Accurate assessment of land surface thermal emissivity|
Maria Mira, Vicente Caselles, Enric Valor and César Coll
The emissivity of
natural surfaces is a required parameter in high accuracy land surface temperature
(LST) determinations. Then, it is of great importance to know how thermal
infrared (TIR) emissivity changes with the different
factors. Soil type influence on emissivity is well known
from experimental studies, whereas the effect of soil moisture (SM) on emissivity
is one of the pending issues in thermal remote sensing. It is seen that the SM
variation may cause a high systematic error in emissivity, e.g., about +0.1 in emissivity
for an increase from 0.04 to 0.10 g/cm3 in SM for sandy soils. This study
reveals a common emissivity increase with soil water
content, higher for larger sand content and in the 8 to 9 μm
range, but always significant considering the experimental uncertainty (δε~±0.5%).
The emissivity increase is understood because of both the quartz contribution to decrease the emissivity
of the material chiefly between 7.7 and 9.7 μm, and the decrease of the reflectivity by the water
film on the soil particles. Emissivity-SM relationships were obtained with the
aim of using them together with SM estimates from remotely sensed data in order to improve the TIR
emissivity estimations and thus, the LST determination.
Furthermore, the study showed the possibility of retrieving SM from TIR emissivities
with a standard estimation error of about ±0.08 m3m-3,
when considering the OM content.