Longwave infrared (LWIR) sensors are widely used for measuring land surface radiation in ground and unmanned aerial vehicle (UAV) remote sensing missions. Although the land surface temperature (LST) retrieval algorithms for thermal in-frared (TIR) satellite sensors with narrow spectral response ranges have achieved good results, they are generally unsuitable for LWIR sensors. At present, the LST retrieval algorithm for LWIR data needs further investigation. In this study, an im-proved radiative transfer (IRT) algorithm based on the segmen-tation of spectral response function (SRF) is proposed for retriev-ing LST from LWIR data. The IRT algorithm is applied to three types of commonly used LWIR sensors. The simulation results show that the root-mean-squared error (RMSE) is lower than 0.1 K when the segmentation width is 0.2 μm. The higher the height of the sensor, the more obvious the fluctuation of the accuracy increases with the segmentation width. Using thein-situdata of the Heihe River basin (HRB) for validation, RMSEs are between 1.1 and 1.8 K, depending on different land cover types. The IRT algorithm can retrieve the relatively high-accuracy LSTs from LWIR data observed by a variety of LWIR sensors, and promote the collaborative application of multi-sensor LSTs, which is of great significance in ecological environment research.

Land Surface Temperature (LST) is a critical parameter for monitoring surface energy balance and evaluating climate and environmental changes. However, LST retrieval from thermal infrared satellite remote sensing often suffers from data gaps in cloud-affected regions. Existing methods for estimating cloud-covered LST do not adequately account for physical mechanisms under complex meteorological and surface conditions, nor do they address dynamic error variations during the fusion process. To address these limitations, this study integrates the numerical weather prediction model (WRF), land surface model (Noah-MP) and satellite observation data. It comprehensively evaluates the accuracy of the LST simulated by the WRF and Noah-MP. Moreover, a data assimilation and fusion method based on the Kalman filter is used to consider the changes of errors, and the dynamic fusion of these LST data is carried out to obtain the hourly LST with a resolution of 1 km. Furthermore, assimilating downward shortwave and longwave radiation into the Noah-MP model improves its simulated LST accuracy to a certain extent. The fused LST is not only spatially continuous but also exhibits improved reliability. Validation within-situmeasurements shows that the Root Mean Square Error (RMSE) under clear-sky conditions is 2.56 K, and the RMSE of the LST under all-weather conditions is 2.88 K. This method has good potential in generating spatially continuous LST with high temporal and spatial resolution, thus promoting relevant research and applications.

Passive microwave brightness temperature (PMW BT) images acquired by PMW imagers onboard polar-orbit satellites suffer from large observations missing between adjacent orbits due to the swath width of images, i.e., PMW BT swath gaps, limiting the spatiotemporal integrity and application potential of PMW BT-generated products. Here, we propose a gap-filling method [i.e., physical indicators-guided CNN-LSTM (PICL)] for PMW BT images by physically guided selection of surface indicators with CNN-LSTM model, which is suitable for special underlying surfaces (e.g., seasonal permafrost and snow) using only BT data to generate spatially gapless PMW BT images. The core of PICL is to use the CNN-LSTM model to capture the relationship of BT time series, thereby filling the missing BT values via historical BT data. PICL is applied to 7, 10, 18.7, 36, and 89 GHz frequencies of Advanced Microwave Scanning Radiometer 2 (AMSR2) for the Tibetan Plateau (TP). Validation results show good accuracy of the PICL filled BT, with the root-mean-squared error (RMSE) ranging from 1.28 to 2.43 K (<89 GHz), and the accuracy decreases as the frequency increases. The reconstructed BT images agree well with the original AMSR2 BT images and show no obvious boundary effect. PICL also has a good ability in capturing the temporal trends and discontinuities caused by snow and seasonal permafrost. PICL only requires historical BT before the missing moment, highlighting its feasibility to be extended to other satellite PMW imagers. It enables the generation of spatially seamless products such as all-weather land surface temperature (LST) and soil moisture.

Earth’s surface temperature (EST), encompassing both land surface temperature (LST) and sea surface temperature (SST), serves as a crucial indicator of climate change. This study introduces a groundbreaking framework for the daily estimation of Earth’s Surface Temperature (EST), integrating reanalysis data with thermal infrared remote sensing data merging (RTM) techniques and employing machine learning methods. The spatial distribution of the generated all-weather EST aligns effectively with MODIS EST, showcasing its capability to recover EST values in cloudy regions and estimate missing values in orbital gap areas. Validation results for all-weather LST and SST demonstrate commendable accuracy, with minimal variations observed under both clear-sky and cloudy conditions. The Root Mean Square Error (RMSE) for LST ranges from 1.69 to 2.84 K, while for SST, it spans from 0.38 °C to 0.59 °C. The framework exhibits adaptability to diverse weather conditions, maintaining consistent relative trends across different geographical locations. In summary, this innovative approach provides a robust solution for generating all-weather ESTs, effectively addressing challenges associated with conventional Thermal Infrared (TIR) data.

While numerous algorithms have been developed for retrieving land surface temperature (LST) and various LST products have been released for satellite thermal infrared (TIR) remote sensing, capturing thermal details on finer scales remains challenging due to limitations in revisit period and spatial resolution. Unmanned aerial vehicle (UAV) TIR remote sensing, on the other hand, proves capable of obtaining LST at high to super-high spatial resolutions, thereby supporting studies such as evapotranspiration estimation and precision agriculture. However, challenges arise from the operational characteristics of UAVs and the inherent defects in UAV-borne TIR imagers, which leads to issues in the obtained data. Moreover, the lack of methods to convert LST between ground, UAV, and satellite scales hampers the validation of LST products and impacts our understanding of LST variation from regional to global scales. Therefore, a MUlti-Scale Observation Experiment on land Surface temperature using UAV remote sensing (MUSOES-UAV) was designed and implemented in the middle reaches of the Heihe River basin. MUSOES-UAV provides a research basis for obtaining reliable, high-accuracy LST, offering new insights into the spatiotemporal changes of LST.

Land surface temperature (LST) is a widely focused parameter between the land surface and the atmosphere. Currently, satellite remote sensing is the main approach to obtaining regional and global LST. Validation of satellite-derived LST can promote its application and provide feedback for the retrieval algorithms and parameterization schemes. The current widely used temperature-based method faces many influences, e.g., obtaining the ground truth on the pixel scale and its uncertainty. Here, a comprehensive validation scheme is proposed for validating the satellite-derived LST by combining the near-surface atmospheric correction for longwave-radiation-based in situ LST and considering the validation station’s spatial representativeness, and applying it in the validation of AVHRR-derived LST. The LST “ground truth” of three validation stations in Heihe River Basin, China was obtained, with a mean uncertainty of 0.87 K (range:$0.47\sim 2.96$K), 1.07 K ($0.49\sim 1.81$K), and 0.61 K ($0.47\sim 1.13$K) for A’rou superstation (ARS), daman superstation (DMS), and sidaoqiao superstation (SDQ), respectively. Validation of AVHRR-derived LST against the obtained “ground truth” shows that the random error is lower than 3 K, and the system error is station-dependent, with a range of$- 1.02\sim 3.93$K. Further comparison indicated significant systematic error differences (range:$- 1.5\sim 3.45$K) and inapparent random errors difference ($- 0.84\sim 0.43$K) between the proposed comprehensive scheme and the classic scheme at the selected stations. Since the main influences are considered in the proposed comprehensive validation scheme, the validation results are more objective and credible. The comprehensive validation scheme provides a reference for LST validation and could be extended to the validation of related hydrothermal parameters.

Sea surface temperature (SST) is an important parameter for assessing sea-atmosphere energy interaction and understanding climate change. One of the primary approaches for obtaining global-scale SST is retrieving from satellite thermal infrared remote sensing data. However, it is challenging to accurately retrieve large-scale SST due to the complexity of retrieving the key intermediate parameter, i.e., sea surface emissivity (SSE), using the standard theoretical model. In this study, we proposed a simplified SSE estimation model based on the satellite view zenith angle (VZA) and wind speed and compared it with three commonly used SSE estimation models. Then, the retrieved SSTs based on those SSEs were validated againstin-situSSTs. Results show that the SSE from the proposed estimation model shows the highest consistency and the lowest biases with the theoretical values compared to the other three estimation models, especially in large VZAs.In-situobservation-based SST validation results show that the SST retrieved using the proposed SSE estimation model also achieves the highest accuracy compared to the other three SSTs, with a mean bias error of 0.08 K, and a root-mean-square error of 0.30 K, which is close to the official SST products. In conclusion, the proposed SSE estimation model shows good performance both in SSE estimating and SST retrieving. Furthermore, the proposed model has the potential to estimate SSE on large scales that can serve as a reference for obtaining SST from other similar sensors to promote the development of marine remote sensing.

In-situ land surface temperature (LST) measuring plays a crucial role in quantitative remote sensing, as well as many environmental and climate studies. However, selecting representative sites for in-situ LST measuring is often challenging, as the spatial representativeness of LST measurements must be considered. Therefore, a site selection method for radiation-based in-situ LST measuring was proposed based on the ground station’s spatial representativeness evaluation method. This paper presents a case study of the site pre-selection for a meteorological research station located at Chengdu, China. The related results can provide a basis for the subsequent selection and construction of the station.

Time series analysis of land surface temperature (LST) is one of the most important topics in climate change-related research. As the third pole of the Earth and the water tower of Asia, the temperature change of the Qinghai-Tibet Plateau will inevitably affect the rapid response of the surrounding environment. Currently, many studies analyzed the spatio-temporal variation of LST in this area. However, due to cloudy weather conditions, the time series of the clear-sky LST may introduce large errors in their conclusions. Therefore, in this study, a newly released spatiotemporal seamless satellite all-weather LST product (TRIMS LST), as well as MODIS LST (MYD21), is employed to model and analyze the LST time series under all-weather conditions on the Qing-Tibet Plateau. Results show that the tendency of LST variation from clear-sky LST is weaker than that from all-weather LST. The all-weather LST indicates a warming trend on the Qing-Tibet Plateau.

Acquiring the satellite land surface temperature (LST) with high spatiotemporal resolutions is pressing in the land surface biophysical process. However, most current LST products hardly satisfy this requirement. LST Downscaling provides an effective way to solve this issue by introducing driving factors, but existing methods usually ignore the weights of descriptors. In this letter, based on the Geographically Weighted Regression (GWR) and Random Forest (RF), a new downscaling method (i.e., WGWR) considering the weights of LST descriptors is proposed. To examine the performance of WGWR, the 100-m Landsat-8 TIRS and Terra ASTER LSTs are aggregated to 1000 m as the simulated coarse LSTs, and then the coarse LSTs are downscaled to 100 m using WGWR, RF, and GWR. Meanwhile, the original 100-m LSTs are used as validation references. Results indicate that the proposed WGWR outperforms RF and GWR: for RF (GWR), the RMSEs can be reduced by 0.34 K (0.26 K) in Zhangye and 0.22 K (0.1 K) in Beijing. Compared to RF and GWR, WGWR also yields better image quality: the downscaled LST images have neither obvious smoothing effect nor boundary effect and maintain the details of the image at high spatial resolution. Validation based onin-situLST indicates that the downscaled LST based on WGWR has better agreement with thein-situLST, and the RMSE is reduced by 0.57 K. The proposed WGWR contributes to obtain high spatio-temporal resolution LSTs and promote hydrological, meteorological, and ecological studies.

It is urgently needed to obtain the hourly near-real-time all-weather land surface temperature (NRT-AW LST) for immediately monitoring the disaster and environmental changes. Nevertheless, studies on estimating hourly NRT-AW LST are in the preliminary stage. In this study, we proposed a Spatio-TEmporal Fusion (STEF) method for fusing the reanalysis dataset derived from China Land Surface Data Assimilation System (CLDAS) and thermal infrared (TIR) data derived from the Chinese Fengyun-4A (FY-4A) geostationary satellite to estimate the hourly NRT-AW LST with 0.04° resolution. STEF method can produce NRT-AW LST without relying on the data after the target moment. STEF is tested in the Tibetan Plateau. Validation results on DOY 215-366 of 2020 indicate that STEF has good accuracy: RMSEs (MBEs) under clear-sky, cloudy-sky, and all-weather conditions vary from 2.74 K (-1.06 K) to 3.77 K (0.14 K), from 3.31 K (-1.40 K) to 4.46 K (-0.22 K), and from 3.10 K (-1.11 K) to 3.87 K (-0.22 K), respectively. STEF method can improve the accuracies of FY-4A LST, and RMSEs are reduced by about 0.77 K to 1.82 K. The NRT-AW LSTs estimated by STEF have better accuracies than CLDAS LSTs under all-weather conditions. The SETF also exhibited similar results in 2021. We believe that the proposed STEF method can meet the requirements of NRT-AW LST estimation and contributes to improving the timeliness of region monitoring and related parameter estimations.

Glacial debris flows are a common geological hazard in theglacial region of the Tibetan Plateau. This study analyzed the relationship between land surface temperature (LST) and glacial debris flow in the southeastern part of the Tibetan Plateau. LST showed a year-to-year upward trend, which was more pronounced in the glacial region, throughout the study area. After analyzing the causes of eight glacial debris flows, we found that the sudden increase of LST and the long-term high LST in the early period are the main causes besides the rainfall. The results of the study show that LST can be an effective parameter for monitoring and forecasting glacial debris flows.

Aerosol optical depth (AOD) is closely related to atmospheric pollutants. However, a large number of missing values in satellite AOD severely limits its application. We proposed a spatiotemporal random forest (RF) model to estimate the missing AHI AOD in this study. In addition to the commonly used meteorological and topographic parameters, the spatiotemporal data and MERRA-2 AOD were introduced as the model inputs. Specifically, the training data was divided into multiple subsets based on the land cover types and local times to explicitly characterize the spatiotemporal variation of AOD. The validation results indicated that the RF model achieved promising results with RMSE of 0.03 to 0.17, MBE of −0.01 to 0.02, and R of 0.87 to 0.97 at different land cover types and local times.

Land surface temperature (LST) plays a key role in various land surface processes. Limited to the balance between the spatial resolution and revisit interval, it is difficult to obtain the high spatiotemporal resolution LST via satellite remote sensing. Downscaling is an economical approach to achieve it, and it has been relatively mature associated with a large number of land surface parameters. However, it still needs to address issues such as the physical meaning of the downscaling method. This study implemented a practical LST downscaling method that combined the spatial and temporal information to obtain a higher spatial resolution LST over an oasis in the Heihe River basin. The downscaled 500-m and 250-m LST shows more details, especially the boundary between the oasis and the desert. The validation against in-situ LST also shows that the downscaled LST has similar accuracy and precision with the original MODIS LST, with a RMSE of 1.78 K and 1.70 K at daytime, 1.36 K and 1.40 K at nighttime, for 500-m and 250-m, respectively.

Pedestrian detection from UAV images is vital for many fields. Given that visible images are susceptible to bad illumination, thermal images with the ability to characterize the temperature of an object can provide auxiliary information. It is useful to fuse the visible and thermal images to improve the pedestrian detection performance. Unfortunately, studies on pedestrian detection with UAV visible-thermal image pairs are still rare. Therefore, we propose a method for Multi-source Pedestrian Detection based on Feature Fusion (MPDFF). With the registered visible and thermal image pairs as the input, MPDFF can achieve better characterization of pedestrians by concatenating the features from the two images. MPDFF performs much better than the methods that use only single-source images, which demonstrates that visible and thermal images are complementary in pedestrian detection.

Near surface air temperature (NSAT) products are required for environment-related researches and applications. Existing NSAT products vary in spatial-temporal resolution and data quality. Thus, it is necessary to evaluate and investigate the difference of different NSAT products to provide an overall assessment to help researchers and users to choose and use among the many NSAT products. In this study, Tibetan Plateau was selected as our study area, and six released NSAT products were collected for comparison and evaluation. The NSAT products were compared with in situ NSAT from China Meteorological Administration stations (CMA) respectively. The evaluation process was conducted from daily and monthly scale and gave out the accuracy ranking of the six NSAT products.

Unmanned aerial vehicle (UAV) thermal infrared (TIR) remote sensing is an important way to obtain land surface temperature (LST) with high spatial and temporal resolutions. Due to wide spectral response function (SRF) ranges of UAV thermal imagers, currently available LST retrieval methods suitable for satellite sensors may induce significant uncertainty when applied to UAV sensors. Despite that some methods have been proposed to retrieve LST from UAV remote sensing, studies considering the adverse effect caused by the SRF ranges are still rare. Here, we present a so-called Temperature Retrieval for UAV Broadband thermal imager data (TRUB) method to retrieve LST from UAV broadband thermal imager data. TRUB’s core includes two parts: 1) a simple lookup table (LUT) algorithm for reducing the uncertainty induced by the wide SRF ranges; and 2) models suitable for UAV remote sensing for estimating the atmospheric parameters. Validation from the Heihe River Basin shows that the LST retrieved by TRUB, of which the root mean square error (RMSE) and mean bias error (MBE) is 1.71 and −0.02 K, respectively, is highly consistent with thein situLST. TRUB is helpful to reduce the uncertainty caused by the wide SRF ranges of UAV thermal imagers and quantify the influence of atmosphere, thus can obtain UAV remote-sensing LST with better accuracy in large-area operating missions.

Thermal infrared (TIR) land surface temperature (LST) products derived from geostationary satellites have a high temporal resolution in a diurnal cycle, but they have many missing values under cloudy-sky conditions. Therefore, it is pressing to obtain all-weather LST (AW LST) with a high temporal resolution by filling the gap of TIR LST. In this study, a method integrating reanalysis data and TIR data from geostationary satellites (RTG) was proposed for reconstructing hourly AW LST. Then, taking the Tibetan Plateau, which is a focus of climate change as a case, RTG was applied to the Chinese Fengyun-4A (FY-4A) TIR LST and China Land Surface Data Assimilation System (CLDAS) data. Validation based on thein-situLST shows that the accuracy of the AW LST is better than the FY-4A LST and CLDAS LST under clear-sky, cloudy-sky, and all-weather conditions. The mean RMSEs are 3.02 K for clear-sky conditions, 3.94 K for cloudy-sky conditions, and 3.57 K for all-weather conditions. Uncertainty and coarse resolution of the original FY-4A and CLDAS data affect the accuracy of the obtained AW LST. The results of the LST time series comparison also show that the reconstructed AW LST is consistent within-situLST. The reconstructed AW LST also has good image quality and provides reliable spatial patterns. RTG is practical in obtaining high temporal resolution AW LST from the Chinese FY-4A to satisfy related applications. It can also be extended to other geostationary satellites and reanalysis datasets.

Land Surface Temperature (LST) is a key factor in the land surface energy budget. The accuracy of the LST is affected by topographical fluctuations, observation time, and other factors. Thus, it is necessary to validate the retrieved LST products. In this study, the Fengyun-4A (FY-4A) LST was evaluated against the in-situ LST, which is collected from four ground sites in the Heihe River basin from August 1st, 2019 to December 31st, 2019. The results show that the root-mean-square error(RMSE) varies from 2.39 K to 4.07 K. Therefore, it is considered that FY-4A LST has good correlations with the in-situ LST. However, the FY-4A LST product has large systematic errors over some sites, e.g Jingyangling. The main reason is that the longwave radiometer has a scale mismatch between the pixels of FY-4A, and the scale mismatch can affect the representatives of measurements at pixel scales.

Land surface temperature (LST) is an important parameter for series land surface processes, models and applications. The accuracy of LST directly influenced its application. Therefore, a reasonable validation method is meaningful to assess the accuracy of LST datasets. In this study, an in-situ observation representativeness assessment method was proposed. Based on this method, the JPSS VIIRS LST product was validated against in-situ LST at 7 ground sites over the Heihe River Basin during the HiWATER experiments period. Results show that about 70%, 28%, 43%, 68%, 42%, 35% and 25% of the FOV LST for ARS, DMS, DSL, EBO, HHL, JYL, and SDQ are able to well represent the corresponding LST of VIIRS pixels, respectively and determined the representativeness period of each site. The VIIRS LST has a high correlation with the in-situ LST with an accuracy of 2.30 K - 5.76 K at daytime and 1.26 K-2.68 K at nighttime, respectively.

Land Surface Temperature (LST) is an important input parameter for many land surface models. The accuracy of satellite LST products directly affect its application; therefore, it is necessary to evaluate LST products. In this study, two satellite remotely sensed LST products, i.e. AMSR2 LST and MODIS LST, were evaluated against the in-situ LSTs at 17 ground sites in Heihe River Basin in 2014. Results show that both AMSR2 and MODIS LSTs have good correlations with the in-situ LST, with R2 from 0.80 to 0.98 except at AR2 site at daytime. However, both of these two products have large systematic errors compared with the in-situ LST. The possible main reason is the scale mismatch between the FOV of the longwave radiometer and the AMSR2 and MODIS pixels.