[U-04] Geospatial Applications for Societal Benefits

[U04-P04] Estimation of water vapor distribution by near-infrared imaging from small satellites

*Haruki Sugiyama¹, Yukihiro Takahashi¹ (1.Department of Cosmosciences, Graduate School of Science, Hokkaido University)

Keywords:Water Vapor, multispectral imaging, sattelite

In recent years, torrential rains caused by well-developed cumulonimbus clouds in the Southeast Asia region have caused significant damage. In order to reduce these damages, it is necessary to predict the development of cumulonimbus clouds that bring torrential rains earlier and more accurately. Although numerical forecasting has been widely used in the past, it is difficult to predict mesoscale phenomena such as torrential rains, and as shown by Bryan et al, However, the time and spatial resolutions of observation equipment are insufficient for the lidar observations and satellite microwave observations that have been conducted so far. Therefore, observations with a high spatial resolution of about 100 m would be effective in improving forecasts. In this study, near-infrared multi-wavelength imaging observations around Manila, Republic of the Philippines, were conducted on March 19, 2023, using the Spaceborne Multispectral Imager (SMI) onboard the DIWATA-2B satellite. This time, we focused on the water vapor absorption band between 710 and 744 [nm], which is weaker than the general water vapor observation band around 1.4 [μm], but it allows us to observe the distribution of water vapor near the ground without saturation. The equivalent width of water vapor absorption was calculated from the imaging data in this band and the observations in the bands around the absorption. In order to calculate the equivalent width, it is necessary to accurately estimate the continuum component of the ground surface reflection spectrum as a reference. In this study, images taken at 671 [nm], 710 [nm], 749 [nm], and 780 [nm], which are absorption-free bands around water vapor absorption, were aligned and a quadratic polynomial was fitted to the value of each pixel using the least-squares method. component was calculated. Data from the P-POTEKA observation network, which has 39 stations around Manila, was used to link the equivalent width to the water vapor content. The absolute humidity distribution on the ground obtained by this method was compared with the distribution of equivalent widths obtained from satellites, and an attempt was made to calibrate the amount of water vapor. In this presentation, we will present the results of the analysis to date and future tasks.