Space-based vegetation health for wheat yield modeling and prediction in Australia

Abstract

, "Space-based vegetation health for wheat yield modeling and prediction in Australia," J. Appl. Remote Sens. 12(2), 026002 (2018), doi: 10.1117/1.JRS.12.026002. Abstract. An early warning of crop losses in response to weather fluctuations helps farmers, governments, traders, and policy makers better monitor global food supply and demand and identifies nations in need of aid. This paper discusses the utility of vegetation health (VH) indices, derived from the advance very high-resolution radiometer (AVHRR) and visible infrared imaging radiometer suite (VIIRS), as a proxy for modeling Australian wheat from the National Oceanic and Atmospheric Administration (NOAA) operational afternoon polar-orbiting satellites. These models are used to assess wheat production and to provide an early warning of drought-related losses. The NOAA AVHRR-and VIIRS-based VH indices were used to model wheat yield in Australia. A strong correlation (≥0.7) between wheat yield and VH indices was found during the critical reproductive stage of development (enhanced crops sensitivity to weather), which starts 2 to 3 weeks before and ends 2 to 3 weeks after wheat heading. The results of modeling and independent testing proved that the VH indices (especially those estimating thermal and health conditions) are a good proxy providing 1 to 2 months before harvest yield prediction (with 3% to 6% error). With the new generation of NOAA-20 operational polar-orbiting satellites, launched in November 2017, the VH method will be improved considerably both in an advanced crop/pasture prediction, spatial resolution, and accuracy. © The Authors. Published by SPIE under a Creative Commons Attribution 3.0 Unported License. Distribution or reproduction of this work in whole or in part requires full attribution of the original publication, including its DOI.

Enhancing Crop Yield Prediction Utilizing Machine Learning on Satellite-Based Vegetation Health Indices

“…However, yield estimates are challenging due to complex interactions between crop growth and yield-influencing natural factors, such as weather [5][6][7], soil conditions ( [7,8])), disease [9], and anthropogenic factors such as irrigation, fertilizers, tillage, rotation, and seed varieties [9]. Although some crop yield models estimate the yield reasonably well for subregions, e.g., wheat [10][11][12][13][14][15][16][17][18], rice [19][20][21], potato [22,23], soybean [3,4,24,25], maize [26][27][28][29], corn [25,[30][31][32], cotton [33], barley [15,17,34], cereal [35], coffee [36], canola [15,37], and sugarcane [17], better performance for yield prediction is still desirable [17].…”

“…Conventional statistics can determine the regression model for calculating yearly crop productivity if the historical annual crop yield and the vegetation condition index (VCI), as well as the thermal condition index (TCI) time series during the same period, are available, see, e.g., [11,22,32,[47][48][49]. This approach requires a good understanding of the relationship between the annual crop yield (dependent variable) and the VCI and TCI data (independent variables), and this is often not satisfied in practice.…”

“…This approach requires a good understanding of the relationship between the annual crop yield (dependent variable) and the VCI and TCI data (independent variables), and this is often not satisfied in practice. Assumptions about the relationship are consequently used for creating the models ( [11,21,22,32,[47][48][49]). On the contrary, ML is based on a learning approach that can develop models without any assumption regarding the distribution and interconnections of input variables [50].…”

“…Satellite data are cheaper, have higher coverage, and are more accessible than in situ data [21]. Therefore, the rapid development of satellite technology coupled with modern ML methods has recently driven the empirical approach to become more efficient and hence is currently the preferred approach (see, e.g., [3,4,[10][11][12][14][15][16][17][18][19][20][22][23][24]27,28,[31][32][33][34]36,44]).…”

“…However, VCI and TCI have two theoretical advantages. Firstly, they are weather-related components of the environment in each ecosystem, and therefore indicate the cumulative weather effects on the annual crop yield fluctuation around the trend [11]. Therefore, if the predictors are VCI/TCI, the responses will be yield deviations from the trend.…”