Incorporation of high accuracy surface modeling into machine learning to improve soil organic matter mapping

Digital soil mapping approaches related to soil organic matter (SOM) are crucial to quantify the process of the carbon cycle in terrestrial ecosystems and thus, can better manage soil fertility. Recently, many studies have compared machine learning (ML) models with traditional statistical models in...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Zong Wang, Zhengping Du, Xiaoyan Li, Zhengyi Bao, Na Zhao, Tianxiang Yue
Formato:	article
Lenguaje:	EN
Publicado:	Elsevier 2021
Materias:	Soil organic matter Dongzhi Loess Tableland Gradient boosting modeling Random forest High accuracy surface modeling Ecology QH540-549.5
Acceso en línea:	https://doaj.org/article/56f5059ed42145c7b13d73dc757c01df
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:56f5059ed42145c7b13d73dc757c01df
record_format	dspace
spelling	oai:doaj.org-article:56f5059ed42145c7b13d73dc757c01df2021-12-01T04:57:15ZIncorporation of high accuracy surface modeling into machine learning to improve soil organic matter mapping1470-160X10.1016/j.ecolind.2021.107975https://doaj.org/article/56f5059ed42145c7b13d73dc757c01df2021-10-01T00:00:00Zhttp://www.sciencedirect.com/science/article/pii/S1470160X21006403https://doaj.org/toc/1470-160XDigital soil mapping approaches related to soil organic matter (SOM) are crucial to quantify the process of the carbon cycle in terrestrial ecosystems and thus, can better manage soil fertility. Recently, many studies have compared machine learning (ML) models with traditional statistical models in digital soil mapping. However, few studies focused on the application of hybrid models that combine ML with statistical models to map SOM content, especially in loess areas, which have a complicated geomorphologic landscape. In this study, the trend prediction used two ML models, i.e., gradient boosting modeling and random forest (RF), and a traditional stepwise multiple linear regression plus interpolated residuals generated from two classic geostatistical models, i.e., ordinary kriging and inverse distance weighting, and a high accuracy surface modeling (HASM) were implemented to map SOM content in the Dongzhi Loess Tableland area of China. A total of 145 topsoil samples and heterogeneous environmental variables were collected to develop the hybrid models. Results showed that 18 variables related to soil properties, climate variables, terrain attributes, vegetation indices, and location attributes played an important role in SOM mapping. The models that incorporate ML algorithms and interpolated residuals to predict SOM variation were found to have a better ability to handle complex environment relationships. The HASM model outperformed traditional geostatistical models in interpolating the residuals. In contrast, RF combined with HASM residuals (RF_HASM) gave the best performance, with the lowest mean absolute error (1.69 g/kg), root mean square error (2.30 g/kg), and the highest coefficient of determination (0.57) and concordance correlation coefficient (0.69) values. Moreover, the spatial distribution pattern obtained with RF_HASM yielded a spatial distribution of SOM that better fit the actual distribution pattern of the study area. In conclusion, these results suggest that RF_HASM is particularly capable of improving the mapping accuracy of SOM content at the regional scale.Zong WangZhengping DuXiaoyan LiZhengyi BaoNa ZhaoTianxiang YueElsevierarticleSoil organic matterDongzhi Loess TablelandGradient boosting modelingRandom forestHigh accuracy surface modelingEcologyQH540-549.5ENEcological Indicators, Vol 129, Iss , Pp 107975- (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	Soil organic matter Dongzhi Loess Tableland Gradient boosting modeling Random forest High accuracy surface modeling Ecology QH540-549.5
spellingShingle	Soil organic matter Dongzhi Loess Tableland Gradient boosting modeling Random forest High accuracy surface modeling Ecology QH540-549.5 Zong Wang Zhengping Du Xiaoyan Li Zhengyi Bao Na Zhao Tianxiang Yue Incorporation of high accuracy surface modeling into machine learning to improve soil organic matter mapping
description	Digital soil mapping approaches related to soil organic matter (SOM) are crucial to quantify the process of the carbon cycle in terrestrial ecosystems and thus, can better manage soil fertility. Recently, many studies have compared machine learning (ML) models with traditional statistical models in digital soil mapping. However, few studies focused on the application of hybrid models that combine ML with statistical models to map SOM content, especially in loess areas, which have a complicated geomorphologic landscape. In this study, the trend prediction used two ML models, i.e., gradient boosting modeling and random forest (RF), and a traditional stepwise multiple linear regression plus interpolated residuals generated from two classic geostatistical models, i.e., ordinary kriging and inverse distance weighting, and a high accuracy surface modeling (HASM) were implemented to map SOM content in the Dongzhi Loess Tableland area of China. A total of 145 topsoil samples and heterogeneous environmental variables were collected to develop the hybrid models. Results showed that 18 variables related to soil properties, climate variables, terrain attributes, vegetation indices, and location attributes played an important role in SOM mapping. The models that incorporate ML algorithms and interpolated residuals to predict SOM variation were found to have a better ability to handle complex environment relationships. The HASM model outperformed traditional geostatistical models in interpolating the residuals. In contrast, RF combined with HASM residuals (RF_HASM) gave the best performance, with the lowest mean absolute error (1.69 g/kg), root mean square error (2.30 g/kg), and the highest coefficient of determination (0.57) and concordance correlation coefficient (0.69) values. Moreover, the spatial distribution pattern obtained with RF_HASM yielded a spatial distribution of SOM that better fit the actual distribution pattern of the study area. In conclusion, these results suggest that RF_HASM is particularly capable of improving the mapping accuracy of SOM content at the regional scale.
format	article
author	Zong Wang Zhengping Du Xiaoyan Li Zhengyi Bao Na Zhao Tianxiang Yue
author_facet	Zong Wang Zhengping Du Xiaoyan Li Zhengyi Bao Na Zhao Tianxiang Yue
author_sort	Zong Wang
title	Incorporation of high accuracy surface modeling into machine learning to improve soil organic matter mapping
title_short	Incorporation of high accuracy surface modeling into machine learning to improve soil organic matter mapping
title_full	Incorporation of high accuracy surface modeling into machine learning to improve soil organic matter mapping
title_fullStr	Incorporation of high accuracy surface modeling into machine learning to improve soil organic matter mapping
title_full_unstemmed	Incorporation of high accuracy surface modeling into machine learning to improve soil organic matter mapping
title_sort	incorporation of high accuracy surface modeling into machine learning to improve soil organic matter mapping
publisher	Elsevier
publishDate	2021
url	https://doaj.org/article/56f5059ed42145c7b13d73dc757c01df
work_keys_str_mv	AT zongwang incorporationofhighaccuracysurfacemodelingintomachinelearningtoimprovesoilorganicmattermapping AT zhengpingdu incorporationofhighaccuracysurfacemodelingintomachinelearningtoimprovesoilorganicmattermapping AT xiaoyanli incorporationofhighaccuracysurfacemodelingintomachinelearningtoimprovesoilorganicmattermapping AT zhengyibao incorporationofhighaccuracysurfacemodelingintomachinelearningtoimprovesoilorganicmattermapping AT nazhao incorporationofhighaccuracysurfacemodelingintomachinelearningtoimprovesoilorganicmattermapping AT tianxiangyue incorporationofhighaccuracysurfacemodelingintomachinelearningtoimprovesoilorganicmattermapping
_version_	1718405661226696704

Incorporation of high accuracy surface modeling into machine learning to improve soil organic matter mapping

Ejemplares similares