A Review of Reconstructing Remotely Sensed Land Surface Temperature under Cloudy Conditions

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.

Applying deep learning to manage urban ecosystems in arid Abha, Saudi Arabia: Remote sensing-based modelling for ecological condition assessment and decision-making

This study aims to quantitatively and qualitatively assess the impact of urbanisation on the urban ecosystem in the city of Abha, Saudi Arabia, by analysing land use changes, urbanisation processes and their ecological impacts. Using a multidisciplinary approach, a novel remote sensing-based urban ecological condition index (RSUSEI) will be developed and applied to assess the ecological status of urban surfaces. Therefore, the identification and quantification of urbanisation is important. To do so, we used hyper-tuned artificial neural network (ANN) as well as Land Cover Change Rate (LCCR), Land Cover Index (LCI) and Landscape Expansion Index (LEI). For the development of (RSUSEI), we have used four advanced models such as fuzzy Logic, Principle Component Analysis (PCA), Analytical Hierarchy Process (AHP) and fuzzy Analytical Hierarchy Process (FAHP) to integrate various ecological parameters. In order to obtain more information for better decision making in urban planning, sensitivity and uncertainty analyses based on a deep neural network (DNN) were also used. The results of the study show a multi-layered pattern of urbanisation in Saudi Arabian cities reflected in the LCCR, indicating rapid urban expansion, especially in the built-up areas with an LCCR of 0.112 over the 30-year period, corresponding to a more than four-fold increase in urban land cover. At the same time, the LCI shows a remarkable increase in 'built-up' areas from 3.217% to 13.982%, reflecting the substantial conversion of other land cover types to urban uses. Furthermore, the LEI emphasises the complexity of urban growth. Outward expansion (118.98 km2), Edge-Expansion (95.22 km2) and Infilling (5.00 km2) together paint a picture of a city expanding outwards while filling gaps in the existing urban fabric. The RSUSEI model shows that the zone of extremely poor ecological condition covers an area of 157-250 km2, while the natural zone covers 91-410 km2. The DNN based sensitivity analysis is useful to determine the optimal model, while the integrated models have lower input parameter uncertainty than other models. The results of the study have significant implications for the management of urban ecosystems in arid areas and the protection of natural habitats while improving the quality of life of urban residents. The RSUSEI model can be used effectively to assess urban surface ecology and inform urban management techniques.

Estimating daily air temperature and pollution in Catalonia: A comprehensive spatiotemporal modelling of multiple exposures

Environmental epidemiology studies require models of multiple exposures to adjust for co-exposure and explore interactions. We estimated spatiotemporal exposure to surface air temperature and pollution (PM2.5, PM10, NO2, O3) at high spatiotemporal resolution (daily, 250 m) for 2018-2020 in Catalonia. Innovations include the use of TROPOMI products, a data split for remote sensing gap-filling evaluation, estimation of prediction uncertainty, and use of explainable machine learning. We compiled meteorological and air quality station measurements, climate and atmospheric composition reanalyses, remote sensing products, and other spatiotemporal data. We performed gap-filling of remotely-sensed products using Random Forest (RF) models and validated them using Out-Of-Bag (OOB) samples and a structured data split. The exposure modelling workflow consisted of: 1) PM2.5 station imputation with PM10 data; 2) quantile RF (QRF) model fitting; and 3) geostatistical residual spatial interpolation. Prediction uncertainty was estimated using QRF. SHAP values were used to examine variable importance and the fitted relationships. Model performance was assessed via nested CV at the station level. Evaluation of the gap-filling models using the structured split showed error underestimation when using OOB. Temperature models had the best performance (R2 =0.98) followed by the gaseous air pollutants (R2 =0.81 for NO2 and 0.86 for O3), while the performance of the PM2.5 and PM10 models was lower (R2 =0.57 and 0.63 respectively). Predicted exposure patterns captured urban heat island effects, dust advection events, and NO2 hotspots. SHAP values estimated a high importance of TROPOMI tropospheric NO2 columns in PM and NO2 models, and confirmed that the fitted associations conformed to prior knowledge. Our work highlights the importance of correctly validating gap-filling models and the potential of TROPOMI measurements. Moderate performance in PM models can be partly explained by the poor station coverage. Our exposure estimates can be used in epidemiological studies potentially accounting for exposure uncertainty.

High-resolution soil moisture mapping using PSO-based optimized cerebellar model articulation controller (CMAC)

A good knowledge in eco-hydrological processes requires significant understanding of geospatial distribution of soil moisture (SM). However, SM monitoring remains challenging due to its large spatial variability and its dynamic time response. This study was performed to assess the performance of a particle swarm optimization (PSO)-based optimized Cerebellar Model Articulation Controller (CMAC) in generating high-resolution surface SM estimates using sentinel-2 imagery over a Mediterranean agro-ecosystem. Furthermore, the results were compared with those of PSO-based optimized group method of data handling (GMDH) as a more common data-driven method. Two different modeling approaches i.e. modeling in homogenous clusters (local approach) and modeling in entire area as an entity (global approach) were examined. Candidate predictors namely sentinel-2 spectral bands, normalized difference vegetation index (NDVI) and normalized difference water index (NDWI), digital elevation model (DEM), slope and aspect were used as the input variables to estimate SM. An intensive field survey had been done to gather in-situ SM data using a time-domain reflectometer (TDR). K-fold validation based on in-situ SM measurements demonstrated the reasonability of the SM estimation of the proposed methodology. Detecting homogeneous areas was done using genetic and particle swarm optimization algorithms. Synthesized SM product of PSO-GMDH showed a mean Normalized Root-Mean-Square Error (NRMSE) of 13.6 to 8.91 for global and local approaches in the test phase. PSO-CMAC method with an average NRMSE of 12.47 to 8.72 for global and local approaches shows the highest accuracy and outperforms the PSO-GMDH method at both local and global approaches. Overall, results revealed that clustering study area prior to running machine learning (ML) models coupled with optical satellite imagery and geophysical properties, boosts their predictive performance and can lead to more accurate mapping of SM with more heterogeneity. The results also showed that the global approach had a moderate performance in capturing the SM heterogeneity.

Reconstruction of Gap-Free Land Surface Temperature at a 100 m Spatial Resolution from Multidimensional Data: A Case in Wuhan, China

Land surface temperatures (LST) are vital parameters in land surface-atmosphere interactions. Constrained by technology and atmospheric interferences, LST retrievals from various satellite sensors usually return missing data, thus negatively impacting analyses. Reconstructing missing data is important for acquiring gap-free datasets. However, the current reconstruction methods are limited for maintaining spatial details and high accuracies. We developed a new gap-free algorithm termed the spatial feature-considered random forest regression (SFRFR) model; it builds stable nonlinear relationships to connect the LST with related parameters, including terrain elements, land coverage types, spectral indexes, surface reflectance data, and the spatial feature of the LST, to reconstruct the missing LST data. The SFRFR model reconstructed gap-free LST data retrieved from the Landsat 8 satellite on 27 July 2017 in Wuhan. The results show that the SFRFR model exhibits the best performance according to the various evaluation metrics among the SFRFR, random forest regression and spline interpolation, with a coefficient of determination (R²) reaching 0.96, root-mean-square error (RMSE) of 0.55, and mean absolute error (MAE) of 0.55. Then, we reconstructed gap-free LST data gathered in Wuhan from 2016 to 2021 to analyze urban thermal environment changes and found that 2020 presented the coolest temperatures. The SFRFR model still displayed satisfactory results, with an average R² of 0.91 and an MAE of 0.63. We further discuss and discover the factors affecting the visual performance of SFRFR and identify the research priority to circumvent these disadvantages. Overall, this study provides a simple, practical method for acquiring gap-free LST data to help us better understand the spatiotemporal LST variation process.

Satellite Altimetry: Achievements and Future Trends by a Scientometrics Analysis

Scientometric reviews, facilitated by computational and visual analytical approaches, allow researchers to gain a thorough understanding of research trends and areas of concentration from a large number of publications. With the fast development of satellite altimetry, which has been effectively applied to a wide range of research topics, it is timely to summarize the scientific achievements of the previous 50 years and identify future trends in this field. A comprehensive overview of satellite altimetry was presented using a total of 8541 publications from the Web of Science Core Collection covering the years from 1970 to 2021. We begin by presenting the fundamental statistical results of the publications, such as the annual number of papers, study categories, countries/regions, afflictions, journals, authors, and keywords, in order to provide a comprehensive picture of satellite altimetry research. We discuss the co-occurrence of the authors in order to reveal the global collaboration network of satellite altimetry research. Finally, we utilised co-citation networks to detect the development trend and associated crucial publications for various specific topics. The findings show that satellite altimetry research has been changed immensely during the last half-century. The United States, France, China, England, and Germany made the most significant contributions in the field of satellite altimetry. The analysis reveals a clear link between technology advancements and the trend in satellite altimetry research. As a result, wide swath altimetry, GNSS-reflectometry, laser altimetry, terrestrial hydrology, and deep learning are among the most frontier study subjects. The findings of this work could guide a thorough understanding of satellite altimetry’s overall development and research front.

Outlier Reconstruction of NDVI for Vegetation-Cover Dynamic Analyses

The normalized difference vegetation index (NDVI) contains important data for providing vegetation-cover information and supporting environmental analyses. However, understanding long-term vegetation cover dynamics remains challenging due to data outliers that are found in cloudy regions. In this article, we propose a sliding-window-based tensor stream analysis algorithm (SWTSA) for reconstructing outliers in NDVI from multitemporal optical remote-sensing images. First, we constructed a tensor stream of NDVI that was calculated from clear-sky optical remote-sensing images corresponding to seasons on the basis of the acquired date. Second, we conducted tensor decomposition and reconstruction by SWTSA. Landsat series remote-sensing images were used in experiments to demonstrate the applicability of the SWTSA. Experiments were carried out successfully on the basis of data from the estuary area of Salween River in Southeast Asia. Compared with random forest regression (RFR), SWTSA has higher accuracy and better reconstruction capabilities. Results show that SWTSA is reliable and suitable for reconstructing outliers of NDVI from multitemporal optical remote-sensing images.

All-sky 1 km MODIS Land Surface Temperature Reconstruction Considering Cloud Effects Based on Machine Learning

A high spatio-temporal resolution land surface temperature (LST) is necessary for various research fields because LST plays a crucial role in the energy exchange between the atmosphere and the ground surface. The moderate-resolution imaging spectroradiometer (MODIS) LST has been widely used, but it is not available under cloudy conditions. This study proposed a novel approach for reconstructing all-sky 1 km MODIS LST in South Korea during the summer seasons using various data sources, considering the cloud effects on LST. In South Korea, a Local Data Assimilation and Prediction System (LDAPS) with a relatively high spatial resolution of 1.5 km has been operated since 2013. The LDAPS model’s analysis data, binary MODIS cloud cover, and auxiliary data were used as input variables, while MODIS LST and cloudy-sky in situ LST were used together as target variables based on the light gradient boosting machine (LightGBM) approach. As a result of spatial five-fold cross-validation using MODIS LST, the proposed model had a coefficient of determination (R2) of 0.89–0.91 with a root mean square error (RMSE) of 1.11–1.39 °C during the daytime, and an R2 of 0.96–0.97 with an RMSE of 0.59–0.60 °C at nighttime. In addition, the reconstructed LST under the cloud was evaluated using leave-one-station-out cross-validation (LOSOCV) using 22 weather stations. From the LOSOCV results under cloudy conditions, the proposed LightGBM model had an R2 of 0.55–0.63 with an RMSE of 2.41–3.00 °C during the daytime, and an R2 of 0.70–0.74 with an RMSE of 1.31–1.36 °C at nighttime. These results indicated that the reconstructed LST has higher accuracy than the LDAPS model. This study also demonstrated that cloud cover information improved the cloudy-sky LST estimation accuracy by adequately reflecting the heterogeneity of the relationship between LST and input variables under clear and cloudy skies. The reconstructed all-sky LST can be used in a variety of research applications including weather monitoring and forecasting.

Surface urban heat island detected by all-weather satellite land surface temperature

Since the existing satellite thermal infrared (TIR) land surface temperature (LST) is susceptible to cloud contamination and other factors, surface urban heat island (SUHI) studies based on TIR LST are limited to clear-sky conditions and are not representative of SUHI under all-weather conditions, which result in a possible clear-sky bias for SUHI. This study introduces a newly released 1-km all-weather LST product (TRIMS LST), which is spatiotemporally seamless, to investigate the real SUHI under all-weather conditions for five megacities (i.e. Harbin, Beijing, Shanghai, Guangzhou, and Chengdu) in China. Firstly, this study compares TRIMS SUHI with MODIS SUHI under clear-sky, partial-cloudy, and cloudy conditions. Secondly, the extent of the influence of cloudiness on SUHI is quantified. Finally, the monthly TRIMS SUHI is used to analyze the clear-sky bias that is caused by using only clear-sky data for the SUHI. Results indicate that (i) the absence of pixel data leads to negative offsets in the SUHI intensities calculated by MODIS LST, and these offsets expand gradually with increases in the number of missing-pixel data, causing the maximum offset to reach -1.83 °C under cloudy conditions in Chengdu; (ii) cloud can mitigate the SUHI for most cities: when the cloud coverage in Guangzhou reaches 90-100%, the daytime SUHI intensity decreases from 2.66 °C for clear-sky conditions to 1.70 °C; the mitigating effect differs at daytime and nighttime; and (iii) clear-sky bias varies significantly across climate zones and seasons, with a varying range of -1.6-1.2 °C for the five selected cities.