نقشه‌برداری رقومی کربنات کلسیم معادل و درصد رس خاک با استفاده از تصاویر ماهواره‌های لندست8 و پریسما توسط الگوریتم جنگل تصادفی

نوع مقاله : مقاله پژوهشی

نویسندگان

1 گروه علوم ومهندسی خاک دانشکدگان کشاورزی و منابع طبیعی دانشگاه تهران، کرج، ایران

2 عضو هیأت علمی گروه مهندسی علوم خاک، پردیس کشاورزی و منابع طبیعی دانشگاه تهران

چکیده

نقشه‌برداری خصوصیات خاک با استفاده از تصاویر ماهواره‌ای ابرطیفی و چند طیفی در کنار رویکردهای آماری و با استفاده از مدل‌های یادگیری ماشین از جمله جنگل‌های تصادفی پیشرفت زیادی در دقت و صحت نقشه‌های تهیه­شده داشته است. این تحقیق برای بررسی عملکرد تصاویر پریسماو لندست8 در مدل‌سازی کربنات‌کلسیم معادل و درصد رس با مدل جنگل تصادفی، در بخشی از اراضی شهرستان آبیک استان قزوین از مهرماه سال 1399 تا مهرماه 1401 انجام شده است. در ابتدا، با استفاده از 229 داده‌ که از خاک سطحی جمع‌آوری‌شده در منطقه آبیک استان قزوین به مساحت 60 هزار هکتار اندازه­گیری شد، در مرحله بعد مجموعه‌داده‌های طیفی دو ماهواره پریسماو لندست8 ، استخراج  و داده‌های بازتاب خاک به دست آمدند. در این تحقیق از شاخص­های طیفی، شاخص­های مدل رقومی ارتفاع و تجزیه مولفه­های اصلی به عنوان متغییر کمکی استفاده شد. در مرحله بعدی، مدل رگرسیون جنگل تصادفی جهت تخمین ویژگی‌های خاک با استفاده از 80% از داده­ها آموزش داده­شد و از 20% داده­ها برای آزمون مدل استفاده شد. نتایج نشان داد که بهترین دقت در بازیابی ویژگی‌های خاک­سطحی توسط داده‌های پریسما، با استفاده از مجموعه‌داده‌های کمکی تجزیه مولفه‌های اصلی، شاخص‌های طیفی و شاخص‌های مستخرج از مدل رقومی ارتفاع به دست آمد. به طور دقیق‌تر، استفاده از این سه­دسته داده، بیشترین ضریب تبیین، انحراف پیش‌بینی باقی‌مانده و نسبت عملکرد به فاصله بین چارکی و کمترین ریشه میانگین مربعات خطا  و ریشه میانگین مربعات خطا  نرمال شده را برای تخمین کربنات کلسیم معادل و درصد رس خاک نشان داد. بهترین مدل برای تخمین درصد رس، با مدل جنگل‌های تصادفی، شاخص‌های آماری (ضریب تبیین: 90/0؛ریشه میانگین مربعات خطا : 91/4؛ ریشه میانگین مربعات خطا  نرمال شده : 23/0؛ نسبت دامنه بین چارکی: 8/0؛ نسبت انحراف عملکرد: 29/2) و بهترین مدل برای تخمین کربنات کلسیم معادل، با مدل جنگل‌های تصادفی، شاخص‌های آماری (ضریب تبیین: 61/0؛ریشه میانگین مربعات خطا : 72/0؛ ریشه میانگین مربعات خطا  نرمال شده : 20/0؛ نسبت دامنه بین چارکی: 77/0؛ نسبت انحراف عملکرد: 27/1) به دست آمد.

کلیدواژه‌ها

موضوعات


عنوان مقاله [English]

Digital mapping of soil properties (Calcium Carbonate and soil clay percentage) using landsat 8 and Prisma satellite images by the random forest algorithm

نویسندگان [English]

  • sajjad teimouri bardyani 1
  • Fereydoon Sarmadian 2
1 Department of Soil Science and Engineering, Faculty of Agriculture and Natural Resources, University of Tehran, Karaj, Iran
2 soil science department< faculty of agricultural engineering and technology, university of Tehran
چکیده [English]

Mapping soil properties using hyperspectral and multispectral satellite images, along with statistical approaches, and machine learning models such as Random Forests (RF), has shown great progress in accurately preparing agricultural maps. This study aimed to compare the performance of PRISMA and Landsat 8 images in modeling calcium carbonate and clay percentage using a Random Forest model. Firstly, Surface soil data was collected from Abik region of Qazvin province from October 2020 to October 2022. Furthermore, PRISMA and Landsat 8 spectral datasets were extracted from images downloaded from the websites of these two satellites, and soil reflectance data were obtained. The Random Forest regression model was then calibrated to estimate soil properties. The results of this study showed that the best accuracy in estimating soil characteristics using PRISMA data was obtained by using Auxiliary Variables such as principal components analysis, spectral indices, and indices extracted from the digital elevation model. The use of these three data sets provided the uppermost value for following statistical indices for estimating calcium carbonate and soil clay percentage: coefficient of determination (R2), and Ratio of Performance to Inter Quartile range (RPIQ), Ratio Performance Deviation (RPD) and the lowest Root Mean Squared Error (RMSE) and Normalized Root Mean Squared Error (NRMSE). The best model for estimating clay percentage, using the Random Forest model and statistical indices, had an R2 of 0.90, RMSE of 4.11, NRMSE of 0.18, RPIQ of 0.95, and RPD of 2.29. The best model for estimating calcium carbonate, using the Random Forest model and statistical indices, had an R2 of 0.62, RMSE of 0.72, NRMSE of 0.20, RPIQ of 0.77, and RPD of 1.27. The results supported the expectation of the good ability of the PRISMA imager to estimate surface soil properties.

کلیدواژه‌ها [English]

  • Random Forest
  • clay percentage
  • Calcium carbonate
  • PRISMA hyperspectral satellite

Digital Mapping of Soil Properties (Calcium carbonate and soil clay percentage) Using Landsat 8 and PRISMA Satellite Images by the Random Forest Algorithm

EXTENDED ABSTRACT

Introduction

Land management is recognized as a major challenge in solving global issues such as food, water, energy, environment, biodiversity and human health. To have accurate information about soil, soil mapping and assessment are essential. Imaging sensors including Landsat 8 and PRISMA with a wide range of spatial, temporal and spectral resolutions are important tools for soil mapping. In this research, the zoning of soil properties using PRISMA and Landsat 8 data and the effect of different inputs on the accuracy and correctness of maps have been investigated using the RF algorithm.

Material and Methods

In this study, soil properties in a region of Abik City in the Qazvin province of Iran were delineated using an RF algorithm based on laboratory data and remote sensing images. L2D surface images from PRISMA and Landsat 8 satellites, as well as a 30-meter digital elevation model, were employed for soil mapping. The remote sensing data underwent preprocessing steps, including the removal of problematic bands, geometric correction, and image calibration using ENVI 5.6 software. Additionally, 12 spectral indices, such as NDVI, and 20 indices were derived by combining two or more bands in the images with the assistance of SAGA GIS software and the digital elevation model.To analyze the spatial diversity of hyperspectral images, the principal component analysis (PCA) method was used and also the effect of different inputs on the RF algorithm were investigated. The random forest algorithm has very high accuracy in predicting soil properties due to its ability to model non-linear relationships between auxiliary variables and the target soil. This algorithm provides the possibility to determine the relative importance of environmental auxiliary variables simply. In this research, to validate the performance of the random forest model using data obtained from laboratory analysis and remote sensing images, RMSE, RPD, RPIQ and R2 statistics were calculated from the measured values of soil properties.

Results and discussion

The results show that the simultaneous use of spectral indices and principal component analysis does not improve accuracy, and the use of Digital elevation model indices has a great effect on improving the results. In general, using principal component analysis to improve the quality of generated calcium carbonate maps is more effective than using spectral indices. But it cannot be said which of the maps obtained from the analysis of the main components of satellite images or the maps obtained from spectral indices extracted from satellite images are more accurate and precise. Also, the results did not show that the derivatives of Landsat 8 or PRISMA images are more effective for preparing maps of soil CaCO3 values, but the predicted values using PRISMA and Landsat 8 satellite spectra were consistent with the trend of soil calcium carbonate changes in the region.

Also, the results showed that soil clay percentage maps that PRISMA-based models provide better results compared to Landsat8-based models. The accuracy of PRISMA satellite images for preparing the clay percentage map in the study area is higher than that of the Landsat8 satellite, due to the typical absorption characteristics in the SWIR spectral region. The use of the digital height model and the indicators derived from it increases the performance of the models. Also, the maps obtained from the analysis of the main components of satellite images are less accurate than the maps obtained from spectral indices extracted from satellite images. The values of clay percentage predicted using PRISMA and Landsat8 satellite spectra were consistent with the trend of changes in soil texture and soil clay percentage in the region.

Conclusion

The results showed that PRISMA images are more accurate than Landsat 8 due to their higher number of bands and greater spectral resolution. However, the high cost of PRISMA and its limited access to complete time series images are limitations of this method. The combined use of spectral indices, principal component analysis, and digital height models significantly improved the model's performance. However, a large number of inputs can limit the model's inputs, and optimizing the selection of features from each category can identify the best features and improve the model's performance.

 

Akpa, S. I., Odeh, I. O., Bishop, T. F., and Hartemink, A. E. (2014). Digital mapping of soil particle‐size fractions for Nigeria. Soil Science Society of America Journal78(6), 1953-1966.‏
Alicandro, M., Candigliota, E., Dominici, D., Immordino, F., Masin, F., Pascucci, N., ... and Zollini, S. (2022). Hyperspectral PRISMA and Sentinel-2 Preliminary Assessment Comparison in Alba Fucens and Sinuessa Archaeological Sites (Italy). Land, 11(11), 2070.
Anderson, M. C., Neale, C. M. U., Li, F., Norman, J. M., Kustas, W. P., Jayanthi, H., and Chavez, J. O. S. E. (2004). Upscaling ground observations of vegetation water content, canopy height, and leaf area index during SMEX02 using aircraft and Landsat imagery. Remote sensing of environment, 92(4), 447-464.
Angelopoulou, T., Chabrillat, S., Pignatti, S., Milewski, R., Karyotis, K., Brell, M., ... and Zalidis, G. (2023). Evaluation of airborne hyspex and spaceborne PRISMA hyperspectral remote sensing data for soil organic matter and carbonates estimation. Remote Sensing, 15(4), 1106.
ASI Prisma Products Specification Document Issue 2.3. [(accessed on 12 March 2020)]; Available online: http://prisma.asi.it/missionselect/docs/PRISMA%20Product%20Specifications_Is2_3.pdf:
Baig, M. H. A., Zhang, L., Shuai, T., and Tong, Q. (2014). Derivation of a tasselled cap transformation based on Landsat 8 at-satellite reflectance. Remote Sensing Letters, 5(5), 423-431.
Ben-Dor, E., Chabrillat, S., Demattê, J. A. M., Taylor, G. R., Hill, J., Whiting, M. L., and Sommer, S. (2009). Using imaging spectroscopy to study soil properties. Remote sensing of environment, 113, S38-S55.‏‏
Bouzekri, S., Lasbet, A. A., and Lachehab, A. (2015). A new spectral index for extraction of built-up area using Landsat-8 data. Journal of the Indian Society of Remote Sensing43, 867-873.
Broge, N. H., and Leblanc, E. (2001). Comparing prediction power and stability of broadband and hyperspectral vegetation indices for estimation of green leaf area index and canopy chlorophyll density. Remote sensing of environment, 76(2), 156-172.
Carre, F., McBratney, A. B., Mayr, T., and Montanarella, L. (2007). Digital soil assessments: Beyond DSM. Geoderma, 142(1-2), 69-79.‏
Casa, R., Pignatti, S., Pascucci, S., Castaldi, F., and Marrone, L. (2023). Estimation of agronomic soil properties from multitemporal PRISMA satellite imaging spectroscopy. In Precision agriculture'23 (pp. 839-845). Wageningen Academic.
Casa, R., Pignatti, S., Pascucci, S., Castaldi, F., and Marrone, L. (2023). Estimation of agronomic soil properties from multitemporal PRISMA satellite imaging spectroscopy. In Precision agriculture'23 (pp. 839-845). Wageningen Academic.
Castaldi, F., Casa, R., Castrignanò, A., Pascucci, S., Palombo, A., and Pignatti, S. (2014). Estimation of soil properties at the field scale from satellite data: a comparison between spatial and non‐spatial techniques. European Journal of Soil Science, 65(6), 842-851.‏
Colovic, M., Yu, K., Todorovic, M., Cantore, V., Hamze, M., Albrizio, R., and Stellacci, A. M. (2022). Hyperspectral vegetation indices to assess water and nitrogen status of sweet maize crop. Agronomy, 12(9), 2181.‏
Cutler, D. R., Edwards Jr, T. C., Beard, K. H., Cutler, A., Hess, K. T., Gibson, J., and Lawler, J. J. (2007). Random forests for classification in ecology. Ecology88(11), 2783-2792.‏
Dash, P. K., Panigrahi, N., and Mishra, A. (2022). Identifying opportunities to improve digital soil mapping in India: A systematic review. Geoderma Regional, 28, e00478.
De Santana, F. B., and Daly, K. (2022). A comparative study of MIR and NIR spectral models using ball-milled and sieved soil for the prediction of a range soil physical and chemical parameters. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy, 279, 121441.
Dharumarajan, S., and Hegde, R. (2022). Digital mapping of soil texture classes using Random Forest classification algorithm. Soil Use and Management38(1), 135-149.‏
Diek, S., Schaepman, M. E., and De Jong, R. (2016). Creating multi-temporal composites of airborne imaging spectroscopy data in support of digital soil mapping. Remote Sensing8(11), 906.
Dindaroglu, T., Tunguz, V., Babur, E., Alkharabsheh, H. M., Seleiman, M. F., Roy, R., and Zakharchenko, E. (2022). The use of remote sensing to characterise geomorphometry and soil properties at watershed scale. International Journal of Global Warming, 27(4), 402-421.‏
Du, Q., and Yang, H. (2008). Similarity-based unsupervised band selection for hyperspectral image analysis. IEEE geoscience and remote sensing letters, 5(4), 564-568.‏
Duan, M., Song, X., Liu, X., Cui, D., and Zhang, X. (2022). Mapping the soil types combining multi-temporal remote sensing data with texture features. Computers and Electronics in Agriculture, 200, 107230.
Fang, Y., Xu, L., Wong, A., and Clausi, D. A. (2022). Multi-temporal landsat-8 images for retrieval and broad scale mapping of soil copper concentration using empirical models. Remote Sensing14(10), 2311.
Galvao, L. S., dos Santos, J. R., Roberts, D. A., Breunig, F. M., Toomey, M., and de Moura, Y. M. (2011). On intra-annual EVI variability in the dry season of tropical forest: A case study with MODIS and hyperspectral data. Remote Sensing of Environment, 115(9), 2350-2359.‏
Galvao, L. S., Formaggio, A. R., and Tisot, D. A. (2005). Discrimination of sugarcane varieties in Southeastern Brazil with EO-1 Hyperion data. Remote sensing of Environment, 94(4), 523-534.‏
Gasmi, A., Gomez, C., Chehbouni, A., Dhiba, D., and El Gharous, M. (2022). Using PRISMA hyperspectral satellite imagery and GIS approaches for soil fertility mapping (FertiMap) in Northern Morocco. Remote sensing14(16), 4080.‏
Grunwald, S. (2010). Current state of digital soil mapping and what is next. In digital soil mapping (pp. 3-12). Springer, Dordrecht.
Haboudane, D., Miller, J. R., Pattey, E., Zarco-Tejada, P. J., and Strachan, I. B. (2004). Hyperspectral vegetation indices and novel algorithms for predicting green LAI of crop canopies: Modeling and validation in the context of precision agriculture. Remote sensing of environment, 90(3), 337-352.
Huang, S.Y., Zhang, S.H., Liu, L.L., Zhu, W.Q. and Cheng, Y.M.,( 2021). Efficient slope reliability analysis and risk assessment based on multiple Kriging metamodels. Computers and Geotechnics, 137, p.104277.
Inoue, Y., Peñuelas, J., Miyata, A., and Mano, M. (2008). Normalized difference spectral indices for estimating photosynthetic efficiency and capacity at a canopy scale derived from hyperspectral and CO2 flux measurements in rice. Remote Sensing of Environment112(1), 156-172.
Ishwaran, H., and Kogalur, U. B. (2010). Consistency of random survival forests. Statistics and probability letters80(13-14), 1056-1064.‏
Jurgens, C. (1997). The modified normalized difference vegetation index (mNDVI) a new index to determine frost damages in agriculture based on Landsat TM data. International Journal of Remote Sensing, 18(17), 3583-3594.‏
Karray, E., Elmannai, H., Toumi, E., Gharbia, M. H., Meshoul, S., Aichi, H., and Ben Rabah, Z. (2023). Evaluating the Potentials of PLSR and SVR Models for Soil Properties Prediction Using Field Imaging, Laboratory VNIR Spectroscopy and Their Combination. Comput. Model. Eng. Sci136, 1399-1425.‏
Khaledian, Y. and Miller, B.A., (2020). Selecting appropriate machine learning methods for digital soil mapping. Applied Mathematical Modelling, 81, pp.401-418.
Li, P., Xiao, C., and Feng, Z. (2018). Mapping rice planted area using a new normalized EVI and SAVI (NVI) derived from Landsat-8 OLI. IEEE Geoscience and Remote Sensing Letters, 15(12), 1822-1826.‏
Lucieer, A., Malenovský, Z., Veness, T., and Wallace, L. (2014). HyperUAS—Imaging spectroscopy from a multirotor unmanned aircraft system. Journal of Field Robotics, 31(4), 571-590.
Lymburner, L., Beggs, P. J., and Jacobson, C. R. (2000). Estimation of canopy-average surface-specific leaf area using Landsat TM data. Photogrammetric Engineering and Remote Sensing, 66(2), 183-192.‏
Mallah, S., Delsouz Khaki, B., Davatgar, N., Scholten, T., Amirian-Chakan, A., Emadi, M., ... and Taghizadeh-Mehrjardi, R. (2022). Predicting Soil Textural Classes Using Random Forest Models: Learning from Imbalanced Dataset. Agronomy12(11), 2613.
Marques, K. P., Demattê, J. A., Miller, B. A., and Lepsch, I. F. (2018). Geomorphometric segmentation of complex slope elements for detailed digital soil mapping in southeast Brazil. Geoderma Regional, 14, e00175.
Martin, J. B. (2017). Carbonate minerals in the global carbon cycle. Chemical Geology449, 58-72.‏
McDowell, M. L., Bruland, G. L., Deenik, J. L., Grunwald, S., and Knox, N. M. (2012). Soil total carbon analysis in Hawaiian soils with visible, near-infrared and mid-infrared diffuse reflectance spectroscopy. Geoderma, 189, 312-320.
Mishra, M., Singh, K. K., Pandey, P. C., Devrani, R., Pandey, A. K., Raju, K. P., ... and Pandey, M. (2022). Spectral Indices Across Remote Sensing Platforms and Sensors Relating to the Three Poles: An Overview of Applications, Challenges, and Future Prospects. Advances in Remote Sensing Technology and the Three Poles, 83-116.
Mosleh, Z., Salehi, M. H., Jafari, A., Borujeni, I. E., & Mehnatkesh, A. (2016). The effectiveness of digital soil mapping to predict soil properties over low-relief areas. Environmental monitoring and assessment, 188, 1-13.
Mousavi, S., Sarmadian, F., Omid, M., & Bogaert, P. (2021). Modeling the Vertical Soil Calcium Carbonate Equivalent Variation by Machine Learning Algorithms in Qazvin Plain. Water and Soil, 35(5), 719-734. doi: 10.22067/jsw.2021.71748.1076. (in Persian)
Mousavifard, S. M., Momtaz, H., Sepehr, E., Davatgar, N., and Sadaghiani, M. H. R. (2013). Determining and mapping some soil physico-chemical properties using geostatistical and GIS techniques in the Naqade region, Iran. Archives of Agronomy and Soil Science59(11), 1573-1589.
Mzid, N., Casa, R., Pascucci, S., Tolomio, M., and Pignatti, S. (2022, July). Assessment of the Potential of PRISMA Hyperspectral Data to Estimate Soil Moisture. In IGARSS 2022-2022 IEEE International Geoscience and Remote Sensing Symposium (pp. 5606-5609). IEEE.
Padarian, J., Minasny, B., McBratney, A.B., (2020). Machine learning and soil sciences: a review aided by machine learning tools. Soil 6, 35–52. https://doi.org/10.5194/soil-6-35-2020.
Panahi, M., Gayen, A., Pourghasemi, H. R., Rezaie, F., and Lee, S. (2020). Spatial prediction of landslide susceptibility using hybrid support vector regression (SVR) and the adaptive neuro-fuzzy inference system (ANFIS) with various metaheuristic algorithms. Science of the Total Environment, 741, 139937.
Pôças, I., Rodrigues, A., Gonçalves, S., Costa, P. M., Gonçalves, I., Pereira, L. S., and Cunha, M. (2015). Predicting grapevine water status based on hyperspectral reflectance vegetation indices. Remote sensing, 7(12), 16460-16479.
Pu, R., Gong, P., and Yu, Q. (2008). Comparative analysis of EO-1 ALI and Hyperion, and Landsat ETM+ data for mapping forest crown closure and leaf area index. Sensors, 8(6), 3744-3766.‏
Rahmani, A., Sarmadian, F., & Arefi, H. (2022). Digital Mapping of Top-soil Thickness and Associated Uncertainty Using Machine Learning Approach in Some Part of Arid and Semi-arid Lands of Qazvin Plain. Iranian Journal of Soil and Water Research, 53(3), 585-602. doi: 10.22059/ijswr.2022.338007.669195.
Rahmani, A., Sarmadian, F., Mousavi, S. R., & Khamoshi, S. E. (2020). Application of Geomorphometric attributes in digital soil mapping by using of machine learning and fuzzy logic approaches. Journal of Range and Watershed Managment73(1), 105-124. doi: 10.22059/jrwm.2020.288580.1418. (in Persian)
Riihimäki, H., Kemppinen, J., Kopecký, M., and Luoto, M. (2021). Topographic Wetness Index as a Proxy for Soil Moisture: The Importance of Flow‐Routing Algorithm and Grid Resolution. Water Resources Research57(10), e2021WR029871.
Roy, D. P., Kovalskyy, V., Zhang, H. K., Vermote, E. F., Yan, L., Kumar, S. S., and Egorov, A. (2016). Characterization of Landsat-7 to Landsat-8 reflective wavelength and normalized difference vegetation index continuity. Remote sensing of Environment, 185, 57-70.‏
Sahoo, R. N., Ray, S. S., and Manjunath, K. R. (2015). Hyperspectral remote sensing of agriculture. Current Science, 848-859.
Schlerf, M., Atzberger, C., and Hill, J. (2005). Remote sensing of forest biophysical variables using HyMap imaging spectrometer data. Remote Sensing of Environment, 95(2), 177-194.‏
Skentos, Athanasios. "Topographic position index-based landform analysis of Messaria (Ikaria Island, Greece)." Acta Geobalcanica 4.1 (2018): 7-15.
Tagliabue, G., Boschetti, M., Bramati, G., Candiani, G., Colombo, R., Nutini, F., ... and Panigada, C. (2022). Hybrid retrieval of crop traits from multi-temporal PRISMA hyperspectral imagery. ISPRS Journal of Photogrammetry and Remote Sensing, 187, 362-377.‏
Tan, C., Samanta, A., Jin, X., Tong, L., Ma, C., Guo, W., ... and Myneni, R. B. (2013). Using hyperspectral vegetation indices to estimate the fraction of photosynthetically active radiation absorbed by corn canopies. International Journal of Remote Sensing, 34(24), 8789-8802.‏
Thenkabail, P. S., Smith, R. B., and De Pauw, E. (2002). Evaluation of narrowband and broadband vegetation indices for determining optimal hyperspectral wavebands for agricultural crop characterization. Photogrammetric engineering and remote sensing, 68(6), 607-622.‏
Toth, C., and Jóźków, G. (2016). Remote sensing platforms and sensors: A survey. ISPRS Journal of Photogrammetry and Remote Sensing, 115, 22-36.
Tu, Y. H., Phinn, S., Johansen, K., and Robson, A. (2018). Assessing radiometric correction approaches for multi-spectral UAS imagery for horticultural applications. Remote Sensing10(11), 1684.
Vaughn Ihlen, Landsat 8 (L8) Data Users Handbook, 2019, Department of the Interior U.S. Geological Survey.
Vaze, J., Teng, J., and Spencer, G. (2010). Impact of DEM accuracy and resolution on topographic indices. Environmental Modelling and Software25(10), 1086-1098.
Wang, C., Li, W., Yang, Z., Chen, Y., Shao, W., and Ji, J. (2015). An invisible soil acidification: Critical role of soil carbonate and its impact on heavy metal bioavailability. Scientific reports5(1), 12735.‏
Wu, C., Niu, Z., Tang, Q., and Huang, W. (2008). Estimating chlorophyll content from hyperspectral vegetation indices: Modeling and validation. Agricultural and forest meteorology, 148(8-9), 1230-1241.
Zhao, D., Wang, J., Zhao, X., and Triantafilis, J. (2022). Clay content mapping and uncertainty estimation using weighted model averaging. Catena209, 105791.‏
Zolfaghari Nia, M., Moradi, M., Moradi, G., and Taghizadeh-Mehrjardi, R. (2022). Machine Learning Models for Prediction of Soil Properties in the Riparian Forests. Land12(1), 32.
Zeraatpisheh, M., Ayoubi, S., Jafari, A., Tajik, S., & Finke, P. (2019). Digital mapping of soil properties using multiple machine learning in a semi-arid region, central Iran. Geoderma, 338, 445-452.
Hengl, T., Miller, M. A., Križan, J., Shepherd, K. D., Sila, A., Kilibarda, M., ... & Crouch, J. (2021). African soil properties and nutrients mapped at 30 m spatial resolution using two-scale ensemble machine learning. Scientific Reports, 11(1), 6130.
Shan, M., Liang, S., Fu, H., Li, X., Teng, Y., Zhao, J., ... & Ma, Z. (2021). Spatial prediction of soil calcium carbonate content based on Bayesian maximum entropy using environmental variables. Nutrient Cycling in Agroecosystems, 120, 17-30.
Lagacherie, P., Baret, F., Feret, J. B., Netto, J. M., & Robbez-Masson, J. M. (2008). Estimation of soil clay and calcium carbonate using laboratory, field and airborne hyperspectral measurements. Remote Sensing of Environment, 112(3), 825-835.
Gabriels, D., & Moldenhauer, W. C. (1978). Size distribution of eroded material from simulated rainfall: Effect over a range of texture. Soil Science Society of America Journal, 42(6), 954-958.‏