Since air temperature maps are essential tools in various fields, there are several studies to estimate surface air temperature at unobserved locations. Air temperature maps have also practical use in a variety of agricultural fields. Practical methods for estimation of surface air temperature are currently not physical but statistical methods. Recent studies on air temperature map development using statistical methods are classified according to method type: (1) rasterizing point data using general techniques for interpolation within computer software such as geographic information systems, i.e., "general rasterization methods"; (2) considering geographical functions based on several factors, i.e., "geographical function methods"; and (3) interpolating anomalies between observation data and long-term normals, i.e., "anomaly methods." Geographical function or anomaly methods should be more suitable when interpolating current conditions. It is difficult, however, to select a single method, because the most suitable method depends on the area. Practical types of spatial resolution for agriculture are suggested, i.e., high resolution less a few hundred meters, herein called "precise data," and low resolution greater than that scale, called "coarse data." Substantial data at specific sites are required for efficient management toward productivity improvement, but current resolutions in agriculture are primarily coarse data. A verified estimation method for surface air temperature, combining statistical and physical techniques, is described as one for acquiring practical and precise data. The method continuously estimates temperature from existing observation data using a new meteorological scale, a radiative cooling scale (RCS). RCS values were calculated using numerical weather prediction model outputs to estimate daily values. Daily temperature at six sites was estimated to have a root-mean-square error of 0.47 K and a mean error of -0.02 K, validating the values for complex terrain such as hilly areas.