Land surface temperature retrieval from MSG1-SEVIRI data

doi:10.1016/j.rse.2004.06.009

Remote Sensing of Environment

Volume 92, Issue 2, 15 August 2004, Pages 247-254

https://doi.org/10.1016/j.rse.2004.06.009 Get rights and content

Abstract

We have developed a physical-based split-window Land Surface Temperature (LST) algorithm for retrieving the surface temperature from SEVIRI/MSG1 (Spinning Enhanced Visible and Infrared Imager/Meteosat Second Generation1) data in two thermal infrared bands (IR 10.8 and IR 12.0). The proposed algorithm takes into account the SEVIRI angular dependence. The numerical values of the split-window coefficients have been obtained from a statistical regression method, using synthetic data. The look-up tables for atmospheric transmission, path radiance, and downward thermal irradiance are calculated with the MODTRAN3 code. The new LST algorithm has been tested with simulated SEVIRI/MSG1 data over a wide range of atmospheric and surface conditions. Comprehensive sensitivity and error analyses have been undertaken to evaluate the performance of the proposed LST algorithm and its dependence on surface properties, the ranges of atmospheric conditions and surface temperatures, and on the noise-equivalent temperature difference. The results show that the algorithm is capable of producing LST with a standard deviation lower than 1.5 K for viewing zenith angles lower than 50°. Since MSG1 is becoming fully operational in 2004, the proposed algorithm will allow MSG1 users to obtain surface temperatures immediately.

Introduction

The Meteosat Second Generation 1 (MSG1) satellite, renamed Meteosat-8, was launched on August 2002 and is becoming fully operational in early 2004. The objective of the mission is to improve weather forecasting and get a better understanding of the planet’s climate and its potential changes. MSG1 is a meteorological geostationary satellite developed by the European Space Agency (ESA) in close cooperation with the European Organization for the Exploitation of Meteorological Satellites (EUMETSAT), situated at 36000 km altitude and centered at 0° latitude above the Equator. Fig. 1 shows the MSG1 field of view compared with other geostationary satellites. Two sensors are placed on this satellite; the Geostationary Earth Radiation Budget (GERB), and the Spinning Enhanced Visible and Infrared Imager (SEVIRI). The GERB is a scanning radiometer with two broadband channels, one covering the solar spectrum, the other covering the entire electromagnetic spectrum. The objective of this sensor is to retrieve radiative fluxes of reflected solar radiation and emitted thermal radiation. The Spinning Enhanced Visible and Infrared Imager (SEVIRI) is the main payload on board MSG1. This optical imaging radiometer consists of three visible and near infrared channels centered at 0.6, 0.8 and 1.6 μm, eight infrared channels centered at 3.9, 6.2, 7.3, 8.7, 9.7, 10.8, 12.0 and 13.4 μm and finally one visible broadband channel at 0.5–0.9 μm called the High Resolution Visible channel (HRV). In the following, SEVIRI channels will labeled according to their specifications given in the EUMETSAT technical information, namely, VIS0.6, VIS0.8, IR 1.6, IR 3.9, WV6.2, WV7.3, IR 8.7, IR 9.7, IR 10.8, IR 12.0, IR 13.4, and HRV.

The MSG1 improves the first generation of Meteosat satellites in terms of spatial resolution, frequency of image acquisition, and number of observation bands. The spatial resolution for HRV is 1 km and for the other infrared and visible channels is 3 km. This compares with 5 km for previous Meteosat satellites. There is one image acquisition every 15 min, compared with 30 min per acquisition for previous Meteosat satellites. The new Meteosat satellite has four visible channels that study the sunlight reflected from the Earth's surface and clouds, in comparison with the previous Meteosat camera that has only one visible channel. Eight infrared channels enhance the two thermal infrared bands on the first Meteosat. Four of them are absorption bands of total atmospheric water vapour (W), CO₂ and O₃, and the remaining four allow the land surface, sea surface, and clouds temperature to be determined (Schmid, 2000).

Having all these advantages in mind, the aim of this paper is to provide a simplified algorithm to obtain the land surface temperature (LST) from SEVIRI sensor data. To this end, Section 2 describes the theory associated with the LST retrieval and presents the simulations made using MODTRAN 3.5. The results are explained in Section 3 where the numerical values of the coefficients for the proposed two-channel algorithm are given. Finally we include a sensitivity analysis and, by considering a simulated dataset different from that one used to obtain the algorithm, we test the algorithm.

Section snippets

Two-channel algorithm

The radiative transfer equation applied to the infrared region of the electromagnetic spectrum gives, for a cloud free atmosphere under local thermodynamic equilibrium, the radiance measured from the sensor in channel i under a zenith observation angle θ according to: $B(T_{iθ})=ε_{iθ} B_{i} (T_{s})τ_{iθ} +R_{atiθ}^{↑} +R_{i} (ref)τ_{iθ}$

Three contributions are taken into account: (a) the surface emission that is attenuated by the atmosphere, (b) the upwelling atmosphere emission towards the sensor, and (c) the downwelling

Two-channel algorithm coefficients

In this section the numerical values of the coefficients given by Eq. (3) after each simulation are presented, as are the brightness temperature calculations. In order to determine these values the Levenberg-Marquardt regression method was used, which is a modification of Gauss-Newton algorithm for non-linear minimizations. The standard deviations of the minimizations for all the possible SEVIRI thermal infrared channel combinations and angles are shown in Table 1. The outputs in this table

Conclusions

In this paper a general and operational algorithm to retrieve LST from MSG1-SEVIRI sensor data is presented. IR 10.8 and IR 12.0 channels are used to develop a split-window algorithm as a function of the satellite zenith observation angle. This work is original in that it addresses the question of the necessity of new algorithms for these new generation data. From simulations it is shown that the LST can be obtained from SEVIRI data with accuracy lower than 1.5 K for the most of the situations

Acknowledgements

The authors wish to thank Dr. Gail Andersson from the Air Force Research Laboratory, Hanscom (USA) for providing us with the MODTRAN 3.5 program, Maisoon Al-Jawad for the grammar supervision and The European Union (EAGLE, project SST3-CT-2003-502057) and the “Ministerio de Ciencia y Tecnologı́a” (project REN2001-3105/CLI) for financial support. This work has been carried out while Mireia Romaguera was in receipt of a grant from the “Agència Valenciana de Ciència i Tecnologia”.

References (16)

J.W. Salisbury et al.
Emissivity of terrestrial materials in the 8–14 μm atmospheric window
Remote Sensing of Environment
(1992)
Abreu, L. W., & Anderson, G. P., (Eds.). (1996). The MODTRAN 2/3 Report and LOWTRAN 7 MODEL, Modtran Report, Contract...
D.M.A. Aminou et al.
Meteosat Second Generation. A comparison of on-ground and on-flight Imaging and Radiometric performances of SEVIRI on MSG-1
P. Dash et al.
Potential of MSG for surface temperature and emissivity estimation: considerations for real-time applications
International Journal of Remote Sensing
(2002)
E. Hurtado et al.
Daily net radiation estimated from air temperature and NOAAAVHRR data: a case study for the Iberian Peninsula
International Journal of Remote Sensing
(2001)
Z.-L. Li et al.
A new approach for retrieving precipitable water from ATSR2 split-window channel data over land area
International Journal of Remote Sensing
(2003)
J. Labed et al.
Angular variation of Land surface spectral emissivity in the thermal infrared: Laboratory Investigations on Bare Soils
International Journal of Remote Sensing
(1991)
E.P. McClain et al.
Comparative performance of AVHRR-based multichannel sea surface temperatures
Journal of Geophysical Research
(1985)

There are more references available in the full text version of this article.

Cited by (130)

A normal form for synchronous land surface temperature and emissivity retrieval using deep learning coupled physical and statistical methods
2024, International Journal of Applied Earth Observation and Geoinformation
The innovative normal form, named deep learning-couple-physical and statistical methods (DL-C-PS), has been developed for synchronously retrieving land surface temperature and emissivity (LST&E) from Japan's geostationary meteorological satellite Himawari-8 carrying the Advanced Himawari Imager (AHI). First, geophysical logical reasoning (GLR) and expert knowledge were used to formulate radiative transfer equations (RTEs). Then, a hybrid approach integrating physical and statistical methods was employed to derive the solution, with deep learning (DL) optimizing the solution process. Three band combination schemes were designed to assess the effectiveness of the DL-C-PS normal form. Simulation data from the MODerate spectral resolution atmospheric TRANsmittance mode (MODTRAN) yielded promising results during validation. Root mean square error (RMSE) values were below 1 K for LST and below 0.008 for LSE when using band combinations of four thermal infrared (TIR) bands or at least three TIR bands combined with water vapor information. Cross-validation and in situ validation showed consistent findings with simulation validation. Compared to MODIS LST&E products (MYD21), in most cases, the RMSE values for LST&E were approximately 2 K and less than 0.015 during daytime, and below 1.3 K and 0.017 during night, respectively. Validated against in situ observations, nighttime RMSE values for LST were approximately 1.5 K with correlation coefficient (R) values better than 0.93. The higher RMSE observed in daytime compared to nighttime can be attributed to the influence of the sun's illumination angle on satellite scanning imaging. Overall, this research presented a normal form for multi-parameter estimation by leveraging the optimal calculation of DL and incorporating physical interpretations.
Strengthening weather forecast and dissemination capabilities in Central Africa: Case assessment of intense flooding in January 2020
2023, Climate Services
The first dekad of January 2020 was characterised by heavy precipitation in the capital of the Republic of Congo, Brazzaville, which led to several localised landslides. Satellite-derived rainfall estimates and rain-gauge totals illustrate a strong wet spell between the 6th and 9th January 2020 across southern Congo. This study highlights the generation and implementation of user-driven weather and climate forecast bulletins, developed at the Economic Communities of Central African States (ECCAS) Climate Application and Prediction Centre (CAPC), to reduce the impacts associated with intense precipitation during this dekad. Through doing so, we document the current state of regional-scale climate services across Central Africa. Advisories and outlooks generated by CAPC use risk matrices developed by the World Meteorological Organisation (WMO) and are produced at hourly, daily, sub-seasonal, and seasonal timescales. To develop them, meteorologists and climate scientists at CAPC combine information from a wide range of meteorological observations and forecasts. Regional-scale forecasts are downscaled to individual countries to improve accessibility and relevance.
Central African users have reported that bulletins provide support for mitigating against the impacts of extreme weather and have requested more reliable sub-seasonal and seasonal forecast products. In this paper we take the opportunity to discuss the resources obtained through the Satellite and Weather Information for Disaster Resilience (SAWIDRA) framework, which are often taken for granted in developed nations, including the procurement of a high performance computing system, satellite data and numerical models outputs receiving stations. This study is the first to highlight the current state of regional-scale climate services across Central Africa and motivates further co-production of climate services across the region.
Thermal infrared remote sensing data downscaling investigations: An overview on current status and perspectives
2023, Remote Sensing Applications: Society and Environment
Land surface temperature (LST) retrieved from moderate resolution or downscaled from coarse thermal infrared (TIR) data is one of key environment parameters. Over the last four decades, most advanced remote sensing sensors/systems can acquire TIR data at a low spatial resolution but high temporal resolution. However, per different application purposes, both high spatial and temporal resolution TIR data are needed. Given that many investigations on downscaling LST (DLST) processes have been done and findings have been reported in the literature, it necessitates to have an updated review on DLST investigations of the status, trends, and challenges and to recommend future directions. An overview is provided on various polar orbits and geostationary orbits' satellite TIR sensors/systems and on scaling factors’ determination and selection techniques/methods suitable for DLST processes. Existing various techniques/methods for DLST processes are presented and assessed, and limitations and future research directions are identified and recommended. In this review, several concluding remarks were made, including (1) most investigations on DLST processes used coarse spatial resolution but high temporal resolution MODIS TIR data; (2) compared to fusion-based method, the kernel-driven processes are the most frequently used thermal downscaling methods; (3) machine-learning methods have demonstrated their excellent performance and robustness in improving DLST accuracy; (4) more advanced spatiotemporal fusion-based methods consider synergic powers by combining a kernel-driven process with a fusion-based process method. The three future research directions for DLST processes are recommended: further reducing uncertainties of DLST results, developing novel DLST models and algorithms, and directly reducing the spatial scaling effect in DLST processes.
Combining GOES-R and ECOSTRESS land surface temperature data to investigate diurnal variations of surface urban heat island
2022, Science of the Total Environment
The surface urban heat island (SUHI) phenomenon is characterized by both high spatial and temporal variability, while its diurnal (i.e., diel) variations have rarely been investigated because traditional satellites and sensors flying on polar orbits (e.g., Landsat, MODIS) have no diurnal sampling capability. Here we combined land surface temperature (LST) data from the Geostationary Operational Environmental Satellites (GOES-R) and the Ecosystem Spaceborne Thermal Radiometer Experiment on Space Station (ECOSTRESS) to explore the diurnal variations of SUHI and thermal differentiation among various land covers over the Boston Metropolitan Area. With the combined use of the LST data from GOES-R and ECOSTRESS, we took advantage of the strengths of both GOES-R (i.e., high frequency in each day and night) and ECOSTRESS (i.e., much finer spatial resolution). The SUHI intensity of the urban-core and suburban areas both exhibited clear diurnal patterns for different seasons: a continuous increase in the SUHI intensity from sunrise to noon and a decrease thereafter to sunset, followed by a relatively low and constant intensity during nighttime. The LST contrasts among different land cover types were clearly larger in the daytime than at nighttime and peaked around midday. At noon in summer, the LST of ‘Developed, High Intensity’ was 2.6 °C higher than that of ‘Developed, Medium Intensity’, and about 4.6 °C higher than that of “Developed, Open Space” and “Developed, Low Intensity”. Controlling the percent impervious surface in construction land at a relatively low level (e.g., below ~49%) could effectively alleviate the impacts of SUHI. Compared with GOES-R data, ECOSTRESS LST is suitable for monitoring the diurnal variations of intracity thermal environment at the subdistrict (or neighborhood) scale. Our study highlights the value of the combined use of geostationary satellite and ECOSTRESS LST in exploring the diurnal cycling of the SUHI, and can help inform urban planning and land-based climate mitigation policies in the context of climate change.
Detection of land use/land cover and land surface temperature change in the Suha Watershed, North-Western highlands of Ethiopia
2022, Environmental Challenges
Human-induced land use land cover changes resulted in adverse impacts on the environment at various spatial and temporal scales. The Highland regions of Ethiopia are typical examples of these phenomena. The objective of this study was to analyze spatiotemporal changes in land use/ land cover and their impacts on land surface temperature in the suha watershed, northwestern highlands of Ethiopia. Multi-temporal Landsat images (1985–2019) were used to analyze LU/LC changes and LST using GIS and remote sensing techniques. Image preprocessing, supervised classification, accuracy assessment, and change detection were conducted to identify LU/LC classes, area coverage, and their transitions. Thermal bands of these satellite images were also used to extract LST.
Significant changes in land use/land cover (spatial and temporal) were observed in the suha watershed during the study periods. Agricultural land has got the largest proportion in all study periods. The barren land and built area also expanded greatly in 35 years period. Agricultural land has increased by 15417.6 ha (34.8%) and bare land has expanded by 5297.2 ha (373.6%). However, grazing and shrub lands were reduced by18568.4 ha (72.1%) and 3544.2 ha (47.6%), respectively. Spatial and temporal variation of LST was also observed in the same period. The highest mean values were found on impervious surfaces (built-up areas and bare land), and the lowest values were recorded on forest land. A negative correlation was found between LST and NDVI. Undesirable land use and land cover changes put greater pressure on environmental resources, resulting in an adverse effect on them. Therefore, to reverse this situation and create a balanced ecosystem, management strategies should be applied that mainly focus on soil conservation technologies in steep slope areas, improve afforestation and apply proper land-use policies. The outcomes of this research are useful in designing and implementing appropriate strategies that can address critical social and environmental problems. It also provides new knowledge that helps us better understand the spatiotemporal land use land cover changes and their impacts on LST.
Quantifying the local cooling effects of urban green spaces: Evidence from Bengaluru, India
2021, Landscape and Urban Planning
Rapid unplanned urbanization has led to a deterioration in green cover in Indian cities and an increase in urban temperatures due to the urban heat island (UHI) effect. With India’s urban population set to double from 400 million in 2011 to 800 million by 2050, it becomes critical to understand the role of urban green spaces (UGS) in mitigating the UHI. In this study, we have used high-resolution Landsat and Google Earth data and integrated it with spatial statistical analysis to quantify the cooling effects provided by UGS beyond their boundaries. We analyzed cooling effects at the level of individual UGS for 262 UGS in the megacity of Bengaluru, India. Our results showed that the average UGS provided local cooling effects till points 347 m (95% CI: 318 m to 376 m) beyond its boundary. The average UGS was 2.23 °C (95% CI: 2.13 °C to 2.33 °C) cooler than the point where it ceased to provide cooling effects. Cooling effects reduced with distance from the UGS, and were impacted by the greenness, size, and shape of the UGS. The findings of this study are important in the context of India’s Smart Cities Mission that has been criticized for an inadequate focus on urban greening. Our study addresses a concern that most previous studies have used a small sample of UGS for their analysis. To the best of our knowledge, this is the first study to quantify the role of UGS in localized surface temperature reduction for a large Indian city.

View all citing articles on Scopus

View full text

Land surface temperature retrieval from MSG1-SEVIRI data

Abstract

Introduction

Section snippets

Two-channel algorithm

Two-channel algorithm coefficients

Conclusions

Acknowledgements

Remote Sensing of Environment

Meteosat Second Generation. A comparison of on-ground and on-flight Imaging and Radiometric performances of SEVIRI on MSG-1

Potential of MSG for surface temperature and emissivity estimation: considerations for real-time applications

International Journal of Remote Sensing

Daily net radiation estimated from air temperature and NOAAAVHRR data: a case study for the Iberian Peninsula

International Journal of Remote Sensing

A new approach for retrieving precipitable water from ATSR2 split-window channel data over land area

International Journal of Remote Sensing

Angular variation of Land surface spectral emissivity in the thermal infrared: Laboratory Investigations on Bare Soils

International Journal of Remote Sensing

Comparative performance of AVHRR-based multichannel sea surface temperatures

Journal of Geophysical Research