It's well-known that temperature measurements are affected by environmental conditions of meteorological observing stations. Buildings or trees around the stations may reduce air flow or project shades, thus generate measurement errors. Asphalt pavement surfaces such as roads or car parks may act as heat sources. The surroundings of some surface observing stations in Japan have been deteriorated over the years as more and more buildings are being constructed and trees are growing in the vicinity of the sites. According to the siting classification of WMO, the main factors which adversely affect temperature measurements are unnatural surfaces and shading. It is necessary to clarify to how much extent each factor can cause measurement errors in order to keep observing stations in a good state in the long run.

This investigation was focused on the effect of artificial surfaces, specifically asphalt car roads and aimed to measure it quantitatively by field tests.

In the experiment, the sample thermometers with screens were installed on one side of the asphalt road at different distances of 0.8 m, 3.2 m, 6.9 m and 10.0 m from the road. At each distance, three thermometers were mounted at different heights of 0.5 m, 1.5 m and 2.5 m from the ground. The reference temperature was measured at the opposite side of the road at a distance of 10.0 m. The reference side was selected to be situated in the windward side considering the wind directions that are prevailing differently in summer and winter in the test site. As the wind blows over the road to the sample thermometers, it is expected that the air heated by the road causes some bias from the reference in the readings of temperature.

From this experiment, it was shown that the bias of +0.3 to +0.5 degrees in summer and +0.2 to + 0.4 degrees in winter existed at the height of 0.5 m from the ground depending on the distances from the road, the closer the distance, the bigger the bias. The bias was slightly diminished to +0.1 degrees at 1.5m-high, while significant differences from the reference were not observed at the height of 2.5m. When the wind was strong, the bias was too small to observe, indicating that the effect of the asphalt road as a heat source wasweaken by the air flow and exchange.

The field experiment has proved to be very useful in estimating the effects of heat sources on temperature measurement at observing stations. In order to further explain the mechanism of the temperature distribution of the site, it is necessary to consider other factors such as terrain roughness, heat balance, soil moisture and vapor from the vegetation and surfaces in the neighborhood. Another field experiment is now going on to estimate the effect of surrounding obstacles that bring changes in the exposure of a thermometer to the wind. More knowledge about the method of evaluating the environmental conditions of observing stations is expected from such kind of field experiments as well as computer simulations.