Modeled conterminous United States Crop Cover datasets for 2004

Crop cover maps have become widely used in a range of research applications. Multiple crop cover maps have been developed to suite particular research interests. The National Agricultural Statistics Service (NASS) Cropland Data Layers (CDL) are a series of commonly used crop cover maps for the conterminous United States (CONUS) that span from 2008-2013. In this investigation we wanted to expand the temporal coverage of the NASS CDL archive back to 2000 by creating yearly NASS CDL-like crop cover maps derived from a classification tree model algorithm. We used over 11 million crop sample records to train a classification tree algorithm and to develop a crop classification model (CCM). The model was used to create crop cover maps for the CONUS for years 2000-2013 at 250 meter spatial resolution. The CCM and the maps for years 2008-2013 were assessed for accuracy relative to downscaled NASS CDLs to 250 meter. The CCM performed well against a withheld test dataset with a prediction accuracy of over 90 percent. The assessment of the crop cover maps indicated that the model performed well spatially, placing crop cover pixels within their known domains. However, the model did show a bias toward the “Other” crop cover class which caused frequent misclassifications of pixels around the periphery of large crop cover patches and of pixels that form small, sparsely dispersed crop cover patches.

作物覆盖地图已在众多研究领域得到广泛应用，目前已有多款适配特定研究需求开发的专用作物覆盖地图。美国国家农业统计服务局（National Agricultural Statistics Service, NASS）推出的耕地数据图层（Cropland Data Layers, CDL）是一套常用于美国本土（conterminous United States, CONUS）的主流作物覆盖地图数据集，时间跨度为2008年至2013年。本研究旨在通过分类树模型算法构建生成与NASS CDL风格一致的年度作物覆盖地图，将NASS CDL档案的时间覆盖范围回溯至2000年。研究团队使用超过1100万条作物样本记录训练分类树算法，构建了作物分类模型（Crop Classification Model, CCM）。利用该模型，我们生成了2000年至2013年美国本土范围内、空间分辨率为250米的作物覆盖地图。针对2008年至2013年的模型生成地图与降尺度至250米分辨率的原始NASS CDL，我们开展了精度评估。结果显示，模型在预留测试数据集上表现优异，预测准确率超过90%。对作物覆盖地图的评估表明，模型在空间分布上表现良好，能够将作物覆盖像元准确归类至其对应的已知类别域中。不过，模型存在针对"Other"作物覆盖类别的分类偏差，这导致在大型作物覆盖斑块的边缘区域，以及小型、稀疏分散的作物覆盖斑块像元上频繁出现误分类情况。