Quantifying the Clear‐Sky Bias of Satellite Land Surface Temperature Using Microwave‐Based Estimates

Abstract

Most available long‐term databases of land surface temperature (LST) derived from space‐borne sensors rely on infrared observations and are therefore restricted to clear‐sky conditions. Hence, studies based on such data sets may not be representative of all‐weather conditions and may be considered as “biased” toward clear sky. An assessment of the impact of this restriction is made using 3 years of LST derived from passive microwave observations that are not affected by most clouds. A systematic analysis in space and time is performed of the “clear‐sky bias,” defined as the difference between average clear‐sky and average all‐weather LSTs. The amplitude of the bias is closely related to the fraction of clear‐sky days, and therefore, arid regions are associated to very low values of bias whereas midlatitudes present the highest values. During daytime, the input of solar radiation for clear‐sky situations leads to higher LST values, and therefore, the bias is generally positive (e.g., 2–8 K over the midlatitudes) whereas, during nighttime, the bias is generally negative although with lower amplitude (around −2 K), because of the increased radiative cooling for clear‐sky situations. The clear‐, cloudy‐, and all‐sky LSTs are also compared with near‐surface air temperature. Although LST is generally higher than air temperature, the contrast between the two may be strongly influenced by local weather conditions. Both the clear‐sky bias and differences between LST and air temperature are also analyzed at the local scale taking into account the predominant cloud regime.

Towards a Unified and Coherent Land Surface Temperature Earth System Data Record from Geostationary Satellites

“…The need for controlled experiments to facilitate discussion on sources of discrepancies between methods has been recognized by the scientific community and is conducted frequently. Examples are numerical model evaluation as conducted at Lawrence Livermore National Laboratory (https://pcmdi.llnl.gov/?projects/amip/0), while controlled experiments to estimate errors due to aerosols is described in Randles et al [68]. The objective of the current study is to present a credible methodology to generate long term time series of LST at best available spatial and temporal resolution (that currently are possible with a long term outlook), and evaluate it against best available satellite products and ground observations.…”

“…Future improvement would be possible by satellite observations of higher spatial resolution, the incorporation of higher temporal resolution of surface emissivity and improved/innovative methodologies to remove cloud contamination [68] and by accounting for anisotropy in emissivity Pinheiro et al [69], Ermida et al [70].…”

An All-Weather Land Surface Temperature Product Based on MSG/SEVIRI Observations

“…The first kind of methodologies usually rely on ancillary information such as land cover, elevation, day of year, a diurnal cycle model, or data from another sensor [22,23]. However, they generally provide LST estimates corresponding to clear sky situations and, therefore, are affected by the so-called clear-sky bias [24]. MW products are more common since there have been multiple operational MW instruments for decades, allowing the production of long-term data records.…”

Cold Bias of ERA5 Summertime Daily Maximum Land Surface Temperature Over Iberian Peninsula

“…One of the limitations of TIR-based LST data is its dependence on clear-sky measurements. Absence of LST data will occur for pixels classified as totally or partially cloudy during the observation period, and therefore such satellite LST products will be biased towards clear-sky conditions [16]. This implies that any evaluation of model LST using TIR-based products must be preceded by a careful cloud screening in the model dataset to ensure the compatibility of model and satellite variables.…”