Various terms such as “canopy temperature”, “surface temperature”, and “skin temperature” have been used in the literature and the following list is an atempt to identify more precise meanings for various terms.

• Brightness temperature is the land surface temperature obtained on the satellite level by thermal infrared (TIR) sensors.
• Radiometric surface temperature is the land surface temperature sensed on (or immediately above) the ground. It reflects the temperature and emissivity of each horizontal layer of the canopy and soil visible to the radiometer. It tends to vary with changes in the radiometer zenith view angle.
• Aerodynamic temperature is the temperature obtained by extrapolating the air temperature profile to an apparent canopy height given by the displacement height plus the roughness length, which is typically about 3/4 the canopy height.
• Canopy temperature is a nonspecific term referring to aerodynamic temperature or directional infrared temperature, which usually is made from a view angle oblique to the surface so that vegetation dominates the IRT field of view minimizing the effect of soil.
• Kinetic temperature is weighted-average thermodynamic temperature of all the elements comprising the surface. A mean kinetic temperature may be defined as the arithmetic average of the temperatures of all the objects, or an area weighted mean.
• Skin temperature is a nonspecific term that may refer to the directional brightness or infrared temperature for a sensor with particular wavelength and view characteristics.
• Surface temperature is a general, non-specific term referring to the aggregate temperature of all objects, comprising the surface.

The spectral radiance of the ground surface is usually measured by a sensor installed in an airplane or satellite which is far from the ground, so the transmission of the emitted spectral radiance through the atmosphere to the sensor is affected by a number of factors, which make the retrieval of ground surface temperature from the remote sensing data more complicated. In the thermal wavelength, usually the atmosphere has three very important effects on the spectral radiation transmission: absorption, upward atmospheric radiance and bi-directional reflection of the downward atmospheric radiance. Therefore, the spectral radiance reaching the sensors is not only that emitted by ground objects and attenuated by atmospheric absorption, but also includes the radiance emitted by the atmosphere and the reflected component of the downward atmospheric radiance. At the same time, different viewing angles of the sensors and the characteristics of the ground objects also have significant effects on the observed radiance from space (Figure 1).

The theoretical basis for the remote sensing of LST is that total radiative energy emitted by ground surface increases rapidly with increase in temperature. The spectral distribution of the energy emitted by a ground object also varies with temperature. According to Wien’s Displacement Law (Figure 2) on the relationship between spectral radiance and wavelength, for the Earth with an ambient temperature of 300 K, the peak of its spectral radiance occurs at about 9.6 um. Therefore, theoretically, the thermal energy in relation to the physical temperature of the ground surface can be remotely observed by using sensors operating at wavelength around 10 um, which have been defined as thermal channels in remote sensing systems.

On the other hand, the spectral characteristics of the atmosphere indicates that there is an atmospheric window in the spectral region 8-14 um (Figure 3), where atmospheric absorption is minimum and through which the energy source of ground surface can transmit with least losses. Thus, ground surface temperature can be remotely sensed using the channels within the atmospheric window at the thermal spectral wavelength.

Unlike remote sensing of reflected light from surfaces in which only the topmost layers (a few molecular layers thick) are involved, thermal remote sensing includes energy variations extending to varying shallow depths below the ground surface. This takes time and is the normal consequence of heating during the day and cooling at night. The most critical consideration in analyzing and interpreting thermal data and imagery is that of knowing the physical and temporal conditions that heat the near surface layers. Over the seasons, minor shifts in the mean temperature in bedrock can occur to depths of 10 m (33 ft) or more. Solar radiation and heat transfer from the air significantly heat materials at and immediately below the surface during the day. Temperatures usually drop at night primarily by radiative cooling (maximum radiative cooling occurs under cloudless conditions), accompanied by some conduction and convection. During a single daily (diurnal) cycle, the near surface layers (commonly, unconsolidated soils) experience alternate heating and cooling to depths typically between 50 and 100 cm (20-40 in). The daily mean surface temperature is commonly near the mean air temperature. Observed temperature changes are induced mainly by changes during the diurnal heating cycle, but seasonal differences (averages and range) in temperature and local meteorological conditions also affect the cycle response from day to day.

The methods that are popularly used in correction of satellite observations of surface radiance for atmospheric effects fall into two categories, the direct methods which use atmospheric soundings of temperature and moisture from balloon-borne sonde, and the indirect approaches which attempt to accomplish atmospheric corrections with satellite observations only. Direct methods combinein situ measurements of temperature and moisture with atmospheric radiative transfer models, such as LOWTRAN and MODTRAN, which calculate the atmospheric transmittance and path radiance as a function of wavelength. Indirect methods derive vertical soundings of the atmosphere from satellite observations and use atmospheric radiative transfer models. In addition, split-window algorithms using channels 4 and 5 of the NOAA Advanced Very High Resolution Radiometer (AVHRR) satellite instrument, or channels 31 and 32 of NASA Earth Observing System Moderate Resolution Imaging Spectrometer (EOS/MODIS) have been widely applied in the atmospheric and emissivity correction of LST for different purposes. After proper corrections, LST can be computed from corrected radiance by reverse analysis of the Planck equation

A simple regression calibration model was developed to estimate the land surface temperature using Landsat Enhanced Thematic Mapper plus (ETM+) thermal band combined with classification-based surface emissivity and solar zenith angle. The brightness temperatures derived from satellite thermal data for 15 weather stations (Figure 4) in southeast New England area were calibrated by the ground truth temperatures observed while the Landsat-7 overpasses. In this study, the brightness temperatures for each station were derived from the band 6 of the ETM+ data (Figure 5). Surface emissivities of each station were derived from the conventional image classification results (Figure 6). Solar zenith angles were calculated from station latitude and date and time of the observations using an algorithm developed by Mark P. Cresswell at University of Liverpool, UK. The correlation between the land surface temperature and brightness temperature was increased significantly after the surface emissivity and solar zenith angle were added to the model. The determination of coefficient increased from 0.33 to 0.85. This indicates that the Landsat ETM+ thermal infrared data can be reasonably calibrated if appropriate ground truth and ancillary data are available.

See more detail of this research, please visit:

Proceedings of Huangshan International Thermal Infrared Remote Sensing Workshop, July 14-17 Huangshan, Anhui, P.R.China.

Land surface temperature (LST) of a forest obtained from space is actually canopy temperature. However in some environmental study (e.g., tick study), we prefer to know the ground temperature of a forest or woodland. In this study, the vertical temperature every five meters from ground to the canopy top were observed in a woodland in Connecticut in 2001 summer for a continuous 3 months (Figure 8). The reason to choose this site is because there is an good observation tower which is about 30 meters tall in a mixed forest land. Thermocouple were used to measure the temperature in a tube with a motor at one end to avoid the effect of rain and sun shine (Figure 9). Measurements of temperature were collected in 21x Micro data logger and stored in BM 716 Storage Module (Figure 10). The outcome of this study will provide accurate surface and ground temperature for a woodland in natural resources management and environmental studies