Yazhen Jiang

The maximum surface relative humidity (h_s,max) is a key factor that affects surface vapor pressure (e_s), which governs evapotranspiration (ET) rates. Existing empirical parameterizations lack a robust physical basis and exhibit limited generalizability. To address these limitations, we propose a remote sensing–based framework for the first global-scale estimation of h_s,max, which integrates physical constraints and data-driven methods. The framework first back-calculated h_s,max by leveraging coupling mechanisms in the surface energy balance equation, using ET and meteorological data from 195 global flux tower sites. Subsequently, XGBoost–based inversion models were constructed separately for different land surface types, with site-level meteorological variables and the back-calculated h_s,max as inputs. The models were applied to global remote sensing and reanalysis datasets, enabling the global-scale estimation of monthly h_s,max from 2001 to 2020. Results demonstrate that the h_s,max models for different land surface types achieved high accuracy, with a mean root mean square error (RMSE) of 0.079 and a correlation coefficient (R) of 0.92. A contribution analysis using SHAP (Shapley Additive Explanations) reveals that relative humidity (RH) is the dominant predictor, while secondary factors vary by land surface type. The global monthly h_s,max estimates for 2001–2020 exhibit distinct climate-driven spatial patterns. Temporal variability is low (CV < 10%) in humid regions but exceeds 20% in arid regions. In conclusion, the proposed parameterization framework shows both the physical consistency and high accuracy in global scale h_s,max estimation, providing essential parameter support for studies of land–atmosphere interactions.

Latent heat flux (LE) plays a crucial role in the water-energy cycle. Eddy covariance (EC) measurements of LE frequently suffer from data gaps due to weather or equipment malfunctions. This study presents a novel framework combining median-adjusted full-factorial and iterative methods to fill LE data gaps. The findings indicate that the framework exhibits stable gap-filling across land surfaces, showing RMSEs of 14.69–19.51 W/m² for isolated gaps and averaging 24.41 W/m² for long-term missing periods (5–50 days). Gap-filled LE data closely matched temporal dynamics and fluctuations of adjacent measurements. Based on gap-filled LE data, all sites demonstrated strong correlations between available energy and turbulent fluxes, with RMSE values ranging 28.41–38.01 W/m² and achieving an average energy closure of 0.79 across sites. Compared to ERA5-Land and GLEAM products, the gap-filled data significantly outperformed both LE products. Overall, the filled LE data are of high quality and significantly improve the usability of observed data. The complete time-series data from 265 sites (1509 site-years) are valuable for LE model validation and water demand assessments.

Evapotranspiration (ET) plays a crucial role in the regional water-energy cycle, illustrating intricate interactions among climate, vegetation and soil. Eddy covariance (EC) technology is a primary method for measuring ET. However, incomplete data due to adverse weather conditions and equipment malfunctions are common. This study utilizes the full-factorial method to address the ET gaps at 339 sites across multiple global flux networks, and the filled ET data are compared with three ET products: the Land component of the Fifth Generation of European Reanalysis (EAR5-Land), Global Land Evaporation Amsterdam Model (GLEAM), and Breathing Earth System Simulator (BESS). Results indicate high consistency between filled ET data and three ET products at 264 out of 339 sites. The absolute average mean error (|MAE|) and root mean system error (|RMSE|) are 0.32 mm/d and 0.92 mm/d, respectively. Among the remaining 75 sites, 49 sites exhibit better agreement between filled ET and measured ET data than ET products in terms of seasonal variations and numerical ranges. Further verification is needed for the reliability of filled ET data at the other 26 sites due to a scarcity of measured ET data. Overall, the gap-filled ET data from 313 sites (2210 site-years) demonstrate high-quality. These sites exhibit a strong correlation between available energy and turbulent fluxes, with R², MAE, and RMSE for different surface types ranging from 0.84 to 0.94, 21.49 to 28.67 W/m², and 28.37 to 36.91 W/m², respectively. And the average energy balance closure rate is 0.73, indicating a relatively high degree of closure in the energy balance. These 313 sites with high-quality filled ET data can be utilized for ET model validation, ET product verification, water demand assessment, and other related tasks. The filled ET dataset can be publicly accessed at https://doi.org/10.57760/sciencedb.11651 (Wang & Jiang, 2024).

Top-of-atmosphere (TOA) outgoing longwave radiation (OLR), a key component of the Earth’s energy budget, serves as a diagnostic of the Earth’s climate system response to incoming solar radiation. However, existing products are typically estimated using the traditional two-step method, which may bring extra uncertainties. This paper presents a direct machine learning method to estimate TOA OLR by directly linking Himawari-8/Advanced Himawari Imager (AHI) TOA radiances with TOA OLR determined by Clouds and the Earth’s Radiant Energy System (CERES) and other information, such as the viewing geometry. Models are built separately under clear- and cloudy-sky conditions using a gradient-boosting regression tree. Independent test results show that the root mean square errors (RMSEs) of the clear-sky and cloudy-sky models for estimating instantaneous values are 7.46 W/m² (3.0%) and 11.61 W/m² (5.8%), respectively. Daily results are obtained by averaging all the instantaneous results in one day. Intercomparisons of the daily results with CERES TOA OLR data show that the RMSE of the estimated AHI OLR is ~6 W/m² (3%). The developed high-resolution AHI TOA OLR dataset will be beneficial in analyzing the regional energy budget.

The evolution of land use functions and structures in ecologically fragile watersheds have a direct impact on regional food security and sustainable ecological service supply. Previous studies that quantify and simulate land degradation in ecologically fragile areas from the perspective of long-term time series and the spatial structure of watersheds are rare. This paper takes the Huangshui Basin of the Qinghai-Tibet Plateau in China as a case study and proposes a long-time series evolution and scenario simulation method for land use function using the Google Earth Engine platform, which realizes the simulation of land use function and structure in ecologically fragile areas by space–time cube segmentation and integrated forest-based prediction. This allows the analysis of land degradation in terms of food security and ecological service degradation. The results show that: (1) the land use function and structure evolution of the Huangshui watershed from 1990 to 2020 have a significant temporospatial variation. In the midstream region, the construction land expanded 151.84% from 1990 to 2004, driven by urbanization and western development policy; in the middle and downstream region, the loss of farmland was nearly 12.68% from 1995 to 2005 due to the combined influence of the policy of returning farmland to forest and urban expansion. (2) By 2035, the construction land in the watershed will be further expanded by 28.47%, and the expansion intensity will be close to the threshold in the upstream and midstream areas and will continue to increase by 33.53% over 2020 in downstream areas. (3) The evolution of land use function and structure will further induce land degradation, causing a 15.30% loss of farmland and 114.20 km² of occupation of ecologically vulnerable areas, seriously threatening food security and ecological protection. Accordingly, this paper proposes policy suggestions to strengthen the spatial regulation for land degradation areas and the coordination of upstream, midstream, and downstream development.

Pedestrian origin–destination (O–D) estimates that record traffic flows between origins and destinations, are essential for the management of pedestrian facilities including pedestrian flow simulation in the planning phase and crowd control in the operation phase. However, current O–D data collection techniques such as surveys, mobile sensing using GPS, Wi-Fi, and Bluetooth, and smart card data have the disadvantage that they are either time consuming and costly, or cannot provide complete O–D information for pedestrian facilities without entrances and exits or pedestrian flow inside the facilities. Due to the full coverage of CCTV cameras and the huge potential of image processing techniques, we address the challenges of pedestrian O–D estimation and propose an image-based O–D estimation framework. By identifying the same person in disjoint camera views, the O–D trajectory of each identity can be accurately generated. Then, state-of-the-art deep neural networks (DNNs) for person re-ID at different congestion levels were compared and improved. Finally, an O–D matrix based on trajectories was generated and the resident time was calculated, which provides recommendations for pedestrian facility improvement. The factors that affect the accuracy of the framework are discussed in this paper, which we believe could provide new insights and stimulate further research into the application of the Internet of cameras to intelligent transport infrastructure management.

Accurate estimation of terrestrial water storage anomalies (TWSA) is crucial for the sustainable management of water resources and human living. In this study, long-term TWSA estimates are reconstructed by integration of multiple meteorological products and the water balance (WB) method at 0.5° × 0.5° resolution, generating a total of 12 combinations of different meteorological data. This scheme is applied to 10 river basins (RBs) within China and validated against GRACE observations and GLDAS simulations from 2003 to 2020. Results indicate that similar seasonal characteristics can be observed between different precipitation and evapotranspiration products with the average correlation coefficient and Nash–Sutcliffe efficiency coefficient metrics larger than 0.96 and 0.90, respectively. Three GRACE solutions indicate similar seasonal variations and long-term trends of TWSA over 10 RBs, with the correlation above 0.90. Similar performance can also be observed concerning the root mean square error and mean absolute error metrics. Nevertheless, WB-based TWSA estimates represent larger discrepancies compared to GRACE observations and GLDAS simulations. Specifically, the variation amplitude and long-term trend of WB-based results are much larger than that of the GRACE observations, which is mainly caused by the inaccuracy of remote sensing products and the neglect of anthropogenic activities. Comparable TWSA estimates independently computed from the WB method can only be achieved in 4 out of 10 RBs. This study can provide insightful suggestions for an enhanced understanding of TWSA estimates and improving the performance of the water balance method.