1:45 PM - 2:00 PM
[HTT21-01] Map representation of wind distribution model in Kobe city, Japan
Keywords:Wind distribution model, Kriging, Vector field, QGIS, GNU R, gstat
Wind distributions are important information when examining the concentration distribution of air pollutants such as PM2.5. Since the number of wind observatories is limited, it is common to use numerical simulation to obtain a wind distribution model (vector field) on a city-wide scale, but in this research, the wind distribution model in Kobe City Hyogo Prefecture, Japan was estimated from the observed values using the Kriging method, one of the modeling methods of geostatistics. Then, a method of map representation of the obtained wind distribution model with GIS was examined.
The target area was Kobe City, and the wind observation data used was the 2018 data (monthly average) released by the Japan Meteorological Agency. For geostatistical analysis, GNU R (R core team, 2022) and gstat (Pebsma, 2004), a package of the Kriging method, were used. The estimation grid was the 3rd cartographic mesh codes (1000m pitch) for the target area (Statistics Bureau, Ministry of Internal Affairs and Communications, 2022). QGIS (QGIS Development team, 2022) was used for map representation.
The spatial dependence of the wind observation data was estimated by the variogram, then the wind distribution model and its variance model in the entire Kobe city were obtained by the Kriging method. These models were mapped using QGIS. As a result, the singularity of the wind distribution could be found in the western part of Kobe city. In addition, when these models and JPGIS DEM (Geospatial Information Authority of Japan, 2022) were combined to examine a PM2.5 distribution model in Kobe City at the same time, it was suggested that the PM2.5 concentration distributions in the western, eastern and northern parts of the city were based on each different geoscientific mechanism.
The map representation of the wind distribution model using GIS facilitated qualitative comparisons and examinations with other geographic information or geostatistical models. Since the map representation of vector fields are widely needed in the field of earth and planetary science, it is expected that this method will become widespread.
The target area was Kobe City, and the wind observation data used was the 2018 data (monthly average) released by the Japan Meteorological Agency. For geostatistical analysis, GNU R (R core team, 2022) and gstat (Pebsma, 2004), a package of the Kriging method, were used. The estimation grid was the 3rd cartographic mesh codes (1000m pitch) for the target area (Statistics Bureau, Ministry of Internal Affairs and Communications, 2022). QGIS (QGIS Development team, 2022) was used for map representation.
The spatial dependence of the wind observation data was estimated by the variogram, then the wind distribution model and its variance model in the entire Kobe city were obtained by the Kriging method. These models were mapped using QGIS. As a result, the singularity of the wind distribution could be found in the western part of Kobe city. In addition, when these models and JPGIS DEM (Geospatial Information Authority of Japan, 2022) were combined to examine a PM2.5 distribution model in Kobe City at the same time, it was suggested that the PM2.5 concentration distributions in the western, eastern and northern parts of the city were based on each different geoscientific mechanism.
The map representation of the wind distribution model using GIS facilitated qualitative comparisons and examinations with other geographic information or geostatistical models. Since the map representation of vector fields are widely needed in the field of earth and planetary science, it is expected that this method will become widespread.