Article
Environmental Sciences
Xibo Xu, Zeqiang Wang, Xiaoning Song, Wenjie Zhan, Shuting Yang
Summary: The selection of predictor variables is crucial in building a digital mapping model for potentially toxic elements (PTEs) in soil. Traditionally, spatial and spectral parameters have been used as predictor variables, but the temporal dimension is often overlooked. This study demonstrates the value of incorporating temporal indices in the model, leading to significant performance improvements. The temporal-spatial-spectral covariate combinations used in a random forest (RF) algorithm achieve satisfactory mapping accuracy and outperform other methods.
ENVIRONMENTAL RESEARCH
(2024)
Article
Soil Science
Fellipe A. O. Mello, Jose A. M. Dematte, Rodnei Rizzo, Andre C. Dotto, Raul R. Poppiel, Wanderson de S. Mendes, Clecia C. B. Guimaraes
Summary: Studies on soil maps using digital mapping techniques, considering drainage network info, evaluate its contribution to predicting soil classes. By calibrating models and cross-validation to optimize model selection, the performance of the models was validated.
Article
Soil Science
Qingliang Li, Cheng Zhang, Wei Shangguan, Lu Li, Yongjiu Dai
Summary: Accurate mapping of soil texture is crucial for agricultural development and environmental activities. This study proposes a novel soil texture prediction model (LGD-LSTM) that utilizes multiple inputs and long short-term memory models to enhance prediction accuracy, outperforming other digital soil mapping methods.
Article
Soil Science
Kamal Nabiollahi, Ruhollah Taghizadeh-Mehrjardi, Aram Shahabi, Brandon Heung, Alireza Amirian-Chakan, Masoud Davari, Thomas Scholten
Summary: In a study conducted in Kurdistan Province, Iran, a combination of random forests and covariate data was used to assess the spatial variability of salinity and sodicity in agricultural salt-affected land. The results showed that optimization algorithms helped improve the accuracy of predictions, and identified groundwater table, categorical maps, salinity index, and multi-resolution ridge top flatness as important covariates for predicting soil salinity and sodicity.
Article
Soil Science
Maiara Pusch, Alessandro Samuel-Rosa, Paulo Sergio Graziano Magalhaes, Lucas Rios do Amaral
Summary: The study aimed to evaluate the use of covariates in sample optimization for characterizing the spatial variability of soil chemical properties. Covariates were used in sample planning optimization and building predictive models. Different criteria were compared and three predictive methods were evaluated. The inclusion of covariates in sample planning was found to be beneficial in some areas, especially when the cropping system is complex.
Article
Soil Science
Renan Storno Nalin, Ricardo Simao Diniz Dalmolin, Fabricio de Araujo Pedron, Jean Michel Moura-Bueno, Taciara Zborowski Horst, Ricardo Bergamo Schenato, Matheus Flesch Soligo
Summary: The study aimed to account for the spatial variation of available phosphorus in southern Brazilian farm. It was found that models incorporating soil covariates could better predict soil phosphorus content.
Article
Soil Science
Stephan van der Westhuizen, Gerard B. M. Heuvelink, David P. Hofmeyr
Summary: In digital soil mapping, traditional univariate methods neglect the dependence structure between soil properties, while multivariate machine learning models can capture complex non-linear relationships and maintain the dependence structure. This study compares the performance of a multivariate random forest model with two separate univariate random forest models, and finds that the multivariate model outperforms in maintaining the dependence structure and producing more realistic results.
Article
Agronomy
Wenjie Zhang, Liang Zhu, Qifeng Zhuang, Dong Chen, Tao Sun
Summary: Nitrogen and phosphorus are crucial indicators of soil nutrients in agriculture, and their accurate management is essential for ensuring food security. This study compared the capabilities of ZH-1 and Sentinel-2 satellite data in mapping soil nutrients using machine learning algorithms, and found that both datasets performed well. The results showed that Sentinel-2 data performed best in computing total nitrogen content, while the XGBoost model with ZH-1 data performed better for soil Olsen-P content.
Article
Soil Science
S. Dharumarajan, Rajendra Hegde, M. Lalitha
Summary: This study evaluated four data mining algorithms for mapping soil field capacity and permanent wilting point along soil depth in Andhra Pradesh, South India. The Random Forest model outperformed others, explaining 39% of variation and achieving the best predictive results for these soil hydraulic properties.
Article
Environmental Sciences
Abdelaziz Htitiou, Abdelghani Boudhar, Abdelghani Chehbouni, Tarik Benabdelouahab
Summary: This study automated the extraction of cropland phenological metrics on GEE and used them with machine-learning models to produce high-resolution cropland and crop field-probabilities maps in Morocco. The classification product showed an overall accuracy of 97.86% for the nominal year 2019-2020, and the cropland probabilities maps accurately estimated sub-national SAU areas with an R-value of 0.9.
Article
Environmental Sciences
Zichen Guo, Yuqiang Li, Xuyang Wang, Xiangwen Gong, Yun Chen, Wenjie Cao
Summary: In this study, a reliable deep learning model was established for soil organic carbon density inversion in the North China agro-pastoral zone, providing a reference and framework for the establishment of soil organic carbon inversion models in other regions.
Article
Environmental Sciences
Hua Jin, Xuefeng Xie, Lijie Pu, Zhenyi Jia, Fei Xu
Summary: This study accurately predicts the soil organic matter (SOM) content in a dryland agroecosystem by collecting soil samples and using machine learning models. The random forest model is determined as the optimal model. The study identifies alkali-hydrolyzable nitrogen, available potassium, mean annual precipitation, and pH as the main controlling factors affecting the spatial distribution of SOM.
Article
Geosciences, Multidisciplinary
Sanaz Zare, Ali Abtahi, Seyed Rashid Fallah Shamsi, Philippe Lagacherie
Summary: This study fills the gap in comparing and developing methods for integrating new sources of soil data for DSM by mapping electrical conductivities using real measurements and EM38MK2 measurements. The results show the utility of EM38MK2 data as surrogate input data for mapping soil salinity, with regression co-kriging identified as the best integration method. The impact of EM38MK2 data on performance gains increases as the sizes of real soil salinity measurements decrease, indicating a promising way to tackle constraints of DSM in areas where soil sensing as alternative data is accessible.
Article
Soil Science
Subramanian Dharumarajan, Rajendra Hegde
Summary: Soil texture is a crucial factor that affects water holding capacity, nutrient availability, and crop growth. This study used digital soil mapping to map soil textural classes in Andhra Pradesh, India at a high spatial resolution, providing valuable information for crop planning and management.
SOIL USE AND MANAGEMENT
(2022)
Article
Soil Science
Rafael G. Siqueira, Cassio M. Moquedace, Marcio R. Francelino, Carlos E. G. R. Schaefer, Elpidio I. Fernandes-Filho
Summary: Soil texture is a crucial soil property in Antarctica, as it influences various ecological processes and is affected by climate changes and human disturbances. This study aimed to predict the distribution of sand, silt, and clay in the main ice-free areas of Maritime Antarctica and Northern Antarctic Peninsula. Machine learning models, particularly Random Forest, were used to analyze legacy soil texture data and generate soil texture maps. The accuracy of clay prediction was the highest, especially in the topsoil. The final maps showed good spatial consistency, reflecting factors such as geomorphology, parent material, and pedogenetic development. These findings contribute to decision-making regarding Antarctic soils and provide valuable data for global environmental models.
Article
Geosciences, Multidisciplinary
Chuan-Peng Zhao, Cheng-Zhi Qin
Summary: China has been restoring mangroves for their ecological and societal values with a focus on small fragmented mangrove patches. They utilized 10-m-resolution Sentinel-1 and -2 images for classification and post-processing to create the first detailed publicly accessible mangrove map, enhancing it with submeter data from Google Earth images. The map outperformed existing ones in identifying small mangrove patches and improving the accuracy of boundaries, providing important information for ecosystem services and carbon stock assessment.
GEOSCIENCE DATA JOURNAL
(2022)
Article
Computer Science, Information Systems
Zhi-Wei Hou, Cheng-Zhi Qin, A-Xing Zhu, Yi-Jie Wang, Peng Liang, Yu-Jing Wang, Yun-Qiang Zhu
Summary: This study proposes a novel method to formalize the parameter constraints of geoprocessing tools, based on SHACL and geoprocessing ontologies, which is comparatively easier and more efficient than existing methods. The proposed method not only covers more types of parameter constraints but also includes application-context-matching constraints that have been ignored by other methods.
ISPRS INTERNATIONAL JOURNAL OF GEO-INFORMATION
(2021)
Article
Computer Science, Information Systems
Yu-Jing Wang, Bei-Bei Ai, Cheng-Zhi Qin, A-Xing Zhu
Summary: This paper proposes a load-balancing strategy of data domain decomposition in parallel programming libraries for raster-based geocomputation, aiming to improve parallel performance by characterizing the distribution of computational intensity based on geocomputation characteristics. The proposed strategy has shown significant improvements in load balance and parallel performance compared to the previously adopted data domain decomposition strategy in the parallelization of typical geocomputation algorithms using PaRGO V2.
INTERNATIONAL JOURNAL OF GEOGRAPHICAL INFORMATION SCIENCE
(2022)
Article
Chemistry, Multidisciplinary
Zhi-Wei Hou, Shijun Yu, Tao Ji
Summary: This paper proposes a forecasting model focused on the trip distribution of tourists who travel with the suburban tourist railway, intensively studying the characteristics of tourists' trips and using the trip chain method to analyze the frequency, order, distance, and visiting volume of stay points. A case study conducted in H city shows that the model can reflect the real trip distribution characteristics of tourists very well, demonstrating the applicability and effectiveness of the proposed model.
APPLIED SCIENCES-BASEL
(2021)
Article
Environmental Sciences
Kingsley John, Yassine Bouslihim, Kokei Ikpi Ofem, Lahcen Hssaini, Rachid Razouk, Paul Bassey Okon, Isong Abraham Isong, Prince Chapman Agyeman, Ndiye Michael Kebonye, Chengzhi Qin
Summary: This study examines the influence of predictive models' choice and sample ratios selection on soil organic matter (SOM) prediction. The results indicate that the accuracy of SOM prediction is sensitive to both predictive models and sample ratios. Certain models perform better at specific sample ratios. The findings are important for cost-effective spatial estimation of SOM in other locations and can serve as a baseline study for future research.
INTERNATIONAL SOIL AND WATER CONSERVATION RESEARCH
(2022)
Article
Geography
Yan-Wen Wang, Cheng-Zhi Qin, Wei-Ming Cheng, A-Xing Zhu, Yu-Jing Wang, Liang-Jun Zhu
Summary: The article introduces an automatic crater detection approach using random forest classifiers and spatial structural information, which outperforms existing methods and can extract implicit expert knowledge on spatial structures of real craters from legacy crater maps.
ANNALS OF THE AMERICAN ASSOCIATION OF GEOGRAPHERS
(2022)
Article
Environmental Sciences
Qing Xia, Ting-Ting He, Cheng-Zhi Qin, Xue-Min Xing, Wu Xiao
Summary: A mangrove forest mapping method based on SMRI was developed in this study using Sentinel-2 images. It can accurately identify submerged mangrove forests and differentiate Spartina alterniflora, thus improving the accuracy of mangrove forest mapping and providing data for coastal wetland monitoring.
Article
Geography, Physical
Chuanpeng Zhao, Cheng-Zhi Qin, Zongming Wang, Dehua Mao, Yeqiao Wang, Mingming Jia
Summary: This study aims to delineate the detailed national-scale distribution of the exotic mangrove species Sonneratia apetala in coastal China. The authors derived samples and used a Random Forest classifier with Sentinel-1 and -2 imagery on Google Earth Engine to generate a map of Sonneratia apetala. The accuracy of the map was evaluated using multiple datasets and achieved high overall accuracies. The information provided in this study can support the management and control of Sonneratia apetala, and the developed approach can be applied to other vegetation species in broad latitudinal areas.
ISPRS JOURNAL OF PHOTOGRAMMETRY AND REMOTE SENSING
(2022)
Article
Engineering, Civil
Tong Wu, Liang-Jun Zhu, Shen Shen, A-Xing Zhu, Mingchang Shi, Cheng Zhi Qin
Summary: This study proposes using landscape positions along hillslopes as identification units for priority management areas (PMAs), which can accurately determine the spatial distribution of PMAs and require less area to achieve the same management goals compared to subbasin-based PMAs. Landscape position units can better represent hillslope processes and the spatial heterogeneity of underlying surface environments within subbasins, leading to the better effectiveness of identifying PMAs.
JOURNAL OF HYDROLOGY
(2023)
Article
Environmental Sciences
Shen Shen, Cheng-Zhi Qin, Liang-Jun Zhu, A-Xing Zhu
Summary: Optimizing the spatial configuration of diverse best management practices (BMPs) can provide valuable decision-making support for comprehensive watershed management. This study proposes a new simulation-optimization framework for determining the implementation plan of BMPs based on economic costs and time-varying effectiveness. The framework was demonstrated in an agricultural watershed case study, showing its effectiveness in providing feasible BMP scenarios with lower investment burden and slight loss of environmental effectiveness.
WATER RESOURCES RESEARCH
(2023)
Article
Environmental Sciences
Yang Yan, Jiajie Yang, Baoguo Li, Chengzhi Qin, Wenjun Ji, Yan Xu, Yuanfang Huang
Summary: This study aims to test the feasibility of using UAV hyperspectral data to map soil organic matter at a 1 m resolution. The results show that the random forest model based on UAV hyperspectral data can successfully predict soil organic matter with high accuracy, compared to other prediction methods.
Article
Soil Science
Xingchen Fan, Naiqing Fan, Cheng-Zhi Qin, Fang-He Zhao, Liang-Jun Zhu, A-Xing Zhu
Summary: Digital soil mapping based on environmental similarity is a useful method for predicting soil properties at unvisited locations. However, the current method lacks consideration of the First Law of Geography and spatial variation of sample distribution. In this paper, a new large-area DSM method is proposed, which considers the spatial distance to samples by using adaptive distance-decay parameter values. Evaluation results show that this method achieves higher accuracy compared to other environmental-similarity-based DSM methods and a large-area DSM method not based on environmental similarity.
Article
Environmental Sciences
Shen Shen, Cheng-Zhi Qin, Liang-Jun Zhu, A-Xing Zhu
Summary: This study designed a user-friendly web-based participatory watershed planning system to assist diverse stakeholders in reaching a consensus on optimal roadmaps. The system design separates the participatory process of non-expert stakeholders from the specialized modeling process and provides an easy-to-use interface. The experimental results demonstrate the effectiveness of the system in capturing stakeholders' varying perspectives and facilitating consensus-building.
JOURNAL OF ENVIRONMENTAL MANAGEMENT
(2023)
Article
Remote Sensing
Nai-Qing Fan, Fang-He Zhao, Liang-Jun Zhu, Cheng-Zhi Qin, A-Xing Zhu
Summary: The effective use of environmental covariates is crucial for digital soil mapping. Traditional approaches ignore the varying applicability of covariates in characterizing soil-environment relationships. This study proposes an adaptive method that quantifies covariate applicability based on terrain conditions and integrates it into the iPSM method, outperforming other methods in predicting soil organic matter content.
INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION
(2022)
Article
Remote Sensing
Chuanpeng Zhao, Cheng-Zhi Qin
Summary: This study introduces a binary classification approach for identifying large-area mangrove distribution, which reduces the sample size while achieving comparable performance. By adjusting the decision surface through the addition of subclasses similar to the mangrove class, the accuracy of the classification is improved.
INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION
(2022)