The Heat Island Effect refers to the gradual increase in recorded temperatures as a result of the urbanization that occurs in the vicinity of the temperature measuring instruments. Primarily this is due to the paving of the land. The water that fell around the weather stations at one time moistened the ground and was absorbed by plant life. Some of the solar radiation then went into evaporating water and transpiring water from the plants. The rest went into raising the temperature of the land. When land was paved the water ran off and so the radiation energy went entirely into raising the surface temperature. Also asphalt paving increased the absorption of the radiative energy. The net result is that the recorded temperature at the weather stations increased over time without their being any climatic change responsible for it.

When there is a temperature trend due to a climatic change then the trend in the recorded temperature is the sum of the heat island effect and the climatic trend. The climatic trend in temperature is estimated as the difference between the trend in recorded temperatures and the estimated trend due to the heat island effect. If the heat island effect is underestimated then the climatic trend will be overestimated, and vice versa.

It is not easy to obtain accurate estimates of the heat island effect. In principle if there were weather stations far from any urbanization their temperature readings would be uncorrupted by the heat island effect, but there are not enough of such stations.

The website smalldeadanimals.com has some interesting illustrations from Pielke Research Group of the location of the temperature monitoring sensors.

The

Although the vagaries of the siting of the temperature measuring sensors is notable, it should also be noted that these problems do not necessarily affect the measurement of the trend in temperature if the conditions around the sensor are unchanging. However the sites are sometimes changed, as in the case of the Fort Morgan site, and this introduces spurious trends into the data.