Large-scale digital mapping of topsoil total nitrogen using machine learning models and associated uncertainty map

Parsaie, Farzaneh; Firouzi, Ahmad Farrokhian; Mousavi, Sayed Roholla; Rahmani, Asghar; Sedri, Mohammad Hossein; Homaee, Mehdi

doi:10.1007/s10661-021-08947-w

Cited by 21 publications

(8 citation statements)

References 54 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…RF is a model that integrates multiple decision trees to invert water quality parameter predictions, and the combination of decision trees it constructs increases the ability of the hyperspectral reflectance model to predict the target variable [63][64][65]. Bagging is a prototype of the parallel integrated learning method that is directly based on the selfsampling method taking randomized bootstrapping with put-back sampling [66].…”

Section: Modeling Of Total Nitrogen Monitoringmentioning

confidence: 99%

“…However, the variation among the modeling results is substantial, and the optimal result was obtained with XGBoost, which can best describe the quantitative relationship between the reflectivity of the sensitive band and TN in water. The modeling accuracy is comparatively balanced for the bagging model; in the RF model based on the reflectance of the sensitive band, higher-sensitivity information is underestimated, although the RF are valid for this nonlinear issue [63,65]. Therefore, the XGBoost model is recommended to explain the relationship between the TN concentration in water and hyperspectral reflectance.…”

Section: The Potential Of the Developed Modelmentioning

confidence: 99%

See 1 more Smart Citation

Evaluation of Total Nitrogen in Water via Airborne Hyperspectral Data: Potential of Fractional Order Discretization Algorithm and Discrete Wavelet Transform Analysis

Liu

Ding

et al. 2021

Remote Sensing

View full text Add to dashboard Cite

Controlling and managing surface source pollution depends on the rapid monitoring of total nitrogen in water. However, the complex factors affecting water quality (plant shading and suspended matter in water) make direct estimation extremely challenging. Considering the spectral response mechanisms of emergent plants, we coupled discrete wavelet transform (DWT) and fractional order discretization (FOD) techniques with three machine learning models (random forest (RF), bagging algorithm (bagging), and eXtreme Gradient Boosting (XGBoost)) to mine this potential spectral information. A total of 567 models were developed, and airborne hyperspectral data processed with various DWT scales and FOD techniques were compared. The effective information in the hyperspectral reflectance data were better emphasized after DWT processing. After DWT processing the original spectrum (OR), its sensitivity to TN in water was maximally improved by 0.22, and the correlation between FOD and TN in water was optimally increased by 0.57. The transformed spectral information enhanced the TN model accuracy, especially for FOD after DWT. For RF, 82% of the model R2 values improved by 0.02~0.72 compared to the model using FOD spectra; 78.8% of the bagging values improved by 0.01~0.53 and 65.0% of the XGBoost values improved by 0.01~0.64. The XGBoost model with DWT coupled with grey relation analysis (GRA) yielded the best estimation accuracy, with the highest precision of R2 = 0.91 for L6. In conclusion, appropriately scaled DWT analysis can substantially improve the accuracy of extracting TN from UAV hyperspectral images. These outcomes may facilitate the further development of accurate water quality monitoring in sophisticated global waters from drone or satellite hyperspectral data.

show abstract

Section: Modeling Of Total Nitrogen Monitoringmentioning

confidence: 99%

Section: The Potential Of the Developed Modelmentioning

confidence: 99%

Evaluation of Total Nitrogen in Water via Airborne Hyperspectral Data: Potential of Fractional Order Discretization Algorithm and Discrete Wavelet Transform Analysis

Liu

Ding

et al. 2021

Remote Sensing

View full text Add to dashboard Cite

show abstract

“…Identifying the most important factors affecting SOCS variation in the landscape related to choose the optimal environmental covariates, therefore, is necessary to provide set of environmental variables that are achieved in low cost or easily accessible (Brungard et al, 2015;Miller et al, 2015). Various feature selection approaches were used by previous researchers in DSM such as Parsaie et al (2021) using Boruta supervised method for selecting appropriate covariates to estimate spatial distribution of surface total nitrogen in Iran, Matinfar et al (2021) applied unsupervised principal component analysis (PCA) for modeling surface soil organic carbon in semi-arid region in west of Iran, or Taghizadeh-Mehrjardi et al (2021) and Tayebi et al (2021) successfully using recursive feature elimination (RFE) for modeling soil salinity and SOCS in Iran and Brazil, respectively. Recursive feature elimination like to Boruta is a type of backward supervised feature selection method that avoids fitting multiple models at each stage (Guyon et al, 2002;Kuhn & Johnson, 2013).…”

Section: Introductionmentioning

confidence: 99%

Spatial prediction of soil organic carbon stocks in an arid rangeland using machine learning algorithms

Rostaminia

Rahmani

Mousavi

et al. 2021

Environ Monit Assess

Self Cite

View full text Add to dashboard Cite

method. Robust and popular random Forest (RF), cubist (CB) along with random forest-ordinary kriging (RF-OK), and cubist-ordinary kriging (CB-OK) hybrid ML models were applied to the prediction of SOCS. Ten-fold CV was implemented for modeling performance and uncertainty map. According to data analysis, the maximum, minimum, and average values of SOCS are 44.50, 10.50, and 20.50 (ton. ha −1 ) at the surface depth (0-30 cm), respectively. In general, normalized and standardized height covariates had a higher effect related to other predictors. On the other hand, two remote sensing (RS) indices, including salinity ratio (salinity) and GNDVI index, had a better impact on SOCS variability. The external validation of model performance indicated that RF-OK with (R 2 = 0.75, RMSE = 6.33 ton. ha −1 ) with the high and low uncertainty range (3.33-9.50 ton. ha −1 ) was the outperformed ML model in compare with other models as RF (R 2 = 0.65, RMSE = 7.38 ton. ha −1 ), CB-OK (R 2 = 0.56, RMSE = 9.22 ton. ha −1 ), and CB (R 2 = 0.33, RMSE = 10.42 ton. ha −1 ). In general, the hybrid models improved the accuracy of RF and CB with increased 0.11 until 0.23 of R 2 , and 1.05 to 1.2 (ton. ha −1 ) decreased RMSE of model's prediction. Hence, we conclude that the topographic attributes (especially normalized and standardized height) were the most critical factors in controlling surface SOCS in arid rangelands when combining with robust RF ML model, and optimized soil sampling methods like RF-cLHS can prepare acceptable soil properties maps.

show abstract

“…Previous studies have shown that RF and Cubist are methods that perform well in mapping various soil properties (Table 5) (Fathololoumi et al 2020;Kaya et al 2022;Parsaie et al 2021;Saidi et al 2022;Suleymanov et al 2023;Zeraatpisheh et al 2019). This fact justifies why these methods are two of the most used in soil property mapping (Khaledian and Miller 2020).…”

Section: Discussionmentioning

confidence: 90%

Looking for Optimal Maps of Soil Properties at the Regional Scale

Barrena-González,

Lavado Contador,

Repe

et al. 2024

Int J Environ Res

View full text Add to dashboard Cite

Around 70% of surface in Extremadura, Spain, faces a critical risk of degradation processes, highlighting the necessity for regional-scale soil property mapping to monitor degradation trends. This study aimed to generate the most reliable soil property maps, employing the most accurate methods for each case. To achieve this, six different machine learning (ML) techniques were tested to map nine soil properties across three depth intervals (0–5, 5–10 and > 10 cm). Additionally, 22 environmental covariates were utilized as inputs for model performance. Results revealed that the Random Forest (RF) model exhibited the highest precision, followed by Cubist, while Support Vector Machine showed effectiveness with limited data availability. Moreover, the study highlighted the influence of sample size on model performance. Concerning environmental covariates, vegetation indices along with selected topographic indices proved optimal for explaining the spatial distribution of soil physical properties, whereas climatic variables emerged as crucial for mapping the spatial distribution of chemical properties and key nutrients at a regional scale. Despite providing an initial insight into the regional soil property distribution using ML, future work is warranted to ensure a robust, up-to-date, and equitable database for accurate monitoring of soil degradation processes arising from various land uses.

show abstract

Large-scale digital mapping of topsoil total nitrogen using machine learning models and associated uncertainty map

Cited by 21 publications

References 54 publications

Evaluation of Total Nitrogen in Water via Airborne Hyperspectral Data: Potential of Fractional Order Discretization Algorithm and Discrete Wavelet Transform Analysis

Evaluation of Total Nitrogen in Water via Airborne Hyperspectral Data: Potential of Fractional Order Discretization Algorithm and Discrete Wavelet Transform Analysis

Spatial prediction of soil organic carbon stocks in an arid rangeland using machine learning algorithms

Looking for Optimal Maps of Soil Properties at the Regional Scale

Contact Info

Product

Resources

About