This document describes a methodology to develop a moderate resolution irrigated area map for South Asia using advanced remote sensing techniques. The methodology involves a two-level process of 1) segmenting high resolution imagery into homogeneous objects and classifying them using machine learning, and 2) analyzing MODIS time series data to identify irrigation intensity. The resulting map identifies irrigated areas in South Asian countries at a resolution of 250m, calculating total irrigated areas of 206.74 million hectares across the region. Automating parts of the process using open source tools could help speed up localized irrigated area mapping.
Presented by IWMI's Lal Muthuwatta at the 3rd International Conference on the Status of Future of the World's Large Rivers, April 18 - 21, 2017, New Dheli, India.
Presented by IWMI's Lal Muthuwatta at the 3rd International Conference on the Status of Future of the World's Large Rivers, April 18 - 21, 2017, New Dheli, India.
Remote sensing (RS) and geographical information systems (GIS) are revolutionising irrigation management. They represent a relatively cheap and rapid method of acquiring up-to-date information over a large geographical area. and are the only practical way to obtain data from inaccessible regions. At small scales, regional phenomena which are invisible from the ground can be clearly visible. Presented at the 1st World Irrigation Forum, 2013, Mardin, Turkey.
Securing Water for Food, Livelihoods and Ecosystems to face Climate Change
Smakhtin, V., de Fraiture, C., Bossio, D., Molden, D, Hoanh C., Noble, A., Giordano, M., McCartney, M., Shah, T.
International Water Management Institute, Colombo, Sri Lanka
Floods can be hugely destructive, but they also offer opportunities for farmers and fisherfolk. If their frequency and extent can be measured, then we will be better able to mitigate costs and maximise benefits. Digital geospatial flood inundation mapping is a powerful new approach for flood response that shows floodwater extent and depth on the land surface. IWMI research will evaluate this new technology and develop a prototype flood inundation map for South Asia. Also discussed is a project to flood map and model in a spate irrigation system in Sudan.
Bharat R Sharma, L. Rebelo, G. Amarnath, I. Miltenburg
5th CRS ICT4D Conference
This presentation is based on the initial results of an IWMI-lead Project "Use of Smart ICT for Weather and Water Information and Advice to Smallholders in Africa". The 3-year project (2011-2014)is funded (US$ 1.8 m) by IFAD and jointly implemented by IWMI, eLeaf-The Netherlands and a number of national partner institutions in Egypt, Ethiopia, Sudan and Mali.
Sasumua: linking a landscape and institutional mosaic to climate change in KenyaCIFOR-ICRAF
Presentation by Meine van Noordwijk & Thomas Yatich, ICRAF
Landscape approaches to mitigation and adaptation, Forest Day 3
Sunday, 13 December 2009
Copenhagen, Denmark
By Poolad Karimi at the "Water in the Anthropocene: Challenges for Science and Governance. Indicators, Thresholds and Uncertainties of the Global Water System" conference in Bonn, Germany May 2013
Presentation by Dr. Meine van Noordwijk, World Agroforestry Centre (ICRAF) at Forest Day 3, 13 December 2009, Copenhagen. Learning event "Landscape approaches to Adaptation and Mitigation"
Improving infield water management - Simon Turner (Agri Tech Ltd)Farming Futures
This presentation formed part of the Farming Futures workshop 'Irrigation in a changing climate: save water, save money, get fit for the future'.
17th November 2009
Improving Agricultural Water Productivity in the Indira Gandhi Nahar Pariyoja...ICARDA
11-14 February 2019. Jodhpur, India. The 13th International Conference on Dryland Development,
Presentation at the session of 12 February: TU - ICARDA Satellite Symposium : Crop Improvement for Sustainable Production.
Improving Agricultural Water Productivity in the Indira Gandhi Nahar Pariyojana (IGNP). A joint project between ICARDA and CAZRI
Vinay Nangia, N.D. Yadava, M.L. Soni and V.S. Rathore
Presented by Presented by Jeremy Bird, Director General - IWMI, (on behalf of IWMI researcher Paul Pavelic) at the 8th Global Environment Facility (GEF) Biennial International Waters Conference (IWC-8) held in Negombo, Sri Lanka, on May 9-13, 2016.
Remote sensing (RS) and geographical information systems (GIS) are revolutionising irrigation management. They represent a relatively cheap and rapid method of acquiring up-to-date information over a large geographical area. and are the only practical way to obtain data from inaccessible regions. At small scales, regional phenomena which are invisible from the ground can be clearly visible. Presented at the 1st World Irrigation Forum, 2013, Mardin, Turkey.
Securing Water for Food, Livelihoods and Ecosystems to face Climate Change
Smakhtin, V., de Fraiture, C., Bossio, D., Molden, D, Hoanh C., Noble, A., Giordano, M., McCartney, M., Shah, T.
International Water Management Institute, Colombo, Sri Lanka
Floods can be hugely destructive, but they also offer opportunities for farmers and fisherfolk. If their frequency and extent can be measured, then we will be better able to mitigate costs and maximise benefits. Digital geospatial flood inundation mapping is a powerful new approach for flood response that shows floodwater extent and depth on the land surface. IWMI research will evaluate this new technology and develop a prototype flood inundation map for South Asia. Also discussed is a project to flood map and model in a spate irrigation system in Sudan.
Bharat R Sharma, L. Rebelo, G. Amarnath, I. Miltenburg
5th CRS ICT4D Conference
This presentation is based on the initial results of an IWMI-lead Project "Use of Smart ICT for Weather and Water Information and Advice to Smallholders in Africa". The 3-year project (2011-2014)is funded (US$ 1.8 m) by IFAD and jointly implemented by IWMI, eLeaf-The Netherlands and a number of national partner institutions in Egypt, Ethiopia, Sudan and Mali.
Sasumua: linking a landscape and institutional mosaic to climate change in KenyaCIFOR-ICRAF
Presentation by Meine van Noordwijk & Thomas Yatich, ICRAF
Landscape approaches to mitigation and adaptation, Forest Day 3
Sunday, 13 December 2009
Copenhagen, Denmark
By Poolad Karimi at the "Water in the Anthropocene: Challenges for Science and Governance. Indicators, Thresholds and Uncertainties of the Global Water System" conference in Bonn, Germany May 2013
Presentation by Dr. Meine van Noordwijk, World Agroforestry Centre (ICRAF) at Forest Day 3, 13 December 2009, Copenhagen. Learning event "Landscape approaches to Adaptation and Mitigation"
Improving infield water management - Simon Turner (Agri Tech Ltd)Farming Futures
This presentation formed part of the Farming Futures workshop 'Irrigation in a changing climate: save water, save money, get fit for the future'.
17th November 2009
Improving Agricultural Water Productivity in the Indira Gandhi Nahar Pariyoja...ICARDA
11-14 February 2019. Jodhpur, India. The 13th International Conference on Dryland Development,
Presentation at the session of 12 February: TU - ICARDA Satellite Symposium : Crop Improvement for Sustainable Production.
Improving Agricultural Water Productivity in the Indira Gandhi Nahar Pariyojana (IGNP). A joint project between ICARDA and CAZRI
Vinay Nangia, N.D. Yadava, M.L. Soni and V.S. Rathore
Presented by Presented by Jeremy Bird, Director General - IWMI, (on behalf of IWMI researcher Paul Pavelic) at the 8th Global Environment Facility (GEF) Biennial International Waters Conference (IWC-8) held in Negombo, Sri Lanka, on May 9-13, 2016.
Similar to Irrigated Area Mapping, South Asia (20)
Feature Matching using SIFT algorithm; co-authored presentation on Photogrammetry studio by Sajid Pareeth, Gabriel Vincent Sanya, Sonam Tashi and Michael Mutale
1. Developing a Moderate Resolution Irrigated Area Map for
South Asia using segmentation and time series analysis
Photo: David Brazier/IWMI
Water for a food-secure world
www.iwmi.org
2. Why Irrigated Area Mapping?
• Perspective of achieving food security by
increasing irrigation
• Though 70-85 % of water used
• Especially with current situation of
population, urbanization , climate change
etc.
• Important to assess the spatial distribution,
intensity, water use etc.
Water for a food-secure world
www.iwmi.org
3. Is it new?
• Many products available globally - FAO, IWMI
• Also national products- CBIP, India
• Global Irrigated Area Map(GIAM) –
developed by IWMI in 2006
• GIAM -Resolution of 10km and datasets from
1990 -1999, AVHRR
• Very course product with detailed
classification
Water for a food-secure world
www.iwmi.org
4. Global Irrigated Area Mapping
• Product from IWMI - developed using multiple global
datasets
• Different datasets were used at
– Segmentation/Localization of landscape
– Classification into different units
– Time series analysis to identify irrigation
intensity
• Nominal resolution of 10KM
• Datasets used were from 1990 – 2000
Water for a food-secure world
www.iwmi.org
5. Need/Opportunity to update GIAM
• Data available from 250m spatial resolution
• Highly capable HW/SW available for data intensive
processes
• Good temporal coverage
• Extensive change in the landscape would have
happened in 12 years
• New algorithms in image classification – „object based
image analysis‟
• Updating the irrigated area map for South East Asia
Water for a food-secure world
www.iwmi.org
6. Datasets - comparison
Dataset - GIAM Resolution Proposed Resolution Availability Role
type dataset Dataset
NDVI AVHRR 10KM MODIS 250m Free Time series
/Reflectance analysis
NDVI/Reflec SPOT 1KM IRS P6 - 56m Purchase Single date
tance AWIFS classification
into objects
DEM GTOPO 1KM SRTM 90m Free conditional
segmentation
Temperature AVHRR 10km MODIS 1KM Free conditional
segmentation
Precipitation CRU 0.5 degree WorldClim 1KM Free conditional
segmentation
Water for a food-secure world
www.iwmi.org
7. Updated
Level 1
Methodology
Entire processing on minimum
Mapping unit – like admin
boundaries, climatic zones etc.
Level 2
Water for a food-secure world
www.iwmi.org
8. Level1 – Segmentation and HR
Land cover map
Water for a food-secure world
www.iwmi.org
9. Optimal segmentation
• Region growing algorithm – SPRING open source
software
• Main parameters; “Similarity” and “Area”
• Objective function based on spatial auto correlation
to determine best parameters
• Optimal segmentation > good classification
• Another factor > size of the image
• Bigger the size > more mix in clustering results
• Optimal size found from trial runs 250km by 200km
Water for a food-secure world
www.iwmi.org
10. Image classification steps
Original Image
Segmented Image
ISOCLASS Classified
Image
Recoded Image
Water for a food-secure world
www.iwmi.org
11. Level 2 – Time series on MODIS
250m NDVI
Water for a food-secure world
www.iwmi.org
12. MODIS Path/row for South Asia
Water for a food-secure world
www.iwmi.org
13. Class – flow diagram
Agriculture
Irrigated Rain fed Water source
Surface water Ground water Conjunctive Irrigation type
Continuous
Single crop Double crop From MODIS
crop
Irrigation intensity
Example class:
Irrigated, surface water, double crop
Water for a food-secure world
www.iwmi.org
14. Irrigated area calculated
Country Irrigated Area (million ha)
Nepal 4
Pakistan 21
Sri Lanka 1.6
India 169
Bhutan 0.2
Bangladesh 10
Total irrigated area calculated for entire South Asia is
206.74 million hectares.
Water for a food-secure world
www.iwmi.org
15. India
Water for a food-secure world
www.iwmi.org
16. Pakistan
Water for a food-secure world
www.iwmi.org
17. Sri Lanka
Water for a food-secure world
www.iwmi.org
21. Speeding up the localized approach
• Use of customizable open source tools
• Developing a R package to manage the segmentation
• Program in R to control
• Dicing the imageries
• Segmentation – SPRING software
• Classification
• Extracting the agc
• Time series on agc
• Localizing based on secondary datasets
• Class assignment based on irrigation intensity
• Time consuming/Manual
• Class assignment at both classification levels
• Comments?
Water for a food-secure world
www.iwmi.org
22. Conclusions
• High resolution global datasets available now
• Introducing a localized approach to avoid mixes
• Key is to identify MMU with homogeneous pattern
• Scope for semi automating the process using R scripting
• Can‟t avoid the manual interventions though…
Water for a food-secure world
www.iwmi.org