Field Crops Research 118 (2010) 13–20 Contents lists available at ScienceDirect Field Crops Research journal homepage: www.elsevier.com/locate/fcr Satellite evidence for yield growth opportunities in Northwest India David B. Lobell a,b,∗ , J. Ivan Ortiz-Monasterio c , Anna S. Lee b a Department of Environmental Earth System Science, Stanford University, Stanford, CA 94305, United States b Program on Food Security and Environment, Stanford University, Stanford, CA 94305, United States c International Maize and Wheat Improvement Center (CIMMYT), Apdo. Postal 6-641, 06600 Mexico D.F., Mexico article info Article history: Received 2 December 2009 Received in revised form 26 March 2010 Accepted 26 March 2010 Keywords: Wheat Yield gap Remote sensing Irrigation Sowing date abstract Improving crop yields in major agricultural regions is one of the foremost scientific challenges for the next few decades. In Northwest India, the stagnation of wheat yields over the past decade presents a distressing contrast to the tremendous yield gains achieved during the Green Revolution. One commonly proposed way to raise yields is to reduce the often considerable gap between yield potential and average yields realized in farmers’ fields, yet the likely effectiveness of different strategies to close this gap has been poorly known. Here we use a unique, decade long satellite-based dataset on wheat yields to examine various options for closing the yield gap in the south of Punjab. Persistent spatial differences in sowing dates and distance from canal are found to be significant sources of yield variation, with the latter factor suggesting the importance of reliable access to irrigation water for yield improvement in this region. However, the total yield gains achievable by addressing persistent factors are only a small fraction of yield losses in farmers’ fields. The majority of the yield gap is found to arise from factors unrelated to field location, such as interactions between management and weather. Technologies that improve farmers’ ability to anticipate or adjust to weather variations, or that improve stability of genotype performance across different weather conditions, therefore appear crucial if average crop yields are to approach their genetic potential. © 2010 Elsevier B.V. All rights reserved. 1. Introduction The Green Revolution transformed cropping systems through- out much of the world, with many areas switching from low input, low yielding agriculture in the 1950s to highly intensive commer- cial farms by the 1970s. Nowhere was this transition more dramatic or complete than in the western Indo-Gangetic Plains (IGP), where irrigated and heavily fertilized wheat fields now dominate the win- ter landscape. The more than doubling of wheat production in Indian states such as Punjab and Haryana over this time led to sub- stantial improvements in food security, and currently the two states alone provide more than a third of all wheat produced in India on a small fraction of the total area (Indian Ministry of Agriculture, 2009). A similar story is evident for the world as a whole, with yield increases responsible for more than three quarters of global cereal production increases since 1960 (Bruinsma, 2003). Currently more than 40% of all food is produced on the small fraction (<3%) of the Earth’s land surface that is irrigated (FAO, 2002). Sustaining yield progress in these regions will be critical if future cereal demand, ∗ Corresponding author at: Department of Environmental Earth System Science, Stanford University, 473 via ortega, Stanford, CA 94305, United States. E-mail address: dlobell@stanford.edu (D.B. Lobell). which is expected to rise 50% in the next 30 years (FAO, 2006), will be met without an acceleration of deforestation rates. However, yield progress has slowed in recent years for wheat in IGP, and in Punjab in particular (Fig. 1a), similar to several crops in many key regions (Cassman, 1999). In IGP, the slowing yield trends have led to concerns that there is limited scope for future improve- ments (Ladha et al., 2003; Chatrath et al., 2007; Joshi et al., 2007) or, even more ominously, that even current yield levels may be unsus- tainable because of soil degradation or climate trends (Pathak et al., 2003). The primary source of historical yield increases was the com- bination of expanded irrigation, adoption of semi-dwarf modern wheat varieties, and greater use of fertilizers (Evenson and Gollin, 2003). However, the main opportunities for further yield improve- ments are less clear (Chatrath et al., 2007; Joshi et al., 2007). Given that experimental yields are still well above average district yields, there would appear to be some room for further yield growth with existing cultivars. Indeed, field-based and modeling studies have indicated poten- tial gains from higher fertilization (Ladha et al., 2003; Khurana et al., 2008), more reliable access to irrigation water (Tyagi et al., 2004; Sarkar et al., 2009), adoption of conservation agriculture (Hobbs et al., 2008), and more timely sowing of wheat (Timsina et al., 2008), among other factors. Yet because studies are typically limited to single factors and to spatial scales significantly smaller than an entire district, there is little consensus on the overall scope for yield 0378-4290/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.fcr.2010.03.013