In order to identify the areas vulnerable for flooding, then it is necessary to calculate total runoff of the area or quantum of water flow coming during a particular season. To calculate this, it would be easy to adopt Curve Number method and areas vulnerable for flooding can be mapped by delineating the low lying areas from DEM. Many recent R&D studies are available as published papers online for reference.
You will need all the necessary data to estimate flood Like catchment characteristics, constants, Discharge values from river guaging stations, catchment area, Rainfall data, Evaporation and evapotranspiration values. Moreover you will need extreme flood values to decide the extent of flood plain using DEM. Uptill now I have not used ARC GIS but the requirement of these parameters shall be necessry. Refer GRAMM++ and HYMOS as well
You can use the DEM information along with Rainfall in the catchment (means representative value for rainfall), and observed runoff from catchment to build a rainfall-runoff model for the catchment. There are many rainfall-runoff models like physical models (distributed models) or conceptual model (global/point models) or statistical models which can be useful to determine flood estimates corresponding to certain extreme rainfall event. Once the model is built (calibrated) for observed rainfall and runoff events, same model can be used to find flood estimates by using specific rainfall inputs (PMP storms, standard storms etc).
In your case you want to map flood plain, so better choice for rainfall-runoff model would be a distributed physical rainfall-runoff model. One of the distributed rainfall-runoff models is SWAT. Further you can use SWAT model in ArcGIS framework which is famously known as Arc-SWAT. Apart from DEM information and LULC information, you may need to consider atmospheric variables such as temperature, radiation, vapour pressure and so on to feed into the SWAT model. These weather variables are required to form weather generator for the catchment or study area. In addition the variables would be useful to compute evaporation and evapotranspiration etc. The atmospheric variables can be obtained from NCEP data. As per my knowledge, weather generators for Indian study areas to use in ArcSWAT model are already available online. These weather generators are constructed based on the same NCEP data that I have described previously. In ArcSWAT, you can opt for various methods for losses such as evaporation and evapotranspiration which are available inherently in the model.
- ArcGIS is not a software for flood inundation mapping. You can use ArcGIS extension of HEC-GeoRAS for DEM processing and perpetration only. The main engine for flood plain mapping in this case will be the HEC-RAS which is independent from ArcGIS.
- You can also choose other software like Flood Modeler Pro or many others which can provide 1D and 2D solution for flood plain mapping.
Sir, after collecting the rainfall datas you need to find the runoff and discharge values,then only you can proceed further.For obtaining those values curve number is a good method.Before that soil map and land use map is a must.Then for the area selected different values of CN can be found which would help in runoff and discharge calculation.Then you can use FLO-2D model for hydraulic modelling further.
I also suggest you to see the trend analysis of flow data at least for thirty years which also help us to find the extreme conditions. The LU/LC also can be taken for consideration....
you can follow this paper for landscape vulnerability to flood hazard assessment https://www.researchgate.net/publication/294089869_Assessing_flood_inundation_extent_and_landscape_vulnerability_to_flood_using_geospatial_technology_A_study_of_Malda_district_of_West_Bengal_India
Article Assessing flood inundation extent and landscape vulnerabilit...