Combined Use of Multi-Temporal Landsat-8 and Sentinel-2 Images for Wheat Yield Estimates at the Intra-Plot Spatial Scale

Abstract

The objective of this study is to address the capabilities of multi-temporal optical images to estimate the fine-scale yield variability of wheat, over a study site located in southwestern France. The methodology is based on the Landsat-8 and Sentinel-2 satellite images acquired after the sowing and before the harvest of the crop throughout four successive agricultural seasons, the reflectance constituting the input variables of a statistical algorithm (random forest). The best performances are obtained when the Normalized Difference Vegetation Index (NDVI) is combined with the yield maps collected during the crop rotation, the agricultural season 2014 showing the lower level of performances with a coefficient of determination (R2) of 0.44 and a root mean square error (RMSE) of 8.13 quintals by hectare (q.h−1) (corresponding to a relative error of 12.9%), the three other years being associated with values of R2 close or upper to 0.60 and RMSE lower than 7 q.h−1 (corresponding to a relative error inferior to 11.3%). Moreover, the proposed approach allows estimating the crop yield throughout the agricultural season, by using the successive images acquired from the sowing to the harvest. In such cases, early and accurate yield estimates are obtained three months before the end of the crop cycle. At this phenological stage, only a slight decrease in performance is observed compared to the statistic obtained just before the harvest.

Remote Sensing for Precision Agriculture: Sentinel-2 Improved Features and Applications

“…Sentinel-2-derived research has approached yield predictions in two different fashions. On the one hand empirical models using fitting techniques of the obtained data sets (VIs and climatological data correlated with actual yields) show robust results with machine learning and random forest techniques [55,56] as well as relatively high correlation, with an R 2 over 0.70, when using simple regressions with the most suitable VIs (i.e., NDVI) [57,58]. On the other hand, the radiative transfer models approach uses the calculated biophysical parameters such as LAI or FAPAR to model crop growth and predict yields [59][60][61][62].…”

Within-Field Relationships between Satellite-Derived Vegetation Indices, Grain Yield and Spike Number of Winter Wheat and Triticale

“…The highest prediction accuracy for wheat was using the model which includes NDVI, NDWI and WETNESS. Prediction of wheat yield based on satellite-derived NDVI was improved if auxiliary data such as grain yield from previous seasons were included in models [36]. Dempewolf et al (2013) [12] obtained the highest R 2 values of 0.964 for NDVI and the yield of wheat grown in Pakistan six weeks before harvest.…”

Relationship between MODIS Derived NDVI and Yield of Cereals for Selected European Countries

“…Remote sensing data from satellites more often are used as input data in crop models for grain yield forecasts. AVHRR and MODIS are the most popular satellite sensors to obtain within-season information to forecast final yield at regional scale [6][7][8], while for smaller areas, the most commonly used data are derived from Landsat and Sentinel satellites [9][10][11]. Multispectral data in the range of visible light and near-infrared are used for calculation of vegetation indices which are used as predictors of cereal grain yield.…”