Prediction of Soil Nutrients Based on Topographic Factors and Remote Sensing Index in a Coal Mining Area, China

Round 1

Reviewer 1 Report

Abstract – correct

Introduction

Can be mention about using GIS

 

Material and methods

Please add references about systematic os soils according to WRB

Brown and hard coal ...

Why not top layer 0-30 cm, only 0-20 cm, according to .......  

Please give extractant for available potassium ...  

 

Results

Can be give sand, silt, clay of soils and nutrients in these soils and correlations ?

Discussion

Conclusion

This is normal conclusion, farming practices in better index than industry activivities

Please give information which model and index are the best for monitoring to nutrients in the soils

Author Response

Review 1

Abstract: Correct

Response: Changes have been made in the revised manuscript in Lines 12-28 as follows:

Abstract: (1) Background: Coal mining operations caused severe land subsidence and altered the distributions of soil nutrients that influenced by multiple environmental factors at different scales. However, the prediction performances for soil nutrients based on their scale-specific relationships with influencing factors remains undefined in the coal mining area. The objective of this study was to establish prediction models of soil nutrients based on their scale-specific relationships with influencing factors in a coal mining area. (2) Methods: Soil samples were collected based on the 1 × 1 km regular grid, and contents of soil organic matter, soil available nitrogen, soil available phosphorus, and soil available potassium were measured. The scale components of soil nutrients and the influencing factors collected from remote sensing and topographic factors were decomposed by 2D-EMD, and the predictions for soil nutrients were established using the methods of MLSR_Ori, PLSR_Ori, PLSR_BIMF, and 2D-EMD_PM. (3) Results: The correlation types and correlation coefficients between soil nutrients and influencing factors were scale-dependent. The variances of soil nutrients at smaller scale were stochastic and non-significantly correlated with influencing factors, while their variances at the larger scales were stable. The prediction performances in the coal mining area were better than those in the non-coal mining area, and 2D-EMD_PM had the most stable performance. (4) Conclusions: The scale-dependent predictions can be used for soil nutrients in the coal mining areas.

 

Introduction: Can be mention about using GIS

Response: Thanks for your comment. In the revised manuscript, we mentioned the geospatial methods (i.e., geospatial autocorrelation, geostatistical models, inverse distance weighted), which were closely related to GIS. Thus, we thought that adding the content about GIS would be redundant. The correlated content in Line 38-43 as follows:

A number of methods, including geostatistical models [1], inverse distance weighted [2], trend surface analysis [3] and the like, resolve the spatial distribution of soil properties based on the geospatial autocorrelation [4, 5]. However, these methods are limited by sampling density and the methods either failure to consider the environmental effects on soil nutrients or consider the environmental effects only at the sampling scale, causing their heterogeneous distribution at different scale in different intensity [6, 7].

 

Material and methods:

Please add references about systematic soils according to WRB

Response: Thanks for your comment, and we added the references in Line 67 as follows:

The dominant soil type in the study area is calcaric cambisols [23], and the major crops in the arable land are maize, potato, and wheat.

 

Brown and hard coal ...

Response: Changes have been made according to your comment in Line 70 as follows:

In the western part, there are many coal mines containing hard coal.

 

Why not top layer 0-30 cm, only 0-20 cm, according to .......  

Response: The top layer 0-20 cm was collected according to the depth of ploughing in the cultivated land here. Changes have been made in Line 79 as follows:

At each point, three soil samples within 5 m radius were collected using a soil auger and mixed together for one sample at top layer of 0-20 cm according to the ploughing depth in the cultivated land of the study area.

 

Please give extractant for available potassium ...

Response: Thank you for your comment, and changes have been made in Line 82-83 as follows:

The SOM was measured by the dichromate oxidation method [24], SAN was determined by the diffusion method [25], SAP was obtained by the Olsen extraction method using alkaline sodium bicarbonate as the extractant in a 20:1 ratio [26], and SAK was extracted by the flame photometer [27].

 

Results

Can be give sand, silt, clay of soils and nutrients in these soils and correlations?

Response: Thanks for your comment. We added the test method in the part of Materials and Methods (Line 84-85) as follows:

Contents of sand (0.050–2.000 mm), silt (0.002–0.050 mm) and clay (<0.002 mm) were determined by the pipette method [28].

We added the data and description of sand, silt, clay, and their correlation with soil nutrients in the part of Result (Line 173-192) as follows:

Obviously, the mean sand in coal mining area (0.23) was higher than that in the non-coal mining area (0.19); the mean silt in the coal mining area (0.48) was lower than that in the non-coal mining area (0.52); the mean clay in the coal mining and non-coal mining areas were similar. The spatial differentiation between coal mining area and non-coal mining area were significant for sand and silt, while it was non-significant for clay.

Table 1. Basic statistics of soil nutrients (SOM, SAN, SAP, and SAK) and soil texture under coal mining and non-coal mining areas, respectively.

Soil nutrients	Areas	Min	Mean	Max	Std	CV	ANOVA
								F	p-value
SOM (g/kg)	Coal mining area	1.10	17.42	36.42	7.79	44.73	a	4.04	0.04
	Non-coal mining area	4.72	19.88	32.63	5.58	28.09	b
SAN (mg/kg)	Coal mining area	7.42	32.44	57.48	12.04	37.12	a	10.12	0.00
	Non-coal mining area	5.56	39.40	66.75	11.92	30.26	b
SAP (mg/kg)	Coal mining area	0.95	7.95	16.91	3.83	48.15	a	2.60	0.11
	Non-coal mining area	1.52	9.21	22.30	5.11	55.48	a
SAK (mg/kg)	Coal mining area	60.30	152.50	311.55	40.75	26.73	a	2.42	0.12
	Non-coal mining area	70.35	163.14	244.55	33.83	20.74	a
Sand (-)	Coal mining area	0.09	0.23	0.64	0.09	40.06	a	4.94	0.03
	Non-coal mining area	0.09	0.19	0.67	0.09	47.93	b
Silt (-)	Coal mining area	0.22	0.48	0.66	0.10	21.37	a	6.13	0.01
	Non-coal mining area	0.18	0.52	0.70	0.10	18.56	b
Clay (-)	Coal mining area	0.14	0.29	0.48	0.08	26.98	a	0.19	0.67
	Non-coal mining area	0.13	0.28	0.56	0.08	27.25	a

CV represents coefficient of variance. ANOVA represents analysis of variance, and different letters between coal mining area and non-coal mining area represent the significant difference of the soil nutrient at P < 0.05 level.

The relationships between soil nutrients and soil texture under the coal mining area, non-coal mining area, and the entire Watershed are shown in Table 2. Despite the non-significant correlation between SOM and soil texture, SOM was positively related to silt in the coal mining area, and positively correlated with clay in the non-coal mining area. Although Silt had non-significant relationship with SAN in the non-coal mining area, silt had positive effect on SAN in the coal mining area and the entire area. The relationships between clay and SAP in the non-coal mining area and the entire area were significant. Notably, sand had significantly negative effect on SAK in the coal mining area, non-coal mining area, and the entire study area.

Table 2. The correlations between soil nutrients (SOM, SAN, SAP, and SAK) and soil texture under coal mining area, non-coal mining area, and the entire area, respectively

Soil texture	Coal mining area				Non-coal mining area				The entire area
	SOM	SAN	SAP	SAK	SOM	SAN	SAP	SAK	SOM	SAN	SAP	SAK
Sand	-0.10	-0.07	-0.01	-0.32*	0.01	0.08	-0.06	-0.44**	-0.09	-0.05	-0.07	-0.39**
Silt	0.27*	0.34**	-0.12	0.32*	-0.23	-0.10	-0.15	0.16	0.10	0.18*	0.10	0.27**
Clay	-0.22	-0.37**	0.20	-0.07	0.29*	0.03	0.26*	0.32*	0.02	0.18	0.22*	0.09

^* significant at P < 0.05; ^** significant at P < 0.01.

 

Conclusion

This is normal conclusion, farming practices in better index than industry activities. Please give information which model and index are the best for monitoring to nutrients in the soils.

Response: Thanks for your comment. The specific numbers of latent variables (LVs) for PLSR and models for MLSR were provided in Table 5. The specific method and some important index were provided in the conclusion of Lines 342-346 as follows:

The method of 2D-EMD_PM, which decomposed the soil nutrients and influencing factors into 4 scale components, predicted soil nutrients at each scale components from the corresponding scale of influencing factors, and predicted soil nutrients at the sampling scale from the predicted soil nutrients at each scale except BIMF1, was stable in the calibration and validation models. The method could be validated in the other coal mining areas in the world.

Author Response File: Author Response.pdf

Reviewer 2 Report

Dear Authors, I have reviewed the manuscript entitled: Prediction of Soil Nutrients Based on Topographic  Factors and Remote Sensing Index in a Coal Mining Area, China.

The paper is on average length and it is correctly divided in sections.

There are some acronyms like 2DEMD that is not what I usually treat in my work DEM is generally Digital elevation model. I Think the acronym should be changed to avoid confusion with it.

At a certain point I discovered that the methodological approach is based on a previous Geoderma paper of the same authors. It is fine. But is important to understand which is the motivation of the new one and which improvements has the new proposed approach.

Very few detailed are reported for the remote sensing indices, therefore in my opinion, the reader needs to be addressed through the choice of which index is better for this purpose, and how to derive it form raw data, or download as ready product from a provider.

Tha same is for the topographic indices calculations. Please add short descriptions for their derivation, from which DEM and which is the pros and cos of using some of them for the prediction of topsoil organic carbon and nitrogen. Please have a look at Schillaci et al 2015, terrain analysis and landform recognition book chapter for the topographic predictors. And Schillaci et al 2017, modelling SOC stock for the remote sensing indices derivation.

How is the topographic index "aspect" derived and used?

Simulation methods represent indeed an interesting subject, but I think the Authors can briefly summarize with a couple of sentences the state of the art of the research carried out so far about this approach for soil carbon and soil properties. For the long term research, I suggest to read something about systematic mapping of the literature search for agro-environmental topic.

The introduction is well written, but it lacks of a systematic state of the art description, (e.g. Scopus or Web of science search).

Methods are clear, but since there is not a strong proposed evidence that the method will bring so much innovation in the field of soil mapping, the latest development in the discipline has to be briefly stated.

Partitioning strategy probably depends on spatial heterogeneity, so the SOC could be more data specific and then it performs better, please try the two methods in an external dataset to assess whether SOC is really most reliable than k-means.

Table 3, how was the statistical significance of R² calculated?

Model equations are missed, it will be interesting for the sake of reproducibility.

From line 122 seems clear that there is no external validation. So in the discussions, please state that the dataset splitted in 75/25 doesn’t mean that the model could work as good as in the randomly selected validation.

Line 16-> multiscale relationships

Line 20 -> 2DEMD please change the acronym

Line 69 -> established is probably odd, please change

Line 72-76 -> please specify better the methods used for SOC and Nitrogen content determination by citing some protocols that have been followed. It is clear what the authors have done but in order to replicate this analysis the reader would need additional information.

Line 77-93 -> the methods need some more details, period of acquisition, software used

Line 93 -> what the authors meant for “run off”?

Line 106 -> which selected environmental fartors? And why into 3 BIMs

Line 118-> please show this ANOVA table and the significant differences, not only letters but p and F

 

 

I believe that the manuscript needs a major revision.

Though I trust the authors, I ask the editor to check for plagiarism, I cannot check on my own.

Therefore, I wish you best of luck.

Kind regards,

Author Response

Review 2

Dear Authors, I have reviewed the manuscript entitled: Prediction of Soil Nutrients Based on Topographic Factors and Remote Sensing Index in a Coal Mining Area, China. The paper is on average length and it is correctly divided in sections. There are some acronyms like 2DEMD that is not what I usually treat in my work DEM is generally Digital elevation model. I Think the acronym should be changed to avoid confusion with it.

Response: Thanks for your comment. We changed the acronyms 2DEMD to 2D-EMD throughout the revised manuscript.

 

At a certain point I discovered that the methodological approach is based on a previous Geoderma paper of the same authors. It is fine. But is important to understand which is the motivation of the new one and which improvements has the new proposed approach.

Response: Thanks for your comment. The two researches have the same foundation of theory, which is the influencing factors have different effect on soil nutrients at different scale in different intensity. Firstly, the sampling methods for the researches were different. Specifically, soil samples were collected in the spatial 2-dimension for this manuscript, while the soil samples were collected at three sampling transects in the spatial 1-dimension for previously study. Secondly, the study areas were different. Specifically, the study area of Changhe Watershed for this manuscript located in a coal mining area, while Taiyuan Basin for the previous Geoderma paper was not. Thirdly, the method was different. Specifically, the predicting method of this manuscript was the combination of 2D-EMD, PLSR, and MLSR, while the previous research was the combination of discrete wavelet transform and MLSR.

 

Very few detailed are reported for the remote sensing indices, therefore in my opinion, the reader needs to be addressed through the choice of which index is better for this purpose, and how to derive it form raw data, or download as ready product from a provider.

Response: We provided more information about the download website and the software deriving the factors in the revised manuscript of Lines 87-92 as follows:

The products of Moderate Resolution Imaging Spectroradiometer (MODIS), including gross primary production (GPP, MOD17A2), net primary production (NPP, MOD17A3), land surface temperature for daytime and night (DLST and NLST, MOD11A2), evapotranspiration (ET, MOD16), normalized difference vegetation index (NDVI, MOD13Q1) with 1 km resolution (the same with soil sampling scale of 1 km) in 2015 was downloaded from https://ladsweb.modaps.eosdis.nasa.gov/. The DLST, NLST, ET and NDVI in 2015 were used for extracting environmental factors in ENVI 5.3, including the intra-annual variance of the land surface temperature for daytime (DLSTV), the intra-annual variance of the land surface temperature for night (NLSTV), the annual mean evapotranspiration (ETM), the intra-annual standard deviation of evapotranspiration (ETD), the intra-annual variance of evapotranspiration (ETV), the annual mean NDVI (NDVIM), the inter-annual standard deviation of NDVI (NDVID) and the inter-annual variance of NDVI (NDVIV).

 

The same is for the topographic indices calculations. Please add short descriptions for their derivation, from which DEM and which is the pros and cos of using some of them for the prediction of topsoil organic carbon and nitrogen.

Response: We download DEM from the website http://gdem.ersdac.jspacesystems.or.jp/, and the topographic index were derived by the software of SAGA 6.4. The topographic factors have effects on the topsoil organic carbon and nitrogen, because they are related to microclimate and soil erosion, and thus affected the accumulation, decomposition and flowage of soil nutrients. We added some descriptions about this in the revised manuscript of Lines 102-109 as follows:

Digital elevation model (DEM) with 30 m from http://gdem.ersdac.jspacesystems.or.jp/ was downloaded. The DEM was used for extracting topographic factors, including elevation, aspect, slope, convergence index (CI), Channel Network Base Level (CNBL), flow accumulation (FA), LS Factor (LS), topographic wetness index (TWI), valley depth (VD), and vertical distance to channel network (VDCN) in the software of SAGA 6.4, and resampled into the same resolution with remote sensing of 1 km in ArcGIS 10.5. The factors indicated topographic conditions which were related to microclimate and soil erosion, and thus affected the accumulation, decomposition and flowage of soil nutrients.

 

 

Please have a look at Schillaci et al 2015, terrain analysis and landform recognition book chapter for the topographic predictors. And Schillaci et al 2017, modelling SOC stock for the remote sensing indices derivation.

Response: Thanks for your recommendation. Some of the remote sensing indices derived in the study was directly downloaded from MODIS product, and some of them was simply calculated from the MODIS product. The topographic factors were directly obtained in the software of SAGA 6.4 from DEM. The simple method was used for easily reproducibility.

 

How is the topographic index "aspect" derived and used?

Response: The topographic factor “aspect” was also derived from DEM using the software of SAGA 6.4.

 

Simulation methods represent indeed an interesting subject, but I think the Authors can briefly summarize with a couple of sentences the state of the art of the research carried out so far about this approach for soil carbon and soil properties. For the long term research, I suggest to read something about systematic mapping of the literature search for agro-environmental topic. The introduction is well written, but it lacks of a systematic state of the art description, (e.g. Scopus or Web of science search).

Response: Thanks for your recommendation. We added some references about soil mapping and predictions in the introduction of Lines 38-54 as follows:

A number of methods, including geostatistical models [1], inverse distance weighted [2], trend surface analysis [3] and the like, resolve the spatial distribution of soil properties based on the geospatial autocorrelation [4, 5]. However, these methods are limited by sampling density and the methods either failure to consider the environmental effects on soil nutrients or consider the environmental effects only at the sampling scale, causing their heterogeneous distribution at different scale in different intensity [6, 7]. Previously studies indicate that the qualitative and quantitative information of vegetation growth or land surface temperatures can be obtained from the electromagnetic radiation of remote sensing, which indirectly demonstrating the distributions of soil nutrients [8-11]. Meanwhile, topographic factors affect the transposition and redistribution of soil nutrient contents [12], and their effects on soil nutrients vary in different scales [13, 14]. A number of prediction methods for soil properties based on their relationships with topographic factors are proposed [15], including ensemble learning model [16], extended models [17], regression model [18] and so on. Nevertheless, these methods do not consider their multiscale relationships between soil nutrients and the covariates. Additionally, scale-specific correlations between soil properties and environmental factors are explored using geostatistical approach [19], multivariate empirical mode decomposition [20, 21], wavelet transform [13] and so on. However, these methods either fail to realize the prediction of soil properties or establish the validation model

 

Methods are clear, but since there is not a strong proposed evidence that the method will bring so much innovation in the field of soil mapping, the latest development in the discipline has to be briefly stated.

Response: The method was proposed in 2011 by Xu et al.(Xu et al., 2011), and the latest developments were in the previous studies (Huang et al., 2017; Zhu et al., 2019b) referred in the manuscript.

 

Partitioning strategy probably depends on spatial heterogeneity, so the SOC could be more data specific and then it performs better, please try the two methods in an external dataset to assess whether SOC is really most reliable than k-means.

Response: Thanks for your comment. We discussed the problem of the research in the discussion part of Lines 328-330, and will do the research in the future study because of the limited space.

However, 2D-EMD_PM was only applied for the validated dataset which was randomly selected only once, and not applied it for all the possible datasets randomly selected. Meanwhile, 2D-EMD_PM application in other coal mining areas was not included in this study, and further study was needed in the future for better predictions on soil nutrients.

 

Table 3, how was the statistical significance of R² calculated? Model equations are missed, it will be interesting for the sake of reproducibility.

Response: The R² was calculated from the measured and estimated soil nutrients using the equation as follows:

where ,  are the measured and estimated soil nutrients for sample i,  is the mean of the soil nutrients, and n is the number of samples. The statistical significance was indicated by the P-value. Because the number of soil samples for SOM (or SAN) prediction models was 60, and the number of soil samples for SAP (or SAK) prediction model was 120, the statistical significance of R² were different for the models SOM (or SAN) and SAP (or SAK). The specific parameters for PLSR and equations for MLSR were provided in Table 5 in the revised manuscript as follows:

Table 5 The properties of the developed 2D-EMD_PM models for soil nutrients.

Soil nutrients	Procedure	LVs	Calibration accuracy			Validation accuracy
			R2	RMSE	RPD	R2	RMSE	RPD
SOM in coal mining area	BIMF1 (PLSR)	3	0.12*	5.65	1.05	0.13	5.45	1.09
	BIMF2 (PLSR)	22	0.95**	0.69	4.51	0.64**	2.02	1.35
	BIMF3 (PLSR)	23	0.99**	0.21	10.69	0.87**	0.96	2.71
	Residue (PLSR)	41	1.00**	0.00	1496.61	1.00**	0.06	18.04
	SOM (MLSR)	-	0.62**	5.08	1.64	0.64**	4.23	1.42
		-6.31 + 1.64 IMF2’ + 0.83 IMF3’ + 1.33 Residue’
SOM in non-coal mining area	BIMF1 (PLSR)	2	0.11	4.29	0.98	0.16	3.76	0.80
	BIMF2 (PLSR)	15	0.81**	1.20	2.29	0.39**	2.45	1.16
	BIMF3 (PLSR)	38	1.00**	0.05	47.47	0.78**	1.44	1.70
	Residue (PLSR)	35	1.00**	0.01	175.19	0.99**	0.15	9.48
	SOM (MLSR)	-	0.48**	3.95	1.38	0.57**	4.11	1.39
		2.39 + 0.95 IMF2’ + 0.87 IMF3’ + 0.87 Residue’
SAN in coal mining area	BIMF1 (PLSR)	1	0.09*	8.09	1.06	0.04	7.39	1.04
	BIMF2 (PLSR)	18	0.94**	4.02	1.10	0.65**	4.11	1.18
	BIMF3 (PLSR)	40	1.00**	0.01	657.15	0.98**	0.54	14.17
	Residue (PLSR)	42	1.00**	0.00	5327.92	0.99**	0.20	9.83
	SAN (MLSR)	-	0.70**	6.68	1.84	0.61**	7.30	1.60
		-3.09 + 0.44 IMF2’ + 1.36 IMF3’ + 2.69 Residue’
SAN in non-coal mining area	BIMF1 (PLSR)	1	0.03	9.64	1.02	0.41**	7.53	0.86
	BIMF2 (PLSR)	20	0.87**	1.55	2.81	0.35**	5.03	0.87
	BIMF3 (PLSR)	40	1.00**	0.00	1024.42	0.92**	3.70	0.96
	Residue (PLSR)	41	1.00**	0.00	1742.53	1.00**	0.27	13.76
	SAN (MLSR)	-	0.43**	9.52	1.34	0.40**	7.27	1.25
		-0.11 + 0.87 IMF2’ + 1.74 IMF3’ + 0.92 Residue’
SAP in the entire area	BIMF1 (PLSR)	1	0.02	4.01	1.02	0.02	6.55	0.30
	BIMF2 (PLSR)	52	0.95**	0.45	4.40	0.50**	2.55	1.27
	BIMF3 (PLSR)	59	0.99**	0.04	11.08	0.66**	0.28	1.60
	Residue (PLSR)	61	1.00**	0.01	166.18	0.99**	0.13	11.65
	SAP (MLSR)	-	0.23**	4.22	1.15	0.15*	3.43	1.05
		1.88 + 0.88 IMF2’ + 0.96 IMF3’ + 0.78 Residue’
SAK in the entire area	BIMF1 (PLSR)	52	0.69**	17.39	1.80	0.01	103.53	0.26
	BIMF2 (PLSR)	52	0.95**	3.66	4.61	0.56**	13.83	1.30
	BIMF3 (PLSR)	54	1.00**	0.31	25.56	0.98**	1.32	6.92
	Residue (PLSR)	52	1.00**	0.12	73.98	0.99**	0.77	11.75
	SAK (MLSR)	-	0.34**	31.40	1.23	0.20*	33.02	1.05
		-27.27 + 0.96 IMF2’ + 0.60 IMF3’ + 1.16 Residue’

LVs: Number of latent variables; IMF2’, IMF3’,and Residue’ were the predicted IMF2, IMF3, and Residue by PLSR; ^* significant at P < 0.05; ^** significant at P < 0.01.

 

From line 122 seems clear that there is no external validation. So in the discussions, please state that the dataset splitted in 75/25 doesn’t mean that the model could work as good as in the randomly selected validation.

Response: Thanks for your recommendation. We added the discussion part about the predicting performance on the randomly selected validated dataset only once in the study in the revised manuscript.

 

Line 16-> multiscale relationships

Response: We changed the “multiscale relationships” to “scale-specific relationships” in the abstract as follows:

However, the prediction performances for soil nutrients based on their scale-specific relationships with influencing factors remains undefined in the coal mining area. The objective of this study was to establish prediction models of soil nutrients based on their scale-specific relationships with influencing factors in a coal mining area.

 

Line 20 -> 2DEMD please change the acronym

Response: We changed the acronyms 2DEMD to 2D-EMD through the revised manuscript.

 

Line 69 -> established is probably odd, please change

Response: In our experiment, we sampled 120 soil points based on the 1 × 1 km regular grid in the study area, and the number of soil samples was 120.

 

Line 72-76 -> please specify better the methods used for SOC and Nitrogen content determination by citing some protocols that have been followed. It is clear what the authors have done but in order to replicate this analysis the reader would need additional information.

Response: We added the references about the protocols that have been followed in the revised manuscript of Lines 81-84 as follows:

The SOM was measured by the dichromate oxidation method [24], SAN was determined by the diffusion method [25], SAP was obtained by the Olsen extraction method using alkaline sodium bicarbonate as the extractant in a 20:1 ratio [26], and SAK was extracted by the flame photometer [27].

 

Line 77-93 -> the methods need some more details, period of acquisition, software used

Response: We added more details, period of acquisition, software used in the research in the revised manuscript.

 

Line 93 -> what the authors meant for “run off”?

Response: We are sorry for the misstatement, and changed “run off” to flowage in the revised manuscript.

 

Line 106 -> which selected environmental fartors? And why into 3 BIMs

Response: The method of PLSR has the advantage for many covariates, and the selected remote sensing index and topographic factors in the Material and Methods were used in the research. If the redundancy needed to be minimized for other predicting methods, the factors could be selected based on their correlation coefficients.

Three BIMFs, decomposed from soil samples at the sampling scale, could guaranteed each scale component contains at least 5% variance of the original data.

 

Line 118-> please show this ANOVA table and the significant differences, not only letters but p and F

Response: We added the P and F value in Table 1 in revised manuscript as follows:

Table 1. Basic statistics of soil nutrients (SOM, SAN, SAP, and SAK) and soil texture under coal mining and non-coal mining areas, respectively.

Soil nutrients	Areas	Min	Mean	Max	Std	CV	ANOVA
								F	P-value
SOM (g/kg)	Coal mining area	1.10	17.42	36.42	7.79	44.73	a	4.04	0.04
	Non-coal mining area	4.72	19.88	32.63	5.58	28.09	b
SAN (mg/kg)	Coal mining area	7.42	32.44	57.48	12.04	37.12	a	10.12	0.00
	Non-coal mining area	5.56	39.40	66.75	11.92	30.26	b
SAP (mg/kg)	Coal mining area	0.95	7.95	16.91	3.83	48.15	a	2.60	0.11
	Non-coal mining area	1.52	9.21	22.30	5.11	55.48	a
SAK (mg/kg)	Coal mining area	60.30	152.50	311.55	40.75	26.73	a	2.42	0.12
	Non-coal mining area	70.35	163.14	244.55	33.83	20.74	a
Sand (-)	Coal mining area	0.09	0.23	0.64	0.09	40.06	a	4.94	0.03
	Non-coal mining area	0.09	0.19	0.67	0.09	47.93	b
Silt (-)	Coal mining area	0.22	0.48	0.66	0.10	21.37	a	6.13	0.01
	Non-coal mining area	0.18	0.52	0.70	0.10	18.56	b
Clay (-)	Coal mining area	0.14	0.29	0.48	0.08	26.98	a	0.19	0.67
	Non-coal mining area	0.13	0.28	0.56	0.08	27.25	a

CV represents coefficient of variance. ANOVA represents analysis of variance, and different letters between coal mining area and non-coal mining area represent the significant difference of the soil nutrient at P < 0.05 level.

 

Though I trust the authors, I ask the editor to check for plagiarism, I cannot check on my own.

Response: Thanks for your comment, and the manuscript was revised by a native English speaker. The certification of language editing was attached at the end of the revision note.

Author Response File: Author Response.pdf

Reviewer 3 Report

The manuscript entitled: Prediction of Soil Nutrients Based on Topographic 2 Factors and Remote Sensing Index in a Coal Mining 3 Area, China presents the application of a novel technique in digital soil mapping, namely the 2-dimensional empirical mode decomposition which is compared to other classical digital soil mapping techniques. The authors should make more effort to present the methodology behind this procedure. I didn't see the maps achieved for soil nutrients by means of the four procedures described in the text. The authors take into account many environmental factors which are partially redundant. Has any effort been done to minimize this redundancy?

I am not a native English speaker but I believe the article should be English checked, since some phrases were not clear to me probably for this reason. 

More comments are in the revised manuscript.

Comments for author File: Comments.pdf

Author Response

The manuscript entitled: Prediction of Soil Nutrients Based on Topographic 2 Factors and Remote Sensing Index in a Coal Mining 3 Area, China presents the application of a novel technique in digital soil mapping, namely the 2-dimensional empirical mode decomposition which is compared to other classical digital soil mapping techniques. The authors should make more effort to present the methodology behind this procedure.

Response: Thanks for your comment. We presented the detailed procedure of 2D-EMD, and modified the description about the prediction method of 2D-EMD_PM in the revised manuscript.

 

I didn't see the maps achieved for soil nutrients by means of the four procedures described in the text.

Response: Figure 6 was the predicted maps by 2D-EMD_PM, which is the combined method of the 2D-EMD, PLSR, and MLSR. The predicting accuracy of MLSR_Ori, PLSR_Ori, and PLSR_BIMF were presented in Table 4, but their predicting maps were not presented in the manuscript because of the limited space.

 

The authors take into account many environmental factors which are partially redundant. Has any effort been done to minimize this redundancy?

Response: Although there are many environmental factors in the study, these factors can be easily acquired from the vegetation index of MODIS products and DEM. Meanwhile, the method of PLSR could demonstrate its advantage over too many environmental factors.

If the redundancy needed to be minimized for other predicting methods, the factors could be selected based on their correlation coefficients.

 

I am not a native English speaker but I believe the article should be English checked, since some phrases were not clear to me probably for this reason

Response: Thank you for your comment, and the manuscript was revised by a native English speaker. The certification of language editing was attached at the end of the revision note.

 

More comments are in the revised manuscript.

1) Introduction. (Line 31): The authors should make a more thorough analysis on the research context of their study. There is a lot of digital soil mapping literature not cited here.

Response: We added some content about soil prediction and related literatures in the revised manuscript (Lines 38-54) as follows:

A number of methods, including geostatistical models [1], inverse distance weighted [2], trend surface analysis [3] and the like, resolve the spatial distribution of soil properties based on the geospatial autocorrelation [4, 5]. However, these methods are limited by sampling density and the methods either failure to consider the environmental effects on soil nutrients or consider the environmental effects only at the sampling scale, causing their heterogeneous distribution at different scale in different intensity [6, 7]. Previously studies indicate that the qualitative and quantitative information of vegetation growth or land surface temperatures can be obtained from the electromagnetic radiation of remote sensing, which indirectly demonstrating the distributions of soil nutrients [8-11]. Meanwhile, topographic factors affect the transposition and redistribution of soil nutrient contents [12], and their effects on soil nutrients vary in different scales [13, 14]. A number of prediction methods for soil properties based on their relationships with topographic factors are proposed [15], including ensemble learning model [16], extended models [17], regression model [18] and so on. Nevertheless, these methods do not consider their multiscale relationships between soil nutrients and the covariates. Additionally, scale-specific correlations between soil properties and environmental factors are explored using geostatistical approach [19], multivariate empirical mode decomposition [20, 21], wavelet transform [13] and so on. However, these methods either fail to realize the prediction of soil properties or establish the validation model.

 

2) Line 39: I disagree. Geostatistical methods such as regression-kriging include covariables.

Response: Thanks for your comment and we are sorry for the misstatement. we corrected it in the revised manuscript (Lines 40-43) as follows:

However, these methods are limited by sampling density and the methods either failure to consider the environmental effects on soil nutrients or consider the environmental effects only at the sampling scale, causing their heterogeneous distribution at different scale in different intensity [6, 7].

 

3) Line 63: you have a maximum of 1163 m of the map.

Response: We are sorry for the misstatement. The elevation ranges from 726 to 1069 m in the coal mining area and from 739 to 979 m in the non-coal mining area in the original manuscript were the elevation ranges of the sampling points. We modified the ranges in Line 72 as follows：

The range of elevation is from 718 to 1163 m in the coal mining area, while it is from 711 to 1040 m in the non-coal mining area.

 

4) Figure 1: please provide a scale for figure 1b.

Response: Thanks for your comment, and the scale for figure 1b was added in the revised manuscript as follows:

 

5) Line 78: please tell us about the resolution and acquisition dates of MODIS images.

Response: The resolution and acquisition dates of MODIS images was provided in the revised manuscript in Line 90-91 as follows:

The products of Moderate Resolution Imaging Spectroradiometer (MODIS), including gross primary production (GPP, MOD17A2), net primary production (NPP, MOD17A3), land surface temperature for daytime and night (DLST and NLST, MOD11A2), evapotranspiration (ET, MOD16), normalized difference vegetation index (NDVI, MOD13Q1) with 1 km resolution (the same with soil sampling scale of 1 km) in 2015 was downloaded from https://ladsweb.modaps.eosdis.nasa.gov/.

 

6) Line 89: Please tell us the resolution of the DEM and its source.

Response: The resolution of the DEM and its source was added in the revised manuscript in Lines 102-103 as follows:

Digital elevation model (DEM) with 30 m from http://gdem.ersdac.jspacesystems.or.jp/ was downloaded.

 

7) Line 104: The authors should provide a more detailed description of this method, especially since it is quite new in digital soil mapping.

Response: Thanks for your comment, and we added a schematic presentation of 2D-EMD step in Figure 2 as follows:

The main process of 2D-EMD can be followed as Figure 2. More detailed description of 2D-EMD can be found in other studies [30, 31].

 

Figure 2. Schematic presentation of the step of 2D-EMD

 

8) Line 127-129: this combined method is not clear to me. The authors should present it more clearly.

Response: The method includes three steps. Firstly, soil nutrients and the environmental factors were decomposed into BIMFs and residue using 2D-EMD. Secondly, soil nutrients at each BIMF (or residue) were predicted from the environmental factors at the corresponding scale components using PLSR. Thirdly, soil nutrients at the sampling scale were estimated from the predicted soil nutrients at the BIMFs and residue using MLSR. We modified the express in the revised manuscript (Lines 145-149) as follows:

For the method 2D-EMD_PM, specifically, soil nutrients and the environmental factors were decomposed into BIMFs and residue using 2D-EMD; soil nutrients at each BIMF (or residue) were predicted from the environmental factors at the corresponding BIMF (or residue) using PLSR; soil nutrients at the sampling scale were estimated from the predicted soil nutrients at the BIMFs and residue using MLSR.

 

9) Line 146: please explain these differences between coal and non-coal mining areas.

Response: We added the explanation in the revised manuscript (Line 168-169) as follows:

The greater CVs in the coal mining area might be attributed to mining practices in the area.

 

10) Line 165 (Figure 2.): what does the sampling scale represent? It is BIMF or IMF? You should use only one abbreviation

Response: The sampling scale was 1 km. It should be BIMF, and we modified IMF to BIMF throughout the revised manuscript.

 

11) Table 2: Can the sum of these value be over 100%?

Response: Yes, the sum of variance explained by each BIMF or residue can be over 100% or less than 100%. Because the denominator was the variance of original data at the sampling scale, rather than the sum of each BIMF (or residue) variance. The closer of the sum to 100%, the better 2D-EMD performed.

 

12) Table 3: why MLSR perform so poor when it comes to the validation sample? can you explain?

Response: The result indicated the established MLSR models from the calibrating points were not universal in the study area. The result might be attributed to the overfitting occurred in the calibration model. This might demonstrate the phenomenon that MLSR has defect for the prediction from too many environmental factors, and PLSR could remedy the disadvantage.

 

13) Table 4: how can there be a R2 of 1? what is the meaning of this abbreviation (SMLR)?

Response: For soil nutrients at smaller scales (i.e., IMF1), predicting performance was not good. However, for soil nutrients at larger scales (i.e., residue), predicting performance was very good, and could achieve the R² of 1 (Hu and Si, 2013; She et al., 2015; Zhu et al., 2018a).

We are sorry for the mistake of SMLR, and replaced it by MLSR in the table 5 of the revised manuscript.

 

14) Figure 5: where are the maps of soil nutrients produced the four methods?

Response: We only compared the predicting accuracy of the four methods and presented the map predicted by 2D-EMD_PM because of the limited space.

 

15) Line 253-255: I don't understand this explanation.

Response: The homogenously spatial distribution of SAP (or SAK) between coal mining area and non-coal mining area might be attributed to the horizontal flow of soil nutrients.

 

16) Line 258-267: not clear. please reformulate.

Response: We reformulated the expression in the revised manuscript (Lines 296-307) as follows:

The correlations between soil nutrients and the influencing factors were generally non-significant at the scale component of BIMF1, which might be attributed to the stochastic effects (i.e., the effect of farming practices) resulted in the randomly distribution of SOM at small scale of BIMF1. The spatial distributions of BIMF1 varied greatly (Figure 3) and the worst predicting perform at BIMF1 (Table 5) demonstrated the stochastic variation at the small scale of BIMF1. In addition, previous study [29] also proves the stochastic variation of soil nutrients at the small scale of BIMF1. Consequently, predicting models excluded BIMF1 could improve their stability of predicting performance. On the other hand, the total percentage of variance explained by BIMF2, BIMF3, and residue was even less than the percentage of variance explained by BIMF1, which revealed the less effect of the structured variance and the greater effect of stochastic variance affected by human activities. Therefore, the predicting performance based on the structured variance of soil nutrients would not be better even excluded the stochastic variance (BIMF1) in the study area.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Dear Authors

 

Many things from the first revision were amended, but some of them not.

Please see again the past revision and try to amend the things that hasn't been done

look careful at the suggested literature ad references therein

 

Kind regards

 

 

, I have reviewed the manuscript entitled: Prediction of Soil Nutrients Based on Topographic  Factors and Remote Sensing Index in a Coal Mining Area, China.

The paper is on average length and it is correctly divided in sections.

There are some acronyms like 2DEMD that is not what I usually treat in my work DEM is generally Digital elevation model. I Think the acronym should be changed to avoid confusion with it.

At a certain point I discovered that the methodological approach is based on a previous Geoderma paper of the same authors. It is fine. But is important to understand which is the motivation of the new one and which improvements has the new proposed approach.

Very few detailed are reported for the remote sensing indices, therefore in my opinion, the reader needs to be addressed through the choice of which index is better for this purpose, and how to derive it form raw data, or download as ready product from a provider.

The same is for the topographic indices calculations. Please add short descriptions for their derivation, from which DEM and which is the pros and cos of using some of them for the prediction of topsoil organic carbon and nitrogen. Please have a look at Schillaci et al 2015, terrain analysis and landform recognition book chapter for the topographic predictors. And Schillaci et al 2017, modelling SOC stock for the remote sensing indices derivation.

Simulation methods represent indeed an interesting subject, but I think the Authors can briefly summarize with a couple of sentences the state of the art of the research carried out so far about this approach for soil carbon and soil properties. For the long term research, I suggest to read something about systematic mapping of the literature search for agro-environmental topic.

The introduction is well written, but it lacks of a systematic state of the art description, (e.g. Scopus or Web of science search).

Methods are clear, but since there is not a strong proposed evidence that the method will bring so much innovation in the field of soil mapping, the latest development in the discipline has to be briefly stated.

Partitioning strategy probably depends on spatial heterogeneity, so the SOC could be more data specific and then it performs better, please try the two methods in an external dataset to assess whether SOC is really most reliable than k-means.

Table 3, how was the statistical significance of R² calculated?

Model equations are missed, it will be interesting for the sake of reproducibility.

From line 122 seems clear that there is no external validation. So in the discussions, please state that the dataset splitted in 75/25 doesn’t mean that the model could work as good as in the randomly selected validation.

Line 16-> multiscale relationships

Line 20 -> 2DEMD please change the acronym

Line 69 -> established is probably odd, please change

Line 72-76 -> please specify better the methods used for SOC and Nitrogen content determination by citing some protocols that have been followed. It is clear what the authors have done but in order to replicate this analysis the reader would need additional information.

Line 77-93 -> the methods need some more details, period of acquisition, software used

Line 93 -> what the authors meant for “run off”?

Line 106 -> which selected environmental fartors? And why into 3 BIMs

Line 118-> please show this ANOVA table and the significant differences, not only letters but p and F

 

 

I believe that the manuscript needs a major revision.

Though I trust the authors, I ask the editor to check for plagiarism, I cannot check on my own.

Therefore, I wish you best of luck.

Kind regards,

Author Response

Response: Thanks for your recommendation. We studied the references (Schillaci et al 2015 for topographic predictors; Schillaci et al 2017 for soc mapping based on remote sensing index and topographic factors), and cited them in the introduction and method parts as follows:

1. Introduction

The qualitative and quantitative information of vegetation growth or land surface temperatures can be obtained from the electromagnetic radiation of remote sensing (RS), which indirectly demonstrating the distributions of soil nutrients [8-11]. Schillaci eta al. indicated that the application of RS index could enhance the predicting performance of soil properties [12]. Meanwhile, topographic factors affect the transposition and redistribution of soil nutrient contents [13], and their effects on soil nutrients vary in different scales [14, 15].

2.2. Collection of environmental factors

The products of Moderate Resolution Imaging Spectroradiometer (MODIS), including gross primary production (GPP, MOD17A2), net primary production (NPP, MOD17A3), land surface temperature for daytime and night (DLST and NLST, MOD11A2), evapotranspiration (ET, MOD16), normalized difference vegetation index (NDVI, MOD13Q1) with 1 km resolution (the same with soil sampling scale of 1 km) in 2015 was downloaded (https://ladsweb.modaps.eosdis.nasa.gov/). The DLST, NLST, ET and NDVI were used for extracting RS covariates in ENVI 5.3, including the intra-annual variance of the land surface temperature for daytime (DLSTV), the intra-annual variance of the land surface temperature for night (NLSTV), the annual mean evapotranspiration (ETM), the intra-annual standard deviation of evapotranspiration (ETD), the intra-annual variance of evapotranspiration (ETV), the annual mean NDVI (NDVIM), the inter-annual standard deviation of NDVI (NDVID) and the inter-annual variance of NDVI (NDVIV). The index of GPP, NPP, NDVIM, NDVID and NDVIV indicated vegetation information and was selected, because vegetation was the key source of the SOM. The index of DLSTV, NLSTV, ETM, ETD and ETV indicated soil temperature and moisture and was selected, because they were associated with the decomposition rates of SOM.

Digital elevation model (DEM) with 30 m (http://gdem.ersdac.jspacesystems.or.jp/) was downloaded. The DEM was used for extracting topographic factors, including elevation, aspect, slope, convergence index (CI), Channel Network Base Level (CNBL), flow accumulation (FA), LS Factor (LS), topographic wetness index (TWI), valley depth (VD), and vertical distance to channel network (VDCN) in the software of SAGA 6.4 [30], and was resampled into the same resolution with RS of 1 km in ArcGIS 10.5. The factors indicated topographic conditions which were related to microclimate and soil erosion, and thus affected the accumulation, decomposition and flowage of soil nutrients [12].

 

We also changed the conclusion part as follows:

The proposed method was 2D-EMD_PM, which decomposed the soil nutrients and their covariates into 4 scale components, then predicted soil nutrients at each scale components from the corresponding scale of the covariates, and finally predicted soil nutrients at the sampling scale from the predicted soil nutrients at each scale except BIMF1. The method was stable in the calibration and validation models, and should be validated in the other coal mining areas in the world.

Author Response File: Author Response.docx