Land Surface Temperature Estimation Research Articles

Mapeamento da temperatura da superfície terrestre com uso do sensor AVHRR/NOAA

O objetivo deste trabalho foi avaliar a adequação do uso do sensor AVHRR/NOAA (Advanced Very High Resolution Radiometer/National Oceanic and Atmospheric Administration) para mapeamento da temperatura da superfície terrestre (TST) no Estado do Rio Grande do Sul, por meio da comparação entre três algoritmos clássicos. Foram comparados os métodos de Becker & Li, Sobrino et al. e Kerr et al. para estimativa das TST mínimas, utilizando imagens noturnas e logo após o amanhecer. Os dados de emissividade e TST foram obtidos por meio de combinações matemáticas da radiação detectada nas faixas do visível, infravermelho próximo e termal do sensor AVHRR/NOAA. O sensor AVHRR é adequado para o mapeamento de TST para as condições do tipo de cobertura do solo que predominam no Rio Grande do Sul, visto que a TST estimada pelos três métodos testados foi próxima à temperatura do ar medida nos locais selecionados. O método de Sobrino et al. foi o mais adequado.

Evaluating NDVI-based emissivities of MODIS bands 31 and 32 using emissivities derived by Day/Night LST algorithm

Surface emissivity estimation is a significant factor for the land surface temperature estimation from remotely sensed data. For fully vegetated surfaces, the emissivity estimation is performed in a simple manner since the emissivity is relatively uniform. However, for arid land with sparse vegetation, the estimation is more complicated since the emissivity of the exposed soil and rock is highly variable. In this study, mean and difference emissivity for bands 31 and 32 of MODIS sensor have been derived based on NDVI values. First, the NDVI thresholds have been determined to separate bare soil, partially vegetated soil and fully vegetated land. Then regression relations have been derived to estimate mean and difference emissivity of the bare soil samples and partially vegetated surfaces. A constant emissivity is also used for fully vegetated area. Along with the correlations, standard deviations of the regression relations have been examined for a set of representative soil types. Standard deviations smaller than 0.003 in mean emissivity and smaller than 0.004 in difference emissivity are resulted in regression linear relations. Evaluation of the NDVI derived regression relations has been performed using the results of MODIS Day/Night Land Surface Temperature (LST) algorithm on a pair of MODIS images. Using around 45,500 pixels with different soil and land cover types, emissivity of each pixel in bands 31 and 32 have been estimated. The calculated emissivities have been compared with emissivities calculated by MODIS Day/Night LST algorithm. Biases and standard deviations of NDVI-based relations show relatively high agreement for mean and difference emissivity relations with Day/Night method results. It may be concluded that the proposed algorithm can be used as a rather simple alternative to complex emissivity estimation algorithms.

Emissitivity spectra obtained from field and laboratory measurements using the temperature and emissivity separation algorithm

Surface emissivities play an important role in thermal remote sensing, since knowledge of them is required to estimate land surface temperature with enough accuracy. They are also important in other environmental or geological studies. We show the results obtained for the emissivity spectra of different natural surfaces (water, green, and senescent vegetation) by applying the temperature and emissivity separation (TES) algorithm to ground-based measurements collected at the field with a multiband thermal radiometer. The results have been tested with data included in spectral libraries, and rms errors lower than 0.01 have been found, except for senescent vegetation. Two methods are also proposed to apply the TES algorithm to measurements achieved in the laboratory: (i) by heating the sample and (ii) using a box with reflective walls.

Development of a daily long term record of NOAA-14 AVHRR land surface temperature over Africa

We developed a new 6-year daily, daytime and nighttime, NOAA-14 AVHRR based land surface temperature (LST) dataset over continental Africa for the period 1995 through 2000. The processing chain was developed within the Global Inventory Modeling and Mapping System (GIMMS) at NASA's Goddard Space Flight Center. This paper describes the processing methodology used to convert the Global Area Coverage Level-1b data into LST and collateral data layers, such as sun and view geometries, cloud mask, local time of observation, and latitude and longitude. We used the Ulivieri et al. [Ulivieri, C., M.M. Castronuovo, R. Francioni, and A. Cardillo (1994), A split window algorithm for estimating land surface temperature from satellites, Adv. Space Research, 14(3):59–65.] split window algorithm to determine LST values. This algorithm requires as input values of surface emissivity in AVHRR channels 4 and 5. Thus, we developed continental maps of emissivity using an ensemble approach that combines laboratory emissivity spectra, MODIS-derived maps of herbaceous and woody fractional cover, and the UNESCO FAO soil map. A preliminary evaluation of the resulting LST product over a savanna woodland in South Africa showed a bias of < 0.3 K and an uncertainty of < 1.3 K for daytime retrievals (< 2.5 K for night). More extensive validation is required before statistically significant uncertainties can be determined. The LST production chain described here could be adapted for any wide field of view sensor (e.g., MODIS, VIIRS), and the LST product may be suitable for monitoring spatial and temporal temperature trends, or as input to many process models (e.g., hydrological, ecosystem).

Diurnal temperature range over the United States: A satellite view

Diurnal temperature range (DTR) is an important climate change index. Information on this parameter comes primarily from sparse and unevenly distributed observations of shelter air temperature. In this study, five years of GOES‐8 based estimates of land surface temperature (LST) over the United States are used to evaluate DTR at high spatial resolution. The spatial and temporal patterns that emerged show a high degree of consistency with independent satellite estimates of the Normalized Difference Vegetation Index (NDVI). Specifically, the arid regions in the western and central U.S. have larger DTRs than the eastern United States or the northwest coast. When stratified by four major surface types, the western U. S. DTRs over these surface types are larger than over the eastern part. It is also observed that urban areas have the lowest DTRs especially over the polluted eastern U. S. The similarity of the DTR spatial and temporal patterns and variations of the independent satellite based vegetation index are encouraging and suggest that satellite based estimates of DTR carry a strong signal on surface conditions which are responsive to climate change.

Implementation of GOES‐based land surface temperature diurnal cycle to AVHRR

To enable frequent estimates of land surface temperature (LST) from satellite measurements, and to characterize the land surface temperature diurnal (LSTD) cycle, two new LST retrieval algorithms are applied to observations from the Geostationary Operational Environmental Satellite (GOES). Evaluation against the atmospheric radiation measurement (ARM) observations indicates that LST can be determined from the real‐time GOES‐8 observations within r.m.s. accuracy of about 2 K. In order to combine the advantages of geostationary and polar orbiting instruments, the LSTD estimated from GOES can be incorporated into LST retrievals from polar orbiting imager National Oceanic and Atmospheric Administration/Advanced Very High Resolution Radiometer (NOAA/AVHRR) using a newly proposed fitting algorithm, with r.m.s. errors close to those obtained directly from GOES‐8.

A practical split‐window algorithm for retrieving land‐surface temperature from MODIS data

This paper presents a practical split‐window algorithm utilized to retrieve land‐surface temperature (LST) from Moderate‐resolution Imaging Spectroradiometer (MODIS) data, which involves two essential parameters (transmittance and emissivity), and a new method to simplify Planck function has been proposed. The method for linearization of Planck function, how to obtain atmosphere transmittance from MODIS near‐infrared (NIR) bands and the method for estimating of emissivity of ground are discussed with details. Sensitivity analysis of the algorithm has been performed for the evaluation of probable LST estimation error due to the possible errors in water content and emissivity. Analysis indicates that the algorithm is not sensitive to these two parameters. Especially, the average LST error is changed between 0.19–1.1°C when the water content error in the simulation standard atmosphere changes between −80 and 130%. We confirm the conclusion by retrieving LST from MODIS image data through changing retrieval water content error. Two methods have been used to validate the proposed algorithm. Results from validation and comparison using the standard atmospheric simulation and the comparison with the MODIS LST product demonstrate the applicability of the algorithm. Validation with standard atmospheric simulation indicates that this algorithm can achieve the average accuracy of this algorithm is about 0.32°C in LST retrieval for the case without error in both transmittance and emissivity estimations. The accuracy of this algorithm is about 0.37°C and 0.49°C respectively when the transmittance is computed from the simulation water content by exponent fit and linear fit respectively.

Variational estimation of soil and vegetation turbulent transfer and heat flux parameters from sequences of multisensor imagery

Turbulent heat fluxes from the surface do not have a unique signature that can be detected by remotely deployed instruments. In order to retrieve the fluxes, the measurements need to be merged into models that infer fluxes from their space and time patterns. This study is based on variational assimilation of land surface temperature (LST) into a surface energy balance model with dual‐source soil and vegetation flux components. There are two major unknown parameters in the estimation of land evaporation: near‐surface air turbulent conductivity (that scales the magnitude of the fluxes) and evaporative fraction (that partitions the total flux into latent and sensible heat flux). This study advances the data assimilation approach in two major new directions. First and foremost, it recasts the variational assimilation system as a multiscale problem with LST estimates from a constellation of satellites. The assimilation system can ingest measurements with varying scales and overlapping coverages. Second, the remotely sensed LST is treated as a combination of contributions from the canopy and the exposed soil surface. Application to a large area within the U.S. Great Plains is shown. Spatial patterns of the retrieved parameters, their correspondence to observed land use maps, and their consistency with seasonal phenology are demonstrated. Finally, the performance of a combined‐source formulation is compared with the dual‐source model. Remarkably, the spatial patterns of the heat transfer coefficient reflect dominant vegetation patterns, even though there was no vegetation index information used in the combined‐source formulation.

Spatial analysis of the distribution of intestinal nematode infections in Uganda

The spatial epidemiology of intestinal nematodes in Uganda was investigated using generalized additive models and geostatistical methods. The prevalence of Ascaris lumbricoides and Trichuris trichiura was unevenly distributed in the country with prevalence greatest in southwest Uganda whereas hookworm was more homogeneously distributed. A. lumbricoides and T. Trichiura prevalence were nonlinearly related to satellite sensor-based estimates of land surface temperature; hookworm was nonlinearly associated with rainfall. Semivariogram analysis indicated that T. trichiura prevalence exhibited no spatial structure and that A. lumbricoides exhibited some spatial dependency at small spatial distances, once large-scale, mainly environmental, trends had been removed. In contrast, there was much more spatial structure in hookworm prevalence although the underlying factors are at present unclear. The implications of the results are discussed in relation to parasite spatial epidemiology and the prediction of infection distributions.

An analysis of GOES and NOAA derived land surface temperatures estimated over a boreal forest

Precise estimates of land surface temperature (LST) are important for land monitoring investigations. This study compares LST values calculated using different satellite platforms (Geostationary Operational Environmental Satellite-Imager and National Oceanographic and Atmospheric Administration-Advanced Very High Resolution Radiometer) and five different split window algorithms. The analysis includes (1) a fitting test with the reference dataset, (2) a comparison of differences between algorithms, and (3) an inter-sensor comparison. Considering the hypothesis of the Temperature/Vegetation Index (TVX) technique, the reference dataset was made with air temperature measured over dense canopy having maximum Normalized Difference Vegetation Index (NDVI). The first and second analyses show that algorithms used by Becker and Li and Ulivieri et al. have smaller estimation errors (less than 2.3 K) than the other algorithms, for example, best-fit linear regression. Although these algorithms show a good agreement in the paired algorithms analysis, the final analysis presents a considerable difference in the root mean square error between Imager and AVHRR (1.7 K for the Ulivieri et al. algorithm and 5.3 K for the Becker and Li algorithm). Finally we considered that the Ulivieri et al. method is more stable for both satellites.

Estimation of Land Surface Temperature Using Spatial Interpolation and Satellite-Derived Surface Emissivity

This study presents a method to estimate land surface temperature (LST) by calibrating spatial interpolation using satellite-derived surface emissivity. Four spatial interpolation methods including Inverse Distance Weighting (IDW), Spline, Kriging, and Cokriging were tested to interpolate LST in southern New England using ground temperatures measured at national weather stations (NWS) in summer 2001. The performance of each interpolation method was evaluated using field measurements collected in a mixed forest in Connecticut during the study period. Kriging is recommended for LST interpolation when surface emissivity data are not available. By analyzing the field data, we found that the maximum daily interpolation error occurred in the early morning and the minimum error occurred at around hour 18:00. Validation result shows that the accuracy of spatial interpolation of LST was much improved after being calibrated by satellite-derived surface emissivity. The average interpolation error reduced from 10 °C before calibration to 1.56 °C after calibration.

Inverse Problems Theory and Application: Analysis of the Two-Temperature Method for Land-Surface Temperature and Emissivity Estimation

The two-temperature method (TTM) allows the separation of land-surface temperature and land-surface emissivity information from radiance measurements, and therefore, the solution can be uniquely determined by the data. However, the inverse problem is still an ill-posed problem, since the solution does not depend continuously on the data. Accordingly, we have used some mathematical tools, which are suited for analyses of ill-posed problems in order to show TTM properties, evaluate it, and optimize its estimations. Related to this last point, we have shown that it is necessary to constrain the problem, either by defining a region of physically admissible solutions and/or by using regularization methods, in order to obtain stable results. Besides, the results may be improved by using TTM with systems that possess a high temporal resolution, as well as by acquiring observations near the maximum and minimum of the diurnal temperature range.

Land surface temperature and emissivity estimation based on the two-temperature method: sensitivity analysis using simulated MSG/SEVIRI data

A feasibility study of land surface temperature (LST) and land surface emissivity (LSE) estimation is presented using METEOSAT Second Generation (MSG)/Spinning Enhanced Visible and Infrared Imager (SEVIRI) data, based on the two-temperature method (TTM). The performance of TTM was assessed by taking into consideration the noise in SEVIRI channels IR 10.8 and IR 12.0 as well as the errors in the atmospheric profiles. Imposed errors due to uncertainties on atmospheric information were generated based on the background error covariance matrix used in the assimilation scheme of the European Centre for Medium-Range Weather Forecasts (ECMWF) Global Circulation Model. TTM has provided LST values with averaged error bias [root-mean square error (RMSE)] ranging from 0.0 to 0.5 K (from 0.8 to 2.5 K) for geographical–seasonal model atmospheres stored in MODTRAN4 with water vapor content varying between 0.85 and 4.11 g cm −2. Obtained results suggest that TTM may be used as a complementary method to split-window (SW) algorithms over areas where LSE is not well known a priori.

Case Study of Soil Moisture Effect on Land Surface Temperature Retrieval

Land surface temperature (LST) is an important element of the climate system. Remote sensing methods for estimating LST have been developed in the past and several of them have been implemented at large-scales. Geostationary satellites are of particular interest because they depict the diurnal cycle. Soil moisture has a strong effect on the magnitude of surface temperature via its influence on emissivity; yet, information on soil moisture at large scales is meager. It is of interest to estimate what effect soil moisture has on the retrieval accuracy of surface temperature by methods of remote sensing. In this study, newly developed algorithms to estimate land surface temperature (LST) from geostationary satellites will be applied to GOES-8 observations during the Southern Great Plains 1997 Hydrology Experiment (SGP-97) when surface observations of both soil moisture and surface temperature were made. The ground observations were used to first demonstrate the influence of soil moisture on the diurnal cycle of the surface temperature, its amplitude and the lag in LST maxima. Subsequently, it was established that errors in LST as derived from GOES-8 measurements have a negative correlation with soil moisture, namely, increasing with the decrease of soil moisture.

Land Surface Temperature Estimation from the Next Generation of Geostationary Operational Environmental Satellites: GOES M–Q

Abstract The next generation of Geostationary Operational Environmental Satellites (GOES M–Q) will have only one thermal window channel instead of the current two split-window thermal channels. There is a need to evaluate the usefulness of this new configuration to retrieve parameters that presently are derived by utilizing the split-window characteristics. Two algorithms for deriving land surface temperatures (LSTs) from the GOES M–Q series have been developed and will be presented here. Both algorithms are based on radiative transfer theory; one uses ancillary total precipitable water (TPW) data, and the other is a two-channel (3.9 and 11.0 μm) algorithm that aims to improve atmospheric correction by utilizing the middle infrared (MIR) channel. The proposed algorithms are compared with a well-known generalized split-window algorithm. It is found that by adding TPW to the 11.0-μm channel, similar results to those from the generalized split-window algorithm are attained, and the combination of 3.9 and 11....

Estimation of land surface temperature–vegetation abundance relationship for urban heat island studies

Remote sensing of urban heat islands (UHIs) has traditionally used the Normalized Difference Vegetation Index (NDVI) as the indicator of vegetation abundance to estimate the land surface temperature (LST)–vegetation relationship. This study investigates the applicability of vegetation fraction derived from a spectral mixture model as an alternative indicator of vegetation abundance. This is based on examination of a Landsat Enhanced Thematic Mapper Plus (ETM+) image of Indianapolis City, IN, USA, acquired on June 22, 2002. The transformed ETM+ image was unmixed into three fraction images (green vegetation, dry soil, and shade) with a constrained least-square solution. These fraction images were then used for land cover classification based on a hybrid classification procedure that combined maximum likelihood and decision tree algorithms. Results demonstrate that LST possessed a slightly stronger negative correlation with the unmixed vegetation fraction than with NDVI for all land cover types across the spatial resolution (30 to 960 m). Correlations reached their strongest at the 120-m resolution, which is believed to be the operational scale of LST, NDVI, and vegetation fraction images. Fractal analysis of image texture shows that the complexity of these images increased initially with pixel aggregation and peaked around 120 m, but decreased with further aggregation. The spatial variability of texture in LST was positively correlated with those in NDVI and in vegetation fraction. The interplay between thermal and vegetation dynamics in the context of different land cover types leads to the variations in spectral radiance and texture in LST. These variations are also present in the other imagery, and are responsible for the spatial patterns of urban heat islands. It is suggested that the areal measure of vegetation abundance by unmixed vegetation fraction has a more direct correspondence with the radiative, thermal, and moisture properties of the Earth's surface that determine LST.

Emissivity and temperature estimation from MSG SEVIRI data; method validation with simulated and NOAA-14 AVHRR data

Land surface temperature (LST), i.e. the radiometric surface temperature corresponding to the field-of-view of a sensor, is an important input for weather forecasting and climate models. Only satellite-based measurements provide the intrinsic spatial averaging required for LST determination on the scale of these models; therefore, LST also provides the best approximation to the thermodynamic temperature at pixel scale. Meteosat Second Generation (MSG) (launched on 28.08.2002), provides data with increased spatial, spectral, and temporal resolutions compared to earlier Meteosats. Among various applications involving data from MSG's primary payload, i.e. the Spinning Enhanced Visible and Infrared Imager (SEVIRI), emissivity and LST estimation on an operational basis is of a major concern. Here, issues related to emissivity retrieval from infrared radiances are addressed and the suitability of various algorithms with respect to SEVIRI are discussed. A scheme to estimate SEVIRI channel emissivity based on thermal infrared spectral indices (TISI) is adopted and issues concerning its operational use are addressed. A method validation using synthetic at-ground radiances shows that the achievable accuracy independent of surface type is about 0.005–0.009 for emissivity (in a 11–12 μm channel) and 1.5 K for LST. The method is also applied to NOAA-14 AVHRR data over central Europe. Atmospheric variables were calculated using MODTRAN and ECMWF (European Centre for Medium-range Weather Forecasts) reanalyses.

Estimation of land surface temperature from a Geostationary Operational Environmental Satellite (GOES‐8)

Two algorithms are developed and applied to observations from the Geostationary Operational Environmental Satellite (GOES) to enable frequent estimate of Land Surface Temperature (LST) representing the diurnal cycle. The derived LSTs are evaluated against a wide range of ground observations. Both algorithms are based on radiative transfer theory; one is similar to the classical split window approach used for deriving Sea Surface Temperature (SST), while the other is a three‐channel algorithm. The three‐channel LST algorithm aims to improve atmospheric correction by utilizing the characteristics of the middle‐infrared (MIR) band. Effects of both the atmosphere and the surface emissivity are accounted for. The simulations from the proposed algorithms are compared with previously developed generalized split window algorithm, and a split window algorithm with water vapor correction. During daytime, the proposed new split window algorithm gives the best LST retrievals, while during nighttime, the proposed three‐channel algorithm gives the best retrievals, both within a Root Mean Square (RMS) error of less than 1 K and without a significant bias. Evaluations against the Atmospheric Radiation Measurement (ARM) observations of radiometric surface temperatures and Surface Radiation Network (SURFRAD) observations of outgoing long wave (LW) radiation indicate that LST can be determined from the actual GOES‐8 observations within an RMS accuracy of about 1–2 K, standard error of about 1 K, and bias of less than 1 K. When evaluated against the North Carolina Agricultural Research Service (NCARS) soil temperature as observed at depth of 8 in. and against air temperature observations, the amplitude of the retrieved LST is found to be significantly greater than that of the observed soil temperature, lower than the nighttime air temperature, and higher than the daytime air temperature. When the soil observations are “corrected” to account for the depth difference, they are in good agreement with the LST retrieved from the satellite observations. This indicates that observations of soil temperature, which are more readily available than measurements of “skin” temperatures, can be useful in evaluating satellite‐based estimates. The LST retrieved from both of the proposed algorithms and from a NOAA/NESDIS algorithm, are generally very close to the converted skin temperature from the SURFRAD surface outgoing LW radiation. In most cases, the newly proposed algorithm shows better agreement with ground observations.

Determination of land surface temperature and soil moisture from Tropical Rainfall Measuring Mission/Microwave Imager remote sensing data

An analytical algorithm for the determination of land surface temperature and soil moisture from the Tropical Rainfall Measuring Mission/Microwave Imager (TRMM/TMI) remote sensing data has been developed in this study. The error analyses indicate that the uncertainties of the enrolled parameters will not cause serious errors in the proposed algorithm. By applying the proposed algorithm to TRMM/TMI remote sensing data collected during the Global Energy and Water Experiment (GEWEX) Asian Monsoon Experiment (GAME)/Tibet Intensive Observation Period field campaign in 1998 (IOP'98), the temporal and regional distributions of land surface temperature and volumetric soil moisture are evaluated over the central Tibetan plateau area. To validate the proposed method, the ground‐measured surface temperature and volumetric soil moisture are compared to TRMM/TMI‐derived land surface temperature and soil Fresnel reflectivity respectively. The results show that the estimated surface temperature is in good agreement with ground measurements; their difference and correlation coefficient are 0.52 ± 2.41 K and 0.80, respectively. A quasi‐linear relationship exists between estimated Fresnel reflectivity and ground‐measured volumetric soil moisture with a correlation coefficient 0.82. The land surface thermal status can also be clearly identified from the regional distribution of the estimated land surface temperature; the mountainous area and water bodies have a very lower surface temperature, while the river basin shows a higher surface temperature compared to the mountainous area. The southeastern part of the selected area has lower soil moisture, while the river basin exhibits high soil moisture. It is therefore concluded that the proposed algorithm is successful for the retrieval of land surface temperature and soil moisture using TRMM/TMI data over the study area.

Simulation of scaling effects of thermal emission from non-isothermal pixels with the typical three-dimensional structure

This paper suggests approaches to simulate scaling effects of thermal emission from non-isothermal pixels with a typical three-dimensional structure. We simulate effective emissivities of various V-shaped valleys by using our Monte Carlo method under the isothermal assumption. An analytical formula of the effective emissivities is derived based on photon rebounding between surfaces in the valleys. The comparison between the simulated and the analytically modelled effective emissivities shows that the analytical formula is highly accurate. After simulating effective emissivities of the V-shaped valley under several non-isothermal conditions, we find that the structure and component temperature difference of a pixel cause the scaling effects of thermal emission of the pixel. The results, therefore, prove that Planck's law has to be corrected for remote sensing to estimate land surface temperature with high accuracy.