نقشه برداری رقومی کلاس‌های خاک با استفاده از رویکردهای یادگیری ماشین در بخشی از اراضی پایاب سد آزاد (مطالعه موردی: استان کردستان)

نوع مقاله : مقاله پژوهشی

نویسندگان

1 استادیار بخش تحقیقات خاک و آب، مرکز تحقیقات و آموزش کشاورزی و منابع طبیعی استان کردستان، سازمان تحقیقات، آموزش و ترویج کشاورزی،

2 استادیار بخش تحقیقات خاک و آب، مرکز تحقیقات و آموزش کشاورزی و منابع طبیعی استان کرمانشاه، سازمان تحقیقات، آموزش و ترویج کشاورزی،

3 دانش آموخته کارشناسی ارشد، گروه علوم و مهندسی خاک، دانشگاه رازی، دانشکده کشاورزی

چکیده

 
نقشه‌برداری رقومی خاک با برقراری روابط کمی میان متغیرهای محیطی و کلاس‌ها یا ویژگی‌های خاک قادر به پیش‌بینی مکانی ویژگی موردنظر است. در این پژوهش از الگوریتم‌های شبکه عصبی مصنوعی، درخت تصمیم، رگرسیون لاجستیک چند جمله‌ای و جنگل تصادفی برای پیش‌بینی نقشه خاک اراضی پایاب سد آزاد شهر سنندج با وسعت حدود 3/2178 هکتار استفاده شد. در سال 1396 تعداد 84 خاک‌رخ با الگوی تصادفی در منطقه مطالعاتی حفر، تشریح و نمونه‌برداری گردید. بر اساس ویژگی‌های ریخت‌شناختی و داده‌های آزمایشگاهی هر یک از خاکرخ‌ها تا سطح خانواده رده‌بندی شدند. بر اساس سیستم رده‌بندی جامع آمریکایی، دو رده اینسپتی‌سول و انتی‌سول، دو زیر رده، سه گروه بزرگ و پنج زیرگروه و خانواده مشاهده شد. برای محاسبه متغیرهای پیش‌بینی کننده، از مدل رقومی ارتفاع با قدرت تفکیک 10 متر و تصویر ماهواره سنتیل 2-B استفاده شد. برای بررسی صحت پیش‌بینی مدل‌ها از صحت عمومی نقشه، شاخص کاپا و درجه برابر استفاده شد که بهترین نتایج (به ترتیب 65/0، 53/0 و 16/0) برای الگوریتم شبکه عصبی مصنوعی به دست آمد. ضعیف‌ترین پیش‌بینی مربوط به مدل درخت تصمیم با صحت عمومی 38/0، شاخص کاپای 22/0 و درجه برابر 87/0 بود.

کلیدواژه‌ها

موضوعات

عنوان مقاله [English]

Hine learning approaches in downstream lands of Azad dam (case study: Kurdistan province)

نویسندگان [English]

  • Maryam Osat 1
  • Shahrokh Fatehi 2
  • Zaynab Zeynoldini 3
1 1. Assistant professor of Soil and Water Research Department, Kordestan agricultural and natural resources and education center, AREEO, Sanandaj, Iran
2 Assistant professor of soil and water research department, Kermanshah agricultural and natural resource and education center, AREEO, Kermanshah, Iran.
3 Students, M.Sc. Graduate, Department of Science and Soil Engineering, Razi University, Faculty of Agriculture
چکیده [English]

Digital Soil Mapping (DSM) encompasses a variety of methodologies that can yield precise spatial information about soil by establishing quantitative relationships between environmental covariates (predictors) and soil classes or properties. In this study, Artificial Neural Networks (ANNs), Decision Tree (DT), Multinomial Logistic Regression (MLR), and Random Forest (RF) algorithms were used to predict the soil map of downstream lands of Azad dam with an area of approximately 178.3 ha in the northwest of Sanandaj city in Kurdistan province. A random soil sampling method was used to determine the location and distribution of the 84 soil profiles in the study area. After recording soil morphological attributes, sampling of all horizons was conducted for required laboratory analysis. Afterward, the soil profiles were classified up to the family taxonomic level based on US classification system. Based on the soil taxonomy classification system, Inceptisols and Entisols order were observed by frequency, two Suborder, three Great groups, five Subgroups, and Family. To calculate the predictor variables, a digital elevation model (DEM) with a 10 m spatial resolution and Sentinel 2-B satellite images were used in the study area. To check the prediction accuracy of the models the Overall accuracy (OA), Kappa Index (K), and Brier Score (BS) were used. The best result was obtained by the ANN model (OA=0.65, K=0.53, and BS=0.16, respectively). The weakest predictions were found by DT model with OA, K, and BS of 0.38, 0.22, and 0.87, respectively.

کلیدواژه‌ها [English]

  • Artificial Neural Network
  • Decision Tree
  • Multinomial Logistic Regression
  • Random Forest

Hine learning approaches in downstream lands of Azad dam (case study: Kurdistan province)

 

Extended Abstract

 

Objective

All agro-economic and environmental activities usually need detailed information about the distribution of soil classes and their characteristics. Digital soil mapping relies on relating soil classes or properties to a particular set of covariates, which capture inherent soil spatial variation. DSM generates accurate soil spatial information through quantitative relationships between environmental covariates (predictors) and soil classes or properties. In DSM, different machine learning algorithms model The relationships between soil and auxiliary variables (e.g., remotely sensed data) and predict soil classes or properties at unknown points.

Research method

In this study, artificial neural network (ANN), decision tree (C.5), logistic regression (MLR), and random forest (RF) were used to predict soil class maps. The relationships between soil and the environment could be well explained by a practical collection of covariates and a good algorithm. The studied region with approximately 178.3 ha is located downstream lands of the Azad dam ​​ in northwest of Sanandaj in Kurdistan province. Based on the meteorological data, the soil temperature and moisture regimes were estimated as mesic and xeric, respectively. The main landforms of the studied area are Footslope, Convergent Backslope, and Divergent Shoulder and the main land use is farm. A random pattern was used to determine the position and distribution of the 84 profiles in the studied area. At each observation point, a soil profile was dug and accurately described and sampled. Physico-chemical analyses like electrical conductivity, soil pH, soil texture, TNV, and organic carbon were carried out based on the standard methods. Based on morphological characteristics and laboratory data, each profile was classified up to the family level based on soil taxonomy.

Finding and conclusions

Two orders (Inceptisols and Entisols) two Suborder, three Great groups, five Subgroups, and a Family were identified in the studied region. Topographic attributes were derived from a digital elevation model with a 10 m cell size. All types of DEM covariates (such as curvatures, slope, aspect, elevation, convergent index, LS factor, topographic wetness index, catchment area, close depressions, valley depth, and …) were extracted using SAGA GIS from DEM. RS covariates were extracted from atmospherically corrected sentinel-2B images. Algorithms were trained with 70% of randomly selected data and validated with the rest of 30% of the dataset. The Kappa coefficient of agreement (K), overall accuracy (OA), and Brier score were used to assess the performance of each model. The best result was obtained for the ANN model (OA=0.65, k=0.53, and BS=0.16, respectively). The weakest predictions were for the C.5 model with 0.38 Overall accuracy, 0.22 kappa index, and 0.87 Brier score. These results are due to the fact that the performance of machine learning models depends on the study area, the number of observation points, and the type and number of predictor variables.

مراجع

AbbaszadehAfshar, F., Ayoubi Sh., & Jafari, A. (2017). Digital Soil Mapping of Soil Classes using Conventional Soil Maps in the Arid Region Southeastern Iran. J. Water and Soil Sci, 21(1): 239-253. (In Persian)
Adu-Poku, S. (2012). Comparing classification algorithms in data mining. (Thesis). Central Connecticut State University. pp: 24-26.
Ali Ehiai, M. (1993). Description of Soil Chemical Analysis Methods (Volume 1). Technical Journal 893. Soil and Water Research Institute. Agricultural Research and Education Organization, Tehran, Iran. (In Persian)
Ali Ehiai, M., & Behbahanizadeh A. A. (1993). Description of Soil Chemical Analysis Methods (Volume 2). Technical Journal 1024. Soil and Water Research Institute. Agricultural Research and Education Organization, Tehran, Iran. (In Persian)
Andronikov, V. L. & Dorbrolv’skiy, G. V. (1991). Theory and methods for the use of remote sensing in the study of soils. Mapping Sciences and Remote Sensing, 28(2), 92-101. ‏
Angelini, M. E., Heuvelink, G. B. M., Kempen, B., Morrás, H. J. M. )2016(. Mapping the soils of an Argentine Pampas region using structural equation modelling. Geoderma. 281, 102-118.
Banaii, M. H. )2001(. Map of Resources and Talents of Iranian Soils. Iran Soil and Water Research Institute, Tehran. (In Persian)
Barthold F. K., Wiesmeier, M., Breuer, L., Frede, H. G., Wu, J., & Blank, F. B. (2013). Land use and climate control the spatial distribution of soil types in the grasslands of Inner Mongolia. J. Arid Environ. (88), 194-205.
Bouma, J., Booltink, H., & Finke, P. (1996). Use of soil survey data for modeling solute transport in the vadose zone. Journal of environmental quality, 25(3), 519-526.
Breiman L. (2001). Random forests. Machine learning, 45(1), 5-32.
Brungard C. W., Boettinger, J. L., Duniway, M. C., Wills, S. A., & Edwards, Jr, T. C. (2015). Machine learning for predicting soil classes in three semi-arid landscapes. Geoderma, (239), 68-83.
Buol, S. W., Southard, R. J., Graham, R. C., McDaniel, P. A. (2003). Soil Genesis and Classification. Iowa State University Press, 494 p.
Burrough P. A., Beckett, P. H. T., & Jarvis, M. G. (1971). The relation between cost and utility in soil survey. European Journal of Soil Science, 22(2), 368-381.
Camera, C., Zomeni, Z., Noller, J. S., Zissimos, A. M., Christoforou, I. C., Bruggeman, A. )2017(. A high resolution map of soil types and physical properties for Cyprus: A digital soil mapping optimization, Geoderma, Volume 285, Pages 35-49, https://doi.org/10.1016/j.geoderma.2016.09.019.
Campling, P., Gobin, A. & Feyen J. (2002). Logistic modeling to spatially predict the probability of soil drainage classes. Soil Science Society of America Journal, (66), 1390–1401.
Chen, S., Arrouays, D.,LeatitiaMulder, V., Poggio, L., Minasny, B., Roudier, P., Libohova, Z., Lagacherie, P., Shi, Z., Hannam, J., Meersmans, J., Richer-de-Forges, A.C., Walter, C. )2022(. Digital mapping of GlobalSoilMap soil properties at a broad scale: A review. Geoderma, 409. 10.1016/j.geoderma.2021.115567.
Congalton, R.G. )1991(. A review of assessing the accuracy of classifications of remotely sensed data. Remote Sens. Environ. 37, 35–46.
Fernandes Coelho, F., Giasson, E., Campos, A. R., Tiecher, T., Ferreira Costa, J. J., Coblinski, J. A. )2021(. Digital soil class mapping in Brazil: a systematic review. Soil and Plant Nutrition, 78 (5).
Forghani, S. J., Pahlevanrad, M. R., Esfandiari, M., Mohammadi Torkashvand, A. (2022). Digital soil mapping of soil classes in floodplain and low relief lands (Case study: Hirmand county). Water and Soil Resources Conservation, 9(4). (In Persian)
Grabs, T., Seibert, J., Bishop, K., and Laudon, H. )2009(. Modeling spatial patterns of saturated areas: A comparison of the topographic wetness index and a dynamic distributed model. J. Hydrol. 373: 15-23.
Grunwald, S. (2006). Environmental soil-landscape Modeling, Geomorphic Information Technologies and Pedometrics. Taylor and Francis Pub., USA.
Han, X., Liu, J., Shen, X., Liu, H., Li, X., Zhang, J., Wu, P., Liu, Y. )2022(. High relief yield strong topography-soil water-vegetation relationships in headwater catchments of southeastern China. Geoderma (428).
Hengl, T., Rossiter, D. G., & Stein, A. (2003). Soil sampling strategies for spatial prediction by correlation with auxiliary maps. Geoderma, (120), 75–93.
Hengl, T., Toomanian, N., Reuter, H. I., & Malakouti, M. J. (2007). Methods to interpolate soil categorical variables from profile observations: Lessons from Iran. Geoderma, 140(4), 417-427. ‏
Heung, B., Ho, H. C., Zhang, J., Knudby, A., Bulmer, C. E., & Schmidt, M. G. (2016). An overview and comparison of machine-learning techniques for classification purposes in digital soil mapping. Geoderma, (265), 62-77.
Jafari, A., Khademi, H., & Ayoubi, Sh. (2013). Digital mapping of characteristic horizons and large soil groups in Zarand area of ​​Kerman. Journal of Soil and Water Sciences, 16 (62): 177-193. (In Persian)
Johansson, U., König, R., Niklasson, L. (2010). Genetic rule extraction optimizing brier score. In: Pelikan, M., Branke, J. (Eds.), GECCO. ACM, pp. 1007–1014.
Kempen, B., Brus, D. J., Heuvelink, G. B., and Stoorvogel, J. J. )2009(. Updating the 1: 50,000 Dutch soil map using legacy soil data: A multinomial logistic regression approach. Geoderma, 151(3-4): 311-326.
Khaledian, Y., Miller, B. A. (2020). Selecting appropriate machine learning method for digital soil mapping. Applied Mathematical Modelling, (81), 401-418.
Kim, J., Grunwald, S., Rivero, R. G., & Robbins, R. (2012). Multi-scale modeling of soil series using remote sensing in a wetland ecosystem. Soil Science Society of America Journal, 76 (6), 2327-2341.
Kuhn, M. (2018). Caret: classifcation and regression training. Astrophysics Source Code Library.
Lu, D., Mausel, P., Brondizio, E. & Moran, E. (2004). Change detection techniques. International journal of remote sensing, 25(12), 2365-2401. ‏
Luoto, M., & Hjort J. (2005). Evaluation of current statistical approaches for predictive geomorphological mapping. Geomorphology, (67), 299-315.
MacMillan, R. A. (2003). LandMapR Software Toolkit C++ Version. LandMapper Environmental Solution. User Manual.
MacMillan, R. A., Pettapiece, W. W., Nolan, S. C., Goddard, T. W. (2000). Ageneric procedure for automatically segmenting landforms into landform elements using DEMs, heuristic rules and fuzzy logic. Fuzzy Sets and Systems, 113: 81-109.
Maghsodi, M., Rostaminia, M., Faramarzi, M., Keshavarzi, A., Rahmani, A., Mousavi, S. R. (2020). Digital mapping of soil family class using the machine learning approach (a case study: semi-arid lands in the west of Iran). Journal of water and soil science, Vol. 24, No, 2. )In Persian(
Mahler. P. J. (1970). Manual of multipurpose land classification. Rep. No. 212, Soil and Water Research Institute, Iran.
Marquardt, D.W. (1963). An algorithm for least-squares estimation of nonlinear parameters. J. Soc. Ind. Appl. Math (11): 431–441.
McBratney, A. B., Santos, M. M., & Minasny, B. (2003). On digital soil mapping. Geoderma, 117(1), 3-52.
 Minasny, B., McBratney, A. B. (2016). Digital soil mapping: a brief history and some lessons. Geoderma 264, 301–311. https://doi.org/10.1016/j.
Monterio, M. E. C., Avalos, F. P., Pelegrino, M. P., Vilela, R. B., Junior, F. W. A., Bueno, I. T., Li, N., Silva, S. H. G., Giasson, E., Curi, N., Menezes, M. D. (2023). Digital mapping of soil classes in southeast Brazil: environmental covariate selection, accuracy, and uncertainty. Journal of South America Earth Sciences.
Moran, C. J., & Bui, E. N. (2002). Spatial data mining for enhanced soil map modelling. International Journal of Geographical Information Science, 16(6), 533-549.
Mosleh, Z., Salehi, M. H., Jafari, A., Borujeni, I. E. and Mehnatkesh, A. (2016). The effectiveness of digital soil mapping to predict soil properties over low-relief areas. Environmental monitoring and assessment, 188(3), 195.
Mousavi, S. R., Sarmadian, F., Rahmani, A. (2019). Modelling and prediction of soil classes using boosting regression tree and random forests machine learning algorithms in some part of Qazvin plain. Iranian Journal of Soil and Water Research, 50 (10), 2525-2538. (In Persian)
Olaya, V. I. (2004). A gentle introduction to SAGA GIS. The SAGA User Group eV, Gottingen,Germany, 208.‏
Padarian, J., Minasny, B., McBratney, A. B. (2019). Using deep learning for digital soil mapping. Soil 5, 79–89. https://doi.org/10.5194/soil-5-79-2019.
Quinlan, J. R. (2001). Cubist: An Informal Tutorial. http://www.rulequest.com.
Rahimi Mashkaleh, M., Delavar, M. A., Jamshidi, M., Sharififar, A. (2023). Modelling spatial distribution of soil classes using machine learning algorithms in some parts of Zanjan province. Iranian Journal of Soil and water Research, 37 (2). (In Persian)
Rodriguez-Galiano, V., Sanchez-Castillo, M., Chica-Olmo, M., & Chica-Rivas, M. (2015). Machine learning predictive models for mineral prospectivity: An evaluation of neural networks, random forest, regression trees and support vector machines. Ore Geology Reviews, (71), 804-818.
Rossiter, D. (2005). Digital soil mapping: Towards a multiple-use Soil Information System. Análisis Geográficos (Revista del Instituto Geográfico" Augusín Codazzi"), 32(1), 7-15.
Rossiter, D. G. (2004). Digital soil resource inventories: status and prospects. Soil Use and Management, 20(3), 296-301.
Rostaminia, M., Rahmani, A., Mousavi, R., Taghizadeh, R. (2021). Spatial prediction of soil organic carbon stocks in an arid rangeland using machine learning algorithms. Environmental Monitoring and Assessment, 193(185):1-17.
Salehi, M. H., & Khademi, H. (2008). Basics of soil mapping. Published by Isfahan University of Technologoy, pp:212. (In Persian)
Saurette, D. D., Berg, A. A., Laamrani, A., Heck, R., Gillespie, A., Voroney, P. (2022). Effects of sample size and covariate resolution on field-scale predictive digital mapping of soil carbon. Geoderma: 425.
Schoeneberger, P. J., Wysocki, D. A.,  Benham, E. C. (2012). Soil Survey Staff. 2012. Field book for describing and sampling soils, Version 3.0. Lincoln: Natural Resources Conservation Service, National Soil Survey Center.
Scull, P., Franklin, J., Chadwick, O. A., & McArthur, D. (2003). Predictive soil mapping: a review. Progress in Physical Geography, 27(2): 171-197.
Scull, P., Franklin, J. & Chadwick, O. A. (2005). The application of classification tree analysis to soil type prediction in a desert landscape. Ecological Modeling. 181: 1–15.
Sharififar, A., Sarmadian, F., Malone, B. P., Minasny, B. (2019). Addressing the issue of digital mapping of soil classes with imbalanced class observations. Geoderma, 350, 84–92.
Siqueira, D. S., Marques Jr, J., Pereira, G. T., Teixeira, D. B., Vasconcelos, V., Carvalho Junior, O. A., Martins, E. S. (2015). Detailed mapping unit design based on soil-landscape relation and spatial variability of magnetic susceptibility and soil color. Catena, 135, 149-162.
Soil Survey Staff, (2022). Keys to Soil Taxonomy, 13th ed. U. S. Department of Agriculture, Natural Recourses Conservation Service.
Taghizadeh-Mehrjardi, R., Sarmadian, F., Omid, M., Toomanian, N., Rousta, M. J., & Rahimian, M. H. (2015). Digital mapping of soil classes using different data mining techniques in Ardakan region, Yazd province. Journal of Agricultural Engineering. 37(2): 101-115. (In Persian)
Taghizadeh-Mehrjardi, R., Minasny, B., Toomanian, N., Zeraatpisheh, M., Amirian-Chakan, A., Triantafilis, J. (2019). Digital Mapping of Soil Classes Using Ensemble of Models in Isfahan Region, Iran. Soil System, (3, 37). https://doi.org/10.3390/soilsystems3020037
Tiwari, S. K., Saha, S. K., & Kumar, S. (2015). Prediction modeling and mapping of soil carbon content using artificial neural network, hyperspectral satellite data and field spectroscopy. Advances in Remote Sensing, 4(01), 63.
Western, A. W. (2004). Spatial correlation of soil moisture in small catchments and its relationship to dominant spatial hydrological processes. J. Hydrol. 286: 1-4. 113-134.
Witten, I. H., Frank, E., Hall, M. A. (2011). Data mining: Practical Machine Learning Tools and Techniques. Morgan Kaufmann, Burlington.
Zhu, X., Hudson, B., Burt, J., Lubich, K., & Simonson, D. (2001). Soil Mapping Using GIS Expert Knowledge and Fuzzy Logic. Journal of Soil Science Society of America, (65), 1463-1472.
تحقیقات آب و خاک ایران
دوره 54، شماره 11
بهمن 1402
صفحه 1697-1714

فایل ها

هم رسانی

ارجاع به این مقاله

آمار