Following is the Notice of Intent proposal submitted by Dr. Cristina Milesi as PI (Foundation of California State University - Monterey Bay) in collaboration with Dr. Prasad S. Thenkabail (International Water Management Institute) in response to NASA solicitation NNH08ZDA001N - LCLUC. The reproduced proposal below:

Global Irrigated Area Mapping at 30m (GIAM30) using Landsat 2005 and 1990 data to assess the dynamics of agricultural water use under a changing climate

Project Summary: As monitoring global irrigated areas and their water use is imperative for ensuring sustainable food security to the people of the world, this project is aimed at mapping temporal changes in global irrigated areas using Mid-decadal Global Land Survey 2005 (MDGLS2005) and Geocover Landsat data sets for year 1990. Irrigated areas will be mapped through a suite of methods and data involving ideal spectral data bank, spectral matching techniques, decision tree algorithms, space-time spiral curves, and ancillary data such as climate, soils, land cover, national statistics, and extensive use of groundtruthing. The results will be used to analyze the impacts of these changes on agricultural water use and crop yields under ongoing and future climate changes over India, which has the world's second largest annualized irrigated area and which growth in food production has been stagnating over the past decade. Agricultural water use estimates under past, present, and projected climate will be performed making use of the TOPS (Terrestrial Observation and Prediction System) data and modeling software system (http://ecocast.arc.nasa.gov). The proposed work will contribute to the LCLUC Key Science question "Where are land cover and land use changing, what is the extent, and over what time scale?" and to the CCSP LULCC science question "How do climate variability and change affect land use and land cover, and what are the potential feedbacks of changes in land use and land cover to climate?"

Rationale and Background: The world is suddenly facing a food crisis not known to it since 1970s and many have already pronounced that the era of cheap food is over. Rice prices this year alone have already soared by 141 percent (Source: Chicago Board of Trades: Jacksons) and the World Bank President Mr. Robert Zoellick has recently warned that as many as 100 million people can be forced into deeper poverty in World's poorest Countries if spiraling world food prices are not controlled. The recent events have almost caught the World napping. But, looking at hard facts, it is not so much of a surprise. The World population is growing at nearly 100 million per year, irrigated lands that increased so rapidly in the 1970s to 2000s have almost stagnated, the era of "green revolution" has ended, populations of emerging markets are consuming more, croplands are diverted to bio-fuels, water is becoming limited for croplands with increasing demands from ever increasing urbanization and industrialization, and the much anticipated "blue revolution" (growing "more crop per drop") has not even taken off (Thenkabail et al., 2008). As the world?s population grew from 2.2 billion in 1950 to 6.6 billion in 2007 irrigation played a major role in tripling of the world grain harvest from 640 million tons to 1855 million tons (Brown, 2003). However, recent global trends suggest that grain production increases are becoming more difficult to achieve as a result of increasing population, and as the competition for water grows between industries and agriculture (Flavin and Gardner, 2006) as well as diversion of croplands for bio-fuels (Polya, 2008). In the Northern China Plain, which produces over half of China's wheat and a third of corn, the declining water tables and resulting loss of irrigated areas have been attributed to the drop in grain production from its peak of 392 million tons in 1998 to 338 million tons in 2003, a drop equivalent to Canada?s harvest (Brown, 2003). The United States (US) is currently using about 9% of its wheat, 25% of its corn and about 15% of its grain in general to produce biofuel (Polya, 2008). The impact of such a measure on global food bank is just beginning to be felt. However, the greatest uncertainty is expected to come from increasing impacts of climate change. There is increasing evidence that irrigation expansion may have been offsetting the greenhouse forcing on a regional scale providing a cooling effect over cropland areas with extensive irrigation (Lobell et al., 2006; Kueppers et al., 2007). As irrigated areas reach an expansion limit and eventually start declining because of the exhaustion of water resources and the competition from industrial and urban uses, the irrigation cooling effect is expected to stop and cropland areas may start warming at a faster rate (Kueppers et al., 2007). The stress on cropland areas due to the increasing temperatures is expected to derive both from increasing water demand on already strained water resources and from increased heterotrophic respiration (which grows with temperature, Peng et al., 2004). A negative feedback from declining yields can further reduce the irrigated areas, causing further warming of croplands and more stress on crop production.

In order to provide accurate and reliable near-real time spatial information on irrigation, the International Water Management Institute (IWMI) has developed a methodology (Thenkabail et al., 2008, 2007a, 2007b, 2006, 2005; Biggs et al., 2006) to provide Global Irrigated Area Mapping (GIAM) at a nominal resolution of 10-kilometers using multi-resolution monthly time-series AVHRR pathfinder, and SPOT Vegetation NDVI data coupled with 40-year climate data, and extensive use of groundtruth and Google Earth data. The study has lead to distinct sets of GIAM 10km data and products for 198 countries (http://www.iwmigiam.org) that include irrigated area maps, area statistics, class characteristics (e.g., single crop, double crop), and time-series animations of each class over space and time. Our further research has shown that finer the spatial resolution, the greater the accuracy of irrigated area class designations (see also Ozdogan and Woodcock 2006). A study of the irrigated areas of Ogallala aquifer in the United States based on Landsat imagery and 364 groundtruth points showed that 18.5 percent of the actual irrigated areas were not included (errors of omissions) in the GIAM dataset at 10 KM resolution but 20.4% of the non-irrigated areas were included as irrigated areas (errors of commissions) (Kurz and Seelan, 2006). The errors of omissions were because at coarser resolutions more fragmented, smaller patches of irrigated croplands could not be delineated whereas the errors of commissions were as a result of the large pixel sizes of coarse resolutions that at times can map patchy non-irrigated areas surrounding irrigated areas as irrigated areas. In either case, the need for finer spatial resolution to resolve the confusion is a must. This proposal aims to overcome these uncertainties and errors through GIAM30.

Responsiveness: The proposal responds to the IGBP/IHDP Global Land Project and the Monsoon Asia Integrated Regional Study components of the LCLUC program. The proposal also fits into the MGDLS Phase III objective of generating continental scale higher level products.

Expected Outcomes: Global irrigated area maps based on Landsat data of 2005 (Gutman et al., 2008) and Geocover 1990s (Tucker et al. 2005) will build upon the goals of Geocover and MDGLS of developing the capability to generate periodic inventories of global land cover conditions. Global irrigated area maps at 30 m (GIAM30) will be a new product providing irrigated area statistics and maps for every country in the world as well as sub-national units such as district and county. Through an extensive network of collaborations for field data collection and validation the project will create a large user community for the data. Precise estimates of global irrigated areas will help studies such as water use by crops, food production, dynamics of virtual water trade, and scenario modeling. Most recent trends in irrigated and cropland areas are crucial given the present food crisis and will be studied using Mid-decadal global land survey 2005 (MDGLS2005) data (Gutman et al., 2008). We will analyze the recent trends in irrigated areas over India, which has the second largest annualized irrigated area (AIA) fraction (28.5% of the global AIA of 467 Mha), next only to China?s 32.5% AIA (Thenkabail, et al. 2008). India's food production has been stagnating for the past decade despite the continued growth in population, a cause of great food security concern for this large developing economy (Naraynamoorthy, 2007). The stagnation in food production coincides with the plateauing of irrigation expansion. We will compare the spatial distribution of the trends in irrigation expansion with the regional patterns of climatic change to understand how the warming is affecting crop water requirements, what are the feedbacks of increased evapotranspiration on climate, and project these effects on the coming decades. The high resolution irrigated area maps over India will improve our understanding of the impacts of water exploitation on the stagnation of crop yields and compare them to the effects of increasing temperatures.

Technical Approach: The GIAM30 will fuse the 8500-odd Landsat Mosaics each from MDGLS2005 (Gutman et al., 2005) and Geocover 1990 (Tucker et al., 2005) of the world with MODIS 500 meter time-series, AVHRR time-series, and digital elevation from SRTM. Irrigated areas will be discerned through suite of methods a nd data involving ideal spectral data bank, spectral matching techniques (Thenkabail et al., 2007a), decision tree algorithms (DeFries, et al., 1998), space-time spiral curves (Thenkabail et al., 2005), and a host of secondary data such as climate, soils, forests, Google Earth very high resolution "zoom in views", information from National statistics, and extensive use of groundtruth (Thenkabail et al., 2008, 2006). The GIAM30m methodology will: (a) build comprehensive, one of its kind, global ideal spectral data bank on irrigated areas (GISDB-IA) using our vast network of global partners by collecting data from precise locations in standardized pattern, (b) strengthen existing GIAM methodology at coarser resolution building upon development and\or use of quantitative approaches on spectral matching techniques (SMTs), decision tree algorithms (DTAs), and space time spiral curves (ST-SCs) in the GIAM project, (c) explore advances in data fusion and data mining approaches, (d) broaden the scope of the use of Google Earth and Groundtruth data, (e) investigate further on relationships between resolution and areas, (f) conduct critical research on sub-pixel areas (SPA) computation approaches leading to greater certainty in area computation, and (d) emphasize on the need to compare remote sensing derived estimates with national statistics for broader understanding, acceptance, and buy-in. Impact assessment of irrigated area expansion on agricultural water use under present and projected climate for India will make use of TOPS (Terrestrial Observation and Prediction System) data and modeling software system (http://ecocast.arc.nasa.gov/). TOPS will be used to produce gridded climate parameters at 1km for I ndia to estimate actual evapotranspiration for the year 1990 and 2005. We will use IPCC climate projections of temperature and monsoonal precipitation for India (Krishna Kumar et al., 2008) to estimate agricultural water demands by the year 2050 under current and expanded irrigation scenarios. Irrigation water use and climatic trends will be compared to shifts in crop production and crop yields from subnational agricultural statistics (Government of India, 2007).

Anticipated Cooperating Investigators: The work will be lead by Dr. Cristina Milesi as PI (Foundation of California State University Monterey Bay) in collaboration with Dr. Prasad S. Thenkabail (International Water Management Institute). It will be strongly supported by Dr. Roland Geerken (Yale University), Dr. Xiangming Xiao (University of New Hampshire), Dr. Seelan Santhosh Kumar (University of North Dakota), Ms. Bethany Kurz (UND), Dr. Gabriel Senay (USGS EROS), Dr. Eddy de Pauw (International Center for Agriculture Research in Dry Areas), Prof. Songcai You (Chinese Academy of Sciences), and Dr. Obi Reddy Ganguntla (Indian Council for Agricultural Research).

References: