Mapping Flood-Prone Areas Using GIS Through as Geo-Artificial Intelligence (Geo-Ai) Approach in Bengkulu City

ABSTRACT


Introduction
A flood is an inundation of water, from the smallest to the largest.caused by both human and natural factors or high water flow, and not accommodated by the river flow so that the water overflows to more land [1].A Flood is defined as a mass of water that is produced from relatively high and uncontainable groundlevel runoff that overflows naturally and that overflows naturally and causes inundation [2].Flooding is a natural phenomenon that is closely related to the hydrological cycle.River water that flows beyond the capacity of the river will pass through the river and cliffs flooding the area [3].The occurrence of floods is also caused by poor soil permeability so that it is no longer able to absorb water [4].Floods occur as a result of the influence of heavy rains and natural waterways that cannot hold water [5].Areas with low-lying conditions and near rivers tend to have higher flood vulnerability [6].
Flooding has become a frequent natural disaster.This is because climate change has become a concern all over the world in recent years causing flooding [7].Flooding cannot be avoided, but it can be controlled and the impact of flood losses can be reduced [8].It is therefore important to have access to reliable and up-todate information to prevent or at least mitigate the effects of flooding.Potential flood maps are one type of important information [9].Floods can have negative socio-economic impacts, loss of people and property, health-related problems, and ecosystem functions [10].Public awareness about the importance of protecting the environment and behavior in disposing of waste properly are also important factors in preventing flooding [11].Lack of awareness and environmentally unfriendly behaviors can cause drainage channels to clog and exacerbate flooding [12].The ability to handle disasters, including floods, is a crucial issue [13].
This research is based on the phenomenon or disaster of flooding that occurs every year in Bengkulu city.Lately, there have been frequent storms or strong winds suddenly accompanied by heavy rain, which causes flooding.The disaster occurred after heavy rains caused the Bengkulu River to overflow.The challenges of dealing with climate change and rapid urban growth demand greater and more integrated efforts in managing flooding in Bengkulu City.
In general, studies on floods have been carried out by researchers Previously, it showed that areas in Bengkulu City had the potential to experience floods.However, previous research still managed flood disaster data which is still done manually so the data results are less accurate.The parameters used are very few, so mapping research is necessary for flood-prone areas in Bengkulu City.
Therefore, it is necessary effective efforts in identifying flood-prone areas to plan appropriate mitigation measures to reduce its impact.One of the innovative solutions is to utilize Artificial Intelligence methods and Geographic Information Systems (GIS) based on the measured parameters to obtain flood-prone maps.As time goes by, human abilities will be replaced by smart or intelligent machines in various fields.This smart machine is commonly called artificial intelligence or Artificial Intelligence, which is a part of computer science [14].
Geo-AI is a combination of artificial intelligence (AI) and geospatial technology, which aims to use geospatial data and perform intelligent analysis to make decisions or determine the identification of certain patterns related to geographic locations [15].In Geo-AI, artificial intelligence technologies are used to process, analyze, and understand large-scale geospatial data with greater efficiency and accuracy [16].
Geospatial data includes information that information related to a particular location or area, such as maps, satellite data, drone data, sensor data, and others related to the geography and topography of an area [17].Utilization of this AI technology can help process hydrological, topographical, and rainfall data to identify areas with high flood risk and plan effective mitigation strategies [18].With AI technology, spatial data can be processed quickly and accurately to facilitate decision-making in flood mitigation planning in Bengkulu City.A geographic information system is a computer-based system that can handle geo-referenced data [19].The hardware used for GIS is a computer (PC), mouse, digitizer, printer, plotter, and scanner [20].GIS can connect various data at a certain point on earth, combine them, and analyze and map the results.[21].
This research aims to map flood-prone areas by utilizing a Geographic Information System (GIS) through a Geo-Artificial Intelligence (Geo-AI) approach.The results of this research are expected to be the initial basis for flood disaster mitigation in Bengkulu City.

Materials
The data used in this study are DEMNAS data, Bengkulu City Shapefile data, and sentinel channel 1.This dataset provides a detailed global map and associated information on the extent of surface waters around the world.It includes information on lakes, rivers, swamps and other surface waters with a spatial resolution of about 10 meters.The sentinel data used in this research is data from 2019-2022.Data processing and analysis were carried out using Microsoft excel software, GIS software, and Google Earth Engine (GEE).For DEMNAS data, and Shapefile data of Bengkulu city are processed in Arcgis Softwere.

Processing of DEMNAS data and Bengkulu city Shapefile data
For data DEMNAS data, and Shapefile data of Bengkulu City were processed in ArcGis Software.The stages carried out, namely, select Add data throughout the DEMNAS data grid, then Open ArcToolbox, point to Raster > Raster Dataset > Create Raster Dataset.This stage creates raster data with the parameters needed for the digitization process.Classify the data into spatial data, and the data is organized into tables that will be used to process the data.Select Raster Calculator to run the AI-jabar expression that will produce the output value.After the digitization process, the mapping for the research location was done.Shapefiles of the data used are in RAR format, and other data are in XLSX format.After doing the digitization and mapping, it was continued with data processing in Microsoft Exel to determine the value of the parameters that have been created.

Data analysis on Google Earth Engine (GEE)
Data analysis on Google Earth Engine (GEE) to determine flooded areas with the modified Otsu algorithm and AI Script.For the first step, search first on Google Chrome, namely the Google Earth engine, then click the platform then select the code editor if loading is complete.editor if loading is complete point the map to the research location, namely Bengkulu City.If you have clicked on assets then new, then select save file.After that, select the folder, the folder must be zipped, then Open and upload.Navigate to the text section where the upload process is being processed if the upload process is complete mark with the writing viewashed then click and make sure the uploaded file is the correct file if you are sure it is correct directly import the file.After that, it will appear in the new script section then search sentinel-1.At the very top click import after that it will appear in the new script in the table section changed to roi then in the image collection section change to S1 for the next step copy and paste the AI script that has been created if you have clicked Run then select tools console.In this tool, the image that will be used zero image is used.Then for the ID don't forget to copy it next will call one of the Sentinel images on the one that is adjusted to the image that was that was already selected.
If you have finished clicking Run, the next step is making permanent water data by using Hansen data and then filtering and updating the mask.Next, we will plot the chart on the console using the Otsu algorithm.The last step is to classify the image and export data.If you click Run, the results will appear in the form of a display of flood-prone areas in Bengkulu city using the Otsu algorithm.In addition, the console window can be seen in the resulting graphs and the threshold value of flooding in the image used.

Results and Discussion
Utilization of Geo-Artificial Intelligence (Geo-AI), Google Earth Engine (GEE) and Geographic Information System (GIS) to identify flood-prone areas is an innovative and highly effective combination of flood-prone areas is an innovative and highly effective combination for addressing flooding issues and improving mitigation efficiency [22].Geo-AI leverages artificial intelligence technology to analyze geospatial data and discover relevant patterns and trends.GIS handles geo-referenced data, i.e. data entry, and data management, and Google Earth Engine provides a very fast infrastructure for processing geospatial data at scale [23].
Utilization of Geo-AI, GIS, and GEE to identify flood-prone areas provides significant benefits, namely the use of Geo-AI with an algorithm allows the identification and mapping of flood-prone areas with high accuracy, and reduces the risk of errors.Earth Engine provides a powerful infrastructure for processing large-scale geospatial data, enabling fast and efficient analysis.With an AI location-based early warning system, authorities and communities can quickly receive warnings about potential flooding, so that preventive and evacuation measures can be taken quickly.
Geo AI data and analysis can provide a strong basis for better and more informed decision-making in the face of disasters.Geo-AI can monitor and predict floods more accurately based on real-time satellite imagery and weather data, helping proactive and more timely weather data, helping proactive and more timely management [24].The use of Geo-AI in combination with, GIS, and GEE to identify flood-prone areas provides enormous benefits in mitigating and limiting flood risks.With satellite image analysis, flood forecasting, early warning systems, and risk assessment, Geo-AI and Google Earth Engine provide an effective, accurate, and efficient solution to the flooding challenges in the modern era.The result of this research is a map of areas that are prone to flooding in Bengkulu City. Figure 3 is a map of Bengkulu City.In the picture, it can be seen that some are white and some are red.The white part shows that the area is a safe area from flood disasters, while the red part shows areas that are prone to flood disasters.The red color explains that the area is often inundated by water.The map above shows that flood-prone areas in Bengkulu City include Rawa Makmur, Tanjung Agung, Bentiring, Kebun Tebeng, Penurunan, Sukarami, Pekan Sabtu and Air Sebakul.

Conclusion
Research Utilization of Geo-AI for mapping flood-prone areas in Bengkulu City has many benefits in the effort to deal with flood disasters.The research results show that areas that are prone to are those that are colored red.The red color can be interpreted as the area is often inundated by water so the area is prone to flooding.With more accurate and effective data analysis, this research can contribute to mitigating flood disasters in Bengkulu City, protecting the community of the environment from impacts of the environment from the impact of flooding, as well as creating a safer and more sustainable city.sustainable city.In addition, this research can open up opportunities for the development of Geo-AI technology and can serve as a reference for similar research in other areas.for similar research in other areas that are also prone to flooding.

Figure 2 .
Figure 2. Display of data processing in Google Earth Engine.

Figure 3 .
Figure 3. Map of flood prone areas in Bengkulu city.