Devendra Dahal Bruce Wylie Danny Howard 2018 spreadsheet https://doi.org/10.5066/F7B27TG8 Spatially accurate annual crop cover maps are an important component to various planning and research applications; however, the importance of these maps varies significantly with the timing of their availability. Utilizing a previously developed crop classification model (CCM), which was used to generate historical annual crop cover maps (classifying nine major crops: corn, cotton, sorghum, soybeans, spring wheat, winter wheat, alfalfa, other hay/non alfalfa, fallow/idle cropland, and ‘other’ as one class for remaining crops), we hypothesized that such crop cover maps could be generated in near real time (NRT). The CCM was trained on 14 temporal and 15 static geospatial datasets, known as predictor variables, and the National Agricultural Statistics Service (NASS) Cropland Data Layers (CDL) was used as the dependent variable. We were able to generate a NRT crop cover map by the first day of September through a process of incrementally removing weekly and monthly data from the CCM and comparing the subsequent map results with the original maps and NASS CDLs. Initially, our NRT results revealed training error of 1.4% and test error of 8.3%, as compared to 1.0% and 7.6%, respectively for the original CCM. Through the implementation of a new ‘two-mapping model’ approach, we were able to substantially improve the results of the NRT crop cover model. We divided the NRT model into one ‘crop type model’ to handle the classification of the nine specific crops and a second, binary model to classify crops as presence or absence of the ‘other’ crop. Under the two-mapping model approach, the training errors were 0.8% and 1.5% for the crop type and binary model, respectively, while test errors were 5.5% and 6.4% for crop type and binary model, respectively. With overall mapping accuracy for the map reaching 67.18 percent, this approach shows strong potential for generating crop type maps of current year in September. The purpose of the study was to identify the earliest possible month for producing annual crop cover maps with acceptable accuracies. 20100901 publication date Complete None planned -125.85937500000001 -66.09375000000001 49.49667452747045 24.04646399966658 USA None crop cover mapping near real time conus modelling USGS Metadata Identifier USGS:5a8700d7e4b00f54eb3a2b08 Common geographic areas United States None. Please see 'Distribution Info' for details. None. Users are advised to read the data set's metadata thoroughly to understand appropriate use and data limitations. Devendra Dahal (CTR) U.S. Geological Survey, CLIMATE & LAND-USE mailing address

47914 252Nd Street

Sioux Falls SD 57198 United States 605-594-2716 devendra.dahal.ctr@usgs.gov Aaron M. Friesz Bruce K. Wylie Daniel M. Howard 20170103 publication Remote Sensing Letters vol. 8, issue 4 n/a Informa UK Limited ppg. 389-398 http://dx.doi.org/10.1080/2150704x.2016.1271469 Daniel M. Howard Bruce K. Wylie 20140601 publication Photogrammetric Engineering & Remote Sensing vol. 80, issue 6 n/a American Society for Photogrammetry and Remote Sensing ppg. 537-549 http://dx.doi.org/10.14358/PERS.80.6.537-549 No formal attribute accuracy tests were conducted. No formal logical accuracy tests were conducted. Data set is considered complete for the information presented, as described in the abstract. Users are advised to read the rest of the metadata record carefully for additional details. A formal accuracy assessment of the horizontal positional information in the data set has not been conducted. A formal accuracy assessment of the vertical positional information in the data set has either not been conducted, or is not applicable. Previously developed crop classification model (CCM), which was used to generate historical annual crop cover maps, was utilised to generate these annual near real time (NRT) crop cover datasets. The CCM was developed in RuleQuest C5 software (version 2.07 GPL - https://www.rulequest.com/see5-info.html) based on regression tree model.The CCM was trained on 14 temporal and 15 static geospatial datasets, as predictor variables, and the National Agricultural Statistics Service Cropland Data Layers as reference variable. For this study, the CCM model were used by incrementally removing weekly and monthly data upto 1st of September and implementing ‘two-mapping model’ approach. In the 'two-mapping model' approach, we divided the CCM model into one ‘crop type model’ to handle the classification of the nine specific crops and a second, binary model to classify crops as presence or absence of the ‘other’ crop. The specific crop and binary models were ingested to internally developed MapC5 software along with the input spatial datasets to generate output crop maps. MapC5 software also generates confidence map along with the output crop cover map. The confidence maps holds information about what percent of the training observations at each respective prediction rule set which were correctly classified. We, then, created a probability of ‘Other’ map from the binary crop confidence map by subtracting 100 when map class is not ‘Other’ to differentiate the confidence by make distinct range for two classes. Finally, two separate spatial crop cover maps were merged using the recreated confidence map to generate final maps for all mapping years, 2008 – 2016. Pixels from binary map were chosen when new probability of ‘Other’ map is greater than and equal to 75 percent and binary map has ‘Other’ class, otherwise all of the pixels were chosen from pure class map. Threshold of 75 percent was chosen after conducting a number of experiments. By doing so, pixels with higher confidents from binary model were kept as ‘Other’ class, pixels with lesser confidents of ‘Other’ class were binned to separate known types (not ‘Other’). The modeled maps were compared against the NASS CDL with 500,000 randomly sampled points. The sample points were used to extract values from both data products for a statistical analysis. The extracted values were formatted into a confusion matrix for analysis and computed Producer's, User's, overall accuracies using the equations below. Overall accuracy = ( sum of all correctly classified pixels of each crop type)/(Total number of pixels of the area) x100 Producer’s accuracy = (number of correctly classified pixels of a crop type)/(total number of the crop pixels in CDL map ) x100 User’s accuracy = (number of correctly classified pixels of a crop type)/(total number of the crop pixels in classifed map ) x100. 20170801 NRT_Crop_Mapping_CONUS_2010_500k_Confusion_matrix.csv Comma Separated Value (CSV) file containing data. Producer defined User's accuracy NASS CDL Producer defined 69.79% Corn Producer defined 73.72% Cotton Producer defined 63.17% Sorghum Producer defined 68.07% Soybeans Producer defined 62.99% Spring Wheat Producer defined 71.71% Winter Wheat Producer defined 59.97% Alfafla Producer defined 60.37% Other Hay/Non Alfalfa Producer defined 64.01% Fallow/idle cropland Producer defined 64.38% Other Crops Producer defined None None Producer defined Producdr's accuracy Modellled Producer defined 70.54% Corn Producer defined 73.07% Cotton Producer defined 53.50% Sorghum Producer defined 63.31% Soybeans Producer defined 64.28% % Spring Wheat Producer defined 74.01% Winter Wheat Producer defined 42.18% Alfalfa Producer defined 15.55% Other Hay/Non Alfalfa Producer defined 60.90% Fallow/idle cropland Producer defined 73.08% Other Crops Producer defined None None Producer defined 67.18% Overall accuracy Producer defined GS ScienceBase U.S. Geological Survey mailing address

Denver Federal Center, Building 810, Mail Stop 302

Denver CO 80225 United States 1-888-275-8747 sciencebase@usgs.gov Unless otherwise stated, all data, metadata and related materials are considered to satisfy the quality standards relative to the purpose for which the data were collected. Although these data and associated metadata have been reviewed for accuracy and completeness and approved for release by the U.S. Geological Survey (USGS), no warranty expressed or implied is made regarding the display or utility of the data on any other system or for general or scientific purposes, nor shall the act of distribution constitute any such warranty. Digital Data https://doi.org/10.5066/F7B27TG8 None 20200818 Devendra Dahal (CTR) U.S. Geological Survey, CLIMATE & LAND-USE mailing address

47914 252Nd Street

Sioux Falls SD 57198 United States 605-594-2716 devendra.dahal.ctr@usgs.gov FGDC Biological Data Profile of the Content Standard for Digital Geospatial Metadata FGDC-STD-001.1-1999 Record created using USGS Metadata Wizard tool. (https://github.com/usgs/fort-pymdwizard)