Subtitle

Subtitle

High resolution crop type maps are an important tool for improving food security, and remote sensing is increasingly used to create such maps in regions that possess ground truth labels for model training. However, these labels are absent in many regions, and models trained in other regions on typical satellite features, such as those from optical sensors, often exhibit low performance when transferred. Here we explore the use of NASA's Global Ecosystem Dynamics Investigation (GEDI) spaceborne lidar instrument, combined with Sentinel-2 optical data, for crop type mapping. Using data from three major cropped regions (in China, France, and the United States) we first demonstrate that GEDI energy profiles are capable of reliably distinguishing maize, a crop typically above 2m in height, from crops like rice and soybean that are shorter. We further show that these GEDI profiles provide much more invariant features across geographies compared to spectral and phenological features detected by passive optical sensors. GEDI is able to distinguish maize from other crops within each region with accuracies higher than 84\%, and able to transfer across regions with accuracies higher than 82\% compared to 64\% for transfer of optical features. Finally, we show that GEDI profiles can be used to generate training labels for models based on optical imagery from Sentinel-2, thereby enabling the creation of 10m wall-to-wall maps of tall versus short crops in label-scarce regions. As maize is the second most widely grown crop in the world and often the only tall crop grown within a landscape, we conclude that GEDI offers great promise for improving global crop type maps.

Subtitle

Crop type mapping at the field level is critical for a variety of applications in agricultural monitoring, and satellite imagery is becoming an increasingly abundant and useful raw input from which to create crop type maps. Still, in many regions crop type mapping with satellite data remains constrained by a scarcity of field-level crop labels for training supervised classification models. When training data is not available in one region, classifiers trained in similar regions can be transferred, but shifts in the distribution of crop types as well as transformations of the features between regions lead to reduced classification accuracy. We present a methodology that uses aggregate-level crop statistics to correct the classifier by accounting for these two types of shifts. To adjust for shifts in the crop type composition we present a scheme for properly reweighting the posterior probabilities of each class that are output by the classifier. To adjust for shifts in features we propose a method to estimate and remove linear shifts in the mean feature vector. We demonstrate that this methodology leads to substantial improvements in overall classification accuracy when using Linear Discriminant Analysis (LDA) to map crop types in Occitanie, France and in Western Province, Kenya. When using LDA as our base classifier, we found that in France our methodology led to percent reductions in misclassifications ranging from 2.8% to 42.2% (mean = 21.9%) over eleven different training departments, and in Kenya the percent reductions in misclassification were 6.6%, 28.4%, and 42.7% for three training regions. While our methodology was statistically motivated by the LDA classifier, it can be applied to any type of classifier. As an example, we demonstrate its successful application to improve a Random Forest classifier.

Subtitle

Crop yield maps estimated from satellite data increasingly are used to understand drivers of yield trends and variability, yet satellite-derived regional maps are rarely compared with location-specific yields due to the difficulty of acquiring sub-field ground truth data at scale. In commercial agricultural systems, combine harvesters with onboard yield monitors collect real-time yield data during harvest with high spatial resolution, generating a large ground dataset. Here, we leveraged a yield monitor dataset of over one million maize field observations across the United States Corn Belt from 2008 to 2018 to evaluate the Scalable Crop Yield Mapper (SCYM). SCYM uses region-specific crop model simulations and climate data to interpret vegetation indices from satellite observations, thus enabling efficient sub-field yield estimation across large regions and multiple years without reliance on ground data calibration. We used the ground dataset to compare alternative SCYM model implementations, define minimum required satellite observation criteria, and evaluate the sensitivity of satellite-based yield estimates to on-the-ground variation in management, soil, and annual weather. We found that smoothing annual time series data with harmonic regression increased 30 m pixel-level accuracy from r2 = 0.31 to 0.40 and reduced dependency on specific satellite observation timing, enabling robust yield estimation on 97% of annual maize area using only Landsat data. Agreement improved as the assessment was aggregated to field-level (r2 = 0.45) and county-level (r2 = 0.69) scales, demonstrating the need for fine-resolution ground truth data to better assess sub-field level accuracy in high resolution products. We found that SCYM and ground data showed a similar yield response to management and environmental variation, particularly capturing linear and nonlinear responses to sowing density, soil water holding capacity, and growing season precipitation. However, sensitivity to factors like soil quality and planting date was muted for SCYM estimates compared to ground-based yields. Random forest models were able to match SCYM performance when trained on at least 1000 ground observations, but performed poorly when tested on years and locations not represented in the training data. Our results demonstrate that satellite yield maps can provide much-needed information on multidecadal yield trends and inform yield gap analyses.