A Global 250-m Downscaled NDVI Product from 1982 to 2018

Using process-oriented model output to enhance machine learning-based soil organic carbon prediction in space and time

Monitoring and modelling soil organic carbon (SOC) in space and time can help us to better understand soil carbon dynamics and is of key importance to support climate change research and policy. Although machine learning (ML) has attracted a lot of attention in the digital soil mapping (DSM) community for its powerful ability to learn from data and predict soil properties, such as SOC, it is better at capturing soil spatial variation than soil temporal dynamics. By contrast, process-oriented (PO) models benefit from mechanistic knowledge to express physiochemical and biological processes that govern SOC temporal changes. Therefore, integrating PO and ML models seems a promising means to represent physically plausible SOC dynamics while retaining the spatial prediction accuracy of ML models. In this study, a hybrid modelling framework was developed and tested for predicting topsoil SOC stock in space and time for a regional cropland area located in eastern China. In essence, the hybrid model uses predictions of the PO model in unsampled years as additional training data of the ML model, with a weighting parameter assigned to balance the importance of SOC values from the PO model and real measurements. The results indicated that temporal trends of SOC stock modelled by PO and ML models were largely different, while they were notably similar between the PO and hybrid models. Cross-validation showed that the hybrid model had the best performance (RMSE = 0.29 kg m-2), with a 19 % improvement compared with the ML model. We conclude that the proposed hybrid framework not only enhances space-time soil carbon mapping in terms of prediction accuracy and physical plausibility, it also provides insights for soil management and policy decisions in the face of future climate change and intensified human activities.

China's public health initiatives for climate change adaptation

China's health gains over the past decades face potential reversals if climate change adaptation is not prioritized. China's temperature rise surpasses the global average due to urban heat islands and ecological changes, and demands urgent actions to safeguard public health. Effective adaptation need to consider China's urbanization trends, underlying non-communicable diseases, an aging population, and future pandemic threats. Climate change adaptation initiatives and strategies include urban green space, healthy indoor environments, spatial planning for cities, advance location-specific early warning systems for extreme weather events, and a holistic approach for linking carbon neutrality to health co-benefits. Innovation and technology uptake is a crucial opportunity. China's successful climate adaptation can foster international collaboration regionally and beyond.

GEE can prominently reduce uncertainties from input data and parameters of the remote sensing-driven distributed hydrological model

The coupling of multisource remote sensing data and the lack of measured runoff introduce input data and model parameters uncertainties to the remote sensing-driven distributed hydrological model (RS-DHM). The PB satellite remote sensing datasets of the Google Earth Engine (GEE) are widely used in RS-DHM and remote sensing runoff inversion research, but whether GEE can reduce the two abovementioned uncertainties is still unknown. To answer this question, twelve remote sensing data sources provided by GEE were used in this study to drive a typical RS-DHM called the remote sensing-driven distributed time-variant gain model (RS-DTVGM) and the remote sensing runoff inversion technology called remote sensing hydrological station (RSHS), and the contribution of GEE to the improving hydrological model uncertainties was quantitatively analyzed from 2001 to 2020. The results showed that (1) the GEE-based improved data preparation not only effectively reduced the uncertainty in the input data with better spatial-temporal continuity and a 6.20 % reduction in the total area occupied by invalid grids, but also enhanced the operational efficiency by reducing the image number, memory size and data processing time of the satellite remote sensing data by 83.63 %, 99.53 %, and 98.73 %, respectively; (2) the GEE-based RSHS technology provided sufficient data support for parameter adjustment and accuracy validation of the RS-DTVGM, which effectively reduced the uncertainty in the model parameters and increased the Nash efficiency coefficient (NSE) in the calibration and validation period from 0.67 to 0.87 and 0.75, respectively; and (3) the calibrated RS-DTVGM was more reliable and robust, and its runoff and evapotranspiration were consistent with the actual statistical data. In the future, GEE and RSHS technology should be widely adopted to drive the RS-DHM to more quickly and easily provide reliable hydrological processes simulation results for integrated water resource management, therefore achieving win-win results in terms of efficiency and accuracy.