U.S. Geological Survey U.S. Department of Interior GFSAD 30 Cropland Products of Australia @ Nominal 250 m Using MODIS Data and Cropland Mapping Algorithms:

U.S. Geological SurveyU.S. Department of Interior

GFSAD 30 Cropland Products of Australia @ Nominal 250 m Using MODIS Data and Cropland Mapping Algorithms: Results and update

Pardhasaradhi Teluguntla, Prasad Thenkabail, and Jun XiongJun 24-26, 2014

Fourth Workshop on Global Food Security Analysis Data @ 30 m (GFSAD30)

Sioux Falls, SD, USA


GFSAD30 Products for Australia Outline


GFSAD30 Cropland Products of Australia @ nominal 250 mOutline

1. Goals and Objectives

2. Issues

3. Data: BDI for Australia

4. Methods and Preliminary results

a) Classification, identification, and label classes for the 4 products

b) Automated Algorithms\script to automatically compute the 4 products

5. Discussion


GFSAD30 Products for AustraliaGoals and Objectives


❖ Cropland Extent➢ Croplands vs. non-croplands

❖ Cropping Intensity➢ Single, double, triple, continuous cropping

❖ Watering Method➢ Irrigated vs. rainfed

❖ Crop Type➢ Major 8 crops and others

❖ Cropland change over space and time➢ Major 8 crops and others

GFSAD30 Cropland Products of Africa @ Nominal 250 m Goal: to Produce 5 Products of GFSAD30 Project

We will focus on these 3 products for now in this presentation

These products are not part of this presentation


GFSAD30 Products for Australia Current State-of-Art


❖ State-of-Art➢ There are 3 global cropland

products + several land cover land use products in which croplands is a class

❖ Limitations➢ coarse resolution (1 km or higher)➢ Lack of detailed work on croplands

(e.g., where are the irrigated areas, what is the frequency of cropping, what crops are grown and where?, what changes are occuring)

GCE V1.0

GFSAD30 Cropland Products of Australia @ Nominal 250 m Current State-of-Art on Global Croplands: 12 Classes derived from 4 existing products

The 12 classes constitute 12 AOIs. Class 1 = AOI 1, Class 2 = AOI 2….. Class 12 = AOI 12


Cropland Mapping Algorithms (CMAs)

Approach to Producing GFSAD30 Products


We will approach development of GFSAD30 products in two distinct steps:

❖ Synergestic approaches to cropland classification leading to➢ Class identification and labeling;➢ Creation of knowledge base

❖ Automated cropland classification algorithms➢ Development and implementation of Automated algorithm to reproduce

cropland products year after year

Cropland Mapping Algorithms (CMAs) Approach to Automated Cropland Classification Algorithms

We will discuss this today

We will not discuss this today



Datasets and Megafile Datacubes


❖ MODIS Data Composition: 36 layers per year➢ monthly NDVI maximum value composite (MVC):

12 layers per year;➢ monthly B1 Minima: 12 layers per year;➢ monthly B2 Maxima: 12 layers per year;

❖ Secondary data➢ Elevation, slope, Precipitation, surface temperature,

PET etc

Cropland Mapping Algorithms (CMAs) for Australia Datasets in the Megafile for entire Australia: nominal

Year 2000


GFSAD30 Cropland Products of Australia @ nominal 250 m First Start with GCE V1.0 (~ 1 km): 12 Classes

Note:1. We will take each GCE V1.0

class and investigate how much of that is croplands versus non croplands @ 250 m MODIS resolution;

2. In addition, we will establish cropping intensity and irrigated vs. rainfed

3. This leads to GCE V2.0 @ 250 m


MODIS 250 m Mega file data cube (36 Layers ): 12 B1 minima, 12 B2 maxima,

12 NDVI (one band per month)AOI / segments

GCE V1.0 Final Output

GCE V2.0 (illustrated here for AOI 1 which has 5 classes)

Calssify

Cropland Mapping Algorithms (CMAs) for Australia Process for producing GCE V2.0 @ 250 m resolution using MODIS data for

nominal year 2000



Synergestic Cropland

Classification and Identification (SCCI)


Synergestic Cropland Classification and Class Identification (SCCI) involves:

A. Classification❖ unsupervised ISOCLASS clustering of MFDC;

B. Class Grouping and identification❖ Spectral matching technique: SCS R-square grouping;

❖ bispectral tassel cap plots for better understanding of classes;

❖ NDVI phenological plots to determine cropping intensity and class identification

C. Class labeling❖ Very high resolution imagery from ESRI archive

❖ Ground data

Cropland Mapping Algorithms (CMAs) for Africa Synergestic Cropland Classification and Class Identification

(SCCI)


Australia: AOI-1 Unsupervised ISOCLASS classification using MODIS 250m MFDC: 25 classes


Australia: AOI-1 Tassel cap/ Bi-Spectral plots: 25 classes

Example1 : Class 1 (cropland) showing how band 1 (red) and band 2 (NIR) reflectivity varies from month to month


Australia: AOI-1 Tassel cap/ Bi-Spectral plots: 25 classes

Example1 : Class 1 (non-cropland) showing how band 1 (red) and band 2 (NIR) reflectivity varies from month to month


Australia: AOI-1 NDVI time series:25 classes; Class 1 identification process detailed

Non cropland Non cropland

Sub-meter to 5 meter imagery


Australia: AOI-1 NDVI time series:25 classes; Class 12 identification process detailed

Croplands; identified from VHRI (sub-meter to 5 meter imagery)

Croplands; identified from VHRI (sub-meter to 5 meter imagery)


Synergestic Cropland Classification and Identification (SCCI)

re-grouping of classes of AOI 1 for Australia


Spectral Matching Technique: SCS R-square to Group ClassesSCS- R2 Matrices for AOI1; Australia

C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 C20 C21 C22 C23 C24 C25

C1 1.00 0.94 0.98 0.90 0.91 0.94 0.92 0.92 0.84 0.85 0.89 0.81 0.91 0.91 0.94 0.75 0.83 0.89 0.78 0.91 0.84 0.81 0.95 0.56 0.57

C2 0.94 1.00 0.94 0.92 0.81 0.99 0.96 0.92 0.90 0.87 0.92 0.84 0.91 0.90 0.88 0.76 0.83 0.85 0.77 0.84 0.81 0.73 0.87 0.57 0.54

C3 0.98 0.94 1.00 0.83 0.86 0.94 0.88 0.86 0.78 0.77 0.85 0.73 0.85 0.86 0.95 0.67 0.77 0.85 0.72 0.91 0.79 0.79 0.95 0.49 0.54

C4 0.90 0.92 0.83 1.00 0.84 0.91 0.97 0.98 0.98 0.99 0.95 0.98 0.98 0.95 0.81 0.92 0.94 0.91 0.89 0.80 0.89 0.76 0.81 0.72 0.62

C5 0.91 0.81 0.86 0.84 1.00 0.80 0.83 0.84 0.78 0.81 0.83 0.78 0.85 0.86 0.88 0.75 0.79 0.91 0.79 0.88 0.81 0.81 0.89 0.64 0.68

C6 0.94 0.99 0.94 0.91 0.80 1.00 0.94 0.90 0.89 0.87 0.89 0.83 0.90 0.88 0.87 0.76 0.81 0.82 0.76 0.82 0.81 0.73 0.86 0.54 0.49

C7 0.92 0.96 0.88 0.97 0.83 0.94 1.00 0.98 0.98 0.95 0.98 0.94 0.97 0.97 0.87 0.88 0.93 0.92 0.88 0.85 0.88 0.78 0.85 0.71 0.65

C8 0.92 0.92 0.86 0.98 0.84 0.90 0.98 1.00 0.96 0.96 0.99 0.95 0.98 0.99 0.87 0.90 0.97 0.96 0.91 0.87 0.91 0.81 0.87 0.74 0.69

C9 0.84 0.90 0.78 0.98 0.78 0.89 0.98 0.96 1.00 0.98 0.96 0.98 0.97 0.95 0.78 0.93 0.95 0.90 0.90 0.79 0.89 0.75 0.77 0.77 0.67

C10 0.85 0.87 0.77 0.99 0.81 0.87 0.95 0.96 0.98 1.00 0.93 0.99 0.98 0.94 0.77 0.96 0.96 0.91 0.92 0.79 0.92 0.77 0.79 0.78 0.66

C11 0.89 0.92 0.85 0.95 0.83 0.89 0.98 0.99 0.96 0.93 1.00 0.93 0.96 0.99 0.87 0.88 0.96 0.95 0.90 0.87 0.88 0.80 0.86 0.73 0.69

C12 0.81 0.84 0.73 0.98 0.78 0.83 0.94 0.95 0.98 0.99 0.93 1.00 0.98 0.95 0.76 0.97 0.97 0.91 0.94 0.78 0.92 0.78 0.77 0.82 0.70

C13 0.91 0.91 0.85 0.98 0.85 0.90 0.97 0.98 0.97 0.98 0.96 0.98 1.00 0.98 0.87 0.95 0.97 0.95 0.95 0.88 0.96 0.86 0.88 0.79 0.72

C14 0.91 0.90 0.86 0.95 0.86 0.88 0.97 0.99 0.95 0.94 0.99 0.95 0.98 1.00 0.90 0.91 0.98 0.98 0.94 0.91 0.93 0.86 0.90 0.78 0.75

C15 0.94 0.88 0.95 0.81 0.88 0.87 0.87 0.87 0.78 0.77 0.87 0.76 0.87 0.90 1.00 0.74 0.82 0.92 0.82 0.99 0.88 0.93 0.99 0.67 0.74

C16 0.75 0.76 0.67 0.92 0.75 0.76 0.88 0.90 0.93 0.96 0.88 0.97 0.95 0.91 0.74 1.00 0.96 0.90 0.98 0.78 0.96 0.83 0.76 0.91 0.79

C17 0.83 0.83 0.77 0.94 0.79 0.81 0.93 0.97 0.95 0.96 0.96 0.97 0.97 0.98 0.82 0.96 1.00 0.96 0.98 0.85 0.95 0.85 0.83 0.85 0.78

C18 0.89 0.85 0.85 0.91 0.91 0.82 0.92 0.96 0.90 0.91 0.95 0.91 0.95 0.98 0.92 0.90 0.96 1.00 0.95 0.93 0.94 0.91 0.92 0.81 0.81

C19 0.78 0.77 0.72 0.89 0.79 0.76 0.88 0.91 0.90 0.92 0.90 0.94 0.95 0.94 0.82 0.98 0.98 0.95 1.00 0.86 0.98 0.91 0.84 0.93 0.87

C20 0.91 0.84 0.91 0.80 0.88 0.82 0.85 0.87 0.79 0.79 0.87 0.78 0.88 0.91 0.99 0.78 0.85 0.93 0.86 1.00 0.91 0.96 0.99 0.74 0.82

C21 0.84 0.81 0.79 0.89 0.81 0.81 0.88 0.91 0.89 0.92 0.88 0.92 0.96 0.93 0.88 0.96 0.95 0.94 0.98 0.91 1.00 0.95 0.90 0.90 0.86

C22 0.81 0.73 0.79 0.76 0.81 0.73 0.78 0.81 0.75 0.77 0.80 0.78 0.86 0.86 0.93 0.83 0.85 0.91 0.91 0.96 0.95 1.00 0.94 0.86 0.90

C23 0.95 0.87 0.95 0.81 0.89 0.86 0.85 0.87 0.77 0.79 0.86 0.77 0.88 0.90 0.99 0.76 0.83 0.92 0.84 0.99 0.90 0.94 1.00 0.67 0.74

C24 0.56 0.57 0.49 0.72 0.64 0.54 0.71 0.74 0.77 0.78 0.73 0.82 0.79 0.78 0.67 0.91 0.85 0.81 0.93 0.74 0.90 0.86 0.67 1.00 0.95

C25 0.57 0.54 0.54 0.62 0.68 0.49 0.65 0.69 0.67 0.66 0.69 0.70 0.72 0.75 0.74 0.79 0.78 0.81 0.87 0.82 0.86 0.90 0.74 0.95 1.00


6 Highly correlated classes

Australia: AOI-1 re-grouping 25 classes to unique croplands versus non-cropland classes using SCS- R2



12 highly correlated classes







Finally, 25 classes are grouped into 4 unique classes


Australia: AOI-1 Tassel cap Bi-Spectral plot characteristics of the Final 5 re-grouped classes

How to separate / distinguish ClassesExample :• Class 1 month 12• Class 2 month 8 • Class 3 month 10 • Class 4 month 9• Class 5 month 10

Note:We can distinctly distingwish the 5 classes in these particular months


o Stage 1

Croplands vs. Non-croplandso Stage 2

single, double, triple, or continuous croppingo Stage 3

irrigation versus rainfedo Stage 4

8 major crops or others

GFSAD30: Australia Class Naming Convention


AOI_1 _class# Histogram % Group Level1 Level2 Level3Class 1 37337 2.01%G5 Non-Cropland BarrenClass 2 72972 3.94%G1 Cropland Single irrigated ?Class 3 58370 3.15%G1 Cropland Single irrigated ?Class 4 88631 4.78%G1 Cropland Single irrigated ?Class 5 49879 2.69%G1 Cropland Single irrigated ?Class 6 60913 3.28%G1 Cropland Single irrigated ?Class 7 105202 5.67%G1 Cropland Single irrigated ?Class 8 85795 4.63%G2 Cropland Single irrigated Class 9 98458 5.31%G2 Cropland Single irrigated Class 10 119651 6.45%G2 Cropland Single irrigated Class 11 82863 4.47%G2 Cropland Single irrigated Class 12 102273 5.52%G2 Cropland Single irrigated Class 13 82085 4.43%G2 Cropland Single irrigated Class 14 106507 5.74%G2 Cropland Single irrigated Class 15 73016 3.94%G3 Natural vegetationClass 16 78424 4.23%G2 Cropland Single irrigated Class 17 76405 4.12%G2 Cropland Single irrigated Class 18 64859 3.50%G2 Cropland Single irrigated Class 19 59052 3.18%G2 Cropland Single irrigated Class 20 62648 3.38%G3 Natural vegetationClass 21 78739 4.25%G4 Cropland Double(?) Irriagted Class 22 62684 3.38%G4 Cropland Double(?) Irriagted Class 23 40861 2.20%G3 Natural vegetationClass 24 43852 2.36%G4 Cropland Double(?) Irriagted Class 25 62843 3.39%G4 Cropland Double(?) Irriagted

1854319 100.00% Total 9.89Mha1640457 88.47% Total-Cropland 8.75Mha

213862 11.53%Total- Non_Cropland 1.14Mha

Class Naming Convention

Croplands, single, irrigated

Non-croplands

We will use the following data to label this:1. VHRI (sub meter to 5

meter;2. Ground data3. GCE V1.0

We will use the following data to label this:1. NDVI phenology;2. Bispectral plots3. SCS R-square4. dendogram5. VHRI (sub-meter to

5 meter)6. Ground data

We will use the following data to label this:1. NDVI phenology plots;

Note: 1.14 million hectares of AOI 1 (out of 9,89 M ha) is non croplands in GCE V2.0. This was classified as croplands in GCE V1.0


Australia: AOI-1 Class Identification and Labeling Process: Distribution of Random areas for Each Class for selecting VHRI

For each of the 5 classes 20 random areas are chosen for selecting VHRI (sub meter to 5 m)….. A total of 100 VHRI are selected for the 5 classes of AOI 1


Australia: AOI-1 20 very high resolution images (VHRI; sub-meter to 5 meter) for class 1 of AOI 1










Class code Label description

A1C001 AOI1 Class 1 (23.5%) Croplands, single crop, irrigated(?)

A1C002 AOI1 Class 2 (51.5%) Croplands, single crop, irrigated

A1C003 AOI1 Class3 (9.5%) Non croplands, Natural vegetation

A1C004 AOI1 Class 4 (13.4%) Croplands, double crop(?), irrigated

A1C005 AOI1 Class 5(2%) Non croplands, barren lands

GFSAD30: Australia Class Naming Convention for AOI 1, 5 final classes of Australia


GFSAD30: Australia Class Naming Convention for AOI 1, 5 final classes of Australia

MODIS 250 m GCE V2.0 for AOI 1

Note: Omissions: 11.5% (class 4 and 5) of the AOI 1 area of GCE V1.0 @ nominal 1 km is now non-croplands in GCE V2.0 @ 250 m


GFSAD30: Australia Class Naming Convention for AOI 1, 5 final classes of Australia and their Characteristics


AOI-1 of Australia @1km resolutionCropland area = 9.89 M ha (full pixel area )

AOI-1 of Australia @ 250m resolutionCropland area = 8.75 M ha 11.5% of area is omission

GFSAD30: Australia Final classes of AOI 1 in GCE V1.0 (1 km) vs. GCE V2.0 (250 m)


❖ Automated cropland classification algorithms➢ Development and implementation of Automated algorithm to reproduce

cropland products year after year

………the above work is in progress

Cropland Mapping Algorithms (CMAs) Approach to Automated Cropland Classification Algorithms


Thank you


AOI-1 of Australia @1km resolutionCropland area = 9.89 M ha (full pixel area )

AOI-1 of Australia @ 250m resolutionCropland area = 8.75 M ha 11.5% of area is omission

Australia: AOI-1 :GCE V1.0 vs GCEV 2.0


Australia: AOI-1 Tassel cap/ Bi-Spectral plots: integrated picture of 25 classes

U.S. Geological Survey U.S. Department of Interior GFSAD 30 Cropland Products of Australia @ Nominal 250 m Using MODIS Data and Cropland Mapping Algorithms:

Documents

global cropland products

cropland products year

presentationthese products

cropland extentcroplands

products of gfsad30

geological surveyu

africagce v1

land cover land use