Abstract:
Flow accumulation derived from terrain analysis is one of the most important hydraulic application. Detailed elevation model can be acquired with advanced technique e.g. Lidar. Thus, the modern digital elevation model is refined and its amount is often massive. Mass DEM may prohibit computing efficiency and accuracy of the modeling result as well. This research emphasizes development of preparation and processing technique for massive DEM from Lidar. Two kinds of hydraulic software packages name ArcGIS/Hydro and QGIS/GRASSS are used for computing of flow accumulation in the test area. Comparison study is then conducted from those results. In the study original massive DEM has resolution of 1-meter and then is reduced into 2-, 5- and 10-meter. The amount of point clouds for each model is 304.9, 67.3, 10.7 and 2.7 million respectively. During preparation step, detail hydraulic features e.g. elevated structures are then integrated. The reduced resolution DEM, 5- and 10-meter DEM obviously produces high difference and may lead to unusable result. The DEM with 2-meter resolution results flow accumulation nearly the same as the original 1-meter DEM with difference of the flow accumulation error 6.21%. Whereas the discrepancy of the flow lines error measured by root mean square error, RMSE is 151.7 meter. The time for flow accumulation calculation of the 2-meter DEM took 1:45 and 21.24 minutes for QGIS/GRASS and ArcGIS/Hydro respectively. With this test dataset QGIS shows better performance than ArcGIS/Hydro, so far as the amount of data still fit within the computer memory. The resolution of DEM is a direct impact on the results of the resolution. Stream lines generated from both software are also spatial visually analyzed and are quite agreed with satellite imageries shown on Google Earth.