Land Surface Temperature Estimation Research Articles

Solar zenith angle-based calibration of Himawari-8 land surface temperature for correcting diurnal retrieval error characteristics

The geostationary Himawari-8 satellite offers a unique opportunity to monitor sub-daily thermal dynamics over Asia and Oceania, and several operational land surface temperature (LST) retrieval algorithms have been developed for this purpose. However, studies have reported inconsistency between LST data obtained from geostationary and polar-orbiting platforms, particularly for daytime LST, which can arise from variations in viewing geometries and inherent differences in sensor types and LST algorithms. Despite this, previous research has primarily focused on analysing the directionality of LST without thoroughly exploring systematic differences between platforms. Hence, we presented a Solar Zenith Angle-based Calibration (SZAC) method to harmonise the daytime component of a split-window retrieved Himawari-8 LST (referred to here as the baseline) with the MODerate-resolution Imaging Spectroradiometer (MODIS) LST. SZAC describes the spatial heterogeneity and magnitude of diurnal LST discrepancies from different platforms, which is anticipated to complement typical directionality analyses. We evaluated the harmonised LST data, referred to as the Australian National University LST with SZAC (ANUSZAC), against MODIS LST and the Visible Infrared Imaging Radiometer Suite (VIIRS) LST, as well as in-situ LST from the OzFlux network. Two peer Himawari-8 LST products from Chiba University and the Copernicus Global Land Service were also collected for comparison. The median daytime bias of ANUSZAC LST against Terra-MODIS LST, Aqua-MODIS LST and VIIRS LST was 1.52 K, 0.98 K and −0.63 K, respectively, which demonstrated improved performance compared to baseline (5.37 K, 4.85 K and 3.02 K, respectively) and Chiba LST (3.71 K, 2.90 K and 1.07 K, respectively). All four Himawari-8 LST products showed comparable performance of unbiased root mean squared error (ubRMSE), ranging from 2.47 to 3.07 K, compared to LST from polar-orbiting platforms. In the evaluation against in-situ LST, the mean values of bias (ubRMSE) of baseline, Chiba, Copernicus and ANUSZAC LST during daytime were 4.23 K (3.74 K), 2.16 K (3.62 K), 1.73 K (3.31 K) and 1.41 K (3.24 K), respectively, based on 171,289 hourly samples from 20 OzFlux sites across Australia between 01/Jan/2016 and 31/Dec/2020. In summary, the SZAC method offers a promising approach to enhance the reliability of geostationary LST retrievals by incorporating the spatiotemporal characteristics observed by accurate polar-orbiting LST data. SZAC can also reduce LST angular effects due to its capability of quantifying spatial inhomogeneity of surface heat dynamics. Furthermore, it is possible to implement SZAC using LST data acquired by geostationary satellites in other regions, e.g., Europe, Africa and Americas. This could improve our understanding of the error characteristics of coincident imageries from geostationary and polar-orbiting platforms, allowing for targeted refinements and global harmonisations to further enhance applicability.

Multi-scale remote sensing retrieval model for land surface emissivity and its impact on land surface temperature estimation

ABSTRACT As a fundamental parameter of the surface radiation process, land surface emissivity (LSE) is the most direct factor affecting the land surface temperature (LST) retrieval accuracy. Based on the measured data from 2017 to 2023, the characteristics of LSE changes were analysed, different timescale LSE retrieval models were developed, and the impact of LSE on pixel scale observational LST acquisition was evaluated by using unmanned aerial vehicle (UAV) remote sensing images. The results show that with changes in atmospheric environmental conditions, the LSE changes are relatively less pronounced at 3.99 µm and 4.02 µm in the mid-infrared band and at 10.8 µm and 11.8 µm in the thermal infrared band than in the other bands. At the timescale above day, the LSE is significant correlation with the normalized differential vegetation index (NDVI). Vegetation is the main parameter that dominates the LSE change under heterogeneous underlying surfaces. On a daily timescale, LSE exhibits typical diurnal variation characteristics, and the amplitude of diurnal variation is influenced mainly by shallow soil water and heat conditions. The developed LSE model based on Fengyun-4A (FY-4A) satellite remote-sensing data can effectively reproduce the daily variation characteristics of LST and avoid the ‘jumping’ phenomenon that may occur in the NDVI threshold method. The root-mean-square error (RMSE) between the retrieved LST and in situ data is reduced to 3.78°C, and the retrieval accuracy is better than that of the published LST product. When using UAVs to obtain pixel-scale LSTs, dry bare soil is more sensitive to LSE. When the LSE increases from 0.95 to 1.0, the average LST obtained by the UAV deviates by approximately 1.0°C. This study preliminarily provides some spatiotemporal variation patterns of LSE and lays a foundation for the later-modified LST retrieval algorithms.

Estimation of Land Surface Temperature from Chinese ZY1-02E IRS Data

The role of land surface temperature (LST) is of the utmost importance in multiple academic disciplines, such as climatology, hydrology, ecology, and meteorology. To date, many methods have been proposed to estimate LST from satellite thermal infrared data. The single-channel (SC) algorithm can provide an accurate result in retrieving LST based on prior knowledge of known land surface emissivity (LSE). The SC algorithm is extensively employed for retrieving LST from Landsat series data due to its simplicity and its reliance on just one thermal infrared channel. The Thermal Infrared Sensor (IRS) on the Chinese ZY1-02E satellite is a pivotal instrument employed for gathering thermal infrared (TIR) data of land surfaces. The objective of this research is to evaluate the feasibility of a single-channel approach based on water vapor scaling (WVS) for deriving LST from ZY1-02E IRS data because of its wide spectrum range, i.e., 7~12 μm, which is affected strongly by both atmospheric water vapor and ozone. Three study areas, namely the Baotou, Heihe River Basin, and Yantai Sea sites, were selected as validation sites to evaluate the LST inversion accuracy. This evaluation was also conducted via cross-comparison between the retrieved LST and MODIS LST products. The results revealed that the WVS-based method exhibited an average bias of 0.63 K and an RMSE of 1.62 K compared to the in situ LSTs. The WVS-based method demonstrated reasonable accuracy through cross-validation with the MODIS LST product, with an average bias of 0.77 K and an RMSE of 2.0 K. These findings indicate that the WVS-based method is effective in estimating LST from ZY1-02E IRS data.

Estimation of land surface temperature from Landsat-8 OLI: comparative analysis of two periods for Duzce in Turkey

Estimation of land surface temperature from Landsat-8 OLI: comparative analysis of two periods for Duzce in Turkey

A time-continuous land surface temperature (LST) data fusion approach based on deep learning with microwave remote sensing and high-density ground truth observations

Land surface temperature (LST) is a crucial parameter in the circulation of water, exchange of land-atmosphere energy, and turbulence. Currently, most LST products rely heavily on thermal infrared remote sensing, which is susceptible to cloud and rain interference, leading to inferior temporal continuity. Microwave remote sensing has the advantage of being available “all-weather” due to strong penetration capability, which provides the possibility to simulate time-continuous LST data. In addition, the continuous increase in high-density station observations (>10,000 stations) provides reliable measured data for the remote sensing monitoring of LST in China. This study aims to adopt the “Earth big data” generated from high-density station observation and microwave remote sensing data to monitor LST based on deep learning (U-Net family) for the first time. Given the significant spatial and temporal variability of LST and its sensitivity to various factors according to radiation transmission equations, this study incorporated climatic, anthropogenic, geographical, and vegetation datasets to facilitate a multi-source data fusion approach for LST estimation. The results showed that the U-Net++ model with modified skip connections better minimized the semantic discrepancy between the feature maps of the encoder and decoder subnetworks for 0.1° daily LST mapping across China than the U-Net and U2-Net deep learning models. The accuracy of the LST simulation exhibited favorable outcomes in the spatial and temporal dimensions. The station density met the requirements of monitoring air-ground integration monitoring in China. Additionally, the temporal change in the simulation accuracy fluctuated in a W-shape owing to the limited simulation capability of deep learning in extreme scenarios. Anthropogenic factors had the largest influence on LST changes in China, followed by climate, geography, and vegetation. This study highlighted the application of deep learning in remote sensing monitoring against the background of “big data” and provided a scientific foundation for the response of climate change to human activities, ecological environmental protection, and sustainable social and economic development.

Improving monthly mean land surface temperature estimation by merging four products using the generalized three-cornered hat method and maximum likelihood estimation

Land surface temperature (LST) is a critical parameter associated land surface energy and water balance at the land–atmosphere interface. Monthly mean LST (MMLST) is used not only as an input parameter for land surface models, but also as an indicator for general climate trend studies. Four MMLST products (i.e., NESDC, TPDC, ZENODO, and ERA5) were generated by different research teams and institutes using different algorithms. In this study, a new method was presented to improve MMLST estimation by merging these four products using the generalized three-cornered hat (TCH) method and maximum likelihood estimation (MLE). The generalized TCH method was used to simultaneously calculate the relative uncertainties of these four MMLST products during the period 2003–2019. In general, the smallest uncertainties are obtained for the TPDC MMLST product in most areas of the globe, whereas the largest uncertainties are achieved for the ERA5 MMLST product, especially in high-latitude regions. The MLE method was used to merge these four MMLST products and to generate a new MMLST product in consideration of the uncertainty of each individual MMLST product. In situ measurements at 25 stations from the AmeriFlux and SURFRAD networks during the period 2003–2019 were used to evaluate the accuracies of the individual and merged MMLST products. The largest root mean squared error (RMSE) of approximately 2.4 K is obtained for the TPDC MMLST product, whereas the smallest RMSE of around 1.6 K is reached for the merged MMLST product. The NESDC, ZENODO, and ERA5 MMLST products have similar RMSE of approximately 2 K. The results indicate that the merged MMLST product outperforms each individual MMLST product. The merged MMLST product shows promising prospects for climate- and global-scale studies and applications.

Urban land surface temperature forecasting: a data-driven approach using regression and neural network models

The insinuations of the ailments associated with the unrestrained and disorganized proliferation of artificial impervious materials over natural surfaces are prevalent among city dwellers. These impacts can be comprehended by estimating land surface temperature (LST), as it is vital for evaluating urban climate, particularly to explain the intensity of urban heat islands and to define the health and welfare of the planet as well as the living beings. Urbanization-driven landscape changes severely disrupt comfortable living in almost every city, necessitating monitoring and modelling historical, current, and likely future LSTs. This research article proposes two forecasting techniques: Multiple Linear Regression (MLR) and Artificial Neural Network (ANN) models. These models have been widely accepted for the efficient prediction of climatic parameters, including LST, over an urban area. The landscape, elevation, and LST trend served as input to the models for an accurate prediction of LST. The analysis was performed over the Kolkata Metropolitan Area (KMA) with an additional 10 km buffer to understand urban growth and its effect on the LST of the entire region. The two developed models (MLR and ANN) effectively anticipated the LST over the KMA region. A continual increment in the surface temperatures ranging from 1 °C to 4 °C, over existing and likely-predicted urban areas was comprehended. It was anticipated that the regions near the urban areas will also experience severe discomfort and heat waves without proper mitigation measures. This scientific literature provides essential insights for decision-makers, stakeholders, and government officials to articulate new policies and modify the existing ones to create a sustainable and livable urban environment for the inhabitants.

Comparison of multi-factor spatial downscaling models for high-resolution LST estimation in mountainous and hilly open-pit mines

The stripping of soil and vegetation by mining activities in mountainous and hilly areas of southern China has caused drastic surface disturbance and serious ecological and environmental problems. The spatial distribution of Land Surface Temperature (LST) in the mine site can better reflect the ecological disturbances caused by mining and is an important parameter for assessing the ecological environment of the mining sites. The classic Distrad thermal pixel decomposition model was improved by using different regression factors and linear and nonlinear fitting methods in this study, and three multi-factor downscaling models were proposed. Finally, four downscaling models were used to estimate the high spatial resolution LST for three three typical open-pit mining areas and explore the contribution of different surface parameters in the downscaling process from qualitative and quantitative perspectives. The results show that: (1) The MTVET (Extremely Randomized Trees regression model with multitype predictor variables) model exhibits the most accurate prediction accuracy in all three mining areas. The spatial distribution of errors shows that the area share of absolute errors between 0 and 1 K exist in 86.48 %, 81.38 %, and 84.51 % of the areas in Ganzhou rare earth mine, Dexing copper mine and Pingxiang coal mine, respectively. Based on the Root Mean Square Error (RMSE) assessment, the three mines are 0.699 K, 0.790 K, and 0.751 K, respectively, which are all controlled within 1 K; (2) The downscaled results of the MTVET model show rich and clear LST spatial texture features when typical mine sites are selected in each mining areas for detail comparison; (3) It was found that the topographic index has the greatest impact on accuracy enhancement, suggesting that topographic factors have an important influence in the study of LST downscaling in mountainous and hilly areas of southern China.

EXPLORING VERY HIGH-RESOLUTION REMOTE SENSING FOR ASSESSING LAND SURFACE TEMPERATURE OF DIFFERENT URBAN LAND COVER PATTERNS

Abstract. Very-high-resolution thermal infrared data has a good potential to detect and monitor LST variations in urban areas. In this work, an attempt was made to estimate Land Surface Temperature (LST) from very high resolution (VHR) data using machine learning technics in an area of Botanical garden of Yerevan, Armenia; UAV-derived and in-situ measured high precision LST of various land cover (LC) patterns were then compared for September and October. The main purpose of this study was to explore the capabilities of UAV imagery (multispectral/TIR) in assessing LST of the different LC patterns. For this purpose, a UAV survey was performed, and VHR data were collected using multispectral and TIR thermal camera while in-situ measurements of temperature and surveying of the different LC patterns were performed. A significant correlation was detected between LSTs in situ measured and detected by UAV in 06 September (r=0.758; p-value <0.01) and in 27 October (r=0.686; p-value <0.01). When comparing the LSTs of separate land-cover patterns the best results were received for Bare soils (r=0.591; p-value <0.05) and Concrete (r=0.927; p-value <0.01) when the survey and measurements was done in September. Despite its limitations, it can be stated that UAV thermal survey has a good potential to detect and monitor LST variations in urban areas.

A Cloud Coverage Image Reconstruction Approach for Remote Sensing of Temperature and Vegetation in Amazon Rainforest

Remote sensing involves actions to obtain information about an area located on Earth. In the Amazon region, the presence of clouds is a common occurrence, and the visualization of important terrestrial information in the image, like vegetation and temperature, can be difficult. In order to estimate land surface temperature (LST) and the normalized difference vegetation index (NDVI) from satellite images with cloud coverage, the inpainting approach will be applied to remove clouds and restore the image of the removed region. This paper proposes the use of the neural network LaMa (large mask inpainting) and the scalable model named Big LaMa for the automatic reconstruction process in satellite images. Experiments are conducted on Landsat-8 satellite images of the Amazon rainforest in the state of Acre, Brazil. To evaluate the architecture’s accuracy, the RMSE (root mean squared error), SSIM (structural similarity index) and PSNR (peak signal-to-noise ratio) metrics were used. The LST and NDVI of the reconstructed image were calculated and compared qualitatively and quantitatively, using scatter plots and the chosen metrics, respectively. The experimental results show that the Big LaMa architecture performs more effectively and robustly in restoring images in terms of visual quality. And the LaMa network shows minimal superiority for the measured metrics when addressing medium marked areas. When comparing the results achieved in NDVI and LST of the reconstructed images with real cloud coverage, great visual results were obtained with Big LaMa.

Monitoring Land Surface Temperature in the Karajae Watershed as an influential ecosystem for Pare-Pare City

Temperature is an indicator that is often used to assess environmental conditions in an area because it affects food and air resources, hydrology, and all ecosystems on earth. Monitoring the temperature conditions of an area is also carried out from time to time to see trends that occur. Land surface temperature monitoring using remote sensing which has spatial resolution, temporal resolution, and spectral resolution. Monitoring is only in urban areas without looking at the surrounding conditions. However, temperature is influenced by several factors such as geographic location, position relative to the sun, altitude, and biophysical conditions such as land. This study looks at changes in land surface temperature in the Karajae watershed area as an ecosystem that affects urban areas, Pare-Pare City or vice versa. Analysis of land surface temperature monitoring using remote sensing methods in the form of multi-temporal Landsat Imagery data 2004-2021. The land image displays changes in land surface temperature with a single channel algorithm. In addition, land images are the basis for analyzing land cover changes which are the basis for estimating land surface temperature changes with normalized fractional vegetation cover relationships. The results show that changes in land cover must occur, which are estimated to be twice as large in the future. Settlements have experienced a very large increase in area, while the closure of forest land and dry agricultural land has experienced a reduction in area which will cause disruption to ecosystems and food. The trend of increasing land surface temperature continues to increase every period. The increase in settlements as built-up land, rice fields, dry land agriculture as a form of community activity causes an increase in land surface temperature. In addition, agricultural activities have an influence on increasing land surface temperatures in vegetated land cover. Indirectly, the city of Pare-Pare has an impact on increasing land surface temperatures in the upstream area of the Karajae watershed. This is a concern in the future, if urban activities, transportation, to industrial activities are much faster, and will cause a greater increase in land surface temperature.

Improving the Accuracy of Landsat 8 Land Surface Temperature in Arid Regions by MODIS Water Vapor Imagery

Land surface temperature (LST) is a significant environmental factor in many studies. LST estimation methods require various parameters, such as emissivity, temperature, atmospheric transmittance and water vapor. Uncertainty in these parameters can cause error in LST estimation. The present study shows how the moderate resolution imaging spectroradiometer (MODIS) water vapor imagery can improve the accuracy of Landsat 8 LST in different land covers of arid regions of Yazd province in Iran. For this purpose, water vapor variation is analyzed for different land covers within different seasons. Validation is performed using T-based and cross-validation methods. The image of atmospheric water vapor is estimated using the MODIS sensor, and its changes are investigated in different land covers. The bare lands and sparse vegetation show the highest and lowest accuracy levels for T-based validation, respectively. The root mean square error (RMSE) is also calculated as 0.57 °C and 1.41 °C for the improved and general split-window (SW) algorithms, respectively. The cross-validation results show that the use of the MODIS water vapor imagery in the SW algorithm leads to a reduction of about 2.2% in the area where the RMSE group is above 5 °C.

Estimation of Daily Mean Land Surface Temperature over the Qinghai–Tibet Plateau Based on an RTM-DTC Model

Accurately estimating daily mean land surface temperature (LST) is crucial for studying the urban heat island effect, land–atmosphere energy exchange, and global climate change. However, limited research has been conducted on average surface temperature estimation, particularly in high-altitude regions like the Qinghai–Tibet Plateau with extensive cloud cover. In this study, we propose the Reanalysis Data and Thermal Infrared Remote Sensing Data Merging-Diurnal Temperature Cycle (RTM-DTC) model specifically for the Qinghai–Tibet Plateau, successfully estimating mean LST using the model. We apply the RTM method to reconstruct LST under cloud cover from the MODIS LST product and calculate the average temperature using the DTC model. Validation with in situ measurements from seven meteorological stations on the Tibetan Plateau yielded daily scale RMSEs ranging from 1.81 K to 2.021 K and monthly scale RMSEs ranging from 1.77 K to 2.0 K, with an average RMSE of 1.91 K. These results demonstrate the adaptability of the RTM-DTC model and its ability to depict the annual variation curve of the mean surface temperature, and provide further research on RTM-DTC as a valuable approach.

Generating daily 100 m resolution land surface temperature estimates continentally using an unbiased spatiotemporal fusion approach

Fine spatial resolution (i.e., ≤ 100 m) land surface temperature (LST) data are crucial to study heterogeneous landscapes (e.g., agricultural and urban). Some well-known spatiotemporal fusion methods like the Spatial and Temporal Adaptive Reflectance Fusion Model (STARFM) and the Enhanced STARFM (ESTARFM), which were originally developed to fuse surface reflectance data, may not be suitable for direct application in LST studies due to the high sub-diurnal dynamics of LST. Furthermore, the effectiveness of spatiotemporal fusion methods for LST data has not been thoroughly evaluated in previous studies that only focused on relatively small spatiotemporal extents. To address these limitations, we proposed a variant of ESTARFM, referred to as the unbiased ESTARFM (ubESTARFM), specifically designed to accommodate the high temporal dynamics of LST to generate fine-resolution LST estimates. We evaluated ubESTARFM and ESTARFM against in-situ LST and the ECOsystem Spaceborne Thermal Radiometer Experiment on Space Station (ECOSTRESS) LST across 12 regions throughout Australia, encompassing various land covers and environments. Independent validation showed that ubESTARFM had a bias of 2.55 K, unbiased root mean squared error (ubRMSE) of 2.57 K, and Pearson correlation coefficient (R) of 0.95 against the in-situ LST over 11,290 observations at the 12 sites, all of which were considerably better than those calculated for ESTARFM, being a bias of 4.73 K, ubRMSE of 3.80 K and R of 0.92. When compared to ECOSTRESS data, ubESTARFM LST had a bias of −1.69 K, ubRMSE of 2.00 K, and R of 0.70 over 43 near clear-sky scenes, while ESTARFM LST had a bias of 1.79 K, ubRMSE of 2.68 K, and R of 0.59. Overall, our results demonstrated that ubESTARFM can avoid systematic bias accumulation, substantially reduce uncertainty deviation, and maintain a good level of correlation with validation datasets when compared to ESTARFM. A further assessment underscored the potential of ubESTARFM for application using LST data acquired from geostationary platforms (e.g., Himawari-8), with a mean ubRMSE (R) of 2.22 K (0.97) against in-situ LST over 1327 observations at 3 sites from southeast Australia at the overpass time of MODIS/Terra. This promising method leverages reliable numeric values from coarse-resolution LST while borrowing spatial heterogeneity from fine-resolution LST and has the potential to be coupled with energy balance and/or radiative transfer models thus enabling better farm and/or regional-scale water management strategies to be implemented. Furthermore, both the input and generated LST data, encompassing a comprehensive spatial extent over diverse land covers and climatic conditions, are publicly available for benchmarking future algorithmic refinements.

An adaptive synchronous extraction (ASE) method for estimating intensity and footprint of surface urban heat islands: A case study of 254 North American cities

The urban heat island (UHI) effect has attracted great attention due to its potential impacts on rapidly growing urban areas. Using remotely sensed estimates of land surface temperature (LST), a large number of studies have focused on the surface UHI (SUHI) effect, which can be characterized by its two fundamental properties: intensity and footprint. The SUHI intensity reflects the LST difference between the urban area and the background reference area (BRA), and the SUHI footprint indicates the spatial extent influenced by the heat island. Currently, numerous methods have been developed to estimate the SUHI intensity and footprint, but are still greatly challenged by three main issues. Namely, the discrepancy in BRA selection criterion brings great uncertainty to the estimated SUHI intensity, the estimation of SUHI footprint is largely constrained by the predefined models, and the quantification of SUHI effect is potentially influenced by several confounding factors. Here, we proposed an adaptive synchronous extraction (ASE) method, which is capable of adaptively selecting the most optimal BRA while removing the influence of confounding factors, and achieving synchronous estimation of SUHI intensity and footprint. We applied the ASE method to 254 North American cities and conducted an in-depth comparative analysis to discuss its applicability and benefits. The main results include: (1) The ASE method avoids the limitations of existing methods in BRA selection and model presetting, and shows resilience to parameter variations. This makes the ASE method highly applicable to quantify the SUHI intensity and footprint in cities with various thermal characteristics. (2) The ASE method can better highlight the spatial, seasonal and day-night contrasts in the estimated SUHI intensity. This superiority is particularly evident when comparing it to methods based on the equal-area buffer or the simplified urban-extent algorithm. (3) Confounding factors pose non-negligible impacts on the quantification of the SUHI effect. Typically, ignoring the influence of topographic relief or missing LST data can lead to an overall overestimation of the SUHI intensity, while not removing surrounding urban areas will cause some underestimation of the SUHI intensity. Overall, the proposed ASE method provides a new generalizable tool for quantifying the SUHI effect, which has great potentials for future studies and urban climate assessments.

A framework for reconstructing 1km all-weather hourly LST from MODIS data

ABSTRACT Land surface temperature (LST) is an essential parameter in environmental monitoring. However, due to the cloud contamination and the limitations of sensors, existing thermal infrared LST products are challenging to provide all-weather LST with high spatiotemporal resolution. Therefore, this paper presents a framework for reconstructing hourly LST under clouds based on the moderate resolution imaging spectroradiometer (MODIS) LST products. This framework consists of two main steps: (1) Instantaneous LST estimation for MODIS using a modified annual temperature cycle (ATCM) model and (2) Hourly LST reconstruction under all-weather conditions from atmospheric reanalysis data and MODIS instantaneous LST with the linear model (LM). The proposed framework is evaluated with the full year data of 2019 with tile number H26V05. The results show that the spatial distribution of the estimated LST is similar with the MODIS LST. Additionally, it could provide an accurate indication of the spatial and temporal variability of LST. Comparing the MODIS LST of the selected area for tile number H26V05 in 2019 with the estimated LST of the ATCM, the root mean squared errors (RMSEs) of the model are around 1K ~ 3K. The hourly LST is then reconstructed based on the LM method, and the RMSE is around 2K during the daytime and around 1K during the night-time. Finally, to verify the robustness of the hourly LST reconstruction method, the accuracy of the hourly LST was evaluated by the MODIS LST at different situations in the day. The results show that the missing data at one moment will cause the accuracy of the model decreasing by 0.09 ~ 1.67K at the corresponding moment. In general, the proposed framework has the potential to reconstruct the 1 km all-weather hourly LST with high accuracy and a certain level of robustness.

The effect of pixel heterogeneity on surface heat and water vapor flux estimated by the remote sensing-based model coupled with deep learning

Accurate quantification of sensible (H) and latent (LE) heat fluxes are desperately crucial for agricultural irrigation strategy, drought monitoring, water resource management, and climate change. Regional-scale H and LE were derived generally from satellite remote sensing data using relevant modeling methods. However, one of the challenges of modeling methods is to deal with sub-pixel heterogeneity, because the model usually assumes homogeneous conditions within a pixel. In this study, a hybrid model, which coupled deep neural network (DNN) and the remote sensing-based two-source energy balance (TSEB) model, was developed to estimate H and LE at multiple spatial resolutions to further quantify the influence of pixel size on the modeled fluxes. First, the DNN-based land surface temperature (LST) estimation model was developed and validated. LST at multiple spatial resolutions (i.e., 1, 5, 10, 30, 100, and 1000 m) were generated by the DNN method using multi-source satellite data including the GaoFen-2 (GF-2, a Chinese high-resolution optical remote-sensing satellite), Sentinel-2, and Landsat-8 images, respectively. Then, the hybrid DNN-TSEB model was used to estimate H and LE at multi-pixel sizes. Finally, modeled H and LE at 1 ∼ 1000 m were validated by eddy covariance (EC) and optical-microwave scintillometer (OMS) flux observation. Results revealed that the DNN-based LST model was a potential approach to generate LST at multiple pixel sizes. The hybrid DNN-TSEB model can produce more accurate H and LE when different land cover types were discriminated over heterogeneous farmland (e.g., the average bias for H and LE at a spatial resolution of 1 m were −9 and −6 W/m−2, respectively). Besides, the deviation of modeled fluxes from the observations increased as the spatial resolution became coarser, particularly the pixel size was greater than 30 m (average bias for H and LE were −34 and 30 W m−2, respectively). However, the magnitude of bias of fluxes estimates remained unchanged when the pixel sizes of satellite data were between 100 and 1000 m. The more important finding was that the hybrid model was a promising approach, which can obtain accurate estimates of H and LE at a high spatial resolution. In addition, the performance of the model based on GF-2 data was better than that based on Sentinel-2 and Landsat-8 data. The results suggest that pixel heterogeneity should be considered when estimating H and LE over heterogeneous landscapes with various surface types.

Four-decade response of land surface temperature to urban expansion in Beijing

Urban expansion with increased impervious surface area (ISA) can affect the thermal regimes by altering the water and energy balance. Previous studies have indicated that ISA is the primary surrogate of rising land surface temperature (LST). However, urban-induced LST changes are not yet well quantified because of the temporal limitations of the LST datasets and the uncertainties from climate variabilities. In this study, we integrated multiple satellite-based products (vegetation coverage, albedo and radiation forcing) into a physically-based land surface model i.e., Variable Infiltration Capacity (VIC), to detect multi-temporal scale response of LST to the increasing ISA in Beijing. Heat transfer-related parameters in the model were updated along with urbanization processes. The integrated modeling presented high performance for simulating LST, as evaluated with ground-based observations and the MODerate-resolution Imaging Spectroradiometer (MODIS) LST product. We found that the average LST over Beijing increased from 10.3 °C in 1980–1990 to 11.2 °C in 2010–2020, with nearly 33% contributed by the urban growth. The urban-induced thermal effect was particularly strong in summer daytime and winter nighttime. The frequency of heat days (the day with the maximum LST over 40 °C) in a year for Beijing appeared to linearly correlate with the impervious surface fraction (ISF) when ISF > 25%. For the urbanizing area, moreover, the four-decade LST estimates indicated that the annual overheating duration was obviously extended with a rate of about 5 days per decade. We therefore conclude that the urban expansion in Beijing not only amplified heat stress but also altered heat phenology, which have implications for urban planning and the construction of heat mitigation facilities.

A new approach to estimate daily evapotranspiration, based on Landsat data and FAO56 principles

ABSTRACT The accurate estimation of evapotranspiration is crucial for enhancing crop water productivity and effectively managing water resources. This research offers a novel method that integrates satellite data and crop coefficients to calculate ET c and ET a on a daily basis and overcome the limitation of low temporal frequency of non-commercial satellite data. The study was carried out in the southern part of Khuzestan Province, Iran, on sugarcane crops in the Amirkabir Agro-industries area. The method involves obtaining Landsat-8 data with an 8-day temporal resolution, which was used to estimate Land Surface Temperature (LST) using a Single-Channel Algorithm. The estimated LST was then validated with in-situ canopy temperature measurements and used to predict the crop stress coefficient (K s) based on its relationship with the crop water stress index (CWSI). The crop coefficient (K c) was obtained using the Surface Energy Balance Algorithm for Land (SEBAL) algorithm, and both Ks and Kc were utilized to calculate daily ETa by multiplying by the daily reference evapotranspiration (ET 0) obtained from local meteorological data. The results indicated that the crop coefficients of sugarcane in the initial and mid-stages were 12% and 18% higher, respectively, compared to the FAO56 guideline. The aggregated decadal and monthly ET a showed good agreement with the WaPOR datasets, with an RMSE of 8.7 and 1.93 mm, respectively. This approach offers a potential solution to the challenge of obtaining remote sensing data with a higher temporal frequency.

Satellite-based land surface temperature and soil moisture observations accurately predict soil respiration in temperate deciduous and coniferous forests

Soil respiration, defined as the total flux of carbon dioxide (CO2) from the soil to the atmosphere, is a key ecosystem process that affects the regional and global carbon (C) cycles and is highly sensitive to temperature and soil moisture. It is challenging to quantify soil respiration at the ecosystem level from commonly used in-situ soil chamber measurements because of large spatial variability. Methods that provide temporally and spatially continuous estimates of soil respiration at various scales are vital to understand the impact of climate change on soil C stock. In this study, we evaluate three commonly used empirical models and a Random Forest machine learning algorithm applied to satellite derived estimates of land surface temperature (LST) and soil moisture to estimate soil respiration in temperate deciduous and coniferous forests in Canada. The models were calibrated using in-situ soil temperature and moisture and validated against in-situ measurements of soil CO2 fluxes (gCm−2day−1) from automatic soil chambers. We separately evaluate the performance of nighttime and daytime satellite-based LST and soil moisture observations in modeling soil respiration. The soil respiration models were also evaluated at daily and monthly time scales against in-situ measurements. Results indicate that models based on satellite LST, and soil moisture can explain more than 70% of the variability in observed soil respiration. Nighttime LST at a monthly time scale resulted in consistently higher accuracy than daytime LST in estimating soil respiration. Satellite observations resulted in comparable accuracy in estimating soil respiration as in-situ measurements. Satellite LST and soil moisture observations are indispensable data sources to estimate soil respiration at ecosystem level and its upscaling to regional and global scales.