we measured the web browser's heap memory, but did not measure GPU memory consumption. However, because 3DWebGIS uses WebGL for rendering, we believe that a more precise evaluation can be made by measuring GPU memory consumption as well.
In addition, since distributing rendering across more web browsers is expected to further distribute memory load, we plan to conduct experiments by increasing the number of distributed displays.
In this paper, we have shown the limitations of current 3DWebGIS when the data to be displayed increases, and proposed a distributed rendering method as a means to solve this problem, and introduced the view frustum offset API of Three.js, iTowns, a 3DWebGIS to which it can be applied, and ChOWDER, a web-based tiled display driver that incorporates them, as a means to realize this method. Furthermore, we have presented the results of an experiment that shows that memory load distribution is achieved by distributed rendering using these and demonstrated that this method is one solution to the increase in data to be displayed in 3DWebGIS.

References

[1] Limitations. About NVIDIA Mosaic. https://www.nvidia.com/content/Control-Panel-Help/vLatest/en-us/mergedProjects/nvwks/SLI_Mosaic_Mode.htm Accessed July 29, 2024.

[2] Tokyo Digital Twin Project. https://info.tokyo-digitaltwin.metro.tokyo.lg.jp/ Accessed July 29, 2024.

[3] 3DTiles. The open specification for 3D data. https://cesium.com/why-cesium/3d-tiles/ Accessed July 29, 2024.

[4] Kawanabe, T., Nonaka, J., Hatta, K., & Ono, K. (2018, September). ChOWDER: an adaptive tiled display wall driver for dynamic remote collaboration. In International Conference on Cooperative Design, Visualization and Engineering (pp. 11-15). Cham: Springer International Publishing.

[5] ChOWDER GitHub repository. https://github.com/SIPupstreamDesign/ChOWDER Accessed July 29, 2024.

[6] iTowns (in French). https://www.itowns-project.org/ Accessed July 29, 2024.

[7] three.js API Reference. https://threejs.org/docs/#api/en/cameras/PerspectiveCamera.setViewOffset Accessed July 29, 2024.

[8] Kawanabe, T., Hatta, K., & Ono, K. (2020, September). ChOWDER: A New Approach for Viewing 3D Web GIS on Ultra-High-Resolution Scalable Display. In 2020 IEEE International Conference on Cluster Computing (CLUSTER) (pp. 412-413). IEEE.

[9] Project PLATEAU portal site (in Japanese). https://www.geospatial.jp/ckan/dataset/plateau Accessed July 29, 2024.

Reservoirs play a crucial role in global water resource management, hydroelectric power generation, and flood control. However, their construction often entails significant ecological and socio-economic impacts, necessitating thorough environmental assessments. The Mekedatu Reservoir Project, situated on the Cauvery River in the Ramanagar district of Karnataka, India, holds paramount significance. Aimed at supplying the Bengaluru Metropolitan Region and its surroundings with drinking water, the project also endeavors to generate 400 MW of renewable energy annually. Despite its benefits, the project comes with ecological costs, as approximately 5252.40 hectares of revenue, forest, and wildlife land will be submerged. This necessitates a detailed evaluation of its potential environmental consequences.

This study identifies a knowledge gap in the existing literature regarding the ecological implications of the Mekedatu Reservoir Project. It seeks to fill this void by forecasting land use and land cover (LULC) changes for the years 2000, 2010, and 2020 using the Random Forest method, and assessing the submergence area for different levels of the proposed reservoir. Catchment delineation is performed using the Soil and Water Assessment Tool (SWAT). Additionally, the Cellular Automaton-Markov Chain technique is employed to predict land use and land cover changes for the year 2030. Integrating these methodologies, the research provides a holistic understanding of the project's environmental footprint.

The land use and land cover analysis revealed significant shifts from 2000 to 2020, with forest cover decreasing from 71.54% to 60.71% and barren land increasing from 19.55% to 29.56%. The projected land use and land cover for 2030 shows further forest reduction to 58.28% and barren land increasing to 31.11%. These changes highlight a trend towards deforestation and land degradation, posing severe ecological threats. The submergence area at the proposed reservoir Full Reservoir Level is estimated to be 5252.4 hectares, distributed as 6.62% water, 19.55% barren land, 71.54% forest area, and 2.29% built-up area for the year 2000. The inundation of these areas will lead to significant biodiversity loss, affecting numerous plant and animal species.

In line with Sustainable Development Goals, which advocates for sustainable water management, this study emphasizes the importance of informed decision-making and sustainable development practices. The findings underscore the need for new ecologically sensitive areas and the establishment of wildlife corridors, conservation zones, and afforestation programs to mitigate the adverse impacts. Continuous environmental monitoring and research are essential to track biodiversity impacts and adjust conservation strategies accordingly.

Policy implications of this study suggest that due process of law, linked with the principle of natural justice, must be adhered to in ensuring environmental balance. Recommendations from the World Commission on Dams (WCD) highlight the need to reduce the negative impacts of dams by increasing the efficiency of existing assets and minimizing ecosystem impacts. Policymakers must understand the long-term ecological consequences of such mega projects and explore alternatives. Sustainable development models must be based on equality and natural justice.

Future research should focus on the socio-economic impacts of the Mekedatu Reservoir Project, particularly the displacement of local communities. This includes conducting detailed socio-economic assessments, inclusive resettlement planning, livelihood restoration programs, and initiatives to preserve cultural heritage. Continuous monitoring and long-term studies are crucial to ensure the well-being of resettled populations and to balance development with environmental and social sustainability.

In summary, this study advances the understanding of environmental impact assessment in reservoir projects, providing valuable insights for stakeholders and policymakers. It highlights the critical need for sustainable development practices that ensure equitable access to water resources while preserving environmental integrity.

Keywords: Environmental Impact Assessment, Reservoir Project, Machine Learning, Random Forest, Markov Chain, Cellular Automaton, Land Use Changes, Submergence Area, Sustainable Development Goals, Water Resource Management.

In recent years, 3D city models have become crucial for urban planning and research. Japan's Project PLATEAU has led the development of open 3D city models and point cloud data, with over 100 cities releasing Digital Twin data in CityGML format by February 2023. This talk explores the collaboration between Japan's Digital Twin initiatives and the global OpenStreetMap community. Since 2022, Japan's Digital Twin data, following the ODbL license, has been integrated with OpenStreetMap using specially developed tools. This integration aims to promote the global adoption of 3D city models, enhancing urban development through the synergy of Digital Twin technologies and OpenStreetMap.

The ZOO-Project is an open-source processing platform released under the MIT/X11 Licence. It provides the polyglot ZOO-Kernel, a server implementation of the Web Processing Service (WPS) (1.0.0 and 2.0.0), and the OGC API - Processes standards published by the OGC. It contains ZOO-Services, a minimal set of ready-to-use services that can be used as a base to create more useful services. It provides the ZOO-API, initially only available from the JavaScript service implementation, which exposes ZOO-Kernel variables and functions to the language used to implement the service. It contains the ZOO-Client, a JavaScript API that can be used from a client application to interact with a WPS server.

In today's rapidly evolving field of map data visualization, the use of vector tiles is increasingly prevalent. Vector tiles offer flexible and efficient data display, but they require JSON Style data to define their visual representation. JSON Style plays a crucial role in formatting data, ensuring that presentations are both diverse and user-friendly.

Environmental degradation in Southeast Asia, particularly in Cambodia, is an alarming issue that poses significant threats to both ecosystems and human well-being. The region is experiencing rapid deforestation driven by agricultural expansion and urban development, leading to substantial loss of forest cover. This deforestation disrupts ecological balance and biological environments, contributing to habitat fragmentation, biodiversity loss, and alterations in local microclimates. Concurrently, changes in land use and land cover exacerbate these problems, further fragmenting habitats and impacting species distribution. The increasing risk of forest fires, fueled by climate variability, land use changes, and agricultural burning, adds another layer of concern, contributing to air pollution and posing risks to both natural and human systems.

In response to these growing challenges, a comprehensive monitoring tool is essential for systematically observing and analyzing environmental parameters. Such a tool would provide critical data necessary to address these issues, inform policy decisions, and support sustainable land management practices. Recognizing the need for a robust solution, SERVIR Southeast Asia (SEA)—a collaborative initiative of the U.S. Agency for International Development (USAID), the National Aeronautics and Space Administration (NASA), and the Asian Disaster Preparedness Center (ADPC)—has developed the “Biophysical Monitoring and Evaluation (M&E) Dashboard” for Cambodia. This open-access tool is designed to offer comprehensive, near-real-time insights into critical environmental parameters, leveraging advanced technologies to support environmental protection and sustainable management.

The Biophysical M&E Dashboard utilizes a range of cutting-edge technologies to analyze and visualize environmental data. It harnesses the power of Google Earth Engine (GEE), a cloud computing platform capable of processing vast amounts of satellite imagery. GEE is employed to analyze large-scale, open satellite data to map key indicators such as the Enhanced Vegetation Index (EVI), land use and land cover, forest cover, forest fire occurrences, and crop monitoring. The tool integrates these insights with data from GeoServer, an open-source geospatial data publisher, and PostgreSQL, a powerful open-source relational database system. Modern web technologies, including React with NextJS and the Python-based Django framework, are used to develop the user interface and ensure seamless functionality.

The M&E Dashboard is designed to integrate multiple data sources, providing a holistic view of landscape dynamics. It visualizes and analyzes critical aspects such as forest cover changes, land use transformations, vegetation health, and fire hotspots. By offering detailed analytical information at various levels—country, province, district, and designated protected areas—the tool enables users to gain a nuanced understanding of environmental trends. The dashboard’s current capabilities include monitoring forest gain and loss, assessing rice cropping fields, and tracking deforestation and fire hotspots. Future plans include expanding its functionality to support user-defined area levels and the incorporation of socio-economic and vulnerability indicators, further enhancing its adaptability and utility.

In addition to land use and land cover monitoring, the M&E Dashboard incorporates weather and climate information to support sustainable agriculture practices. This integration provides valuable insights into the impacts of climatic conditions on crop health and agricultural productivity, and also a drought assessment framework, aiding in the development of strategies to mitigate adverse effects and enhance resilience.

This paper presents a detailed overview of the architecture, design, and functionality of the Biophysical M&E Dashboard. It outlines how the tool addresses critical environmental challenges in Cambodia, including deforestation, habitat fragmentation, and fire risks. By offering a comprehensive suite of analytical features and visualizations, the dashboard supports informed decision-making and strategic planning for sustainable landscape management. Through its integration of advanced technologies and multi-source data, the Biophysical M&E Dashboard stands as a vital resource for protecting Cambodia’s natural resources and promoting ecological resilience in the face of ongoing environmental pressures.

The SpatioTemporal Asset Catalog (STAC) revolutionizes geospatial applications by providing a standardized framework for cataloging spatiotemporal data. Developed in 2017 through a collaborative effort among various organizations, STAC streamlines the discovery and retrieval of geospatial assets, making it easier for users to access satellite imagery and other spatial data. This open-source specification, which aligns with FAIR principles—Findable, Accessible, Interoperable, and Reusable—promotes interoperability among various data providers and applications, fostering innovation in the geospatial community.

STAC's design allows for automated data retrieval through the STAC API, making it especially useful for applications in environmental monitoring, disaster management, and urban planning. Its JSON-based structure enhances user accessibility, allowing developers to quickly integrate geospatial data into their workflows. Furthermore, STAC's extensibility ensures it can adapt to a wide range of geospatial data types, from remote sensing to 3D point clouds.

The benefits of STAC go beyond theoretical applications. In Thailand, STAC is applied to the GISTDA Decision Support System for Disaster Management Platform. On this platform, STAC catalogs vector data related to flooding areas, thermal activities, and drought indices. As a result, the implemented application can efficiently browse and retrieve data from the STAC catalog, enhancing data retrieval speed and user experience.

As the geospatial landscape continues to evolve, STAC stands out as a remarkable tool for driving innovation, enabling seamless data sharing, and empowering users to harness the full potential of geospatial technologies in addressing complex global challenges.

In the past, the installation of GIS applications often encountered challenges regarding the flexibility of services, which were unable to be scaled to accommodate a growing number of users. The interconnection and exchanging of data across services were constrained, and service separation was not feasible. These issues had a significant impact on overall usability.

The design and deployment of applications in the form of microservices are gaining popularity and widespread adoption. This approach aims to provide flexibility to the installation process, allowing services to be added or reduced as needed to align with usage requirements. It can subdivide services into smaller units to facilitate installation, following the principles outlined in The Twelve-Factor App (https://12factor.net/).

Nowadays, GIS application development has OGC Standards, which are standardized guidelines that define the process of storing and providing geospatial data. These standards encompass a multitude of aspects of geospatial information interoperability. Therefore, the principles of The Twelve-Factor App can be adapted to the design and deployment of GIS applications, ensuring compliance with OGC Standards.

This session will elaborate on how the principles of The Twelve-Factor App can be harmonized with OGC Standards, as well as the various technologies selected for the design and deployment of applications.

Introduction/Background:
The consequences of climate and environmental changes on health are now obvious to communities, institutions and researchers alike. The impact of these changes must now be considered in an operational way in health management, in order to anticipate their effects, prevent them or mitigate them where possible. In practice, there is very little routine real-time use of space observation data in the public health sector, despite the increasing availability of space data. Indeed, space observation technologies have been constantly evolving since the 1970s, and now offer a wide range of data at different spatial, temporal and radiometric resolutions. More recently, access to data acquired by satellite has greatly improved, with free, massive data and easier processing. This offers the possibility of supporting health surveillance at various scales, which will be explored in this presentation using different examples from South-East Asia.

Main Aim/Purpose:
This presentation aims at raising the issue of integrating environmental and climate change indicators into health monitoring, by presenting practical tools and case studies that are operational in Southeast Asia. It will provide an update on the needs for further operational implementation that will benefit health monitoring.

Methodology and Findings:
The presentation will focus firstly on the development of a web platform that aims at modeling suitable climate and environmental conditions for leptospirosis through Earth observation, over the agglomeration of Yangon, Myanmar. Leptospirosis is a bacterial zoonosis that remains rarely diagnosed in Southeast Asia despite a high morbidity as shown in several active investigations. It is strongly associated to water and seasons with epidemics following heavy rainfall and flooding episodes. In the frame of the ECOMORE 2 Project (coordinated by Institut Pasteur and funded by the French Agency for Development - AFD), the locations of leptospirosis confirmed cases (vs non-leptospirosis controls) included in 2019 and 2020 were analyzed retrospectively. Time series of vegetation, water, and moisture indices from Sentinel-2 satellite imagery (available at 10 meters spatial resolution, every 5 days, from the European Space Agency, Copernicus Program) were produced to describe the dynamics of the environment around the locations of residence. This process relies on the use of the Sen2Chain processing chain developed in Python and open (https://framagit.org/espace-dev/sen2chain). The most relevant indices were used to build a spatiotemporal prediction model of positive vs negative locations. This model was spatialized on homogeneous landscape units from the point of view of land use, and describing the whole study area. The acquisition of Sentinel-2 images, their processing and the modelling were then automated as soon as a new image is available (every 5 days). An online platform, named LeptoYangon (https://leptoyangon.geohealthresearch.org/), was developed with R and R-Shiny to display this dynamic mapping of suitable environments and inform the epidemiologists and physicians of the study, in the frame of the ClimHealth project (funded by CNES and accredited by the Space Climate Observatory International Initiative). This fully automated tool allows retrospective consultation at any date since the first Sentinel-2 image was available in March 2016 (over 7 years). By clicking on the map, the user can select a landscape unit and view the temporal dynamics of the risk for that unit (i.e. whether the risk is increasing and decreasing). The user can also view the vegetation, water and moisture indicators to get an idea of the environmental data more specifically. This platform was designed to be used by epidemiologists and physicians to visualize the most at-risk areas and those where the risk is increasing, in order to raise physicians' awareness of leptospirosis (often confused with other fevers).

Discussion:
Implementing this tool in other territories is facing 1) methodological challenges regarding the volume of satellite data to be processed and 2) the need for detailed knowledge of the ecology of leptospirosis and exposure factors to adapt the models in different contexts. However, this already operational tool opens the way to the development of climate and environmental monitoring systems to increase the vigilance of healthcare workers and populations to the risk of leptospirosis. This also shows the relevance of developing specific tools for other diseases associated to climate and environment. At a country or regional scale, it is mainly meteorological variations and climatic anomalies that are relevant to the surveillance of certain diseases, such as dengue fever. The presentation will finally review the development of a national early warning system in Cambodia based on the acquisition of such climatic data.

Southeast Asia (SEA) region leads the world in palm oil production, with Indonesia, Malaysia, and Thailand collectively contributing over 88% of the global production. However, the tropical climate in the SEA region resulted in oil palm trees vulnerable to various diseases such as Fatal Yellowing (FY) and Ganoderma boninense. To keep track of productivity, it is crucial to monitor the varying conditions of oil palm trees—such as healthy, dead, yellow, and small—and apply effective pruning techniques to cure affected trees. Manual approach for oil palm tree detection is expensive, tedious, and prone to inaccuracies. Thus, our study is focused to automate the detection of oil palm trees and their states using a deep learning (DL) algorithm on unmanned aerial vehicle (UAV) imagery. We use YOLOv8, one of the latest open-source models, on the publicly available UAV dataset, named MOPAD, containing training and validation sets. The performance of the model is further compared with other state-of-the-art object detectors. In addition, a prototype web application is developed to demonstrate the robustness of the model in adverse real-world conditions and for potential deployment.

Air pollution in Southeast Asia has reached critical levels, significantly impacting human health across the region. Nearly the entire population lives in areas where air pollution exceeds the World Health Organization’s (WHO) safe air standards. This severe pollution is primarily due to rapid industrialization, urbanization, and deforestation, which have increased the amount of harmful pollutants in the air, particularly fine particulate matter known as PM2.5. Seasonal agricultural burning, a common practice in the region, also contributes significantly by releasing large quantities of smoke and particulate matter into the air. The rapid pace of urbanization has led to increased vehicle emissions and construction activities, further degrading air quality.

To address this issue, SERVIR Southeast Asia (SEA), a joint initiative of the U.S. Agency for International Development, the National Aeronautics and Space Administration (NASA), and the Asian Disaster Preparedness Center (ADPC) — its implementing partner, has developed the "SERVIR Southeast Asia Air Quality Explorer" to monitor air pollution and health impacts using satellite data and atmospheric modeling. The application uses advanced data visualization techniques to present complex datasets in an accessible manner. By harnessing the power of satellite data and predictive models, we hope that the SERVIR Southeast Asia Air Quality Explorer (SEA AQE) serves as a valuable resource for policymakers, researchers, and the general public, empowering them to make informed decisions to mitigate the adverse effects of air pollution.

The Air Quality Explorer features a user-friendly interface accessible on both desktop and mobile devices, allowing users to monitor real-time air pollution levels, including three-day forecasts of PM2.5 with a 5 km resolution and NO2 from Geostationary Environment Monitoring Spectrometer (GEMS). The application also features a fire hotspot map, helping users anticipate changes in air quality. It ranks cities over the Southeast Asia regions based on their PM2.5 levels and integrates PM2.5 data with a health index to translate the data into actionable health recommendations. Additionally, the tool includes six-hourly forecast wind data from NOAA and ground station data from Thailand’s Pollution Control Department (PCD), offering a comprehensive view of air quality dynamics across the region.

This project highlights the potential of combining satellite technology and forecast modeling with web-based platforms to improve environmental monitoring and decision-making in Southeast Asia. SERVIR SEA and collaborators will continue to enhance this tool with new useful data, such as high resolution of forecast PM2.5, fire risk, and fire emission inventory products, enabling users to link and analyze these with air pollution indicators. Additionally, the power of large language models (LLMs) will be applied to this tool, allowing users to input queries in natural language. This feature will translate user input into data retrieval commands, providing users with the desired results and making the tool even more accessible and user-friendly. Furthermore, we will develop “SERVIR SEA AQ API” service to provide air pollution satellite image data and json format data for integration on other platforms.

Spaceborne hyperspectral data can assist in estimating crop yields, predicting crop outcomes, and monitoring crops, which ultimately contributes to loss prevention and food security. Recently, EnMAP hyperspectral imagery has become available, starting from 2022. This study aims to analyze and classify oil palm and nipa palm plantations using hyperspectral images combined with machine learning algorithms. The Random Forest classifier, CatBoost classifier, and LightGBM classifier were utilized to automatically map the oil palm and nipa palm areas. The fully workflows of the hyperspectral imaging process were performed in EnMAP-Box plugin on QGIS software. The overall accuracy of three ML classfiers provides greater than 90% especially in oil palm and nipa palm plantations. Machine learning can find out the hidden information about spectral characteristics.

Geospatial Topology, a fundamental concept in geographic information systems, focuses on the analysis and characterization of spatial relationships between geographic entities without alteration of their intrinsic properties. This presentation examines the implementation of Geospatial Topology rules utilizing Python programming language, facilitating execution in both client-side and server-side environments.
We present an empirical case study from the PWA GIS Department of Thailand, which employs a comprehensive set of over 30 topology rule-based validations to ensure data integrity and consistency across national cartographic operations. The research investigates two primary platforms: QGIS Desktop and Web Applications, both serving as client-side interfaces capable of executing topology scripts and generating inconsistency reports for subsequent rectification.
A critical distinction between QGIS Desktop and Web Application lies in their respective execution paradigms: QGIS Desktop operates within a local environment, while the Web Application leverages server-side processing capabilities.
This presentation will elucidate methodologies for developing custom topology rules and demonstrate techniques for accessing single-algorithm Python scripts across both Desktop and Web Application environments. By bridging the technological gap between these platforms, we aim to enhance the efficacy of geospatial data quality control processes and optimize GIS workflows.
The findings of this research contribute to the broader understanding of cross-platform geospatial topology implementation and offer practical insights for GIS professionals and developers seeking to improve data validation processes in diverse computational environments.

A part of the web map application, ensuring functionality and accuracy is paramount to delivering a reliable user experience. This abstract outlines a systematic approach to expectation testing for web maps, focusing on validating that these applications meet predefined criteria and user expectations.

Expectation testing involves setting specific criteria that web map applications must meet to ensure they perform as intended. This process includes validating map data accuracy, user interface responsiveness, and the effectiveness of interactive features such as zooming, panning, and layer management. The goal is to ensure that the web map not only displays information correctly but also responds appropriately to user interactions and integrates seamlessly with other system components.

By implementing a robust expectation testing framework, developers can identify and address potential issues before deployment, ensuring that web map applications deliver accurate and efficient performance. This process not only enhances the quality of the application but also builds user trust by meeting or exceeding their expectations.

The French National Geographic Institute (IGN) came to Cambodia between 1952 and 1954 to carry out a large-scale photographic project, taking around 11,000 aerial images over a large part of the country.

Today, this collection represents an exceptional archive that takes us back 70 years to the urban and rural landscapes of the time. In addition to these early aerial photographs, there is a higher resolution shot of Phnom Penh in 1993, with incredible details of life in the streets of the capital. This archive is of great interest in many fields, including history, geography, archaeology, urban planning, and ecology. The purpose of the project is to dematerialize this archive and make it accessible and usable free of charge in Cambodia.

All the images were digitized and supplied to the KHmer Earth OBServation (KHEOBS) laboratory to create orthophotographs. Processing has been completed for the municipality of Phnom Penh, for the years 1953 and 1993. In order to make these images viewable by anyone, a web application, GeoCambodia, was developed to visualize Cambodia then (past) and now (present). The user-friendly interface includes an interactive slider to navigate and compare the old 1953 and 1993 aerial orthophotographs with the recent Google Earth images. Also, vector outlines of buildings from 1993, produced by the Atelier Parisien d’Urbanisme (APUR), have been integrated to enable visitors to click on a building and view the APUR’s descriptive architectural sheets. Other functions and the extension of the aerial images to the whole of Cambodia are still to come in this interface.

GeoCambodia.org targets anyone who is interested in aerial and satellite imagery and how Cambodia evolves through time and space, especially geography enthusiasts.

Geospatial big data plays a pivotal role in the context of smart grids, revolutionizing the way modern electrical grids are monitored, managed, and optimized. Smart grids integrate advanced sensing, communication, and control technologies to enhance the efficiency, reliability, and sustainability of electricity distribution. While locational information of the smart meters is pivotal, the consumption patterns combined with other information like consumer type, land use and local weather conditions can really enhance the assessment of energy requirements and usage thus leading to a Spatial Energy Management System (SEMS). This will significantly enrich location-aware decision-making, real-time monitoring, and predictive analysis, thus improving energy resource optimisation and achieving the goal of SDG-7. The SEMS web application represents a critical advancement, facilitating dynamic visualization of temporal clusters, across energy sources and their evolution over time. By overlaying clusters across different time periods, utility personnel gain insights into customer categorizations vs actual utilization, enabling understanding of demand fluctuations, outages and faults, fault localization within the electric network and demand-generation assessment of renewable energy.
The primary objective is the development of a dynamic web-based SEMS application capable of visualizing temporal changes and consumption patterns, while also providing alerts to facilitate proactive management.

The initial phase of the methodology focuses on the Jeedimetla region in Hyderabad, leveraging real-world data encompassing 6,000 households categorized into residential, commercial, and industrial segments. Quantum Geographic Information System (QGIS) software is employed to establish the electric network. To store the spatial and temporal consumption data, the data storage infrastructure employs PostgreSQL, enhanced with the PostGIS extension. This combination is selected due to its robust capabilities in accommodating and managing complex datasets. Additionally, a comprehensive and refined data model has been established to serve as the framework for storing Advanced Metering Infrastructure (AMI) data. Spatial data retrieved from the PostgreSQL database is exposed through a Web GIS Server (GeoServer) as a Web Feature Service, facilitating its integration into applications. This spatial information is utilized within the OpenLayers Software Development Kit (SDK) to visually render data within a JavaScript-based application environment. Non-spatial data is accessed through Application Programming Interfaces (APIs) integrated within a Node server, operating as REST endpoints. Concurrently, the React JavaScript library is employed to present this non-spatial information to end-users in an interactive format.

The SEMS Geospatial Visualization Engine comprises three key components. The first component features two primary views: the Basic View, which presents hourly usage consumption data, and the Combined View, integrating map and graph interfaces. This combined view allows users to select specific time periods, customer locations, transformers, or feeders, with the graph view displaying corresponding hourly consumption data. Both views facilitate the visualization of consumption patterns for customer classes or types, which are pre-defined in the system.
The second component of the SEMS Visualization Engine aggregates data at the daily level, classifying users into low, medium, and high consumption categories. These classifications are visually represented on the map view, providing insights into consumption trends across the study area.
The third component integrates weather data sourced from an open-source Weather API with customer energy consumption data. Weather information is stored in the PostgreSQL database, and the combined view enables the graphical representation of weather data changes alongside energy consumption data. This integration enhances understanding of how weather fluctuations impact customer energy usage patterns.

By leveraging advanced sensing, communication, and control technologies, three components of SEMS Geospatial Visualization engine framework provides a comprehensive platform for understanding and visualizing energy consumption and demand. The spatial view aids in developing spatial clusters, which assist electric utility personnel in comprehending consumption patterns and energy demand requirements for specific areas, facilitating efficient management of power shortage scenarios.

The SEMS visualization engine supports various utility use cases, including energy loss identification, detection of energy overuse, and dynamic reclassification of users based on current consumption data. As a prospective avenue for future research, integrating this Geospatial Visualization Engine framework with machine learning-based prediction models holds promise for forecasting energy consumption and dynamically classifying users. Such advancements are anticipated to further empower utility personnel in decision-making, real-time monitoring, and predictive analytics within smart grid environments.

Species Distribution Modeling (SDM) is a statistical methodology used to predict the spatial and temporal distribution of species based on environmental conditions that are conducive to their survival and reproduction. This modeling approach leverages spatially explicit species occurrence records alongside various environmental covariates, including climate, terrain, and land cover, as input variables, with the aim of quantifying and mapping species-environment interactions. SDM has become a critical tool in ecological research and conservation biology for understanding and predicting species distribution patterns. A range of machine learning and deep learning techniques can be employed in SDM, such as Logistic Regression (LR), Random Forest (RF), Multilayer Perceptron (MLP), Convolutional Neural Network (CNN), and Generative Adversarial Network-CNN (GAN-CNN). Despite the availability of these techniques, there is a lack of a comprehensive application that integrates these algorithms for species distribution modeling. To address this gap, this paper introduces a new tool, kari-sdm, which enables users to perform SDM utilizing a variety of techniques. Kari-sdm supports LR, RF, MLP, CNN, and GAN-CNN algorithms, all based on open-source frameworks PyTorch and scikit-learn. Additionally, it facilitates all necessary preprocessing steps, from data collection, cleaning, transformation, spatial preprocessing, and environmental variable selection, to data splitting. The tool also provides functions for model evaluation, result visualization, and cross-validation. The primary goal of kari-sdm is to assist ecologists in modeling species distributions, interpreting results, and developing informed conservation and management strategies.

Geographic information technology has become an important part of our daily lives, leading to the development of various map applications. The design of these applications requires special attention, particularly in terms of User Experience (UX) and User Interface (UI) design. Effective UX/UI design significantly impacts user satisfaction and ease of use, contributing to the overall efficiency and modernity of map applications. By understanding user needs through User Research and Usability Testing, we can establish principles and guidelines for creative design that enhance usability and ensure consistency between data presentation and user interaction.

Creating a map application that offers a good user experience involves careful design of map elements, such as the layout of basic map tools, the selection of symbols, the arrangement of data, and the design of interactions. The design must be suitable for different types of usage, such as travel applications, survey applications, or application for specific use. Additionally, it must support various devices, including mobile phones, computers, tablets, and other devices with different screen sizes.

Furthermore, techniques that can be applied in the design process are presented in this session to achieve the best results. This includes creating an immersive user experience by strategically using colors, fonts, and layout. Developing engaging and interesting ways to display information will help users feel more connected to the application. It is also important to stay updated with current trends in map application design to ensure that the developed applications are modern and responsive to global changes.

This presentation is suitable for designers, developers, and anyone interested in geographic information technology, especially those involved in developing map applications. It will provide design guidelines that can be applied to various projects, effectively meeting the diverse needs of future users.

In this session, I will introduce a module I developed for clustering icon image markers using MapLibre, with the theme of "finding nearby restaurants that look delicious." This is the first module I’ve ever created, and it’s designed with a very simple structure. Through this session, I hope participants will take away two key messages: that creating what you want to make is enjoyable, and that even if you’re not highly skilled, sharing your work can lead to valuable learning experiences.

Open spatial data plays a transformative role in the landscape of higher education in Thailand, bridging the gap between theoretical learning and practical application in daily life. This keynote address will explore how open spatial data is being integrated into the higher education curriculum, emphasizing its significance in enhancing students' understanding of geography, urban planning, environmental management, and related disciplines.

Illustrating how open spatial data is increasingly influencing daily life in Thailand through community-driven mapping projects, public health initiatives, or sustainable development planning, open spatial data is empowering. The session will also address the challenges and opportunities in the widespread adoption of open spatial data within higher education, including issues of data quality and accessibility and the need for ongoing support and collaboration between academic institutions, government agencies, and the private sector. Ultimately, this session aims to inspire educators, students, and professionals to harness the power of open spatial data, transforming education and society at large in Thailand.

Global Ecosystem Dynamics Investigation (GEDI) and Ice, Cloud, and Land Elevation Satellite 2 (ICESat-2) are earth observation missions from NASA to construct a three-dimensional model of earth surface in space and time empowered by Light Detection and Ranging (LIDAR). GEDI and ICESat-2 data are organized by orbit ID, sub-orbit granule and track, and distributed in HDF5 format, which is optimized for big data storage. However, this approach is inconvenient for extracting spatio-temporal areas of interest, because each file stores a track crossing a huge range of latitude and longitude, while lacking a spatial index.

To facilitate random access to small areas of interest, we propose a data reconstruction process through Apache Parquet. Parquet is an open source column-oriented data format designed for efficient data storage and retrieval. We sequentially stream raw data into spatio-temporal partitioning blocks (5 degree x 5 degree x year). This layout optimizes the number of partitions (n = 3337) and individual file size (~300 MB). Independence of raw data files and a predefined partititoning scheme enables parallel processing, and periodic update while new data is available.

During the reconstruction, we selected essential attributes and applied quality filtering based on scientific literature. We excluded GEDI shots with Quality Flag equal to 0, Degradation Flag larger than 0, or Sensitivity smaller than 0.95; For ICESat-2 ATL08, we first excluded segments where terrain and/or canopy height are in NaN. We then reconstructed individual photons from ATL03 by ph_segement_id, and excluded the ones classified as noise, as well as segments containing more than 28 photons, according to the result from previous research [1].

Data is finally converted to GeoParquet and published on a cloud server under CC-BY 4.0 license. GEDI Level2 has 1.4 TB, and ICESat-2 ATL08 has 3.8TB in total size respectively. GeoParquet supports two levels of predicate push down: first, at the partition level, and second, at the file level. The partitioning of the global LiDAR datasets enables coarse spatial (5 x 5 degree) and temporal (year) filtering.. The footer of each GeoParquet file enables spatial filtering via bounding boxes or geometry features, and temporal filtering using the datetime columns. Further attribute filtering is possible.

The concept of Analysis-Ready Cloud Optimized (ARCO) data has been defined and implemented for raster data, using technologies such as Zarr or Cloud Optimized GeoTiff (COG) [2]. However, corresponding implementations for vector data are scarce. This work delivers two instances of global ARCO vector datasets. It not only adheres to the concept of 4C (complete, consistent, current, and correct), but also tackles the challenge of organizing terabyte-scale geospatial vector data.

Reference

Milenković, M., Reiche, J., Armston, J., Neuenschwander, A., De Keersmaecker, W., Herold, M., & Verbesselt, J. (2022). Assessing Amazon rainforest regrowth with GEDI and ICESat-2 data. Science of Remote Sensing, 5, 100051.
Stern, C., Abernathey, R., Hamman, J., Wegener, R., Lepore, C., Harkins, S., & Merose, A. (2022). Pangeo forge: crowdsourcing analysis-ready, cloud optimized data production. Frontiers in Climate, 3, 782909.

The rapid development of autonomous vehicles (AVs) demands strong perception systems capable of reliably recognizing and classifying objects in complicated urban environments. To improve the reliability and precision of 3D object identification, combining camera and LiDAR sensors has emerged as a potential solution. This research describes a multimodal fusion framework that uses camera pictures and LiDAR point clouds to accomplish high-performance 3D object recognition in urban circumstances. Camera sensors provide precise color and texture information necessary for identifying traffic signs, pedestrians, and cars, whereas LiDAR provides exact depth measurements required for interpreting object geometry and spatial relationships. The indicated fusion technique improves detection accuracy by using these sensors' enhancing strengths, especially in difficult settings such as occlusions and fluctuating illumination conditions taking the KITTI open dataset. Here the OpenPCDet is used for 3D object detection and MMDetection for 2D detection as open library. Beside this the Facebook AI Research, Detectron2 flexible framework for 2D and 3D object detection tasks is more popular too.

Fusion is accomplished via a well built architecture that aligns and combines data from both modalities at various stages of the detection pipeline, such as feature extraction, region proposal, and classification. Advanced deep learning techniques, such as convolutional neural networks, are used to process and integrate multimodal input. Experimental results show that 3D object detection outperforms single-modality techniques in terms of robustness and precision, particularly when recognizing small and partially occluded objects.

This study aims to find the results of using a free disaster mapping application developed on Google Earth, named “Hazmapper”. This tool allows users to create maps and GIS products from Sentinel or Landsat datasets without the high time and cost usually needed for traditional analysis.

The initial design of the HazMapper program used indicators based on the Normalized Difference Vegetation Index (NDVI). Specifically, it developed the relative difference NDVI (rdNDVI) to identify areas where vegetation was removed after natural disasters. Because these indicators rely on vegetation, HazMapper is unsuitable for desert or polar regions, which means appropriate for tropical areas.

Using the rdNDVI indicator for different years in the same area and comparing the average absolute error (MAE) of all results to test the effectiveness of the Hazmapper model in application to flooded areas.

Hydrometeorological data are crucial for effective water resource management, weather forecasting, and climate adaptation. In Asia, a region known for its vast geographic diversity and varied climatic conditions, the lack of high-resolution data has been a significant challenge for addressing local-scale issues. Traditional datasets often have coarse spatial resolution (e.g., 10 – 25 km), limiting their usefulness for detailed, localized analysis. To address this gap, we have developed a pioneering dataset offering 1 km resolution hydrometeorological data for the entire Asian continent. This dataset includes essential variables such as precipitation, surface temperature, radiation, soil moisture, evapotranspiration, groundwater, and surface runoff delivering unprecedented detail and accuracy compared to existing coarse-resolution data. The dataset was created using advanced remote sensing techniques, land surface physics, and sophisticated data assimilation methods, ensuring both enhanced spatial resolution and accurate reflection of local conditions. Our dataset spans from 1940 to the present, providing a comprehensive historical archive and seasonal forecasts extending up to six months into the future. This combination of historical and predictive data makes it an invaluable resource for a variety of applications, including water resources, climate studies, agriculture, and disaster assessment. To validate the accuracy of our high-resolution data, we conducted extensive comparisons with satellite remote sensing products such as MODIS (Moderate Resolution Imaging Spectroradiometer), GRACE (Gravity Recovery and Climate Experiment), and SMAP (Soil Moisture Active Passive). These comparisons confirm that our dataset offers superior accuracy and finer detail compared to publicly available data.

It is crucial for enhancing the efficiency and accessibility of land use and land cover monitoring, supporting informed decision-making in urban planning, environmental management, and sustainable development across diverse geographical contexts. This aims to streamline the process of satellite data acquisition, preprocessing, and seasonal LULC classification through the integration of artificial intelligence (AI) models.

The proposed system will consist of two main components: (1) an automated satellite data fetching and preprocessing module, and (2) an AI-driven LULC classification module. The first component will leverage open-source tools to access and prepare satellite imagery from freely available sources, such as Landsat and Sentinel missions. This module will handle tasks including data download, atmospheric correction, cloud masking, and image compositing.

The second component will employ state-of-the-art machine learning algorithms, particularly deep learning models, to perform seasonal LULC classification. The system will be trained on diverse datasets to recognize and categorize various land cover types across different seasons, accounting for temporal variations in vegetation, urban expansion, and other dynamic landscape features.

By automating these processes, the proposed system aims to significantly reduce the time and expertise required for LULC analysis, making it more accessible to researchers, urban planners, and environmental managers. The use of free and open-source software ensures that the developed tools will be widely available and customizable for different geographical contexts and research needs.

This contributes to the advancement of remote sensing applications and supports informed decision-making in land management, urban planning, and environmental conservation efforts.

This study explores the integration of Natural Language Processing (NLP) and Geographic Information Systems (GIS) to analyze the spatial distribution and sentiment of restaurants based on TikTok data. Data was collected from TikTok using primary and secondary hashtags related to restaurant reviews in Bangkok. The resulting database enabled a detailed analysis of restaurant locations and customer sentiment using Logistic Regression for sentiment analysis. The findings indicate that negative reviews were predicted with the highest accuracy (84%), followed by positive (78%) and neutral (76%) reviews. The spatial analysis identified a dense of restaurants in the inner districts of Bangkok. This integration of NLP and GIS not only mapped the popularity of restaurants as mentioned on TikTok but also provided significant insights into consumer behavior and preferences. The study demonstrates the effectiveness of combining NLP and GIS for geospatial analysis, offering a powerful tool for understanding social media trends and their impact on local businesses. The results underscore the potential for leveraging social media data to inform urban planning and business strategies, particularly in the context of food and hospitality industries.

When considering the sustainability of open source as a future digital public good, the cycle of success and contribution in the business sector becomes crucial.

MIERUNE was born from the FOSS4G community in 2016 and has spent the past eight years solving client challenges as a location-based systems integrator, growing steadily in the Japanese market.
Through these business activities, MIERUNE not only actively gives back to the FOSS4G community—both technically and financially—but also creates local employment for GIS engineers, helping to build a sustainable society.

In this presentation, we will share specific examples of the challenges we have faced to support the growth and development of FOSS4G companies in the Asian region, thereby contributing to the sustainability of the whole community.

Managing Geographic Information Systems (GIS) data with H3 (H3Geo) is an efficient and modern method for handling and analyzing geographic data. H3 uses a hexagonal grid system that offers special features, allowing data to be stored at resolution levels from 0 to 14, which helps in dividing and storing data effectively.

TBA

first ventured into the world of open-source GIS in 2003. Back then, the landscape was much different—GRASS GIS and MapServer were known names, but familiar tools like QGIS and PostGIS were still in their infancy, with limited functionality and a small user base.

It was in this early era that Dr. Venkatesh Raghavan coined the term FOSS4G—a milestone now celebrating its 20th anniversary since that pivotal moment in 2004. Just a few months later, in September, the world’s first FOSS4G conference was held at Chulalongkorn University in Bangkok, where Europe’s GRASS GIS community met the MapServer community from North America. This historic gathering laid the groundwork for the creation of the OSGeo Foundation.

In this keynote, I will take you down memory lane, sharing the story of how FOSS4G was born and how it evolved into the vibrant ecosystem we know today. Join me as I recount these foundational moments through storytelling, from a perspective shaped by those pioneering days.

The GRASS GIS project, a pioneering open-source geographic information system, celebrated its 40th anniversary in 2023. As one of the long-standing contributors, I am honored to reflect on the remarkable journey of this leading open-source geospatial software and community. Over the past four decades, GRASS GIS has grown from a modest project initiated by the U.S. Army Corps of Engineers to a robust, globally recognized platform for geospatial analysis and modeling. This evolution is a testament to the dedication and collaborative spirit of the GRASS community, which has continually driven innovation and excellence.

My personal relationship with GRASS GIS began over thirty years ago when I was a student and first encountered its powerful capabilities. Even then, I was fascinated by its potential to revolutionize spatial analysis and environmental modeling. With the advent of the Internet, we were able to build a passionate community behind the project. Through collaborative efforts, we have significantly expanded the functionality of GRASS GIS, improved its user interface through multiple iterations, and ensured its adaptability to the ever-changing technological landscape.

In this keynote, I will reflect on the milestones that have shaped GRASS GIS from its inception at the U.S. Army Corps of Engineers' Construction Engineering Research Laboratory (USA/CERL) to its current status as a cornerstone of the open-source geospatial ecosystem. The latest releases of GRASS GIS include thousands of changes, including the new single-window GUI layout and enhanced parallelization capabilities. These enhancements underscore our commitment to improving the user experience and computational efficiency. The past decade has also been marked by vibrant community engagement through the OSGeo Foundation. I will highlight key contributions from the global community, showcase groundbreaking research and applications, touch on FOSS business models, and explore the challenges we have overcome along the way.

The future of GRASS GIS is bright as we anticipate further innovation and expanded applications, driven by the same collaborative ethos that has defined our past. Together we will continue to push the boundaries of what is possible in geospatial analysis, ensuring that GRASS GIS remains at the forefront of this dynamic field.

Water does not flow according to geographical boundaries but it follows
elevation. Inland waters from rivers and lakes present crucial natural
resources playing an indispensable role in the global hydrological cycle.
Still, their conventional monitoring is constrained by poor spatial
coverage. Though satellites help improve coverage, the hydraulic properties
of rivers often change at a rate faster than the temporal sampling of
satellites. Through this talk, two novel web applications are introduced
for rivers and lakes built using geospatial datasets and python. While the
first shall seamlessly extract time series of water surface area for
rivers, lakes and reservoirs from Sentinel-1 VV polarized SAR data. The
second application shall integrate dynamic lake water extents for improving
lake water surface temperatures, thereby challenging conventional norms.

Enhancing the efficiency of the Electricity Generating Authority of Thailand (EGAT)’s transmission system through satellite imagery and LiDAR. With over 470 transmission lines spanning 25,000 km, expansion plans aim to provide nationwide coverage. LiDAR is used for aerial mapping to create an accurate initial database, while CCTV systems enable real-time surveillance. Satellite imagery, particularly SAR satellite imagery combined with InSAR and machine learning, ensures precise monitoring of encroachments. Integrated GIS application support right-of-way management, contributing to safe and efficient power distribution.