NASA Logo

NTRS

NTRS - NASA Technical Reports Server

Back to Results
Early Season Large-Area Winter Crop Mapping Using MODIS NDVI Data, Growing Degree Days Information and a Gaussian Mixture ModelKnowledge on geographical location and distribution of crops at global, national and regional scales is an extremely valuable source of information applications. Traditional approaches to crop mapping using remote sensing data rely heavily on reference or ground truth data in order to train/calibrate classification models. As a rule, such models are only applicable to a single vegetation season and should be recalibrated to be applicable for other seasons. This paper addresses the problem of early season large-area winter crop mapping using Moderate Resolution Imaging Spectroradiometer (MODIS) derived Normalized Difference Vegetation Index (NDVI) time-series and growing degree days (GDD) information derived from the Modern-Era Retrospective analysis for Research and Applications (MERRA-2) product. The model is based on the assumption that winter crops have developed biomass during early spring while other crops (spring and summer) have no biomass. As winter crop development is temporally and spatially non-uniform due to the presence of different agro-climatic zones, we use GDD to account for such discrepancies. A Gaussian mixture model (GMM) is applied to discriminate winter crops from other crops (spring and summer). The proposed method has the following advantages: low input data requirements, robustness, applicability to global scale application and can provide winter crop maps 1.5-2 months before harvest. The model is applied to two study regions, the State of Kansas in the US and Ukraine, and for multiple seasons (2001-2014). Validation using the US Department of Agriculture (USDA) Crop Data Layer (CDL) for Kansas and ground measurements for Ukraine shows that accuracies of greater than 90% can be achieved in mapping winter crops 1.5-2 months before harvest. Results also show good correspondence to official statistics with average coefficients of determination R(exp. 2) greater than 0.85.
Document ID
20180000183
Document Type
Reprint (Version printed in journal)
Authors
Skakun, Sergii
(Maryland Univ. College Park, MD, United States)
Franch, Belen
(Maryland Univ. College Park, MD, United States)
Vermote, Eric
(NASA Goddard Space Flight Center Greenbelt, MD, United States)
Roger, Jean-Claude
(Maryland Univ. College Park, MD, United States)
Becker-Reshef, Inbal
(Maryland Univ. College Park, MD, United States)
Justice, Christopher
(Maryland Univ. College Park, MD, United States)
Kussul, Nataliia
(National Academy of Sciences of the Ukraine Kiev, Ukraine)
Date Acquired
January 5, 2018
Publication Date
May 1, 2017
Publication Information
Publication: Remote Sensing of Environment
Volume: 195
ISSN: 0034-4257
Subject Category
Earth Resources And Remote Sensing
Statistics And Probability
Report/Patent Number
GSFC-E-DAA-TN46331
Funding Number(s)
CONTRACT_GRANT: NNX17AJ63A
CONTRACT_GRANT: NNX14AR70A
Distribution Limits
Public
Copyright
Other
Keywords
mixture model
mapping
classification
Winter crop
No Preview Available