Soil water content (SWC) is a crucial variable in the thermal infrared research and is the major control for land surface hydrological processes at the watershed scale. Estimating the surface SWC from remotely sensed data using the triangle method proposed by Price has been demonstrated in previous studies. In this study, a new soil moisture index (Temperature Rising Rate Vegetation Dryness Index—TRRVDI) is proposed based on a triangle constructed using the mid-morning land surface temperature (LST) rising rate and the vegetation index to estimate the regional SWC. The temperature at the dry edge of the triangle is determined by the surface energy balance principle. The temperature at the wet edge is assumed to be equal to the air temperature. The mid-morning land surface temperature rising rate is calculated using Meteosat Second Generation—Spinning Enhanced Visible and Infrared Imager (MSG-SEVIRI) LST products over 4 cloud-free days (day of year: 206, 211, 212, 242) in 2007. The developed TRRVDI is validated by in situ measurements from 19 meteorological stations in Spain. The results indicate that the coefficient of determination (R²) between the TRRVDI derived using the theoretical limiting edges and the in situ SWC measurements is greater than that derived using the observed limiting edges. The R² values are 0.46 and 0.32; respectively (p < 0.05). Additionally, the TRRVDI is much better than the soil moisture index that was developed using one-time LST and fractional vegetation cover (FVC) with the theoretically determined limiting edges.

Trajectory prediction has been a crucial task in building a reliable autonomous driving system by anticipating possible dangers. One key issue is to generate consistent trajectory predictions without colliding. To overcome the challenge, we propose an efficient masked autoencoder for trajectory prediction (Traj-MAE) that better represents the complicated behaviors of agents in the driving environment. Specifically, our Traj-MAE employs diverse masking strategies to pre-train the trajectory encoder and map encoder, allowing for the capture of social and temporal information among agents while leveraging the effect of environment from multiple granularities. To address the catastrophic forgetting problem that arises when pre-training the network with multiple masking strategies, we introduce a continual pre-training framework, which can help Traj-MAE learn valuable and diverse information from various strategies efficiently. Our experimental results in both multi-agent and single-agent settings demonstrate that Traj-MAE achieves competitive results with state-of-the-art methods and significantly outperforms our baseline model. Project page: https://jiazewang.com/projects/trajmae.html.

This work estimated and validated the land surface temperature (LST) from thermal-infrared Channels 4 (10.8 µm) and 5 (12.0 µm) of the Visible and Infrared Radiometer (VIRR) onboard the second-generation Chinese polar-orbiting FengYun-3A (FY-3A) meteorological satellite. The LST, mean emissivity and atmospheric water vapor content (WVC) were divided into several tractable sub-ranges with little overlap to improve the fitting accuracy. The experimental results showed that the root mean square errors (RMSEs) were proportional to the viewing zenith angles (VZAs) and WVC. The RMSEs were below 1.0 K for VZA sub-ranges less than 30° or for VZA sub-ranges less than 60° and WVC less than 3.5 g/cm², provided that the land surface emissivities were known. A preliminary validation using independently simulated data showed that the estimated LSTs were quite consistent with the actual inputs, with a maximum RMSE below 1 K for all VZAs. An inter-comparison using the Moderate Resolution Imaging Spectroradiometer (MODIS)-derived LST product MOD11_L2 showed that the minimum RMSE was 1.68 K for grass, and the maximum RMSE was 3.59 K for barren or sparsely vegetated surfaces. In situ measurements at the Hailar field site in northeastern China from October, 2013, to September, 2014, were used to validate the proposed method. The result showed that the RMSE between the LSTs calculated from the ground measurements and derived from the VIRR data was 1.82 K.

Utilizing a temperature time-series prediction model to achieve good results can help us to accurately sense the changes occurring in temperature levels in advance, which is important for human life. However, the random fluctuations occurring in a temperature time series can reduce the accuracy of the prediction model. Decomposing the time-series data prior to performing a prediction can effectively reduce the influence of random fluctuations in the data and consequently improve the prediction accuracy results. In the present study, we propose a temperature time-series prediction model that combines the seasonal-trend decomposition procedure based on the loess (STL) decomposition method, the jumps upon spectrum and trend (JUST) algorithm, and the bidirectional long short-term memory (Bi-LSTM) network. This model can achieve daily average temperature predictions for cities located in China. Firstly, we decompose the time series into trend, seasonal, and residual components using the JUST and STL algorithms. Then, the components determined by the two methods are combined. Secondly, the three components and original data are fed into the two-layer Bi-LSTM model for training purposes. Finally, the prediction results achieved for both the components and original data are merged by learnable weights and output as the final result. The experimental results show that the average root mean square and average absolute errors of our proposed model on the dataset are 0.2187 and 0.1737, respectively, which are less than the values 4.3997 and 3.3349 attained for the Bi-LSTM model, 2.5343 and 1.9265 for the EMD-LSTM model, and 0.9336 and 0.7066 for the STL-LSTM model.

In a previous study, a daily evaporative fraction (EF) parameterization scheme was derived based on day–night differences in surface temperature, air temperature, and net radiation. Considering the advantage that incoming solar radiation can be readily retrieved from remotely sensed data in comparison with surface net radiation, this study simplified the daily EF parameterization scheme using incoming solar radiation as an input. Daily EF estimates from the simplified scheme were nearly equivalent to the results from the original scheme. In situ measurements from six Ameriflux sites with different land covers were used to validate the new simplified EF parameterization scheme. Results showed that daily EF estimates for clear skies were consistent with the in situ EF corrected by the residual energy method, showing a coefficient of determination of 0.586 and a root mean square error of 0.152. Similar results were also obtained for partly clear sky conditions. The non-closure of the measured energy and heat fluxes and the uncertainty in determining fractional vegetation cover were likely to cause discrepancies in estimated daily EF and measured counterparts. The daily EF estimates of different land covers indicate that the constant coefficients in the simplified EF parameterization scheme are not strongly site-specific.