Monitoring cropland areas using Remote sensing Murali K Gumma 1 Prasad S Thenkabail 2 and Jun Xiong 2 1 International Crops Research Institute for the Semi-Arid Tropics, Patancheru-502324, India 2 U.S. Geological Survey (USGS), Western Geographic Science Center, Flagstaff, AZ 8600, USA Operationalizing the Regional Collaborative Platform to address water consumption, Water productivity and Drought management’ in Agriculture 27-29 October, 2015 Fairmont Nile City Hotel, Cairo
65
Embed
Monitoring cropland areas using Remote sensing, Murali Krishna Gumma
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Monitoring cropland areas using Remote sensing
Murali K Gumma1 Prasad S Thenkabail2 and Jun Xiong2
1International Crops Research Institute for the Semi-Arid Tropics, Patancheru-502324, India2 U.S. Geological Survey (USGS), Western Geographic Science Center, Flagstaff, AZ 8600, USA
Operationalizing the Regional Collaborative Platform to address water consumption, Water productivity and Drought management’ in
Agriculture27-29 October, 2015
Fairmont Nile City Hotel, Cairo
Global Croplands using AVHRR, SPOT Vegetation, and Secondary Data
GFSAD30 Cropland Products of South Asia @ MODIS 250m
Outline of Today’s Presentation
Outline1. Goals and Objectives2. Data: MODIS time-series3. Cropland Knowledge Creation through Ground
survey Data4. MethodsDecision Trees algorithms SMTs for baseline cropland product generation
5. Results Cropland products from SMTs using MODIS
250 m time series;6. Challenges/Way forward
GFSAD30 Cropland Products of South Asia @ MODIS 250m
Goals and Objectives
The overarching goal of this research is to develop and implement spectral matching techniques (SMTs) and automated cropland classification algorithms (ACCA’s) for production of multi-year cropland products that in turn will help address food security issues using MODIS 250m time-series data and Landsat 30m.
Four distinct cropland products will be produced. These are: A. Cropland Extent\Areas, B. Irrigated and Rainfed CroplandsC. Cropping Intensities (single, double, triple, or continuous cropping), and D. Crop Type and\or Dominance. F. Change over space and timeG. Length of growing periodF. Other products (e.g., study drought in rainfed and irrigated)
GFSAD30 Cropland Products of Africa / South Asia @ Nominal 250 mGoals and Objectives:
ACCA- Automated
Decision Tree
DT/SMT- Interpretation
based on knowledge
MODIS NDVI, EVI & LSWI/ Landsat
Crop Extent/Mask Crop Signatures
Baseline Map(2014 for now)
Reference Bank
FAO statistics
Annual Dynamic Map(2003-2014)
VHRI
Approach: Flowchart
Feb 2011_MVC Mar 2011_MVC Apr 2011_MVC May 2011_MVC
• Coordinates: latitude, longitude• Land cover percentages• Land use categories• Irrigated area class types (eg. small scale, large scale)• Crop types• Cropping pattern• Cropping calendar• Watering Method (eg; surface water, ground water, tank)• Others: eg; digital photos, detailed descriptions
Watering method Irrigation type Crop type* Scale Intensity
LULC
SW GW
e.g. Rice, Wheat, Maize, or others
Large scale Small scale
Singlecrop
Doublecrop
Continuouscrop
+ + + +
FragmentConjunctive use
Knowledge Base on Croplands Details on Ground Data
GFSAD30 Cropland Products of Africa / South Asia @ MODIS 250m
METHODSIdeal Spectra Generation for Irrigated
cropland
Knowledge Base on Croplands Irrigated Classes: INITIAL IDEAL SPECTRA
Knowledge Base on Croplands Irrigated classes: INITIAL IDEAL SPECTRA
Grouping Calsses to Unique Categories Irrigated Information Classes: MODIS NDVI Spectral Profiles
Initial 100 Classes or Initial Class Spectra to 20 groups
Grouping Calsses to Unique Categories Rainfed Information Classes: MODIS NDVI Spectral Profiles
Spectral matching techniques: Rainfed-single crop
Grouping Calsses to Unique Categories Irrigated Information Classes: MODIS NDVI Spectral Profiles
Initial 100 Classes or Initial Class Spectra to 20 groups
RESULTSFinal Cropland Classes of Africa &South
AsiaCropland Extent\Areas
Crop intensityIrrigated/rainfed
LGPAbiotic stresses
Total Gross Cropland areas (TGCA) = 282 Mha
Total Net Cropland Area (TNCA) = 257 Mha
SMT Derived Product 4
Mapping crop land areas in Africa(2014)SMTs
Mapping crop land areas in Malawi (2014)
Mapping crop land areas in Africa(2014)
Major crops (2014)01. Rainfed-sc-sorghum02. Rainfed-sc-millets/sorghum03. Rainfed-sc-groundnut04. Rainfed-sc-pigeonpea05. Rainfed-SC-maize/sorghum/millet06. Other crops
Planting dates01. Jun, 1st to 30th02. Jul, 1st to 15th03. Jul, 16th to 30th04. Aug, 1st to 15th05. Aug, 16th to 30th06. Sept, 1st to 15th07. Sept, 16th to 30th
Rice – Fallows (Rabi fallows) in South Asia (2010-11)
Rice – fallows (Rabi fallows) areas(2010-11)
Country
MODIS–based Estimates ('000ha)
Total net rice (area)
Rice-fallows (rainfed)
Rice-fallows (irrigated)
Rice-fallows (rainfed) (%)
Rice-fallows (irrigated) (%)
Bangladesh 6,834 2,270 128 33 6
Bhutan 16 2 0 12 0
India 45,117 11,519 2,205 26 19
Nepal 1,231 109 0 9 0
Pakistan 1,490 23 12 2 52
Sri Lanka 1,139 357 1 31 0
Crop Intensity (rice): Bangladesh
Boro rice: 5.01 Mha Aus rice: 1.1 Mha Aman rice: 5.8 Mha
Drought & Submergence areas: South Asia(Agriculture areas)
/ submergence
Identifying drought villages
a) Abiotic stresses
Abiotic stressesFlood damageMild drought
Moderate droughtSevere drought
RESULTSAutomated Cropland Classification
Algorithm (ACCA) for
Producing Cropland Products for Baseline Year 2010
Mapping crop land areas in Africa(2014)SMT Vs ACCA
Total Gross Cropland areas (TGCA) = 282 MhaTotal Net Cropland Area (TNCA) = 257 Mha
Total Gross Cropland areas (TGCA) = 282 MhaTotal Net Cropland Area (TNCA) = 259 Mha