Characterizing Major Crop Change Trends in Florida from Multiple Sources: A Comparison of USDA Cropland Data Layer and Statewide Random Forest Classifications of the Top Five Crops

Details

• Personal Author:
  Schaper TA
• Description:
  Monitoring crop coverage and crop change in agricultural areas is of paramount importance for the understanding and management of food production, statewide economies, environmental impacts and global environmental change. The USDAs Cropland Data Layer (CDL) is the most commonly used database of crop coverage produced annually for the conterminous US. This research compared the output of the published CDL products to a statewide classification focused on the top five agricultural crops in the state of Florida. Given the prominence of agriculture to the statewide economy and the extensive area of the state in crop production, the consequence of accurate understanding of crop coverage and crop changes is key. The research utilized a multi-date, within season, image stack to produce a five-crop classification using a Random Forest (RF) classifier. 250 testing sites were used to compare both the RF and CDL products, for both 2008 and 2016, and results showed the CDL to have lower than 40% overall accuracy, compared to over 85% overall accuracy for the RF classification, for these five crops. Change in crop coverage statewide decreased about 4% between 2008 and 2016, and for individual crop coverages, over 12% decrease occurred in citrus and over 4% decrease in peanuts. Cotton and strawberry coverages increased significantly, although both are much less significant crops in terms of area statewide. Sugarcane remained stable in area over time. Changes in agricultural production, especially given the position of Florida as the top citrus and sugarcane producing state, and their importance to the statewide economy, are of key concerns to agricultural and land managers alike. [Description provided by NIOSH]
• Subjects:
  Agriculture Occupational Health Safety
• Keywords:
  Algorithms Environmental Health Monitoring Environmental Impacts Foodstuff Machine Learning
• Publisher:
  University of Florida
• Document Type:
  Text
• Funding:
  Cooperative Agreement
• Genre:
  Thesis
• Place as Subject:
  Florida ; OSHA Region 4
• CIO:
  National Institute for Occupational Safety and Health (NIOSH)
• Topic:
  Workplace Safety & Health
• Location:
  North America
• Pages in Document:
  1-37
• NIOSHTIC Number:
  nn:20070668
• Citation:
  Gainesville, FL: University of Florida, 2021 May; :1-37
• Federal Fiscal Year:
  2021
• NORA Priority Area:
  Agriculture, Forestry and Fishing
• Performing Organization:
  University of Florida, Gainesville
• Peer Reviewed:
  False
• Start Date:
  20160930
• Source Full Name:
  Characterizing major crop change trends in Florida from multiple sources: a comparison of USDA cropland data layer and statewide random forest classifications of the top five crops
• End Date:
  20270929
• Collection(s):
  National Institute for Occupational Safety and Health
• Main Document Checksum:
  urn:sha-512:16a3e5fa03fab18e9d2dda7b90efc902b8f6341818e7a32aeb911f848b7fc871c9b01e286713ec14793fd7b0dbf1a16c89a83bebb18543c1247c0dc3e6b4aa3b
• Download URL:
  https://stacks.cdc.gov/view/cdc/208844/cdc_208844_DS1.pdf
• File Type:
  [PDF - 7.29 MB ]