Submit to this Journal Review for this Journal Propose a Special Issue

Article Menu

Share Help Cite Discuss in SciProfiles

Open AccessArticle

Peer-Review Record

Geospatial Evaluation of Cropping Pattern and Cropping Intensity Using Multi Temporal Harmonized Product of Sentinel-2 Dataset on Google Earth Engine

Appl. Sci. 2022, 12(24), 12583; https://doi.org/10.3390/app122412583

by Sonia¹, Tathagata Ghosh^1,*

, Amel Gacem²

, Taghreed Alsufyani³, M. Mujahid Alam⁴

, Krishna Kumar Yadav⁵

, Mohammed Amanullah⁶

and Marina M. S. Cabral-Pinto^7,*

Reviewer 1:

Giovanni Romano

Reviewer 2:

Riccardo Orusa

Appl. Sci. 2022, 12(24), 12583; https://doi.org/10.3390/app122412583

Submission received: 20 November 2022 / Revised: 3 December 2022 / Accepted: 5 December 2022 / Published: 8 December 2022

(This article belongs to the Special Issue Recent Advances in Geospatial Big Data Mining)

Round 1

Reviewer 1 Report

The paper deals with the depiction of cropping pattern and cropping intensity in an arid plain zone of India. Kharif and rabu crop distributions were mapped using Sentinel-2 dataset on Google Earth Engine (GEE) computing platform, while crop intensity map was performed on GIS environment. The analysis attempts were supported by ground truth verification and accuracy assessment.

The Manuscript is well structured and clearly presented. The methods used are suitable for the study aims. The authors could improve the description of the procedures used in the Materials and Methods section. I suggest giving some detail about GEE scripts used.

Furthermore, I only have some minor suggestions, which can help the authors to further improve the Manuscript quality:

Throughout the manuscript, replace:

- google earth engine with Google Earth Engine;

- random forest with Random Forest;

- Sentinel 2 or sentinel 2 with Sentinel-2;

Write the words kharif and rabi with their appropriate initial capital or small letter.

1. Row 92: replace Sentinel 1 and sentinel 2 with Sentinel-1 and Sentinel-2;

2. Row 100: is the sentence “With a root-mean-square error (RMSE) of 3.5 million hectare” a misprint?

3. Row 155: the section Algorithm execution is empty;

4. Row 165-166: why you chose, for the filtered dataset the combined constellation revised frequency instead of the frequency of each single Sentinel-2 satellite?

5. Row 173: replace khrif with kharif;

6. Figure 11: it could be unnecessary since Confusion matrix is well described in Table 2.

Author Response

Point 1: The authors could improve the description of the procedures used in the Materials and Methods section.

Response 1: Procedure used in Materials and method section is thoroughly revised. A detailed workflow diagram is also incorporated (Figure 2). A generalized algorithm of the Google Earth Engine coding is incorporated (Figure 3)

Point 2: I suggest giving some detail about GEE scripts used.

Response 2:

The GEE script used is as follows:

//For Kharif crop:

function maskS2clouds(image) {
var qa = image.select('QA60');
var cloudBitMask = 1 << 10;
var cirrusBitMask = 1 << 11;
var mask = qa.bitwiseAnd(cloudBitMask).eq(0)
.and(qa.bitwiseAnd(cirrusBitMask).eq(0));
return image.updateMask(mask).divide(10000);
}
var dt = table.filter(ee.Filter.eq('DISTNAME','CHURU'));

var coll1 = ee.ImageCollection("COPERNICUS/S2_SR_HARMONIZED")

.filterDate('2020-04-01','2020-04-15')

.filter(ee.Filter.lt('CLOUDY_PIXEL_PERCENTAGE',10))

.map(maskS2clouds);

.filterBounds(dt);

var coll2 = ee.ImageCollection("COPERNICUS/S2_SR_HARMONIZED")

.filterDate('2020-04-16','2020-04-30')

.filter(ee.Filter.lt('CLOUDY_PIXEL_PERCENTAGE',10))

.map(maskS2clouds);

.filterBounds(dt);

var coll3 = ee.ImageCollection("COPERNICUS/S2_SR_HARMONIZED")

.filterDate('2020-05-01','2020-05-15')

.filter(ee.Filter.lt('CLOUDY_PIXEL_PERCENTAGE',10))

.map(maskS2clouds);

.filterBounds(dt);

var coll4 = ee.ImageCollection("COPERNICUS/S2_SR_HARMONIZED")

.filterDate('2020-05-16','2020-05-31')

.filter(ee.Filter.lt('CLOUDY_PIXEL_PERCENTAGE',10))

.map(maskS2clouds);

.filterBounds(dt);

var coll5 = ee.ImageCollection("COPERNICUS/S2_SR_HARMONIZED")

.filterDate('2020-06-01','2020-06-15')

.filter(ee.Filter.lt('CLOUDY_PIXEL_PERCENTAGE',10))

.map(maskS2clouds);

.filterBounds(dt);

var coll6 = ee.ImageCollection("COPERNICUS/S2_SR_HARMONIZED")

.filterDate('2020-06-16','2020-06-30')

.filter(ee.Filter.lt('CLOUDY_PIXEL_PERCENTAGE',10))

.map(maskS2clouds);

.filterBounds(dt);

var coll7 = ee.ImageCollection("COPERNICUS/S2_SR_HARMONIZED")

.filterDate('2020-07-01','2020-07-15')

.filter(ee.Filter.lt('CLOUDY_PIXEL_PERCENTAGE',10))

.map(maskS2clouds);

.filterBounds(dt);

var coll8 = ee.ImageCollection("COPERNICUS/S2_SR_HARMONIZED")

.filterDate('2020-07-16','2020-07-31')

.filter(ee.Filter.lt('CLOUDY_PIXEL_PERCENTAGE',10))

.map(maskS2clouds);

.filterBounds(dt);

var coll9 = ee.ImageCollection("COPERNICUS/S2_SR_HARMONIZED")

.filterDate('2020-08-01','2020-08-15')

.filter(ee.Filter.lt('CLOUDY_PIXEL_PERCENTAGE',10))

.map(maskS2clouds);

.filterBounds(dt);

var coll10 = ee.ImageCollection("COPERNICUS/S2_SR_HARMONIZED")

.filterDate('2020-08-16','2020-08-31')

.filter(ee.Filter.lt('CLOUDY_PIXEL_PERCENTAGE',10))

.map(maskS2clouds);

.filterBounds(dt);

var coll11 = ee.ImageCollection("COPERNICUS/S2_SR_HARMONIZED")

.filterDate('2020-09-01','2020-09-15')

.filter(ee.Filter.lt('CLOUDY_PIXEL_PERCENTAGE',10))

.map(maskS2clouds);

.filterBounds(dt);

var coll12 = ee.ImageCollection("COPERNICUS/S2_SR_HARMONIZED")

.filterDate('2020-09-16','2020-09-30')

.filter(ee.Filter.lt('CLOUDY_PIXEL_PERCENTAGE',10))

.map(maskS2clouds);

.filterBounds(dt);

print(coll1);

var addNDVI = function(img){

var ndvi = img.normalizedDifference(['B8','B4']).rename('NDVI')

return img.addBands(ndvi);

}

var withNDVI1 = coll1.map(addNDVI);

var withNDVI2 = coll2.map(addNDVI);

var withNDVI3 = coll3.map(addNDVI);

var withNDVI4 = coll4.map(addNDVI);

var withNDVI5 = coll5.map(addNDVI);

var withNDVI6 = coll6.map(addNDVI);

var withNDVI7 = coll7.map(addNDVI);

var withNDVI8 = coll8.map(addNDVI);

var withNDVI9 = coll9.map(addNDVI);

var withNDVI10 = coll10.map(addNDVI);

var withNDVI11 = coll11.map(addNDVI);

var withNDVI12 = coll12.map(addNDVI);

var greenest1 = withNDVI1.qualityMosaic('NDVI').select(['B2','B3','B4','B8','B11','B12']).clip(dt);

var greenest2 = withNDVI2.qualityMosaic('NDVI').select(['B2','B3','B4','B8','B11','B12']).clip(dt);

var greenest3 = withNDVI3.qualityMosaic('NDVI').select(['B2','B3','B4','B8','B11','B12']).clip(dt);

var greenest4 = withNDVI4.qualityMosaic('NDVI').select(['B2','B3','B4','B8','B11','B12']).clip(dt);

var greenest5 = withNDVI5.qualityMosaic('NDVI').select(['B2','B3','B4','B8','B11','B12']).clip(dt);

var greenest6 = withNDVI6.qualityMosaic('NDVI').select(['B2','B3','B4','B8','B11','B12']).clip(dt);

var greenest7 = withNDVI7.qualityMosaic('NDVI').select(['B2','B3','B4','B8','B11','B12']).clip(dt);

var greenest8 = withNDVI8.qualityMosaic('NDVI').select(['B2','B3','B4','B8','B11','B12']).clip(dt);

var greenest9 = withNDVI9.qualityMosaic('NDVI').select(['B2','B3','B4','B8','B11','B12']).clip(dt);

var greenest10 = withNDVI10.qualityMosaic('NDVI').select(['B2','B3','B4','B8','B11','B12']).clip(dt);

var greenest11 = withNDVI11.qualityMosaic('NDVI').select(['B2','B3','B4','B8','B11','B12']).clip(dt);

var greenest12 = withNDVI12.qualityMosaic('NDVI').select(['B2','B3','B4','B8','B11','B12']).clip(dt);

var ndvi1 = greenest1.normalizedDifference(['B8','B4']).rename('NDVI1');

var ndvi2 = greenest2.normalizedDifference(['B8','B4']).rename('NDVI2');

var ndvi3 = greenest3.normalizedDifference(['B8','B4']).rename('NDVI3');

var ndvi4 = greenest4.normalizedDifference(['B8','B4']).rename('NDVI4');

var ndvi5 = greenest5.normalizedDifference(['B8','B4']).rename('NDVI5');

var ndvi6 = greenest6.normalizedDifference(['B8','B4']).rename('NDVI6');

var ndvi7 = greenest7.normalizedDifference(['B8','B4']).rename('NDVI7');

var ndvi8 = greenest8.normalizedDifference(['B8','B4']).rename('NDVI8');

var ndvi9 = greenest9.normalizedDifference(['B8','B4']).rename('NDVI9');

var ndvi10 = greenest10.normalizedDifference(['B8','B4']).rename('NDVI10');

var ndvi11 = greenest11.normalizedDifference(['B8','B4']).rename('NDVI11');

var ndvi12 = greenest12.normalizedDifference(['B8','B4']).rename('NDVI12');

var stacked_ndvi = ndvi1.addBands([ndvi2,ndvi3,ndvi4,ndvi5,ndvi6,ndvi7,ndvi8,ndvi9,ndvi10,ndvi11,ndvi12]);

var stacked_ndvi_scaled = stacked_ndvi.multiply(100).add(100).uint8();

Map.addLayer(stacked_ndvi_scaled,{bands:['NDVI9','NDVI5','NDVI7'],min:100, max:200},'Stacked_NDVI');

Map.addLayer(greenest9,{bands:['B8','B4','B3'],min:0,max:5000},'Sen_Feb1FN');

print(ndvi1)

var gt = Crop.merge(other);

var training_data = stacked_ndvi_scaled.sampleRegions({

collection : gt,

properties:['class'],

scale:10,

});

var classifier = ee.Classifier.smileRandomForest(100)

.train({

features:training_data,

classProperty: 'class',

});

var classified = stacked_ndvi_scaled.classify(classifier, 'Classified');

var classified1 = classified.updateMask(classified.eq(1))

Map.addLayer(classified1,{palette:['white','yellow'],min:0,max:1},'Classified');

var areaImage = ee.Image.pixelArea().addBands(classified1);

var areas = areaImage.reduceRegion({

reducer: ee.Reducer.sum().group({

groupField: 1,

groupName: 'class',

}),

geometry: dt,

scale: 10,

maxPixels: 1e13,

});

print('##### CLASS AREA SQ. METERS #####');

print(areas);

var trainAccuracy = classifier.confusionMatrix();

print('Resubstitution error matrix: ', trainAccuracy);

print('Training overall accuracy: ', trainAccuracy.accuracy());

Export.image.toDrive({

image: classified1,

description: 'Kharif_20221-22',

scale: 30,

folder:'2021-22',

region: dt

});

Export.image.toDrive({

image: ndvi1,

description: 'Kharif_20221-22NDVI',

scale: 30,

folder:'2021-22',

region: dt

});

//For Rabi crop:

var dt = table.filter(ee.Filter.eq('DISTNAME','CHURU'));

var coll1 = ee.ImageCollection("COPERNICUS/S2_SR_HARMONIZED")

.filterDate('2020-10-01','2020-10-15')

.filter(ee.Filter.lt('CLOUDY_PIXEL_PERCENTAGE',10))

.map(maskS2clouds);

.filterBounds(dt);

var coll2 = ee.ImageCollection("COPERNICUS/S2_SR_HARMONIZED")

.filterDate('2020-10-16','2020-10-31')

.filter(ee.Filter.lt('CLOUDY_PIXEL_PERCENTAGE',10))

.map(maskS2clouds);

.filterBounds(dt);

var coll3 = ee.ImageCollection("COPERNICUS/S2_SR_HARMONIZED")

.filterDate('2020-11-01','2020-11-15')

.filter(ee.Filter.lt('CLOUDY_PIXEL_PERCENTAGE',10))

.map(maskS2clouds);

.filterBounds(dt);

var coll4 = ee.ImageCollection("COPERNICUS/S2_SR_HARMONIZED")

.filterDate('2020-11-16','2020-11-30')

.filter(ee.Filter.lt('CLOUDY_PIXEL_PERCENTAGE',10))

.map(maskS2clouds);

.filterBounds(dt);

var coll5 = ee.ImageCollection("COPERNICUS/S2_SR_HARMONIZED")

.filterDate('2020-12-01','2020-12-15')

.filter(ee.Filter.lt('CLOUDY_PIXEL_PERCENTAGE',10))

.map(maskS2clouds);

.filterBounds(dt);

var coll6 = ee.ImageCollection("COPERNICUS/S2_SR_HARMONIZED")

.filterDate('2020-12-16','2020-12-31')

.filter(ee.Filter.lt('CLOUDY_PIXEL_PERCENTAGE',10))

.map(maskS2clouds);

.filterBounds(dt);

var coll7 = ee.ImageCollection("COPERNICUS/S2_SR_HARMONIZED")

.filterDate('2021-01-01','2021-01-15')

.filter(ee.Filter.lt('CLOUDY_PIXEL_PERCENTAGE',10))

.map(maskS2clouds);

.filterBounds(dt);

var coll8 = ee.ImageCollection("COPERNICUS/S2_SR_HARMONIZED")

.filterDate('2021-01-16','2021-01-31')

.filter(ee.Filter.lt('CLOUDY_PIXEL_PERCENTAGE',10))

.map(maskS2clouds);

.filterBounds(dt);

var coll9 = ee.ImageCollection("COPERNICUS/S2_SR_HARMONIZED")

.filterDate('2021-02-01','2021-02-15')

.filter(ee.Filter.lt('CLOUDY_PIXEL_PERCENTAGE',10))

.map(maskS2clouds);

.filterBounds(dt);

var coll10 = ee.ImageCollection("COPERNICUS/S2_SR_HARMONIZED")

.filterDate('2021-02-16','2021-02-28')

.filter(ee.Filter.lt('CLOUDY_PIXEL_PERCENTAGE',10))

.map(maskS2clouds);

.filterBounds(dt);

var coll11 = ee.ImageCollection("COPERNICUS/S2_SR_HARMONIZED")

.filterDate('2021-03-01','2021-03-15')

.filter(ee.Filter.lt('CLOUDY_PIXEL_PERCENTAGE',10))

.map(maskS2clouds);

.filterBounds(dt);

var coll12 = ee.ImageCollection("COPERNICUS/S2_SR_HARMONIZED")

.filterDate('2021-03-16','2021-03-31')