Google Earth Engine

Britta Ricker, PhD

2023

(Press space bar to advance slides in order)

Google Earth Engine
Forest Data
Call salb boundaries
Calculate area

Final Result will look something like this

Legend here

Outline

What is Google Earth Engine?
User Interface
Basics of JavaScript
Browse and add Data
- Sentinel (EU), LandSat (USA)
- Land use, Forest cover, Urban, Elevation
- Other?
Clip to your country
Change Colors
Basic calculate area of land use
Export data to use in QGIS
make a legend (on your own)

What is Google Earth Engine?

Computing platform and runs in your internet browser
- with a code editor,
- offers a compiler that runs in the Google cloud,
- hosts data on Google servers.
Integrated Development Environment (IDE)
Read more here

want more tutorials? https://tutorials.geemap.org/Image/image_visualization/

advanced https://courses.spatialthoughts.com/end-to-end-gee.html

Google Earth Engine is a

Data Repository (data as a service)

Software as a Service (SaaS)

Platform as a service (PaaS)

User Interface

{JavaScript}

is an interactive front-end scripting language

Statements - individual instruction or steps

statements end with a ;

Comments - explains what code does

// Displays a map

JavaScript is case sensitive

Variable - tell the system to remember or store data or a value

First Declare a variable

Then assign a value to a variable

Adding Data in Google Earth Engine

data types

ee.Image

ee.ImageCollection

ee.FeatureCollection

Image - is just that - bands

Image Collection - only images but lots of them

Feature- features and elements - rasters and or polygons

Basic of JavaScript

for GEE

Feel free to dig deeper on your own here

https://developers.google.com/earth-engine/tutorials/tutorials

click this link

Global Forest Change

scroll down at the bottom of the page in blue click "Open in Code Editor"

when the code editor opens

then hit run

in the map pane

when you hover over layers - toggle on and off the layers to make them visible

// Load a country border to a region of interest (roi).These are recognized boardings of the USA
var countries = ee.FeatureCollection('USDOS/LSIB_SIMPLE/2017');
var roi = countries.filterMetadata('country_na', 'equals', 'Spain');

//center the map on the coutnry identified by the ROI

Map.centerObject(roi);

# PRESENTING CODE

Now add the following lines of code to the top of the demo code to center the map on your country - change the country name "Spain" to your country

Forest Cover Example - Full script including Clip data to your country - and code comments

# PRESENTING CODE

// Load a country border to a region of interest (roi).These are recognized boardings of the USA

var countries = ee.FeatureCollection('USDOS/LSIB_SIMPLE/2017');
var roi = countries.filterMetadata('country_na', 'equals', 'Ethiopia');

//center the map on the coutnry identified by the ROI

Map.centerObject(roi);

var dataset = ee.Image('UMD/hansen/global_forest_change_2022_v1_10');

//select your country of interest by country
var datasetClip = dataset.clip(roi);

//style the clipped tree cover black no trees and green to trees. If you want global coverage call the dataset instead of clipped variable

var treeCoverVisParam = {
  bands: ['treecover2000'],
  min: 0,
  max: 100,
  palette: ['black', 'green']
};
Map.addLayer(datasetClip, treeCoverVisParam, 'tree cover');

//Style the clipped tree loss layer from yellow to red
var treeLossVisParam = {
  bands: ['lossyear'],
  min: 0,
  max: 22,
  palette: ['yellow', 'red']
};

//only add clipped forest loss area
Map.addLayer(datasetClip, treeLossVisParam, 'tree loss year');

hit run

toggle on and off the layers
you can change the transparency of the layers with the slider to the right of the layer name
scroll around your country
What do you notice?

Ok we added forest cover and loss

Now let's add

Elevation
Urban
Land Cover
Rivers
What else?

not always easy to get everything to run together...

For this next part - let's start with a clean code editor

next to the "reset" button click the down arrow and select clear script.

# CHAPTER 2

About MODIS

Moderate Resolution Imaging Spectroradiometer (MODIS) Land Cover Type (MCD12Q1) Version 6.1 data product provides global land cover types at yearly intervals.

500 M resolution global

2001 to present

Let's add Modis Land Cover Data for the entire world for the year 2019 - is there another year you would like to search?

//Call Modis land cover dataset
var modisLandcover = ee.ImageCollection("MODIS/006/MCD12Q1")
//filter by date
var filtered = modisLandcover.filter(
  ee.Filter.date('2019-01-01', '2019-12-31'))

//landcover for the year 2019
var landcover2019 = ee.Image(filtered.first())
//classify data based on land class type 1 
var classified = landcover2019.select('LC_Type1')
 
 //color palette for the classification - this matches the Modus official color pallet but feel free to play here
var palette = ['05450a', '086a10', '54a708',
 '78d203', '009900', 'c6b044','dcd159', 
 'dade48', 'fbff13', 'b6ff05', '27ff87',
 'c24f44', 'a5a5a5', 'ff6d4c', '69fff8',
 'f9ffa4', '1c0dff']

//add the visualized modis data to the map and label name landcover 2019 the entire world
Map.addLayer(classified,
{min:1, max:17, palette: palette},
'Land cover entire world 2019');

# CHAPTER 2

Save your work

click Save in the code editor.

pick a name you will remember

Code will be saved to your google account

Highlighting Japan

// Load a country border to a region of interest (roi).These are recognized boardings of the USA

var countries = ee.FeatureCollection('USDOS/LSIB_SIMPLE/2017');
var roi = countries.filterMetadata('country_na', 'equals', 'Japan');

//center the map on the coutnry identified by the ROI

Map.centerObject(roi);

var dataset = ee.Image('UMD/hansen/global_forest_change_2022_v1_10');

//select your country of interest by country
var datasetClip = dataset.clip(roi);

//Style the clipped tree loss layer from yellow to red
var treeLossVisParam = {
  bands: ['lossyear'],
  min: 0,
  max: 22,
  palette: ['yellow', 'red']
};
//style the clipped tree cover black no trees and green to trees. If you want global coverage call the dataset instead of clipped variable

var treeCoverVisParam = {
  bands: ['treecover2000'],
  min: 0,
  max: 100,
  palette: ['#edf8fb',  '#b2e2e2', '#66c2a4','#2ca25f','#006d2c']
};
Map.addLayer(datasetClip, treeCoverVisParam, 'tree cover');

//only add clipped forest loss area
Map.addLayer(datasetClip, treeLossVisParam, 'tree loss year');

# PRESENTING CODE

Highlighting Japan - color pallet

// Load a country border to a region of interest (roi).These are recognized boardings of the USA

var countries = ee.FeatureCollection('USDOS/LSIB_SIMPLE/2017');
var roi = countries.filterMetadata('country_na', 'equals', 'Japan');

//center the map on the coutnry identified by the ROI

Map.centerObject(roi);

var dataset = ee.Image('UMD/hansen/global_forest_change_2022_v1_10');

//select your country of interest by country
var datasetClip = dataset.clip(roi);

//Style the clipped tree loss layer from yellow to red
var treeLossVisParam = {
  bands: ['lossyear'],
  min: 0,
  max: 22,
  palette: ['yellow', 'red']
};
//style the clipped tree cover black no trees and green to trees. If you want global coverage call the dataset instead of clipped variable

var treeCoverVisParam = {
  bands: ['treecover2000'],
  min: 0,
  max: 100,
  palette: [
'#77a37a',
'#5f926a',
'#587e60', 
'#485e52',
'#3a4f3f'
    ]
};
Map.addLayer(datasetClip, treeCoverVisParam, 'tree cover');

//only add clipped forest loss area
Map.addLayer(datasetClip, treeLossVisParam, 'tree loss year');

# PRESENTING CODE

Here is the full script with a clip to your country - region of interest (roi) this time using FAO polygons instead of US State Department (does this matter?)

What is commented out here?

//Call Modis land cover dataset
var modisLandcover = ee.ImageCollection("MODIS/006/MCD12Q1")
//filter by date
var filtered = modisLandcover.filter(
  ee.Filter.date('2019-01-01', '2019-12-31'))

//landcover for the year 2019
var landcover2019 = ee.Image(filtered.first())
//classify data based on land class type 1 
var classified = landcover2019.select('LC_Type1')
 
 //color palette for the classification - this matches the Modus official color pallet but feel free to play here
var palette = ['05450a', '086a10', '54a708',
 '78d203', '009900', 'c6b044','dcd159', 
 'dade48', 'fbff13', 'b6ff05', '27ff87',
 'c24f44', 'a5a5a5', 'ff6d4c', '69fff8',
 'f9ffa4', '1c0dff']

//add the visualized modis data to the map and label name landcover 2019 the entire world - commenting this out bc we only want our country
//Map.addLayer(classified,
// {min:1, max:17, palette: palette},
 //'Land cover entire world 2019');
 
 // add roi using FAO countries 

var gaul = ee.FeatureCollection("FAO/GAUL_SIMPLIFIED_500m/2015/level1")
var roi = gaul.filter(
  ee.Filter.eq('ADM0_NAME', 'Spain'))
//Map.addLayer(roi, {color: 'purple'}, 'UN Country Name')
 
 // clip to only your country
 var roiLandcover = classified.clip(roi)
Map.addLayer(roiLandcover,
 {min:1, max:17, palette: palette},
 'ROI Land Cover 2019')
//don't forget to center the map on your country 
Map.centerObject(roi);

Let's clip to your country - region of interest (roi) with the US State dept data

// Load a country border to a region of interest (roi).These are recognized boardings of the USA
var countries = ee.FeatureCollection('USDOS/LSIB_SIMPLE/2017');
var roi = countries.filterMetadata('country_na', 'equals', 'Spain');

//center the map on the coutnry identified by the ROI

Map.centerObject(roi);

//Call Modis land cover dataset
var modisLandcover = ee.ImageCollection("MODIS/006/MCD12Q1")
//filter by date
var filtered = modisLandcover.filter(
  ee.Filter.date('2019-01-01', '2019-12-31'))

//landcover for the year 2019
var landcover2019 = ee.Image(filtered.first())
//classify data based on land class type 1 
var classified = landcover2019.select('LC_Type1')



 // clip to only your country
 var roiLandcover = classified.clip(roi)

 
 //color palette for the classification - this matches the Modus official color pallet but feel free to play here
var palette = ['05450a', '086a10', '54a708',
 '78d203', '009900', 'c6b044','dcd159', 
 'dade48', 'fbff13', 'b6ff05', '27ff87',
 'c24f44', 'a5a5a5', 'ff6d4c', '69fff8',
 'f9ffa4', '1c0dff']

//add the visualized modis data to the map and label name landcover 2019 the entire world
Map.addLayer(roiLandcover,
{min:1, max:17, palette: palette},
'Land cover entire world 2019');

Now add another dataset - this is elevation

read comments carefully

make changes you see fit

 //add elevation data
var dem = ee.Image('CGIAR/SRTM90_V4');
var elevation = dem.select('elevation');

 // clip to only your country
 var elevationclip = elevation.clip(roi)
//add to map you can play with the color pallet and the max numbers to make the varation stick out
Map.addLayer(elevationclip, {min: 0, max: 1000, palette: ['#000004', '#50127b', '#b63679', '#fc8761', '#fdfd65']}, 'Elevation');

All together now

Land Cover

Elevation

Clipped to your ROI

// Load a country border to a region of interest (roi).These are recognized boardings of the USA
var countries = ee.FeatureCollection('USDOS/LSIB_SIMPLE/2017');
var roi = countries.filterMetadata('country_na', 'equals', 'Spain');

//center the map on the coutnry identified by the ROI

Map.centerObject(roi);

//Call Modis land cover dataset
var modisLandcover = ee.ImageCollection("MODIS/006/MCD12Q1")
//filter by date
var filtered = modisLandcover.filter(
  ee.Filter.date('2019-01-01', '2019-12-31'))

//landcover for the year 2019
var landcover2019 = ee.Image(filtered.first())
//classify data based on land class type 1 
var classified = landcover2019.select('LC_Type1')



 // clip to only your country
 var roiLandcover = classified.clip(roi)

 
 //color palette for the classification - this matches the Modus official color pallet but feel free to play here
var palette = ['05450a', '086a10', '54a708',
 '78d203', '009900', 'c6b044','dcd159', 
 'dade48', 'fbff13', 'b6ff05', '27ff87',
 'c24f44', 'a5a5a5', 'ff6d4c', '69fff8',
 'f9ffa4', '1c0dff']

//add the visualized modis data to the map and label name landcover 2019 the entire world
Map.addLayer(roiLandcover,
{min:1, max:17, palette: palette},
'Land cover 2019');

//add elevation data
var dem = ee.Image('CGIAR/SRTM90_V4');
var elevation = dem.select('elevation');

 // clip to only your country
 var elevationclip = elevation.clip(roi)
//add to map you can play with the color pallet and the max numbers to make the varation stick out
Map.addLayer(elevationclip, {min: 0, max: 1000, palette: ['#000004', '#50127b', '#b63679', '#fc8761', '#fdfd65']}, 'Elevation');

{hex colors}

Change colors using hex codes

Color Schemes

Qualitative

Land use classification are qualitative or normative data

1.

2.

Sequential

Useful for ordinal data

GEE-stretched between 2 values

3.

Diverging

Best when their is a known zero, but also can be helpful to show variation

# CHAPTER 2

#e5f5f9
#99d8c9
#2ca25f

Color Schemes

Qualitative

Land use classification are qualitative or normative data

1.

2.

Sequential

Useful for ordinal data

3.

Diverging

Best when their is a known zero, but also can be helpful to show variation

# CHAPTER 2

Calculate area of urban areas based on the modis data

add this to your existing code from last slide

//now we will calculate the area of your country from this tutorial https://spatialthoughts.com/2020/06/19/calculating-area-gee/
 
 
var stateArea = roi.geometry().area()
var stateAreaSqKm = ee.Number(stateArea).divide(1e6).round()
print(stateAreaSqKm)
// Result can be seen in the console on the right: Spain 505195 

//next calculate the area of urban areas

var urban = roiLandcover.eq(13)

//visualize urban areas as blue and everything else as grey name of layer is Urban
Map.addLayer(urban,
 {min:0, max:1, palette: ['grey', 'blue']},
 'Urban')
 //here we calculate the pixel area - changing the pixel value to 0 or 1 urban or not
var areaImage = urban.multiply(ee.Image.pixelArea())
var area = areaImage.reduceRegion({
  reducer: ee.Reducer.sum(),
  geometry: roi.geometry(),
  scale: 500,
  maxPixels: 1e10
  })
var urbanAreaSqKm = ee.Number(
  area.get('LC_Type1')).divide(1e6).round()
print(urbanAreaSqKm)
// Result: Spain = 5584

What happened?

Area Urban

print will calculate a value and show in the Console, in this case it is total urban area

1.

2.

Reducer

Reduced data from 1 is urban and 0 non urban

Use the inspector and click on a pixel and see LC_type 1 value

3.

New Layer

Urban in blue (change the color)

Non-Urban grey (make transparent!)

# CHAPTER 2

// Export a cloud-optimized GeoTIFF.
Export.image.toDrive({
  image: urban,
  description: 'urban',
  region: roi,
  fileFormat: 'GeoTIFF',
  formatOptions: {
    cloudOptimized: true
  }
});

# PRESENTING CODE

Export your new layer

it is stored in your root directory of Google Drive

see this extremely advanced example example https://code.earthengine.google.com/?scriptPath=Examples%3AUser%20Interface%2FForest%20Change