Accuracies, Errors, and Uncertainties of Global Cropland Products - PowerPoint Presentation

Accuracies, Errors, and Uncertainties of Global Cropland Products Kamini Yadav, PhD Student Advisor: Dr. Russell Congalton Natural resource & the environment University of New Hampshire

Introduction Lack of interoperability, insufficient information on accuracy highlighting strengths and weaknesses in existing land cover maps ( Herold et al., 2008 ) Assessing the map accuracy in an objective manner is fundamental to most mapping projects ( Foody 2002; Strahler , 2006 ) With the advent of advanced digital remote sensing classification, there are currently five major global cropland maps: (1) Thenkabail et al. (2009a, b, 2011) , (2) Ramankutty et al. (1998) , (3) Goldewijk et al. (2011) , (4) Portmann et al. (2009) , Siebert and Doll (2009) , and (5) Pittman et al. (2010). The solution to improve the monitoring of global croplands lies in mapping them routinely, rapidly, consistently with sufficient accuracy ( Congalton and Green 2009 ) Significant impediments in crop type mapping is the lack of quality training data ( Shao et al., 2010 ) Similarity in spectral reflectance of different crops, variability in field to field reflectance of the same crop, particular combinations of crops ( Wheeler and Misra , 1980; Buechel et al., 1989 In practical situations cropping environments exhibit smaller average field extent than the pixel size of the imagery ( Wardlow et al., 2007 )

Overarching Goal The overarching goal of this research is to develop methods and approaches as well as conduct comprehensive evaluation of accuracies, errors, and uncertainties of various global cropland products which are produced using a wide array of remotely sensed and geospatial data. This will be performed based on novel approaches using a large volume of reference data from different reliable sources.

Objectives Evaluate and develop methods and approaches of determining accuracies, errors, and uncertainties for global cropland products Organize various distinct global cropland data products (e.g., croplands, irrigation versus rain fed, cropping intensity, crop type), produced at number of different resolutions of remotely sensed data Establish novel global reference data for accuracies, errors, and uncertainties sourced from very high resolution imagery (VHRI), ground or field data, secondary data, and crowd sourced data Measure the accuracy, error and uncertainty of various global cropland products (e.g., croplands, irrigation versus rain fed, cropping intensity, crop type) that are in various resolutions using espoused methods and approaches and the novel reference datasets Accuracy confidence with different spatial and spectral resolution remote sensing data in diverse crop growing environments

Review of existing accuracy methods   Methods References Strengths Limitations 1 Looks right or good Dicks &Lo, 1990 Ease of good looking maps Subjective approach 2 Comparison of areal extent of classes in thematic map (non-site specific)   Objective approach, correct proportions Locational accuracy ignored, report high accuracy, use same data as used by training classifier , incorrect locations 3 Accuracy metrics based on comparison of class labels and ground data for specific locations (Site specific )   Independent datasets, accuracy metrics   Includes only percentage of cases correctly allocated 4 Confusion or error matrix (pattern of class allocation relative to reference data ) Congalton, 1994; Congalton and Green,1999 Measures of accuracy use information content of confusion matrix fully More scope to extend the analysis 5 Spatial distribution of accuracy, errors and uncertainty Kriging ( Steele et al. 1998 ), Geographically weighted Regression and difference measure ( Comber et al., 2012 ) for Boolean and Fuzzy classes    

Organize global cropland products   Products Data used Spatial Resolution C ropland classes Other classes 1Pittman et al 2010MODIS250 m Cropland extent 2Thenkabail et al 2009AVHRR1 KmCropland, Irrigated and Rain fed dominance, Natural vegetation with minor cropland fractions 3FAO Aquastat Inventory data Crop Statistical data 4Portmann et al. 2008National Sub-National statistical data 10 kmIrrigated and Rain fed 5Friedl et al. 2010MODIS 500 m Global Croplands 6Loveland et al. 2000 AVHRR1 kmGlobal Croplands 7Goldewijk et al. 2011Population, Cropland pasture statistics combined with satellite information10 km Cropland statistics 8Siebert & Döll,2009Growing areas and cropping seasonClimate and Soil data 10 kmCrop production in irrigated and rain fed agriculture 9Ramankutty et al 2000Agriculture inventory data and satellite derived land cover data (MODIS & GLC2000)10 km Global agriculture land 10Yu et al 2013Landsat and MODIS30mCropland extentBare Cropland * Best available current state-of-art Global cropland extent maps

Global reference data from very high resolution imagery

Accuracy confidence with different types of remote sensing data in diverse crop growing environments

Perform accuracies, errors, and uncertainties in different continents based on the collected reference data Produce accuracy standards and protocols to perform the accuracy assessment and spatial uncertainty analysis for different cropland products Developing methods and strategies using high quality reference data to achieve best accuracy results in each continent

References Congalton, R. G. (1994). Accuracy assessment of remotely sensed data: future needs and directions. In: Proceedings of Pecora 12 land information from space-based systems ( pp. 383 –388). Bethesda: ASPRS Congalton, R. G., & Green, K. (1999). Assessing the accuracy of remotely sensed data: principles and practices. Boca Raton: Lewis Publishers. Congalton, R.G. and Green, K. 2009. Assessing the accuracy of remotely sensed data: principles and practices, 2nd, London: Taylor and Francis. Comber A., Fisher P., Brundson C., Khmag A.,2012. Spatial Analysis of remote sensing image classification accuracy. Remote Sensing of Environment 127, 237-246.Congalton, R.G. 1988. Using spatial autocorrelation analysis to explore the errors in maps generated from remotely sensed data. Photogrammetric engineering and remote sensing, 54(5): 587–592.Dicks S E., &Lo, T. H.C. 1990. Evaluation of thematic map accuracy in a land use and land cover mapping program. Photogrammetric Engineering & Remote Sensing, 56, 1247-1252.Foody GM, 2002. Status of land cover classification accuracy. Remote Sensing of Environment, 80:185-201.Goldewijk, K., A. Beusen, M. de Vos and G. van Drecht, 2011. The HYDE 3.1 spatially explicit database of human induced land use change over the past 12,000 years, Global Ecology and Biogeography 20(1): 73-86.DOI: 10.1111/j.1466-8238.2010.00587.xHerold M., Mayaux P., Woodcock C.E., Baccini A., Schmullius C., 2008. Some challenges in global land cover mapping: An assessment of agreement and accuracy in existing 1 km datasets. Remote Sensing of Environment 112, 2538-2556McGwire, K. C., & Fisher, P. (2001). Spatially variable thematic accuracy: Beyond the confusion matrix. In C. T. Hunsaker, M. F. Goodchild, M. A. Friedl, & T. J. Case(Eds.), Spatial uncertainty in ecology: Implications for remote sensing and GIS applications (pp. 308–329). New York: Springer–VerlagThenkabail. P., Lyon, G.J., Turral, H., and Biradar, C.M. 2009a. Book entitled: “Remote Sensing of Global Croplands for Food Security” (CRC Press- Taylor and Francis group, Boca Raton, London, New York. Pp. 556

Thenkabail , P.S., Biradar C.M., Noojipady , P., Dheeravath , V., Li, Y.J., Velpuri , M., Gumma , M., Reddy, G.P.O., Turral, H., Cai, X. L., Vithanage, J., Schull, M., and Dutta, R. 2009b. Global irrigated area map (GIAM), derived from remote sensing, for the end of the last millennium International Journal of Remote Sensing. 30(14): 3679-3733Thenkabail, P.S., Hanjra, M.A., Dheeravath, V., Gumma, M. 2011. Book Chapter # 16: Global Croplands and Their Water Use Remote Sensing and Non-Remote Sensing Perspectives. In the Book entitled: “Advances in Environmental Remote Sensing: Sensors, Algorithms, and Applications”. Taylor and Francis Edited by Dr. Qihao Weng. Pp. 383-419Pittman K., Hansen M., Becker-Reshef I., Potapov P., and Justice C., 2010. Estimating Global Cropland Extent with Multi-year MODIS Data, Remote Sensing, 2, 1844-1863; doi: 10.3390/rs2071844Portmann, F., Siebert, S., & Döll, P., 2009. MIRCA2000 – Global monthly irrigated and rainfed crop areas around the year 2000: a new high- resolution data set for agricultural and hydrological modelling. Global Biogeochemical Cycles, 2008GB0003435.Ramankutty, N., and J. A. Foley, 1998, Characterizing patterns of global land use: An analysis of global croplands data, Global Biogeochemical Cycles, 12(4), 667–685, doi: 10.1029/98GB02512Siebert, S., & Döll, P., 2009.Quantifying blue and green virtual water contents in global crop production as well as potential production losses without irrigation. Journal of Hydrology, doi:10.1016/j.jhydrol.Strahler A., Boschetti L., Foody G.M., Friedl M.A., Hansen M.C., Herold M., Mayaux P., Morisette J.T., Stehman S.V. and Woodcock C.E. 2006. Global Land Cover Validation: Recommendations for Evaluation and Accuracy Assessment of Global Land Cover Maps, GOFC-GOLD Report No. 25Steele, B. M., Winne, J. C., & Redmond, R. L. (1998). Estimation and mapping of misclassification probabilities for thematic land cover maps.Remote Sensing of Environment, 66, 192 – 202.Wardlow, B.D., Egbert, S.L., Kastens, J.H., 2007. Analysis of time-series MODIS 250 m vegetation index data for crop classification in the U.S. Central Great Plains. Remote Sens. Environ. 108, 290–310. Wheeler, T., and M. G. Kay, 2010. Food crop production, water and climate change in the developing world, outlook on Agriculture, 39(4): 239-243.