The spatial interpolation of precipitation has been employed in a number of fields, including by spatially downscaling the Global Circulation Model (GCM) to a finer scale. Most precipitation events become more sporadic when the coverage area increases (i.e., a portion of the points experience zero precipitation). However, spatial interpolations of precipitation generally ignore these dry areas, and the interpolated grids are filled with certain precipitation amounts. Subsequently, no delineation of dry and wet regions can be made. Therefore, the current study suggested a novel approach to determine dry areas in spatial interpolations of precipitation events by assigning latent negative precipitation (LNP) to points with observed precipitation values of zero. The LNP-assigned points are then employed in a spatial interpolation. After that, the dry region can be determined using the negative region (i.e., points with zero precipitation). The magnitude of LNP can be defined by multiplying the precipitation values of neighboring stations by a tuning parameter. The LNP method and the tuning parameter are tested on weather stations covering South Korea. The results indicate that the proposed LNP method can be suitable for the spatial interpolation of precipitation events by delineating dry and wet regions. Additionally, the tuning parameter plays a special role in that it increases in value with longer precipitation durations and denser networks. A value of 0.5-1.5 can be suggested for the tuning parameter as a rule of thumb when high accuracy for final products of interpolated precipitation is not critical. For future studies, the LNP model derived herein can be tested over much larger areas, such as the United States, and the model can also be easily adopted for other variables with spatially sporadic values.
© 2021. The Author(s).