Identifying rice stress on a regional scale from multi-temporal satellite images using a Bayesian method

Bayesian risk prediction model: An accessible strategy to predict cadmium contamination risk in wheat grain grown in alkaline soils

Excessive cadmium (Cd) concentration in wheat grain is becoming a widespread concern in China. Considering the complexity of Cd transfer in the soil-wheat system, how the Cd risk in wheat grain be accurately predicted from the limited details available is of great significance for the risk management of Cd. Bayes' theory could leverage existing data by combining prior information and observational data, providing a promising strategy with which to calculate a more robust posterior probability of a grain sample exceeding the food safety standard (FSS) for Cd (0.1 mg kg-1). In the current study, a risk prediction model, based on Bayes' theory, was established to achieve a more accurate prediction of the wheat grain Cd risk from a limited number of soil parameters. The risk prediction model could predict the risk probability of wheat grain with a Cd concentration exceeding the FSS under a given soil concentration of either total Cd or diethylenetriaminepentaacetic acid (DTPA)-extractable Cd. Soil total Cd concentration proved to be a better variable for the model with greater predictive accuracy. The model predicted that fewer than 5% of the wheat grain would have a Cd concentration exceeding the FSS when grown in soil with a total Cd concentration of less than 0.299 mg kg-1. The risk probability rose significantly to 50% when the soil total Cd reached 0.778 mg kg-1. The accuracy of the model was greater than the widely applied multiple linear regression model, whereas previously published data from similar soil conditions also confirmed that the Bayesian model could predict wheat Cd risk with minimal error. The proposed model provides an accurate, accessible and cost-effective methodology for predicting Cd risk in wheat grown in alkaline soils before harvest. The wider application to other soil conditions, crops or contaminants using the Bayesian model is also promising for risk management authorities.

Ensemble learning-based applied research on heavy metals prediction in a soil-rice system

Accurate prediction of heavy metal accumulation in soil ecosystems is crucial for maintaining healthy soil environments and ensuring high-quality agricultural products, as well as a challenging scientific task. In this study, we constructed a dataset containing 490 sets of multidimensional environmental covariate data and proposed prediction models for heavy metal concentrations (HMC) in a soil-rice system, EL-HMC (including RF-HMC and GBM-HMC), based on Random Forest (RF) and Gradient Boosting Machine (GBM) ensemble learning (EL) techniques. To reasonably evaluate the effectiveness of each model, Multiple linear and Bayesian regressions were selected as benchmark models (BM), and mean absolute error (MAE), root mean square error (RMSE), and determination coefficient R2 were selected as evaluation indicators. In addition, sensitivity and spatial autocorrelation (SAC) analyses were used to examine the robustness of the model. The results showed that the R2 values of RF-HMC and GBM-HMC for modeling available cadmium (Cd) concentrations in soil were 0.654 and 0.690, respectively, with an average increase of 48.0 % compared to the BMs. The R2 values of RF-HMC and GBM-HMC for predicting Cd, lead (Pb), chromium (Cr), and mercury (Hg) concentrations in rice ranged from 0.618 to 0.824 and 0.645 to 0.850, respectively, with an average increase of 58.2 % compared with the BMs. The corresponding MAEs and RMSEs of RF-HMC and GBM-HMC had low error levels. Sensitivity analysis of the input features and the SAC of the prediction bias showed that the EL-HMC models have excellent robustness. Therefore, the EL technology-based prediction models for HMCs proposed herein are practical and feasible, demonstrating better accuracy and stability than the traditional model. This study verifies the application potential of EL technology in pollution ecology and provides a new perspective and solution for sustainable management and precise prevention of heavy metal pollution in farmland soil at the regional scale.

Temporal Characteristics of Stress Signals Using GRU Algorithm for Heavy Metal Detection in Rice Based on Sentinel-2 Images

Heavy metal stress, which is a serious environmental problem, affects both animal and human health through the food chain. However, such subtle stress information is difficult to detect in remote sensing images. Therefore, enhancing the stress signal is key to accurately identifying heavy metal contamination in crops. The aim of this study was to identify heavy metal stress in rice at a regional scale by mining the time-series characteristics of rice growth under heavy metal stress using the gated recurrent unit (GRU) algorithm. The experimental area was located in Zhuzhou City, Hunan Province, China. We collected situ-measured data and Sentinel-2A images corresponding to the 2019-2021 period. First, the spatial distribution of the rice in the study area was extracted using the random forest algorithm based on the Sentinel 2 images. Second, the time-series characteristics were analyzed, sensitive parameters were selected, and a GRU classification model was constructed. Third, the model was used to identify the heavy metals in rice and then assess the accuracy of the classification results using performance metrics such as the accuracy rate, precision, recall rate (recall), and F1-score (F1-score). The results showed that the GRU model based on the time series of the red-edge location feature index has a good classification performance with an overall accuracy of 93.5% and a Kappa coefficient of 85.6%. This study shows that regional heavy metal stress in crops can be accurately detected using the GRU algorithm. A combination of spectrum and temporal information appears to be a promising method for monitoring crops under various types of stress.

Mapping leaf metal content over industrial brownfields using airborne hyperspectral imaging and optimized vegetation indices

Monitoring plant metal uptake is essential for assessing the ecological risks of contaminated sites. While traditional techniques used to achieve this are destructive, Visible Near-Infrared (VNIR) reflectance spectroscopy represents a good alternative to monitor pollution remotely. Based on previous work, this study proposes a methodology for mapping the content of several metals in leaves (Cr, Cu, Ni and Zn) under realistic field conditions and from airborne imaging. For this purpose, the reflectance of Rubus fruticosus L., a pioneer species of industrial brownfields, was linked to leaf metal contents using optimized normalized vegetation indices. High correlations were found between the vegetation indices exploiting pigment-related wavelengths and leaf metal contents (r ≤ - 0.76 for Cr, Cu and Ni, and r ≥ 0.87 for Zn). This allowed predicting the metal contents with good accuracy in the field and on the image, especially Cu and Zn (r ≥ 0.84 and RPD ≥ 2.06). The same indices were applied over the entire study site to map the metal contents at very high spatial resolution. This study demonstrates the potential of remote sensing for assessing metal uptake by plants, opening perspectives of application in risk assessment and phytoextraction monitoring in the context of trace metal pollution.