This is the homepage to host the STaPL tillage maps, which are generated based on a scale transfer framework, Scale Transfer with Pseudo-Labeling (STaPL).
The annual STaPL tillage maps include 30-m tillage types (low-intensity and high-intensity) for corn and soybean cropland in 2001-2023. The maps cover 12 states in the U.S. Midwest, including Illinois (IL), Indiana (IN), Iowa (IA), Kansas (KS), Michigan (MI), Missouri (MO), Minnesota (MN), Nebraska (NE), North Dakota (ND), Ohio (OH), South Dakota (SD), and Wisconsin (WI).
The map is generated by year. Each map has three bands:
The first band "Tillage" stores the tillage types, in which 0 means "high-intensity tillage" and 1 means "low-intensity tillage".
The second band "Number" stores the number of available Landsat observations at the pixel during April to June. Pixels with high observation numbers generally have higher accuracy.
The third band "Filled_Flag" indicates whether the pixel is missing originally but filled using the tillage type from the previous and next years.
The map is named by the year. For example, the STaPL map in 2022 is named as "STAPL_2022".
The STaPL maps are available on Google Earth Engine (GEE) after paper publication. To use the map, access the target map in GEE by:
var STaPL_map = ee.Image("projects/lobell-lab/US_tillage_2024/STAPL_map_final/STAPL_" + year)For example, if you want to use the tillage map in 2022, the code should be like:
var STaPL_map_2022 = ee.Image("projects/lobell-lab/US_tillage_2024/STAPL_map_final/STAPL_2022")Please refer to the STaPL paper.
The STaPL paper is available in Remote Sensing of Environment (https://authors.elsevier.com/a/1n93y7qzTAsOD).
We have another research that uses the same scale-transfer idea for 30-m crop yield mapping, which is also available in Remote Sensing of Environment (https://www.sciencedirect.com/science/article/abs/pii/S003442572400453X).
STaPL Tillage Maps will follow CC-BY-NC-SA 4.0. Thus, those compounds are freely available for academic purposes or individual research, but restricted for commercial use.
We appreciate feedback on the product. We are also actively seeking collaboration on utilizing STaPL maps for research and/or applying the scale transfer method to other variables and regions.
Please contact us by sending an email to Dr. Yuchi Ma (yuchima@stanford.edu) and Prof. David Lobell (dlobell@stanford.edu).
If you find it useful, please star this project and cite our papers:
[1] Ma, Y.*, Shen, Y., Swatantran, A., Kelly, C., Lobell, D.B., 2026. STaPL: Scale Transfer with Pseudo-Labelling for satellite-based mapping of agricultural practices. Remote Sensing of Environment 343, 115500.
[2] Ma, Y., Liang, S.Z., Myers, D.B., Swatantran, A. and Lobell, D.B., 2024. Subfield-level crop yield mapping without ground truth data: A scale transfer framework. Remote Sensing of Environment, 315, p.114427.